choosing a data migration solution for emc symmetrix · pdf filechoosing a data migration...

Choosing a Data MigrationSolution for EMC Symmetrix Arrays

Version 3.0

• Selecting a Data Migration Solution

• Data Migration Solutions for EMC Symmetrix Arrays

• Factors Affecting the Choice of Data Migration

Andrew LubeckDonald Fried-Tanzer

Choosing a Data Migration Solution for EMC Symmetrix Arrays2

Copyright © 2008 – 2013 EMC Corporation. All rights reserved.

EMC believes the information in this publication is accurate as of its publication date. The information issubject to change without notice.

THE INFORMATION IN THIS PUBLICATION IS PROVIDED “AS IS.” EMC CORPORATION MAKES NOREPRESENTATIONS OR WARRANTIES OF ANY KIND WITH RESPECT TO THE INFORMATION IN THISPUBLICATION, AND SPECIFICALLY DISCLAIMS IMPLIED WARRANTIES OF MERCHANTABILITY ORFITNESS FOR A PARTICULAR PURPOSE.

Use, copying, and distribution of any EMC software described in this publication requires an applicablesoftware license.

EMC2, EMC, and the EMC logo are registered trademarks or trademarks of EMC Corporation in the UnitedState and other countries. All other trademarks used herein are the property of their respective owners.

For the most up-to-date regulator document for your product line, go to EMC Online Support(https://support.emc.com).

Part number H4380.4

Contents

Preface

Chapter 1 IntroductionIntroduction ....................................................................................... 20Simple selection model .................................................................... 22Data migration................................................................................... 24

Data migration defined..............................................................24Virtualization...............................................................................24Device external identity .............................................................25

Full selection model summary........................................................ 26

Chapter 2 Reasons for Moving DataIntroduction ....................................................................................... 30Data growth ....................................................................................... 30Technical refresh................................................................................ 30Performance improvements ............................................................ 31Information Lifecycle Management (ILM).................................... 33Risk reduction.................................................................................... 33Consolidation, localization, and regulatory compliance ............ 34Complexity and transformational migrations .............................. 34

Chapter 3 Existing and Target EnvironmentIntroduction ....................................................................................... 36Initial vs. complete collection of environment information ....... 38Applications....................................................................................... 39

Storage resource management mapping.................................39Application storage redirection................................................40

Choosing a Data Migration Solution for EMC Symmetrix Arrays 3

Contents

Hosts (servers)................................................................................... 40Host I/O capacity ....................................................................... 40Host CPU capacity and migration application control......... 41Host HBA connectivity.............................................................. 41Host migration appliances ........................................................ 41

Storage network ................................................................................ 42Storage network types ............................................................... 42Storage network I/O and port capacity .................................. 42Migration in the storage network ............................................ 43

Storage array information ............................................................... 43Storage array data capacity and utilization............................ 43Storage array I/O bandwidth................................................... 44Director port capacity ................................................................ 44Tiered storage.............................................................................. 45Storage elements homogeneity................................................. 45

Environmental factors: power, cooling, and floor space............. 45Software attributes............................................................................ 46

Chapter 4 EMC Open Migrator/LMIntroduction....................................................................................... 48Key features ....................................................................................... 51Migrate anywhere in the I/O stack................................................ 53Open Migrator considerations........................................................ 54

Chapter 5 Symmetrix Remote Data Facility (SRDF)Introduction....................................................................................... 56SRDF/Synchronous (SRDF/S) ....................................................... 58SRDF/Asynchronous (SRDF/A).................................................... 59SRDF/Data Mobility (SRDF/DM) and adaptive copy modes .. 60SRDF/Automated Replication (SRDF/AR) ................................. 62Concurrent SRDF.............................................................................. 64Cascaded SRDF................................................................................. 65SRDF/Extended Distance Protection (SRDF/EDP) .................... 66SRDF/Star.......................................................................................... 67SRDF and data migration ................................................................ 68

SRDF device migration.............................................................. 69Four-site SRDF data migration................................................. 69


Contents

Chapter 6 EMC Open Replicator for SymmetrixIntroduction ....................................................................................... 72Definitions .......................................................................................... 73

Using VDEVs as source devices for a cold push ................... 73Interaction rules with SRDF ..................................................... 76

Open Replicator, FLM, PPME, and data migration...................... 77

Chapter 7 Symmetrix Software Used for MigrationIntroduction ....................................................................................... 80Federated Live Migration (FLM) .................................................... 81

Benefits of using Federated Live Migration............................81FLM nondisruptive migration overview.................................82

Symmetrix Virtual Provisioning ..................................................... 84EMC TimeFinder family of local replication products ................ 87

TimeFinder and data migration................................................89Virtual LUN technology................................................................... 91

Symmetrix Optimizer.................................................................92Symmetrix Optimizer Virtual LUN technology.....................93Enhanced Virtual LUN technology..........................................95

Fully Automated Storage Tiering (FAST) ...................................... 98FAST VP .......................................................................................98

Federated Tiered Storage ............................................................... 100Introduction to FTS...................................................................100FTS benefits................................................................................101FTS environment example.......................................................101FTS required components........................................................102Modes of operation...................................................................102FTS and data migration............................................................104Non-migration use cases..........................................................104

Chapter 8 PPME and Management Software for Symmetrix ArraysIntroduction ..................................................................................... 108PowerPath Migration Enabler (PPME) ........................................ 109

Benefits of using PPME ............................................................109PPME with Open Replicator technical overview.................111PPME TimeFinder/Clone........................................................112PPME Host Copy ......................................................................112powermig summary .................................................................113

Solutions Enabler............................................................................. 114Symmetrix Management Console (SMC) .................................... 118ControlCenter .................................................................................. 121

5Choosing a Data Migration Solution for EMC Symmetrix Arrays

Contents

ControlCenter Symmetrix Manager ...................................... 121ControlCenter family add-on licenses................................... 122

Chapter 9 Business FactorsIntroduction..................................................................................... 126Budgetary factors............................................................................ 126Human resources ............................................................................ 127Verification and validation ............................................................ 127Application interruption................................................................ 128Other factors .................................................................................... 128Transformational migrations......................................................... 129

Chapter 10 Applying the ModelUse cases choosing a migration solution for Symmetrix .......... 132

Simple selection model ............................................................ 134Full selection model ................................................................. 134Use case information flow....................................................... 135

Open Replicator for Symmetrix/PPME solution use case ....... 136Simple selection model ............................................................ 136Existing and target environments .......................................... 137Business factors......................................................................... 138Symmetrix VMAX migration software ................................. 138EMC host migration and migration managementsoftware...................................................................................... 139Alternate solution summary................................................... 140

Symmetrix Remote Data Facility (SRDF) solution use case ..... 141Simple selection model ............................................................ 141Existing and target environments .......................................... 142Business factors......................................................................... 143Symmetrix VMAX migration software ................................. 143Symmetrix DMX-4 migration software................................. 144EMC host migration and migration managementsoftware...................................................................................... 144Alternate solution summary................................................... 145

Virtual LUN solution use case ...................................................... 147Simple selection model ............................................................ 147Existing and target environments .......................................... 147Business factors......................................................................... 148Symmetrix VMAX migration software ................................. 148EMC host migration and migration managementsoftware...................................................................................... 149


Contents

Alternate solution summary ...................................................149Federated Live Migration (FLM) solution use case ................... 151

Simple selection model.............................................................151Existing and target environments ..........................................152Business factors .........................................................................153Symmetrix VMAX migration software..................................153EMC host migration and migration managementsoftware ......................................................................................154Alternate solution summary ...................................................155

EMC Open Migrator/LM solution use case................................ 156Host-based migration lowest common denominator..........156Simple selection model.............................................................157Existing and target environments ..........................................157Business factors .........................................................................158EMC host migration software .................................................158Alternate solution summary ...................................................158

Glossary

Index


Contents


Title Page

Figures

1 Migration project steps example .................................................................. 212 I/O stack levels............................................................................................... 383 Storage resource management layers.......................................................... 394 EMC Open Migrator/LM ............................................................................. 485 EMC Open Migrator/LM command line and GUI examples ................. 496 Open Migrator operation .............................................................................. 527 I/O stack layers .............................................................................................. 538 SRDF family base products and additional options ................................. 569 SRDF/S ............................................................................................................ 5810 SRDF/A ........................................................................................................... 5911 SRDF Adaptive Copy mode ......................................................................... 6112 SRDF/Automated Replication (SRDF/AR) single-hop data flow.......... 6213 SRDF/Automated Replication (SRDF/AR) multi-hop data flow .......... 6214 Concurrent SRDF ........................................................................................... 6415 Cascaded SRDF............................................................................................... 6516 SRDF/Extended Distance Protection (SRDF/EDP).................................. 6617 Concurrent SRDF/Star .................................................................................. 6718 Four-site SRDF data migration..................................................................... 7019 Open Replicator hot (or live) push .............................................................. 7320 Open Replicator cold (or BCV) push........................................................... 7421 Open Replicator hot (or live) pull ................................................................ 7522 Open Replicator cold (or point-in-time) pull ............................................. 7523 Open Replicator hot pull data migration.................................................... 7824 FLM operation overview............................................................................... 8325 Symmetrix Virtual Provisioning .................................................................. 8426 Virtual LUN migration.................................................................................. 9127 High-level view of an FTS environment................................................... 10128 PPME operation with pseudo-named devices and Open Replicator ... 111


Figures

29 Optimizer Data Migration Wizard page 3................................................ 11930 EMC ControlCenter topology view........................................................... 12331 Transformational data migrations............................................................. 130


Preface

This TechBook explores the complexity of data migration and how to select adata migration solution for a Symmetrix in an Open Systems environment.This document focuses on identifying all of the factors that may influence thechoice of the best solution. EMC migration solutions will be introduced withan emphasis on key features that relate to the factors affecting the solutionchoice. Examples of applying the model of selecting the best solution for amigration will be provided for multiple use cases. The goal is to enablereaders to utilize this model to find the best solutions for their own migrationchallenges.

As part of an effort to improve and enhance the performance and capabilitiesof its product line, EMC routinely releases revisions of its hardware andsoftware. Therefore, some functions described in this guide may not besupported by all revisions of the software or hardware currently in use. Forthe most up-to-date information on product features, refer to the productrelease notes. If a product does not function properly or does not function asdescribed in this document, please contact your EMC representative.

Note: This document was accurate at publication time. New versions of thisdocument might be released on EMC Online Supporthttps://support.EMC.com. Check to ensure that you are using the latestversion of this document.

Audience The intended audience for this TechBook is storage administrators,system administrators, capacity planners, Chief Information Officers(CIOs), and all members of the team tasked with determiningmigration solutions.


https://support.emc.com/

http://support.EMC.com


12

Preface

Readers of this TechBook are expected to have some familiarity with:

◆ Existing and target application environments (hosts and storageinfrastructure)

◆ Technical (application) limitations on possible migrationsolutions

◆ Business reasons that may influence choice of a migrationsolution

◆ Symmetrix hardware and software including SRDF, TimeFinder,Open Replicator, Federated Live Migration, and VPLEX

◆ EMC software products: PowerPath, PowerPath MigrationEnabler (PPME), Solutions Enabler, ControlCenter, SymmetrixManagement Console (SMC), and Open Migrator

Organization This TechBook is divided into the following chapters:

◆ Chapter 1, “Introduction,” introduces the reader to thecomplexity of choosing the best solution for a particularmigration and all of the factors and features that are explored inmore detail in subsequent chapters.

◆ Chapter 2, “Reasons for Moving Data,” defines eight categories ofreasons for moving data that will primarily drive therequirements of potential migration solutions.

◆ Chapter 3, “Existing and Target Environment,” identifies theinformation that must be collected to assay the existingapplication and storage infrastructure and the desired targetenvironment at the end of the migration.

◆ Chapter 4, “EMC Open Migrator/LM,” details features of theOpen Migrator product.

◆ Chapter 5, “Symmetrix Remote Data Facility (SRDF),” detailsfeatures of the SRDF product with a focus on attributes relevantto data migration.

◆ Chapter 6, “EMC Open Replicator for Symmetrix,” detailsfeatures of the Open Replicator product with a focus on attributesrelevant to data migration.

◆ Chapter 7, “Symmetrix Software Used for Migration,” detailsattributes of five other Symmetrix features used for datamigration: TimeFinder, Federated Live Migration (FLM), VirtualLUN technology, Virtual Provisioning, and Fully AutomatedStorage Tiering (FAST).

Choosing a Data Migration Solution for EMC Symmetrix Arrays

Preface

◆ Chapter 8, “PPME and Management Software for SymmetrixArrays,” introduces three management software applications thatprovide for control and monitoring of Symmetrix featuresdescribed in Chapters 6–8. It also details attributes of the EMCPowerPath Migration Enabler (PPME).

◆ Chapter 9, “Business Factors,” identifies multiple business issuesthat may affect the choice of a migration solution.

◆ Chapter 10, “Applying the Model,” brings together the earlierchapters with examples and use cases of choosing data migrationsolutions for Symmetrix arrays.

Details provided in this TechBook cover key factors and features thatwere up-to-date at the time of printing. However, since EMC iscontinuously improving its products it is essential to review productenhancements subsequent to the printing date of this TechBook. Useof the analysis model described in this TechBook should still apply,but the conclusions reached may change when enhancements areconsidered.

Because this TechBook covers a wide product breadth, it does notinclude detailed product usage examples. This TechBook is the first ina series of Data Migration TechBooks. The rest of the series willexplore specific EMC data migration solutions including bestpractices usage examples. For readers who have access to EMCSupport Online, check for new TechBooks in this Data Migrationseries:


Alternatively, you can visit the Vervante On Demand Publishingwebsite at:

http://www.vervante.com/

AuthorsThis TechBook was originally written by Donald Fried-Tanzer, anemployee of EMC based at headquarters in Hopkinton. Donald hasover 30 years of experience in client-server applications, fault-tolerantservers, and storage management software including: provisioning,remote replication, local replication, migration, configuration andmonitoring. Donald would also like to thank the EMC One TechBookCollaboration community and the many reviewers of this documentfor their contributions.


http://powerlink.emc.com:80/km/appmanager/km/secureDesktop?_nfpb=true&_pageLabel=homePgSecureContentBk

http://powerlink.emc.com:80/km/appmanager/km/secureDesktop?_nfpb=true&_pageLabel=--NULL--&internalId=0b01406680024e1b

http://powerlink.emc.com:80/km/appmanager/km/secureDesktop?_nfpb=true&_pageLabel=--NULL--&internalId=0b01406680024e1b

http://www.vervante.com/


14

Preface

Andrew Lubeck is the current owner of this TechBook. He is aConsulting Corporate Systems Engineer with SymmetrixEngineering and has been with EMC for 14 years. Though he hasworked with most of the core Symmetrix software products andEnginuity features over this time, he specializes in migrationtechnologies, Federated Tiered Storage, and Virtual Provisioning.

Relateddocumentation

Related documents, available at http://support.EMC.com, include:

• EMC Solutions Enabler Installation Guide

• EMC Solutions Enabler Symmetrix Array Management CLIProduct Guide

• EMC Solutions Enabler Symmetrix Array Controls CLI ProductGuide

• EMC Solutions Enabler Symmetrix SRM CLI Product Guide

• EMC Solutions Enabler Symmetrix TimeFinder Family CLIProduct Guide

• EMC Solutions Enabler Symmetrix SRDF Family CLI ProductGuide

• EMC Solutions Enabler Symmetrix Open Replicator CLI ProductGuide

• EMC Symmetrix Management Console Version 7.x Release Notes

• EMC Solutions Enabler Version 7.x Release Notes

• EMC Symmetrix Enginuity Release Notes (multiple release levelsavailable)

• SRDF/Extended Distance Protection for Open Systems FeatureSheet

• EMC PowerPath Migration Enabler User Guide

• White Paper: EMC Symmetrix Optimizer Virtual LUN Technology– A Detailed Review

• Open Migrator/LM for UNIX and Linux CLI Product Guide

• Open Migrator/LM for Windows Product Guide

• White Paper: Data Migration Considerations: A CustomerEngineering Residency – Best Practices Planning

• White Paper: New Features in EMC Enginuity 5875 for SymmetrixOpen Systems Environments

• White Paper: EMC Symmetrix VMAX Virtual Provisioning SpaceReclamation and Application Considerations



Preface

• Best Practices for Nondisruptive Tiering via EMC SymmetrixVirtual LUN Technical Notes

• FAST Theory and Best Practices for Planning and PerformanceTechnical Notes

• Implementing Fully Automated Storage Tiering (FAST) for EMCSymmetrix VMAX Series Arrays Technical Notes

• Implementing Fully Automated Storage Tiering with Virtual Pools(FAST VP) for EMC Symmetrix VMAX Series Arrays TechnicalNotes

• Federated Live Migration Technical Overview Technical Notes

Typographicalconventions

EMC uses the following type style conventions in this document:

IMPORTANT

An important notice contains information essential to software orhardware operation.

Note: A note presents information that is important, but not hazard-related.

Typographical conventionsEMC uses the following type style conventions in this document.

Normal Used in running (nonprocedural) text for:• Names of interface elements, such as names of windows, dialog

boxes, buttons, fields, and menus• Names of resources, attributes, pools, Boolean expressions, buttons,

DQL statements, keywords, clauses, environment variables,functions, and utilities

• URLs, pathnames, filenames, directory names, computer names,links, groups, service keys, file systems, and notifications

Bold Used in running (nonprocedural) text for names of commands,daemons, options, programs, processes, services, applications, utilities,kernels, notifications, system calls, and man pages

Used in procedures for:• Names of interface elements, such as names of windows, dialog

boxes, buttons, fields, and menus• What the user specifically selects, clicks, presses, or types

Italic Used in all text (including procedures) for:• Full titles of publications referenced in text• Emphasis, for example, a new term• Variables


16

Preface

Where to get help EMC support, product, and licensing information can be obtained asfollows.

EMC support, product, and licensing information can be obtained onthe EMC Online Support site as described next.

Note: To open a service request through the EMC Online Support site, youmust have a valid support agreement. Contact your EMC sales representativefor details about obtaining a valid support agreement or to answer anyquestions about your account.

Product informationFor documentation, release notes, software updates, or forinformation about EMC products, licensing, and service, go to theEMC Online Support site (registration required) at:

https://support.EMC.com

Technical supportEMC offers a variety of support options.

Support by Product — EMC offers consolidated, product-specificinformation on the Web at:

https://support.EMC.com/products

The Support by Product web pages offer quick links toDocumentation, White Papers, Advisories (such as frequently usedKnowledgebase articles), and Downloads, as well as more dynamic

Courier Used for:• System output, such as an error message or script• URLs, complete paths, filenames, prompts, and syntax when shown

outside of running text

Courier bold Used for specific user input, such as commands

Courier italic Used in procedures for:• Variables on the command line• User input variables

< > Angle brackets enclose parameter or variable values supplied by theuser

[ ] Square brackets enclose optional values

| Vertical bar indicates alternate selections — the bar means “or”

{ } Braces enclose content that the user must specify, such as x or y or z

... Ellipses indicate nonessential information omitted from the example


https://support.EMC.com


https://support.EMC.com/products

Preface

content, such as presentations, discussion, relevant CustomerSupport Forum entries, and a link to EMC Live Chat.

EMC Live Chat — Open a Chat or instant message session with anEMC Support Engineer.

eLicensing supportTo activate your entitlements and obtain your Symmetrix license files,visit the Service Center on https://support.EMC.com, as directed onyour License Authorization Code (LAC) letter e-mailed to you.

For help with missing or incorrect entitlements after activation (thatis, expected functionality remains unavailable because it is notlicensed), contact your EMC Account Representative or AuthorizedReseller.

For help with any errors applying license files through SolutionsEnabler, contact the EMC Customer Support Center.

If you are missing a LAC letter, or require further instructions onactivating your licenses through the Online Support site, contactEMC's worldwide Licensing team at [email protected] or call:

◆ North America, Latin America, APJK, Australia, New Zealand:SVC4EMC (800-782-4362) and follow the voice prompts.

◆ EMEA: +353 (0) 21 4879862 and follow the voice prompts.

We'd like to hear from you!Your suggestions will help us continue to improve the accuracy,organization, and overall quality of the user publications. Send youropinions of this document to:

[email protected]

Your feedback on our TechBooks is important to us! We want ourbooks to be as helpful and relevant as possible. Send us yourcomments, opinions, and thoughts on this or any other TechBook to:

[email protected]




18

Preface


1

This chapter discusses the complexity of choosing the best solutionfor a particular migration and all of the factors and features that areexplored in more detail in subsequent chapters. The topics are:

◆ Introduction ........................................................................................ 20◆ Simple selection model...................................................................... 22◆ Data migration.................................................................................... 24◆ Full selection model summary......................................................... 26

Introduction

Introduction 19

20

Introduction

IntroductionAlthough a single data migration may have a simple effectivesolution, the number of potential solution methodologies and themyriad factors influencing the selection of a solution make choosingthe best data migration solution very complex. Customerenvironments are so diverse that migration projects are rarely thesame. The process to migrate data is complex because it requiresdetailed planning, multiple steps, and often many different tools,resources, and approaches. A migration method that works well forone host application within a customer environment may beinapplicable for another host application, even within the samestorage array; therefore, larger migrations often include multiplesolutions.

Every data migration challenge has alternate solutions, so there isalways a choice to be made. In practice, data migration efforts do notexist in a vacuum and there are often additional issues that affect thedecision beyond the actual data movement itself. Sometimes thesemultiple issues align closely, but at other times they compete and therelative importance of one issue over another may strongly bias theselection of the best solution. In selecting the final solution, secondaryfeatures sometimes become more central than primary features.

Figure 1 on page 21 summarizes a basic migration flow.


Introduction

Figure 1 Migration project steps example

The single most important feature of a migration solution is that datamust be moved from point A to point B. This means that whenlooking for potential solutions, any product that can move the datamay be a good solution. Therefore, products that are primarily DataReplication or Backup/Restore products could be considered, alongwith products designed specifically for Data Migration.

This TechBook will limit the scope of products by focusing only onmigrations to EMC® Symmetrix® VMAX™ and late model EMCSymmetrix DMX™ arrays. Some mention may be made of moregeneric migration solutions that do not require a Symmetrix VMAXor Symmetrix DMX be present for migration. One of the use caseshighlights a host-based solution agnostic to the type of storage arraysin use.

ICO-IMG-000440

Business Impact Analysis

Source ConfigurationDiscovery

Migration Strategy and Tool Selection

Detailed Mapping and Design

Provisioning

Skill Refreshing Migration Pilot

Migration and Cutover

Sign-off

Introduction 21

22

Introduction

Simple selection modelBefore overwhelming the reader with the myriad choices available, itis important to state that choosing a data migration solution can belooked at less elaborately. The simple selection model may evenapply in 80 percent of the data migrations. The simple model makesan assumption that, for performance reasons, it is best to have themigration of data performed outside the application host. The simplemodel proposes a single or just a few solutions. For example, onepossible simple model would go like this:

A. If the migration is for changing the tier or data protection type ofdata within a Symmetrix VMAX array, use Virtual LUNtechnology (this can also be used for Symmetrix DMX arrays,except when also changing the protection type where EMCPowerPath® Migration Enabler with EMC TimeFinder®/Clonecan be used).

Note: Beginning with EMC Enginuity™ 5874 service release andSolutions Enabler 7.1 Fully Automated Storage Tiering (FAST) can beused to continuously analyze the utilization (busy rate) of Symmetrixarray devices, and automatically move data volumes between tiers tofine tune performance and reduce costs.

B. If the migration is a consolidation or technology refresh of entirestorage arrays that need to be transparent to the host application,use Federated Live Migration (FLM) for pre-qualified stacks.

C. If the migration is from Symmetrix-to-Symmetrix storage arrays,use EMC SRDF®, or Federated Tiered Storage (FTS), a new featureof Enginuity 5876.

D. If the migration involves heterogeneous storage arrays (includingsame vendor but not compatible for B), then use EMC OpenReplicator for Symmetrix or use Federated Tiered Storage FTS tomanage, monitor, migrate, and replicate data residing onSymmetrix VMAX and other supported arrays using familiarEMC software and Enginuity features.

E. If the previous choices cannot be used, then use a host-based tool,for example, EMC PowerPath Migration Enabler with Host Copy,or EMC Open Migrator/LM.


Introduction

An even simpler model can be presented, but the simpler the model,the less optimized the solution may be for the data migration. ThisTechBook is designed to provide enough information to lead tofinding the most optimal solution in all cases, which requirespresenting all of the options, thus the complexity. However, for mostmigrations choosing the strategy can be made simply.

Another aspect of simple migration strategies is that they may ignorethe complexities of ensuring that a data migration does not adverselyaffect production work, and verifying that the migrated data iscomplete and ready to be used with no problems.

The actual steps to ensure these important aspects of a migration isbeyond the scope of this TechBook, but the ability of potential datamigration products to effectively complete these steps is part of thereason that one or more of them will be often be part of the best datamigration solution.

Table 1 on page 132 summarizes when to use the five solutionshighlighted in the simple model and Open Migrator/LM as ahost-based solution.

Simple selection model 23

24

Introduction

Data migrationThis section includes the following information:

◆ “Data migration defined” on page 24

◆ “Virtualization” on page 24

◆ “Device external identity” on page 25

Data migration definedOne definition for data migration is: the one-time movement of datafrom source to target, where the data will subsequently only beaccessed at the target.

The key to this definition is that for this particular piece of data, thisis a one-time movement. This single time movement differentiatesdata migration from replication where the source data is still accessedafter the target copy is created. The one-time movement alsodifferentiates from data mobility where incremental updates to thedata would continue to be applied.

Using this definition, migration relocates the original data. Theapplication(s) which accesses the data must point to the data in itsnew location. Part of the migration solution is the methodology usedto point the application to the new data location. The key question iswhether an application outage is necessary for the application toaccess the data in the new location. Few, if any, applications havebeen designed with the ability to continue processing withoutinterruption while changing the location of the data the application isaccessing. Therefore, either there will be an application outage or themigration must be transparent to the application.

VirtualizationMigration transparency can be achieved through virtualization. Inthis case, a broad definition of virtualization is being used: atechnique for hiding the physical characteristics of computingresources from the way in which another function interacts withthose resources. In this case, virtualization can be effective by hidingthe physical characteristics of a resource in the I/O stack from theabove layer. When using virtualization, the key question is whetherthe I/O stack as it currently exists already has an element which does


Introduction

not point directly to the actual resource beneath it. Using thisavailable indirection, the I/O from the application can transparentlybe redirected to the target in place of the original source, as long asthe migration can occur at or below the level where the indirectionexists. Sometimes, a migration strategy will include a step to add inthe indirection both to make it possible to migrate transparently inthe future and to minimize the application outage needed for thecurrent migration.

Device external identityIn addition to virtualization, there is another way to achievemigration transparency by hiding the physical characteristics of aresource in the I/O stack from the layer above it. This method relieson presenting the migration target device’s external identity toappear identical to the source device identity as seen by theapplication host. This method requires intelligence in the I/O stack toboth present the migrated device identity as well as manage theapplication I/O paths to the source and target devices ensuring dataintegrity. EMC Federated Live Migration (FLM) is an example of aproduct using this methodology.

Data migration 25

26

Introduction

Full selection model summaryThis TechBook develops a model for selecting the best data migrationsolution by embracing the complexity fully. This model issummarized in the following steps:

1. Define the reasons for the data migration, clearly notingmandatory and optional objectives.

2. Inventory the existing environment, identifying storage elementsthat must participate with the chosen data migration solution andresources available to support the migration itself.

3. Inventory the target environment identifying storage elementsthat must participate with the chosen data migration solution andthe resources available to support the migration itself.

This step should include scalability considerations to ensure thetarget environment is not obsolete too soon. Additionally, thesolution might include adding in semi-permanent infrastructureto support future data migrations.

4. Identify potential data migration solutions that can successfullymove the data from the existing environment to the targetenvironment.

5. Identify business factors that limit potential data migrationsolutions due to budget, human resources, and applicationoutage and verification requirements.

6. Compare and evaluate the potential data migration solutionsincluding the criteria identified in step 1 through step 5 .

The steps in the model are more complicated in real use because theboundaries of the steps are indistinct and are actually conducted bothin parallel and in mixed order. Small adjustments, such as theaddition of a resource in the target environment may enable alternatesolutions, making the model into an iterative process. Similarly, thechapters of this TechBook cannot fully make issues sharply distinctbecause there is an overlap of factors in more than one place.

◆ Chapter 2, “Reasons for Moving Data,” defines eight categories ofreasons for moving data that corresponds to model step 1 .

◆ Chapter 3, “Existing and Target Environment,” provides anexplanation of features to inventory in the existing and targetenvironments used in model step 2 through step 3 .


Introduction

◆ Chapter 4, “EMC Open Migrator/LM,” provides details abouteach potential migration solution to be used in model step 4through step 6 .

◆ Chapter 9, “Business Factors,” defines business factors thatwould be applied in model step 5 .

◆ Chapter 10, “Applying the Model,” provides examples ofapplying the model in typical customer scenarios.

Full selection model summary 27

28

Introduction


2

This chapter defines eight categories of reasons for moving data thatwill primarily drive the requirements for potential migrationsolutions. The topics are:

◆ Introduction ........................................................................................ 30◆ Data growth ........................................................................................ 30◆ Technical refresh................................................................................. 30◆ Performance improvements ............................................................. 31◆ Information Lifecycle Management (ILM) ..................................... 33◆ Risk reduction..................................................................................... 33◆ Consolidation, localization, and regulatory compliance.............. 34◆ Complexity and transformational migrations ............................... 34

Reasons for MovingData

Reasons for Moving Data 29

30

Reasons for Moving Data

IntroductionThe first step of the selection model is to define the reasons for thedata migration, clearly noting mandatory and optional objectives.There are many different reasons an organization has for moving datafrom one location to another. The reasons behind a particularmigration will likely affect the choice of the data migration solution.

This chapter defines eight categories of reasons for moving data: datagrowth, technical refresh, performance improvements, InformationLifecycle Management (ILM), risk reduction, consolidation,localization (distribution), and regulatory compliance requirements.

Data growthThe simple growth of data often will lead to a data migration need.When data grows beyond its current storage location, it must beexpanded. This expansion may be accommodated with excesscapacity in the current storage array, by adding additional capacity tothe current array or by using storage in additional arrays. Oncemaximum capacity is reached in an array, it will be necessary tomigrate data to another array.

Technical refreshThe need to migrate due to aging technology is commonly referred toas a technical refresh. The aging of technology in itself may force amigration that otherwise may not be necessary. The most commoncause would be a hardware lease ending, which cannot be renewedor would not be a cost-effective renewal. As products get older, itmay no longer be possible to get upgraded parts or software neededto try to scale the product to current needs. Maintenance costs mayescalate or the manufacturer may no longer support the product evenat an escalated price.



Performance improvementsWhen an application is in need of performance improvements inorder to achieve reduced transaction times or reducing the totalwindow for completing an operation, improvements in the datastorage I/O path may be key to realizing a performance goal.Improvements in the I/O path performance are achieved by one oftwo factors: either performance is improved by spreading I/O moreevenly across existing I/O resources, or additional I/O performanceis added to the mix by actually adding more resources or replacingexisting resources with better performing resources.

Spreading I/O more evenly across existing I/O resources can resultin improvements in both throughput and transaction time. Unevenlyspread I/O means that some resources are busier than others, whichmeans that more I/O is queued up for one resource than another.Since total transaction time includes the wait time for a resource,longer queues result in longer transaction times. Additionally,balancing I/O evenly across resources will minimize the time thatany resource is idle because its queue is empty while others are busy,thus increasing the overall throughput by effectively utilizing allresources. Spreading I/O more evenly can consist of spreading I/Oamongst multiple storage arrays or the way in which I/O is spreadwithin a single array.

Within a single array, the relationship between storage data migrationand balancing I/O resources results from the fact that disk resourcesthat appear to the host to be physical disk spindles are actually logicalemulations in a Cached Disk Array. These logical emulationsultimately reside on physical disk spindles within the Cached DiskArray and therefore can result in less than optimal performance ifplacement results in unbalanced I/O queuing. Migrating the locationof logical disks can alter performance characteristics. Balancinglogical disk placement can be done manually through migration orautomated with a product like EMC FAST, or EMC SymmetrixOptimizer. Additionally, a front-end logical disk can actually bespread over multiple physical disks when using striped meta devices,RAID-5 or RAID-6 striping, and Virtual Provisioning (also called thinprovisioning).

It is important to remember that performance is always a relativemeasure. The key question when tuning for performance is whetherperformance is satisfactory to the end user. If the answer is yes, thenthe current performance measure can be recorded. If the answer is no,

Performance improvements 31

32


or as performance needs change, performance must improve to meetthe new level of satisfaction. Because performance is relative, overallthroughput is not necessarily the key measure. There will often be thedesire to bias performance to improve one application at the expenseof another. This can be acceptable because the performance of theapplication that gives up resources to the prioritized application stillachieves a satisfactory level of performance.

Given a desire for unequal levels of performance, the placement oflogical disk volumes may be made to give high levels of performanceto selected applications and lower, but still acceptable, levels ofperformance to other applications. In this case, the objective is notbalanced I/O and overall throughput, so lower priority applicationsmay utilize resources that have longer queues.

The desire for unequal storage performance characteristics is far moretunable than balancing the I/O requests for differing queue depthson equally performing storage resources. Within some modernstorage arrays, there exists the concept of tiered storage. For example,the physical disk devices within the array may have widely varyingdisk performance characteristics from the highest level flash memoryto a low-level large capacity ATA disk drive. Migrating existing datafrom one tier to another can dramatically change performancecharacteristics.

Tiering of data storage performance can also be obtained withmultiple storage arrays each with different performancecharacteristics. In this case, migration would need to be from onestorage array to another, rather than between disks within a singlearray.



Information Lifecycle Management (ILM)The idea that all information is not created equal and should use themost appropriate storage is a key element of Information LifecycleManagement (ILM). The word Lifecycle is important, because thevalue of information increases or decreases over time, leading to theneed to migrate data after its original prioritization and placement.Based on its changing value, information requires different levels ofaccessibility and protection.

The goal of ILM is to continuously move data to the bestcost/resource use. Lowering TCO (Total Cost of Ownership) can beachieved by optimizing storage to only use the tier of servicenecessary to achieve desired levels of accessibility and performance.Accessibility is a performance related measure but ILM may alsochoose data placement by other criteria including the data protectionavailable in a particular tier.

Risk reductionRisk reduction relates to lowering the likelihood of DataUnavailability (DU) and Data Loss (DL) conditions. Migrations to amore highly available storage platform can effectively mitigate theDU risk. Migrations to storage platforms where data replicationstrategies will be implemented can effectively mitigate the DL risk.

The key factors for how data migration can reduce risk is theimproved features the new platform offers in the areas of higheravailability and more options for data migration. And as introducedin the ILM discussion, there is a trade-off decision that has to bemade, where the best data replication functionality may not bechosen due to cost. In this vein, DL risk reduction can also beachieved through data backup. Beyond traditional tape backup,backup-to-disk creates opportunities where data migration plays arole.

Information Lifecycle Management (ILM) 33

34


Consolidation, localization, and regulatory complianceBesides the organic growth of data, changes in the businessorganization will drive data migrations. Consolidation of storagemay result from a merger, centralization, a data center migration, orheterogeneous storage platform migration. Localization needs fordistributed operations may reverse centralization of data. Regulatorycompliance requirements may mandate fixed period of accessibilityand searchability on archived data. The drivers for these changes arenot directly technical but business factors which lead to datamigrations.

Complexity and transformational migrationsMost migrations will include more than a single reason driving themigration. This mix of reasons will influence the choice of solutionand may lead to using the data migration as an opportunity tooptimize infrastructure, simplify management, reduce risk, lowercost and effectively transform the entire Information Technology (IT)infrastructure. This concept will be discussed more in Chapter 9,“Business Factors,” and Chapter 10, “Applying the Model.”


3

This chapter identifies the information that must be collected to assaythe existing application and storage infrastructure and the desiredtarget environment at the end of the migration. The topics are:

◆ Introduction ........................................................................................ 36◆ Initial vs. complete collection of environment information......... 38◆ Applications........................................................................................ 39◆ Hosts (servers) .................................................................................... 40◆ Storage network ................................................................................. 42◆ Storage array information................................................................. 43◆ Environmental factors: power, cooling, and floor space .............. 45◆ Software attributes ............................................................................. 46

Existing and TargetEnvironment

Existing and Target Environment 35

36

Existing and Target Environment

Introduction“Full selection model summary” on page 26 lists steps 2 and 3 asfollows:

2. Inventory the existing environment, identifying storageelements that must participate with the chosen data migrationsolution and resources available to support the migration itself.

3. Inventory the target environment identifying storage elementsthat must participate with the chosen data migration solution andthe resources available to support the migration itself.

These steps inventory the existing and target environments. Thedecisive criteria for selecting the best data migration solution may bethe result of factors in the existing or target environments. Thischapter describes the information needed to determine potentialmigration strategies.

◆ Physical infrastructure that needs to be described includes: hosts,storage networks, and storage platforms.

◆ Software infrastructure includes the applications andfunctionality available for each physical resource, with attentionto existing release versions and licenses.

In addition to identifying infrastructure, it is crucial to determinecapacity levels.

◆ Excess capacity could mean that data migration using thatparticular resource could be executed without adverse effect onexisting applications.

◆ Constrained capacity might preclude or limit data migrationusing that resource, and might also be the reason that themigration is necessary to resolve the resource constraint.

Absolute performance of a potential data migration strategy mayoften be overshadowed by the excess and constrained capacityrelevant to a particular site. An essential element of capacity isutilization rates. Sometimes the problem may not be overall capacity,but the combination of allocations and utilization rates which resultin excess allocated capacity, not free for other uses.

Information needs to be gathered on both the existing and targetenvironments. Depending on the reasons for the data migration andbusiness decisions relative to it, the relationship between theseexisting and target environments will vary. One scenario has data



migrating within an infrastructure where there is no change in overallphysical resources. On the other end of the spectrum the targetenvironment is a complete replacement of the existing environmentwith no shared components.

Most migrations fall somewhere in between these first two scenarioswhere there are some changed resources in the target infrastructure;since the amount of data is always growing, that usually meansadditional resources. These additional resources in the targetenvironment are often crucial resources that are not just thedestination of the data migration but also may directly support themigration itself.

Introduction 37

38


Initial vs. complete collection of environment informationIt is important to stress that this collection of environmentinformation occurs more than once in an actual data migration flow.It occurs early in the process as input to choosing the migrationsolution, which is the major reason for the inclusion of this chapter inthis TechBook. However, it also occurs again in the detailed mappingand design phase where much more detailed data is needed. It is inthis more detailed phase that interoperability concerns at every levelof the infrastructure must be accounted for. An initial check forinteroperability is necessary to ensure that a chosen solution is indeedviable, but this TechBook does not address the essential, moredetailed later stage collection of environment information.

Figure 2 represents the I/O stack and all of the levels in whichinformation may need to be collected.

Figure 2 I/O stack levels

Intero

perab

ility in th

e I/O stack

Applications

File systems

Volume managers

Volume groups

Logical volumes

Multi-pathing

Host bus adapters

Storage arrays

SAN

Disk driver



ApplicationsFirst, it is necessary to collect information about the applications thatare to be migrated. Since migration relocates the original data, it isnecessary to determine exactly where the original data resides.

Storage resource management mappingFigure 3 depicts the application at the top level as being an RDBMSdata file. The RDBMS data file can reside directly on a raw hostphysical device, or can be layered first on either a host File system, ahost Logical Volume Manager, or both. This mapping view can begenerated by Solutions Enabler Storage Resource Management(SRM) commands: symrslv, symrdb, symhostfs, symlv, and symvg.EMC ControlCenter can also depict these relationships graphically aswell as providing additional information through StorageScope™ andStorageScope File Level Reporter.

Figure 3 Storage resource management layers

Storage device

Extents

Logical volume

Host physical device

RDBMS data file

File system

ICO-IMG-000445

Applications 39

40


Application storage redirectionIt is also important to determine what mechanisms can be used to getthe application(s) to point to the data in its new location after themigration. The horizontal line in Figure 3 on page 39 separates thehost components above the line from the storage components belowthe line. As depicted, there is a one-to-one relationship between thehost physical device and the storage device. With the exception ofDirect Attached Storage (DAS), the storage device itself is a logicalconstruct made up of extents. Any of the logical representations inthis mapping could potentially be used to make the migrationtransparent to the application. Of course, the concept of datamigration means that at least one component will be changed in asimilar mapping depiction after the migration is completed. Thatchange can occur at one or more levels in this picture including theFile system or LVM of the host, or both.

Hosts (servers)The hosts that execute the applications were identified in theapplication description, as well as some of the host softwarecomponents including Relational Database Management Systems(DBMS), File Systems, Logical Volume Managers (LVM), andMultipath I/O (MPIO) drivers. Other host information that must becollected includes: I/O bandwidth, CPU capacity, HBA, and OSversions. HBA and OS version information is necessary to ensurecompatibility with the target environment and the data migrationsolution.

Host I/O capacityIf the data migration solution uses the host I/O path, there must besufficient excess capacity, unused I/O bandwidth, in this path tosupport migration. I/O bandwidth needs to be measured not just atthe total bandwidth level, but also for each port evaluating utilizationlevels, depending on how much control there may be of theapplication or migration data paths or both. On hosts that have amultipathing disk driver, like PowerPath, the driver mayautomatically balance I/Os across multiple paths for equal utilizationand best performance.



Host CPU capacity and migration application controlIf using a host-based data migration tool, there must be sufficientexcess CPU capacity to run it. The ability to limit the data migrationtool to only use both the I/O and CPU capacity available is veryimportant to avoid too great an impact on application performancewhen conducting a host-based migration.

Host HBA connectivityTypically hosts may be fully utilizing the full number of configurableHBAs. In some cases, there may be the ability to add additional HBAsproviding additional ports for connectivity to the data in the targetlocation and bandwidth to be used for the data migration itself.However, unless the connectivity is going to be used after the datamigration is completed, the investment in the hardware may not fitthe business factors around the migration.

Host migration appliancesCertain data migration solutions may require an additional host orrepurpose a host to act as a dedicated data migration applicationappliance.

Hosts (servers) 41

42


Storage networkInformation needed about the storage network includes: directconnect, switched, single or multiple networks, network bridgingtechnology, I/O bandwidth, switch port capacity, and switchsoftware capability.

Storage network typesChanging directly attached storage (DAS) to networked storage(SAN, NAS, or both) may be one of the reasons driving the datamigration. The existence and type of storage network is a factor inenabling certain data migration solutions. Many environments havemore than one type of storage network including: Fibre Channel, IP,FICON, and ESCON. Within each type of network there may be asingle interconnected network or distinct networks; best practiceincludes duplicate paths through each network for fault tolerance.Many environments will include network bridging technologieswhich can connect otherwise distinct networks.

Storage network I/O and port capacityAs with hosts, it is necessary to gather information about the storagenetwork I/O bandwidth, including capacity and utilization as well asthe storage network port capacity.

Storage Network Port Capacity relates directly to connectivity; forexample, how many hosts and storage arrays can be connected to theswitches (and possibly hubs) making up the storage network. Someports may also need to be used for Inter-Switch Links (ISLs)connecting switches to each other. There needs to be enoughconnections available to connect the source and target storage arrayssimultaneously as well as ports used for the migration itself, ifnonshared ports are required.

Depending on the type of migration solution, it may not be necessaryto have full connectivity to both the source and target at all timeswhile implementing the data migration solution, but in other casesadditional capacity may be needed only for the time of the migrationitself. Given data growth patterns, including the addition of newservers for new and growing applications, it is likely that any extracapacity needed for the migration is likely to be needed in the nearfuture.



Migration in the storage networkThere are data migration solutions (for example, RecoverPoint)which run in the storage network itself, and require specific hardwareswitches to support these software applications. If these switches arealready present in the environment, then capacities needed for themigration solution must be considered. If the switches will be new tothe network, there must be sufficient capacity for ISL connectionswith existing switches.

EMC VPLEX storage federation provides a virtualization layerbetween the server and storage that enables movement of databetween the underlying storage elements, completely transparent tothe host environment. This virtualization layer enables seamlessmigration between and within storage systems and allows storageadministrators to leverage new storage capabilities that providegreater levels of efficiency, resiliency, and lower cost, withoutcompromising service levels to end users. It is necessary to collectinformation related to VPLEX limits for the number, size, andcomposition of virtual devices to ensure the ability to support boththe migration and target environment.

Storage array informationStorage platform infrastructure information needed includes: datacapacity and utilization, I/O bandwidth, director port capacity, tiersof storage, homogeneous or heterogeneous storage elements, and thenumber and versions of each storage element. These categories ofinformation are all listed for completeness, but may not all berelevant for every site, and are not distinct and overlap with eachother.

Storage array data capacity and utilizationStorage data capacity is more than a single number. One basicnumber is the total amount of data stored versus the existingcapacity. Site-specific trends for data growth imply the need forenough target capacity to meet the business goals to contain all of thedata at satisfactory performance levels. The amount of existing excessor constrained capacity will determine how much capacity has to beadded, or in some cases reduced, to meet target environment goals.

Storage array information 43

44


Utilization and allocation play important roles in measuring capacity.Storage that is unallocated is free for new uses; however, storage thatis allocated but not used is limited to be used within its existingallocation until it is reallocated. Some data migration solutions mayinclude an intermediate copy of the data requiring temporarycapacity needs during the migration itself. This temporary capacitymay be available as part of the data growth allowance in the targetenvironment, as part of the soon to be end-of-lifed capacity from theexisting environment, or may actually require additional temporarycapacity for the migration. Measuring overall storage capacity mightbe less meaningful than determining these measures in each differenttier of storage (more on tiered storage below).

When looking at ways to meet target data storage capacity, theexpandable capacity limits within the existing environment must bereviewed. For example, within an existing storage array, it may bepossible to add additional disk drives, or replace/upgrade existingdrives to higher capacity drives. These additions and upgrades maybe possible within the existing cabinets or may require additionalcabinets or components in order to add the new drives. If existingarray capacity cannot be expanded, additional or replacement storagearrays will need to be configured as part of the target environment.

Storage array I/O bandwidthStorage I/O bandwidth constrained capacity may be a driving reasonfor the data migration. Data migration solutions that use the storageI/O data path require bandwidth for the migration itself; however,there must be enough excess bandwidth to make this migrationpossible. Like data capacity, I/O bandwidth needs to be measured,not just at the total bandwidth level, but also for each port evaluatingutilization levels.

Director port capacityDirector port capacity relates directly to connectivity; for example,how many hosts can be connected to a storage array port. Thisnumber can be constrained by both actual physical connections andsoftware limitations in a networked storage environment. Datamigration solutions that use the storage I/O data path may requireexclusive or shared use of director ports for the migration itself; theremust be enough excess port configuration capacity to make thismigration possible.



Tiered storageTiered storage may be available from different disk types within asingle storage array or from different storage arrays. Data capacity,I/O bandwidth, and director port capacity may need to be measuredon a per-tier basis in both the source and target environments. Theremust be enough excess capacity in the appropriate tier to meet theperformance objectives of the chosen data migration solution.

Storage elements homogeneityIn a small existing infrastructure, the storage array elements may behomogeneous; in larger environments they are usuallyheterogeneous. At the highest level, this may involve a difference ofvendors in storage arrays, or even different types of storage arrayswithin single vendors, or different elements within each storagearray. Certain data migration solutions only work in a homogeneousenvironment or may not support the full diversity of storageelements in the existing or target environments.Additionally, the version of each storage element is important as tothe features supported in current levels of software or the ability toupgrade to more advanced levels that support added features thatprovide or enable additional data migration solutions.It is important to note that the need to upgrade component softwarelevels adds complexity to the data migration. Depending on theelement, there may not be a nondisruptive upgrade methodology,requiring the upgrade to occur during a scheduled maintenancewindow. The necessity to avoid unplanned application interruptionsoften will extend the time it takes to complete the entire migration.

Environmental factors: power, cooling, and floor spaceAn increasingly important issue for all hardware assets is power,cooling, and floor space. According to research reports, many largedata centers are approaching absolute limits in powering, cooling,and fitting their hardware into their data center. The ability of thetarget environment to provide more capacity or performance or bothwhile using the same or even less power, cooling, and floor spacemay be a driving factor for the data migration.

Environmental factors: power, cooling, and floor space 45

46


Software attributesIn addition to collecting information about the hardware attributes ofthe existing and target environments, information must also becollected about the software architecture. The current levels ofsoftware and existing licenses will define the options available forpotential data migration solutions including critical factors, like thesupported number of simultaneous migrations.

Support for additional software products in the existing environmentor in the target environment, or even temporary support in themigration environment can expand the number of potential datamigration solutions. Business factors will likely limit the choice of adata migration solution, but it is first necessary to fully understandthe existing and target environments in order to identify all thepotential data migration solutions.


4

This chapter details features of the Open Migrator product. Thetopics are:

◆ Introduction ........................................................................................ 48◆ Key features ........................................................................................ 51◆ Migrate anywhere in the I/O stack ................................................. 53◆ Open Migrator considerations ......................................................... 54

EMC Open Migrator/LM

EMC Open Migrator/LM 47

48


IntroductionStep 4 in “Full selection model summary” on page 26, whichidentifies potential data migration solutions needs, and step 6,comparing and evaluating potential migration solutions, requiremore detail than is provided in “Open Migrator/LM” on page 40.Additionally, as a host-based migration solution, the detail providedfor Open Migrator/LM speaks to the detail appropriate to manyother host-based migration solutions.

Open Migrator/LM enables online data migration of MicrosoftWindows, UNIX, or Linux volumes between any source and EMCstorage. Open Migrator/LM host-based software boosts theefficiency of the entire information infrastructure by automating andsimplifying data migration. Whether consolidating servers,upgrading storage, or tuning performance, volumes stay online andfully available to critical applications during migration and hostapplications continue operating at peak performance.

Figure 4 illustrates an EMC Open Migrator/LM data migration withthe copy I/O stream passing through the application host.

Figure 4 EMC Open Migrator/LM

Application



Open Migrator/LM provides an online data migration andcomparison solution for high availability data centers, where datacopying is performed from the production host. Open Migrator isimplemented as a host-based kernel driver and command lineinterface (CLI) UNIX and Linux, or graphical user interface (GUI) forWindows.

Open Migrator is used to migrate data from source to target volumeswith only a single disruption to the server or applications. Migrationbetween two storage arrays (or within a single array) is performedfrom the production host.

Figure 5 illustrates both UNIX CLI and WINDOWS GUI OpenMigrator examples.

Figure 5 EMC Open Migrator/LM command line and GUI examples

Open Migrator provides mirroring and background copy functionsthat are used to synchronize data images on one or more source andtarget volumes, LUNs, or LUN partitions. Data can be migratedbetween source and target volumes of any block device type. Duringmigration the source volume can remain available for input/output(I/O) to production host applications. The target volume is set toread/write disabled. The target volume should also be set as notready to any additional hosts having access to the volume.

Introduction 49

50


Open Migrator operates in sessions to manage multiple volume pairsuniformly as a group. Control operations are performed using thestormigrate CLI command (UNIX and Linux) or the GUI (Windows).The Windows interface is a fully compliant MMC Snap-in.

Data can also be compared between source and target volumes in anactivated session. When comparing, Open Migrator checks if thesource and target volumes are identical. Once the data has beenmigrated, mirroring continues to keep the source and target volumessynchronized, until the session is deactivated. When volumes havebeen successfully migrated and the session has been deactivated,Open Migrator can be uninstalled from the kernel I/O subsystem andthe production host.



Key featuresKey features of Open Migrator include:

◆ Broad open systems support: Windows 2000, 2003, and 2008 andmajor UNIX platforms.

◆ Supports smaller volume to larger volume migrations:

• UNIX environments will not automatically expand the filesystem

• In Windows environments the file system is expandedfollowing a reboot

◆ Data consistency and resiliency:

• Provides mirroring and data copying between source andtarget devices

• Mirrored writes before commit ensures consistent data in bothstorage systems

• Data on source volume protected during and after datatransfer

• Data migrations persist across system reboots

◆ Performance / cost optimization:

• Migrate to different drive type or RAID level• Optimize volume layout on target• Ability to pause and resume migration operation• Ability to tune migration rate• Uniform migration method across multiple platforms

◆ Migrate UNIX / Linux hosts with only a single disruption forcutover:

• Driver is dynamically loaded and unloaded• Outage required to point applications at new devices• Windows host may require additional reboot if filter driver

cannot attach to source and target devices

◆ Minimal dependencies.

• Dependent only on operating system support for install.

Key features 51

52


Figure 6 illustrates the Open Migrator operational flow.

Figure 6 Open Migrator operation

Application

Source Volume

Data RAID 1

Application

Target Volume

Data RAID 5

Application

Data RAID 1

Data RAID 5

OM/LM Filter Driver

Source Volume Target Volume

Remove old storage and restart app

Config targets

and install OM/LM

OM Driver Copy



Migrate anywhere in the I/O stackOpen Migrator provides across volume group data migrationfunctionality. This can be used during host reconfiguration wherehost I/O stack components, such as volume managers andmultipathing can be expanded, re-striped, re-sized or replaced.Examples of data migration from any source to any target blockdevice type in the host I/O stack include:

◆ Introduce a SAN device from direct attach device

◆ Migrate from a LUN device to a LVM device

◆ Migrate from a LVM device to a LVM device

◆ Reconfigure a LUN or LVM layer, re-stripe or resize targets

Note: Migration should always be done at the highest possible layer in theapplication's I/O stack. Conducting replication at a lower layer risksconfusing I/O stack elements like volume managers.

Figure 7 illustrates the I/O stack layers.

Figure 7 I/O stack layers

LUN and Partition Slices

Operating System Drivers

Logical Volumes

Mount Points

Application Data

Physical Layer

MultiPath I/O Layer

Logical Volume Manager Layer

File System Layer

Application Layer

Migrate anywhere in the I/O stack 53

54


Open Migrator considerationsKey considerations to include when planning to use Open Migrator:

◆ The source and target volumes remain synchronized until thesession has been deactivated.

◆ The appropriate I/O stack layer must be chosen.

◆ The target file system must be the same type as the source filesystem.

◆ Target logical volumes, partitions, file systems, and so on, mustbe provisioned prior to beginning the migration.

◆ Once the migration has completed, a maintenance window isrequired to redirect the applications to use the new targetvolumes.

◆ Both the source and the target must be available to the host.

◆ Does not allow the migration of boot disks or root file systems.

Note: A use case selecting Open Migrator as the migration method isdescribed in ”EMC Open Migrator/LM solution use case,” on page 156.


5

This chapter details features of the SRDF product with a focus on theattributes relevant to data migration. The topics are:

◆ Introduction .............................................................................................. 56◆ SRDF/Synchronous (SRDF/S)............................................................... 58◆ SRDF/Asynchronous (SRDF/A) ........................................................... 59◆ SRDF/Data Mobility (SRDF/DM) and adaptive copy modes.......... 60◆ SRDF/Automated Replication (SRDF/AR)......................................... 62◆ Concurrent SRDF ..................................................................................... 64◆ Cascaded SRDF ........................................................................................ 65◆ SRDF/Extended Distance Protection (SRDF/EDP)............................ 66◆ SRDF/Star ................................................................................................. 67◆ SRDF and data migration ....................................................................... 68

Symmetrix RemoteData Facility (SRDF)

Symmetrix Remote Data Facility (SRDF) 55

56

Symmetrix Remote Data Facility (SRDF)

IntroductionStep 4 in “Full selection model summary” on page 26, identifiespotential data migration solutions, but needs more detail on theSRDF family of products than is provided in “Symmetrix RemoteData Facility (SRDF)” on page 30 and Chapter 5, “EMC SymmetrixStorage System.” Details of the different deployment options of SRDFand the ways in which they can be used for migration are neededboth for step 4 and for step 6 when the potential migration solutionsare compared and evaluated in this chapter.

Figure 8 illustrates the SRDF family base products and additionaloptions.

Figure 8 SRDF family base products and additional options

SRDF Family

SRDF/S

SRDF/Star

SRDF/CE

SRDF/AR

Synchronous forzero data exposure

SRDF/AAsynchronous for

extended distances

SRDF/DMEfficient Symmetrix-to-

Symmetrix data mobility

Multi-site replicationoption

Cluster Enableroption

Automated Replicationoption

SRDF/CGConsistency Groups

Cascaded SRDFand SRDF/EDP

Extended DistanceProtection

Concurrent SRDFConcurrent



The SRDF family consists of three base solutions:

◆ SRDF/Synchronous (SRDF/S) — High-performance,host-independent, real-time synchronous remote replication fromone Symmetrix to one or more Symmetrix systems.

◆ SRDF/Asynchronous (SRDF/A) — High-performance extendeddistance asynchronous replication using a Delta Set architecturefor optimal bandwidth utilization and minimal host performanceimpact.

◆ SRDF/Data Mobility (SRDF/DM) — Rapid transfer of data fromsource volumes to remote volumes anywhere in the world,permitting information to be shared and content to bedistributed, or information consolidated for parallel processingactivities.

There are a number of additional options and features that can beadded to the base solutions to solve specific service levelrequirements. These options include:

◆ SRDF/Automated Replication (SRDF/AR) solutions for meetingvery specific, remote replication service-level requirements.

◆ SRDF/Star for advanced multisite failover with continuousprotection.

◆ SRDF/Consistency Groups (SRDF/CG) for data consistency.

◆ SRDF/Cluster Enabler (SRDF/CE) for integration withhost-based clustering products such as Microsoft Cluster Server(MSCS) and VERITAS Cluster Server (VCS).

◆ Cascaded SRDF for a three-site, extended distance replicationusing a dual role device, and SRDF/Extended DistanceProtection (SRDF/EDP) using a diskless dual role device.

Introduction 57

58


SRDF/Synchronous (SRDF/S)SRDF/S is a business continuance solution that maintains a real-time(synchronous) copy of data at the logical volume level in Symmetrixsystems in the same or separate locations. The SRDF/S operation istransparent to the host operating system and host applications. Itdoes not require additional host software for duplicating data on theparticipating Symmetrix arrays.

Figure 9 illustrates that the host I/O to the Symmetrix is notacknowledged until the remote target Symmetrix has acknowledgedit. Therefore SRDF/S has some impact on the production applicationand is limited to a ~200KM distance.

Figure 9 SRDF/S

Source Target

Limited distance

1 2

34



SRDF/Asynchronous (SRDF/A)Beginning with Enginuity level 5670, SRDF/A (Figure 10) is anothermode of remote replication that allows customers to asynchronouslyreplicate data while maintaining a dependent write consistent copyof the data on the secondary (R2) device at all times with noperformance impact on the host. The dependent write consistentcopy of the data at the remote side is typically only seconds behindthe primary (R1) side. SRDF/A session data is transferred to thesecondary Symmetrix system in predefined timed cycles or delta sets,eliminating the redundancy of multiple same track changes beingtransferred over the link, potentially reducing the requiredbandwidth.

Figure 10 illustrates the SRDF/A I/O flow.

Figure 10 SRDF/A

SRDF/A provides a long-distance replication solution with minimalimpact on performance. This level of protection is intended forcustomers who require minimal host application impact whilemaintaining a dependent write consistent, restartable image of theirdata at the secondary site. In the event of a disaster at the R1 site or ifSRDF links are lost during data transfer, a partial delta set of data canbe discarded, preserving dependent write consistency on thesecondary site with a data loss of no more than two SRDF/A cycles.

Source Target

Unlimited distance

1

32 4

SRDF/Asynchronous (SRDF/A) 59

60


SRDF/Data Mobility (SRDF/DM) and adaptive copy modesThe SRDF/DM product offering permits operation in SRDF adaptivecopy mode only and is designed for data replication or migrationbetween two or more Symmetrix systems. SRDF/DM transfers datafrom primary volumes to secondary volumes permitting informationto be shared, content to be distributed, and access to be local toadditional processing environments. Adaptive copy mode enablesapplications using that volume to avoid propagation delays whiledata is transferred to the remote site. SRDF/DM supports allSymmetrix systems and all Enginuity levels that support SRDF, andcan be used for local or remote transfers.

Adaptive copy modes facilitate data sharing and migration. Thesemodes allow the primary and secondary volumes to be more thanone I/O out of synchronization. There are two adaptive copyingmodes: adaptive copy write-pending (AW) mode and adaptive copydisk (AD) mode. Both modes allow write tasks to accumulate on thelocal system before being sent to the remote system.

With adaptive copy write-pending mode, write tasks accumulate inglobal memory. A background process moves, or destages, thewrite-pending tasks to the primary volume and its correspondingsecondary volume on the other side of the SRDF link.

The advantage to this mode is that it is faster to read data from globalmemory than from disk, thus improving overall system performance.An additional advantage is that the unit of transfer across the SRDFlink is the updated blocks rather than an entire track, resulting inmore efficient use of SRDF link bandwidth.

The disadvantage is that global memory is temporarily consumed bythe data until it is transferred across the link. Consequently, adaptivecopy write pending mode should only be used where detailedinformation about the host write workload is fully understood.

Adaptive copy disk mode is similar to adaptive copy write-pendingmode, except that write tasks accumulate on the primary volumerather than in global memory. A background process destages thewrite tasks to the corresponding secondary volume. The advantagesand disadvantages of this mode are opposite from those of theadaptive copy write-pending mode; that is, while less global memoryis consumed it is typically slower to read data from disk than fromglobal memory, additionally, more bandwidth is used because the



unit of transfer is the entire track. In addition, because it is slower toread data from disk than global memory, device resynchronizationtime will increase.

Figure 11 illustrates the SRDF Adaptive Copy mode I/O flow.

Figure 11 SRDF Adaptive Copy mode

Source Target

Unlimited distance

1

32 4

SRDF/Data Mobility (SRDF/DM) and adaptive copy modes 61

62


SRDF/Automated Replication (SRDF/AR)SRDF/AR is an automation solution that uses both SRDF andTimeFinder to provide a periodic asynchronous replication of arestartable data image. Use a single-hop SRDF/AR configuration, asshown in Figure 12, that permits controlled data loss (depending onthe cycle time).

Figure 12 SRDF/Automated Replication (SRDF/AR) single-hop data flow

However, if greater protection is required, a multi-hop SRDF/ARconfiguration can provide long distance disaster restart with zerodata loss at a middle or bunker site.

Figure 13 illustrates the SRDF/AR multi-hop data flow.

Figure 13 SRDF/Automated Replication (SRDF/AR) multi-hop data flow

Compared to traditional disaster recovery solutions with their longrecovery time and high data loss, disaster restart solutions usingSRDF/AR provide remote restart with a short restart time and lowdata loss. SRDF/AR offers data protection with dependent writeconsistency across a distance. This protection is accomplished byusing geographically separated replicas with hardware and softwareproducts from EMC Corporation.

Source Target

Unlimited distance

1

32 4

Source

Synchronous Adaptive

1 2

34

Target

5



The SRDF/AR process can be implemented with TimeFinder/Mirrorin a mainframe z/OS environment, or through UNIX and Windowsenvironments using Solutions Enabler's symreplicate command lineinterface.

SRDF/Automated Replication (SRDF/AR) 63

64


Concurrent SRDFEnginuity 5567 and later support the ability for a single primaryvolume to be remotely mirrored to two secondary volumesconcurrently. This feature is called Concurrent SRDF. ConcurrentSRDF can be both SRDF/S, both SRDF/A, both SRDF AdaptiveCopy, or mixed. Normal operating rules for SRDF also apply toconcurrent SRDF configurations. When operating in synchronousmode, ending status for an I/O is not presented until the remoteSymmetrix system acknowledges receipt of the I/O to the primarySymmetrix system. If both secondary volumes are operating insynchronous mode, ending status is not presented until both volumesacknowledge receipt of the I/O. If only one remote mirror is insynchronous mode, ending status is presented to the host when thesynchronous volume acknowledges receipt of the I/O.

Figure 14 shows a concurrent SRDF configuration in which theprimary volume is communicating with one secondary volume insynchronous mode. Concurrently, the same primary volume iscommunicating with its other secondary volume in SRDF/A mode.

Figure 14 Concurrent SRDF

SRDF/S

Source

SRDF/A

Far city

Near city

ICO-IMG-000431



Cascaded SRDFCascaded SRDF is a three-site, no data loss disaster recoveryconfiguration where data from a primary site is synchronouslyreplicated to a secondary site, and then asynchronously replicated toa tertiary site. The core benefit behind a cascaded configuration is itsinherent capability to continue replicating from the secondary site tothe tertiary sites in the event that the primary site goes down withminimal user intervention. This enables a faster recovery at thetertiary site, provided that is where the customer is looking to restarttheir operation.

Cascaded SRDF uses a dual role SRDF R2/R1 device (R21 device) onthe secondary site, which simultaneously acts as both an R2 to theprimary site and an R1 to the tertiary site. SRDF R2/R1 devicesrequire Enginuity 5773 or later. This requires the secondary siteSymmetrix system to be at this level. The primary or tertiarySymmetrix systems may run other Enginuity levels that caninteroperate with the secondary Symmetrix via SRDF. The CascadedSRDF solution complements and offers additional capabilities andoptions to EMC's existing three-site configurations, Concurrent RDF,SRDF/AR Multi-hop, and SRDF/Star.

Figure 15 illustrates a three-site Cascaded SRDF configuration.

Figure 15 Cascaded SRDF

Primary Secondary

SRDF/S SRDF/A

1 2

34

ICO-IMG-000433

Tertiary

5R1 R21 R2

Cascaded SRDF 65

66


SRDF/Extended Distance Protection (SRDF/EDP)Available at Enginuity level 5874, SRDF/EDP is an extension ofCascaded SRDF that uses a diskless R21 device, referred to as aDLDEV. The DLDEV device has no disk storage and temporarilystores writes received from the R1 and owed to the R2 in Symmetrixcache. The R21 DLDEV device must be on a Symmetrix runningEnginuity 5874. The R1 and R2 devices can be on a Symmetrixrunning either Enginuity 5874 or 5773.

The name Extended Distance Protection denotes how the remote R2 isbeyond the distance where SRDF/S would be used, yet still containsa replica of the R1 at a lower hardware resource cost than otherthree-site SRDF solutions. As far as disk storage goes, ExtendedDistance Protection is effectively a two-site SRDF solution. Using aDLDEV R21 in place of a standard R21 device alters the migrationenvironment requirements and capabilities affecting the options forchoosing a data migration solution. A DLDEV R21 device does notrequire any associated disk storage, thereby freeing up the R21 siteSymmetrix from needing that resource. However, DLDEV R21devices will utilize more cache and the SRDF mode between the R21and R2 must be either SRDF/A or Adaptive Copy Write Pending(Adaptive Copy Disk Mode is not supported). In addition, secondaryuse of the DLDEV R21 device by a host local to the Symmetrix thatcontains it is not possible because DLDEV devices cannot bepresented to a host.

Note: The secondary Symmetrix does require a minimum of disk drives inorder to support vault and SFS devices.

Figure 16 illustrates a three-site SRDF/EDP configuration.

Figure 16 SRDF/Extended Distance Protection (SRDF/EDP)

Primary Secondary

SRDF/S SRDF/A

1 2

3 4

Tertiary

5 R1 R2

DLDEVR21



SRDF/StarAvailable at Enginuity level 5x71, SRDF/Star provides advancedmultisite business continuity protection available for mainframe andopen systems environments. It enables concurrent SRDF/S withconsistency groups and SRDF/A with MSC operations from the samesource volumes with the ability to incrementally establish anSRDF/A session between the two remote sites in the event of aprimary site outage, a capability only available through SRDF/Starsoftware.

This capability takes the promise of concurrent or cascaded(including SRDF/EDP) synchronous and asynchronous operations(from the same source device) to its logical conclusion. SRDF/Starallows you to quickly re-establish protection between the two remotesites in the event of a primary site failure, and then just as quicklyrestore the primary site when conditions permit.

With SRDF/Star, enterprises can quickly resynchronize the SRDF/Sand SRDF/A copies by replicating only the differences between thesessions, allowing for much faster resumption of protected servicesafter a source site failure. Used in a data migration scenario, aplanned switch can be used to quickly maintain disaster recoveryprotection at the migrated site.

Figure 17 illustrates a three-site concurrent SRDF/Star configurationshowing the standby recovery links between the B and C sites as adashed line.

Figure 17 Concurrent SRDF/Star

SRDF/S SRDF/A

Source (A)

SRDF/A

Far city (C)

Near city (B)

SRDF/Star 67

68


SRDF and data migrationTwo additional features of SRDF are important to understand fordata migration. First is understanding that SRDF allows theswapping of the R1 and R2 roles. Second is that data will be remotelyaccessed from the R2 device if it is not available on the R1 device.

SRDF has always supported data migrations with the followingseries of operations where SRDF is configured in a way that the new,target Symmetrix pulls data from the original Symmetrix. Availablewith Solutions Enabler 7.1 and later, the SRDF device migrationfeature provides a simpler and more robust method to replace anexisting R1 or R2 device with a new device in an SRDF pair.Additionally, four-site SRDF data migration can be used toimmediately have SRDF remote protection at a newly migrated tosite. The series of operations to configure SRDF for the new, targetSymmetrix to pull data from the original Symmetrix are:

1. Bring in the new array and configure it to act as the R2 remote ofthe production R1 devices:

• Because of Concurrent SRDF, it is possible to do this even ifthe production R1 devices already have corresponding R2devices for disaster recovery.

• This step is not absolutely required, but for performancereasons and to ensure correct configuration, it is almostalways done.

• Due to large amounts of data it may take extensive time for thesynchronization to the new devices to complete, forperformance reasons, this initial synchronization is usuallyconducted in Adaptive Copy Disk mode.

2. Swap the personality of the R1 and R2 devices:

• This is conducted similar to a disaster failover, where theapplications must be restarted at the disaster site.

• This swap cannot occur for only one leg of a concurrent SRDFconfiguration. If it is necessary to maintain disaster protectionvery soon in the new site, then the SRDF device migrationfeature, or the four-site SRDF data migration feature, or anSRDF/Star planned switch can be used.



• It is not necessary to complete the synchronization in step 1before moving on to step 2 , because any data not yet in thenew R1 can be accessed remotely from the R2 (old R1).Though, in order to avoid loss of data from multiple failures, abest practice is to wait for synchronization before the swap.

3. When the synchronization is complete for all the devices, then theSRDF definition can be removed and the old Symmetrix array canbe disconnected.

Note: A use case selecting SRDF as the migration method is described in”Symmetrix Remote Data Facility (SRDF) solution use case,” on page 141.

SRDF operations can be managed with Symmetrix ManagementConsole (SMC) and the Solutions Enabler CLI command symrdf.

SRDF device migrationAvailable with Solutions Enabler 7.1 and later, the SRDF devicemigration feature provides the ability to replace an existing R1 or R2device with a new device in an SRDF pair. During migration, aconcurrent SRDF relationship is established to transfer data from anexisting R1 device to a new device in adaptive copy disk mode. Afterthis data transfer, the R1 device or the R2 device is replaced with thenewly-populated device in the SRDF pair. More information on thesteps to perform either an R1 or R2 device migration can be found inthe EMC Solutions Enabler Symmetrix SRDF Family CLI Product Guide(version 7.1 or later).

Four-site SRDF data migrationA four-site SRDF configuration can be thought of as a two-site datamigration of an existing two-site SRDF configuration to a newtwo-site SRDF configuration. Previously, customers wanting tomigrate an existing two-site SRDF configuration would useconcurrent SRDF to replicate (migrate) from the existing source arrayto the new source array, sever the existing source array pair-to-newsource array pair relationships, and then fully replicate between thenew source and new target array. During the latter cycle theproduction data would either be unprotected, or productionapplications would need to remain down until the new source andtarget arrays were fully synchronized providing full remoteprotection.

SRDF and data migration 69

70


With four-site SRDF data migration, replication from the existingsource array to both the new source array and new target array isperformed simultaneously. This enables a cutover to the new two-siteconfiguration with minimal application downtime and with noproduction data exposure. Four-site SRDF data migration is availableas the Dual-Site SRDF Migration Service from EMC.

Figure 18 illustrates four-site SRDF data migration.

Figure 18 Four-site SRDF data migration

Production

Existing

Production

New

Recovery Recovery

Adaptive copy disk

Sync orAsync

Adaptivecopy disk

R11

R2R2

R21


6

This chapter details features of the Open Replicator product with afocus on the attributes relevant to data migration. The topics are:

◆ Introduction ........................................................................................ 72◆ Definitions........................................................................................... 73◆ Open Replicator, FLM, PPME, and data migration ...................... 77

EMC Open Replicatorfor Symmetrix

EMC Open Replicator for Symmetrix 71

72

EMC Open Replicator for Symmetrix

Introduction“Full selection model summary” on page 26 model step 4, whichidentifies potential data migration solutions needs more detail than isprovided in “Open Replicator for Symmetrix” on page 31, andChapter 5, “EMC Symmetrix Storage System.” Details of the differentdeployment options of Open Replicator and the ways in which theycan be used for migration are needed both for step 4 and for step 6when the potential migration solutions are compared and evaluated.

EMC Open Replicator for Symmetrix enables remote point-in-timecopies to be used for data mobility, remote vaulting, and migrationbetween EMC Symmetrix VMAX or DMX and qualified storagearrays with full or incremental copy capabilities. Open Replicatorcan:

◆ Pull from source volumes on qualified remote arrays to aSymmetrix VMAX or DMX volume

◆ Push any live source Symmetrix VMAX or DMX volume to atarget volume on a qualified array with incremental updates

◆ Perform online data migrations from qualified storage toSymmetrix VMAX or DMX with minimal disruption to hostapplications



DefinitionsThe Symmetrix VMAX or DMX, where Open Replicator is beingmanaged, and its devices are always referred to as the control side ofthe copy operation. Other Symmetrix arrays, CLARiiON arrays, orthird-party arrays on the SAN will always be referred to as the remotearray/devices. Open Replicator has two modes of operation – cold(offline) and hot (online). Online or offline refers to the state of theSymmetrix VMAX or DMX resident devices (control devices). In bothscenarios, the remote devices should be offline to the host connectedto the remote array. Open Replicator supports two types of copyoperations, push and pull. A push operation copies data from thecontrol device to the remote device. A pull operation copies data tothe control device from the remote device.

Open Replicator can push data volumes out from a Symmetrix eitherin a live mode (hot), or from a static copy or source volume (cold). Fora live push, no local point-in-time copies of the volumes are required:The Symmetrix creates logical point-in-time copies without having toallocate additional disk space, and I/O is permitted against thesource volume during the transfer. Figure 19 illustrates an OpenReplicator hot (or live) push.

Figure 19 Open Replicator hot (or live) push

Using VDEVs assource devices for acold push

When an Open Replicator cold push operation is used to migratedata, usually a point-in-time copy of production devices is created sothat the migration can proceed while the production devices are stillin use. In the past this copy needed to be a full copy created usingeither TimeFinder/Clone or TimeFinder/Mirror. With Enginuity5874, Solutions Enabler 7.0, and SMC 7.0, a TimeFinder/Snap VDEV

Start: 6:00 a.m.

End: 6:15 a.m.

Image: 6:00 a.m.

TargetSTD

ICO-IMG-000434

Definitions 73

74


(Virtual Device) can be used as a source volume instead. Unlike thetarget devices used to hold a TimeFinder/Clone orTimeFinder/Mirror copy, which must be equal (or greater) in sizethan the corresponding production data devices, VDEV devices arelogically defined as equal in size but should use significantly lessspace on disk. (Best practice is for VDEV devices to consume less than30 percent of the amount of space needed to store a full copy of theproduction volumes.) The difference in resource requirements for thepoint-in-time copy of the production devices might affect the choiceof data migration solution.

For cold push, up to 16 remote copies of the local volume can bemade, and those remote copies can be incrementally updated. Figure20 illustrates an Open Replicator cold (or BCV) push.

Figure 20 Open Replicator cold (or BCV) push

In pull operations, the Symmetrix volume can again be in a live stateduring the copy process, which makes either restoring remotelyvaulted volumes or migrating from other storage platforms very fastand efficient. The local hosts and applications can begin to access thedata as soon as the session begins, even before the data copy processhas completed. A process referred to as Copy on First Access is used toensure the appropriate data is available to a host operation when it isneeded. Open Replicator hot pull is a technology that supportsFederated Live Migration (FLM). Figure 21 on page 75 illustrates anOpen Replicator hot (or live) pull.

Target

STD

BCV

ICO-IMG-000435

PS0 PS1 PS2 PS3 PS4 SMB0 SMB1

SB0

SB1

SB2

SB3

SB4

SB5

SB6

SB7

SB8

SB9

SB10

SB11

SB12

SB13

SB14

SB15

Target

Target



Figure 21 Open Replicator hot (or live) pull

Of course, the pull can also be in cold mode to a static SymmetrixVMAX or DMX volume. Figure 22 illustrates an Open Replicator cold(or point-in-time) pull.

Figure 22 Open Replicator cold (or point-in-time) pull

To protect against potential data loss due to a SAN failure or otherconnectivity issue during a hot pull operation, the –donor_updateoption can be used. When enabled, the donor update feature enablesarrays to propagate (update) writes to the local device back to theremote device (donor) as data is being pulled from the remote device.When enabled, donor update ensures consistent data between thelocal and remote device. As a result, no new data written to the local

STD

ICO-IMG-000436


SB0

SB1

SB2

SB3

SB4

SB5

SB6

SB7

SB8

SB9

SB10

SB11

SB12

SB13

SB14

SB15

PiTCopy

PiTCopy

STD

STD

Target

Target

ICO-IMG-000437


SB0

SB1

SB2

SB3

SB4

SB5

SB6

SB7

SB8

SB9

SB10

SB11

SB12

SB13

SB14

SB15

STD

STD

Target

Definitions 75

76


device will be lost if an Open Replicator session has to be restartedbefore completing. Federated Live Migration (FLM) sessions usingOpen Replicator hot pull always use donor update.

Interaction ruleswith SRDF

Prior to Enginuity 5874 and Solutions Enabler 7.0, certain interactionsbetween Open Replicator and SRDF operations were blocked. If theexisting environment required certain SRDF operations to continueunimpeded regardless of Open Replicator operations or would notpermit the SRDF link state to change from the RW (Read Write orReady) state to allow an Open Replicator pull operation, then OpenReplicator was not an available choice for the data migrationsolution.

With Enginuity 5874 and Solutions Enabler 7.0 the following SRDFoperations are no longer blocked when an Open Replicator controldevice is in the Created or Recreated state:

• SRDF establish or resume when the R2 is an Open Replicatorcontrol device

• SRDF restore when the R1 is an Open Replicator controldevice

The following Open Replicator operation is no longer blocked whenthe SRDF link is in the RW state:

• Open Replicator pull operation



Open Replicator, FLM, PPME, and data migrationThe typical scenario for using Open Replicator for data migration isusually a hot pull:

1. Configure Symmetrix VMAX or DMX between hosts and originalstorage.

FLM can be used together with Open Replicator to avoid anapplication outage when redirecting the application I/O to thenew migrated devices. More information can be found in ”FLMnondisruptive migration overview,” on page 82, and ”FederatedLive Migration (FLM) solution use case,” on page 151.

PPME can be used together with Open Replicator to avoid anapplication outage when redirecting the application I/O to thenew migrated devices. More information can be found in ”PPMEwith Open Replicator technical overview,” on page 111,and ”Open Replicator for Symmetrix/PPME solution use case,”on page 136.

2. Data migration occurs while applications remain online.

Host I/O passes through Symmetrix

3. When all data is migrated, unplug and remove original storage.

Figure 23 illustrates an Open Replicator hot pull data migration fromheterogeneous arrays to Symmetrix VMAX or DMX.

Open Replicator, FLM, PPME, and data migration 77

78


Figure 23 Open Replicator hot pull data migration

EMC Open Replicator operations can be managed with SymmetrixManagement Console (SMC), the Solutions Enabler CLI commandsymrcopy, and the PPME CLI command powermig.

ICO-IMG-000438

CLARiiON

Symmetrix

Non-EMC


7

This chapter details attributes of some of the Symmetrix features usedfor data migration. The topics are:

◆ Introduction ........................................................................................ 80◆ Federated Live Migration (FLM) ..................................................... 81◆ Symmetrix Virtual Provisioning ...................................................... 84◆ EMC TimeFinder family of local replication products ................. 87◆ Virtual LUN technology.................................................................... 91◆ Fully Automated Storage Tiering (FAST) ....................................... 98◆ Federated Tiered Storage ................................................................ 100

Symmetrix SoftwareUsed for Migration

Symmetrix Software Used for Migration 79

80

Symmetrix Software Used for Migration

IntroductionThis chapter details attributes of five Symmetrix features which mayplay a role in data migration:

◆ Federated Live Migration (FLM)

◆ Virtual Provisioning

◆ TimeFinder

◆ Virtual LUN technology

◆ Fully Automated Storage Tiering (FAST)



Federated Live Migration (FLM)Federated Live Migration (FLM) is a method by which a user cannondisruptively move data from older storage into a new VMAXarray. The host application cutover to use the new VMAX devices ismade transparent by a combination of presenting the VMAX devicesas additional paths to the old devices and managing which paths areactive through a multipath IO (MPIO) driver on the host. PowerPathand other MPIO drivers are supported.

FLM supports moving the data with Open Replicator SAN-basedreplication. Unlike application host-based PPME, control of themigration and cutover is managed through the VMAX array. FLMgreatly simplifies migrations by requiring no remediation whenmigrating pre-qualified stacks.

FLM, like Open Replicator, now supports front-end zero detection,which allows Enginuity to detect all zero thin extents on the source(or remote) device and not copy those extents to the target (or control)volume. This allows thick-to-thin migrations, as well as thin-to-thinmigrations, leaving the control volume thinly provisioned withoutany zero extents taking unnecessary space in the thin pool.

. Note: For information on supported operating systems, file systems, clustersolutions, and logical volume managers, refer to the EMC Federated LiveMigration Simple Support Matrix available at https://elabnavigator.EMC.com,Simple Support Matrix tab.

Benefits of using Federated Live MigrationFederated Live Migration (FLM) greatly reduces migrationcomplexity and risk by eliminating the need for operationalmigration steps on applications hosts. Support for pre-qualifiedstacks, including older versions typical in a migration scenario, caneliminate remediation steps. FLM eliminates the need for atraditional cutover step and provides the ability to failback from themigration without data loss.

Eliminating application downtime during a migration greatlysimplifies the planning for large-scale migrations. Migration windowflexibility is important to administrators because it simplifies themigration planning process. Administrators do not have to rely onothers to follow their directions, nor do they need to work off theirshift, which is when migration data movement operations are

Federated Live Migration (FLM) 81

http://powerlink.emc.com

http://powerlink.emc.com

82


frequently scheduled. With FLM, the pressure to correctly completecritical and complex migration tasks during outages or off hours iseliminated.

IMPORTANT

Data migrations are often complex operations and require carefulplanning and execution of predetermined procedures. Failure toidentify and perform necessary steps or work within supportedconfigurations can result in data unavailability or loss.

FLM nondisruptive migration overviewFLM uses standard Open Replicator commands with additionaloptions specific to Federated Live Migration sessions. Using OpenReplicator as the underlying technology, the hot pull does the bulkdata copying between the arrays through the SAN, and donor updatekeeps source and target devices synchronized. Mirroring the devicesallows testing of the migration and the option to return to the initialstate with no downtime.

Standard auto-provisioning command(s) between the FLM create andactivate operations enable the new VMAX devices to be presentedwith the identity of the old DMX devices. The MPIO driver sees pathsfor the same device on both the old DMX and new VMAX. FLMcontrols which set of paths are active to the MPIO driver. Figure 24 onpage 83 illustrates an overview of FLM operation.



Figure 24 FLM operation overview

Application

Data RAID 1

Application

Data RAID 1

Data RAID 5

Application

Data RAID 5

ICO-IMG-000906

Create andactivate FLM

session

TerminateFLM sessionremove DMX

DMX WWN &geometry

DMX WWN &geometry

DMX WWN &geometry

DMX WWN &geometry

PowerPath MPIO PowerPath MPIO PowerPath MPIO

Open Replicator

Old DMX Old DMX New VMAX New VMAX

Donor update

Federated Live Migration (FLM) 83

84


Symmetrix Virtual ProvisioningStarting with Enginuity version 5773, Symmetrix VirtualProvisioning, an enterprise class implementation of thinprovisioning, can be used to provide ease of use, efficient tiering, andenhanced performance. Symmetrix Virtual Provisioning allowsstorage administrators to create a device that presents an applicationwith more capacity than is physically allocated to it in the storagearray. Although not strictly a data migration tool, the use of virtualprovisioning may preclude the need for some future data migrationsdue to improvements in storage capacity utilization. A Symmetrixthin device can be the target of a data migration.

Figure 25 illustrates the storage perceived by the application as largerthan the physically consumed storage.

Figure 25 Symmetrix Virtual Provisioning

Virtual Provisioning provides simplified storage management. SLVprovisioning becomes decoupled from specific physical storage andthe re-provisioning steps required to support capacity growth aregreatly reduced.

With Virtual Provisioning:

◆ Capacity utilization is improved with the reduction ofallocated-but-unused storage. This has additional powerefficiency benefits.

Application 1

Application 2

Application 3

Consumed

App 1

App 2 App 3

ConsumedConsumed

UnallocatedUnallocatedUnallocated

Common Storage Pool

ICO-IMG-000451

Capacity on demand



◆ Application workloads that benefit from wide striping will showperformance improvements.

Symmetrix thin devices are logical devices used in many of the sameways that Symmetrix devices have traditionally been used. Unliketraditional Symmetrix devices, thin devices do not need to havephysical storage completely allocated at the time the device is createdand presented to a host. The physical storage used to supply diskspace to thin devices comes from a thin storage pool, which iscomprised of devices called data devices created for this purpose.

When a write is performed to a portion of the thin device for whichphysical storage has not yet been allocated, the Symmetrix arrayallocates only the physical storage for that portion of the thin device.The minimum possible amount of additional physical storage isallocated, reducing unnecessary consumption. The minimum amountof physical storage that can be allocated at a time for the dedicateduse of a thin device is referred to as a thin device extent. This isknown more commonly in the industry as a chunk.

When more storage is required to service existing or future thindevices, data devices can be added to existing thin storage pools.New thin devices can also be created and associated with existingstorage pools. Thin devices can be used with many Symmetrix localand remote replication products. This includes TimeFinder/Snap andSRDF:

◆ With TimeFinder/Snap, the target of a snap operation with a thinsource device will be a virtual device (VDEV), just as it is withstandard, fully allocated Symmetrix devices.

◆ Thin devices may be either SRDF source (R1) or SRDF target (R2)devices. SRDF pairs can be configured with thin devices on bothsides of an SRDF link.

◆ Beginning with Enginuity 5875, SRDF supports migration from athick source (R1) to a thin target (R2). Also, zero spacereclamation on the SRDF target (R2) side only destages non-zerodata to disk.

◆ Thin devices may be used as sources or targets of Open Replicatorcopy operations:

• If a thin device is the source of the data, zeroes are sent to thetarget device for any unmapped portions of the source device.

• If a thin device is the target of an Open Replicator copyoperation:

Symmetrix Virtual Provisioning 85

86


– Beginning with Enginuity 5875 zero space reclamationonly destages non-zero data to disk.

– For Enginuity releases before 5875, all data must be sent tothe thin device, leaving the target thin device fullyallocated.

Note: One of the goals when deploying Virtual Provisioning is to ensure theapplications and migration tools used to store or to move data from standardor fully provisioned environments to thin environments do not causeunneeded storage allocations.

When migrating from a thick to a thin device, contiguous series of zeros maybe written to represent available or initialized but unused space for avolumes or files in a file system. To help address this specific concern,Symmetrix Virtual Provisioning enables Symmetrix VMAX users toautomatically reclaim "chunks" containing all zeros. For more informationsee the EMC Symmetrix VMAX Virtual Provisioning Space Reclamation andApplication Considerations White Paper.

For Symmetrix DMX Virtual provisioning it may be necessary to use ahost-based migration if preserving the thin allocation is desired.



EMC TimeFinder family of local replication productsThe EMC TimeFinder family of software is the most powerful suite oflocal storage replication solutions available. Fully leveraging theindustry-leading high-end Symmetrix hardware architecture, it offersunmatched deployment flexibility and massive scalability to deliver awide range of in-the-system data copying capabilities to meet mixedservice level requirements without operational impact. Thefield-proven TimeFinder family is the most widely deployed set ofhigh-end replication solutions in the industry, with tens of thousandsof installations in the most demanding environments. And, only theTimeFinder family can provide cross-volume and cross-storagesystem consistency, tight integration with industry leadingapplications, and simplified usage through automated management.

The EMC TimeFinder family of local replication allows users tonondisruptively create and manage point-in-time copies of data toallow operational processes, such as backup, reporting, andapplication testing to be performed independent of the sourceapplication to maximize service levels without impactingperformance or availability.

TimeFinder is a family of products that enable LUN-based replicationwithin a single Symmetrix system. Data is copied from Symmetrixdevices using array-based resources without using host CPU or I/O.The source Symmetrix devices remain online for regular I/Ooperations while the copies are created. The TimeFinder familyconsists of two base replication products and several componentoptions. The TimeFinder components complement one another toensure maximum replication coverage for the entire data processingenvironment:

Note: The TimeFinder family is used in environments configured with EMCSymmetrix VMAX with Enginuity arrays, Symmetrix DMX arrays, andearlier Symmetrix family models. Symmetrix VMAX arrays requireEnginuity release level 5874 or later; Symmetrix DMX arrays requireEnginuity 5773 and earlier.

◆ Base replication products:

• TimeFinder/Clone provides clone copy sessions that createpoint-in-time copies of full volumes or individual datasets.TimeFinder/Clone enables users to make copies of datasimultaneously on multiple target devices from a single source

EMC TimeFinder family of local replication products 87

88


device. The data is available to a target’s host immediatelyupon activation, even if the copy process has not completed.Data may be copied from a single source device to as many as16 target devices. A source device can be either a Symmetrixstandard device or a TimeFinder BCV device.

• TimeFinder/Snap provides snap copy sessions that createeconomical, pointer-based replicas simultaneously onmultiple target devices from a single source device where onlythe pre-images of changed data are written. TimeFinder/Snapenables users to configure special devices in the Symmetrixarray called virtual devices (VDEVs) and save area devices(SAVDEVs). The data is available to a target’s hostimmediately upon activation.

• TimeFinder VP Snap allows multiple replication sessions toshare capacity allocations within a Virtual Provisioning thinpool, thus reducing the storage that must be dedicated tosaving changed tracks. TimeFinder VP Snap is available inEnginuity 5876 and Solutions Enabler 7.4 and higher andcrates point in time copies that are conceptually similar tothose created by TimeFinder/Snap. With VP Snap, both thesource and target must be thin devices, therefore, the copieddata will be allocated space within a thin pool.

VP Snap sessions are unique because thin pool allocations canbe shared by two or more target devices. For example, sourceupdates that are new to multiple point-in-time copies aresaved in a single set of allocations that are one target, only asingle shared copy resides in the thin pool, which providescost-effective space savings.

◆ Component options:

• TimeFinder/Mirror is another family component that worksin Symmetrix DMX environments running Enginuity 5773 andearlier. Starting with Enginuity 5874, there is a singletechnology that is in use for full volume copies, which is cloneemulation. In environments running Enginuity 5874 and later,all TimeFinder/Mirror scripts are executed in CloneEmulation mode. Therefore, there are no changes in the operationaldetails provided for integrating TimeFinder/Mirror with OpenReplicator.



Note: Enginuity 5874 does not support Virtually Provisioned BCVs.Therefore, VP devices can only use TimeFinder/Clone for fullvolume operations.

• TimeFinder/Consistency Groups (TimeFinder/CG) enablesother TimeFinder family products to coordinate cross-volumeand cross-system consistency for application restartability.

• TimeFinder/Exchange Integration Module(TimeFinder/EIM) automates and simplifies the process ofcreating and managing TimeFinder replications of a MicrosoftWindows Exchange Server environment.

• TimeFinder/SQL Integration Module (TimeFinder/SIM)automates (TimeFinder/SIM) simplifies the process ofcreating and managing TimeFinder replications of a MicrosoftWindows SQL Server environment.

TimeFinder and data migrationTimeFinder plays a role in data migration in two principle ways:

◆ First, TimeFinder/Clone or TimeFinder/Mirror can be used tocreate an independent full point-in-time copy within a singlestorage array.

Redirecting an application to point to the replicated data in itsnew location is effectively a data migration. This method can beused when the necessity of bringing the application offline topoint to the new LUN is acceptable, or PowerPath MigrationEnabler (PPME) TimeFinder/Clone support can be utilized tomake TimeFinder/Clone migrations nondisruptive; see”PPMETimeFinder/Clone,” on page 112. One key use of this technologycombines with Enginuity 5874 and Solutions Enabler 7.1 supportfor TimeFinder/Clone replication between a fully allocated(“thick”) device and a virtually provisioned (thin) device withadditional capability to reclaim space from the thin device whichis all zeros; for more information see the White Paper: EMCSymmetrix VMAX Virtual Provisioning Space Reclamation andApplication Considerations.

◆ Second, TimeFinder is used to create a local point-in-time copy,which is used as a staged copy for migration to a remote array, asseen in”SRDF/Automated Replication (SRDF/AR),” on page 62.

EMC TimeFinder family of local replication products 89

90


TimeFinder operations can be managed with SymmetrixManagement Console (SMC) and the Solutions Enabler CLIcommands symclone, symsnap, and symmir.



Virtual LUN technologyEMC has offered virtual LUN technology that enables data migrationwithin an array without host or application disruption sinceEnginuity 5772. Virtual LUNs enable you to bring your InformationLifecycle Management (ILM) strategy (introduced in Chapter 2,“Reasons for Moving Data,”) to life by easily moving informationthroughout the storage system as its value changes over time. It canassist in system reconfiguration, performance improvement, andconsolidation efforts all while helping maintain vital service levels.

There are currently two versions of Virtual LUN technology whichrun on either the Symmetrix VMAX or the Symmetrix DMX-3 andDMX-4 arrays. The Symmetrix VMAX technology is often referred toas “Enhanced Virtual LUN technology,” and the DMX version as“Symmetrix Optimizer Virtual LUN technology.” The EnhancedVirtual LUN technology available for Symmetrix VMAX is superiorin including the ability to:

◆ Migrate to a different RAID protection type.

◆ Migrate leaving all local and remote replication intact.

◆ Perform Symmetrix control (for example, replication orconfiguration) operations while a migration is in progress. (Allreplication operations can be concurrent and the Symmetrixconfiguration file is locked only for a short time.)

◆ Migrate multiple Symmetrix Logical Volumes (SLVs) in a session

◆ Concurrently migrate up to four migration sessions in a singleSymmetrix.

◆ Starting with Enginuity 5875, migrate VP devices to a target thinpool.

Figure 26 illustrates the Virtual LUN migration process.

Figure 26 Virtual LUN migration

ATARAID 1

7200 rpm500GB

00100100100

Tier 1 Volume Tier 2 Volume

FibreChannelRAID 1

15.000 rpm

Virtual LUN technology 91

92


Note: A use case selecting Virtual LUN technology as the migration methodis described in ”Virtual LUN solution use case,” on page 147.

Symmetrix OptimizerSymmetrix Optimizer improves array performance by continuouslymonitoring access patterns and swapping devices (Symmetrix logicalvolumes) to achieve balance across the disks in the array. Thisautomated process utilizes user-defined parameters and iscompletely transparent to end users, hosts, and applications in theenvironment.

Swapping is performed with constant data availability andprotection. A typical swap would move a highly accessed devicefrom a busy disk spindle to a less busy disk spindle, and at the sametime move a less highly accessed device from the less busy diskspindle to the busy disk spindle. The result of the swap improvesoverall disk performance by increasing the number of disk requeststhat can be processed simultaneously because requests are betterbalanced across the multiple disk spindles. Note that this automatedprocess will usually preclude the need for user directed, manualmigrations for the purpose of balancing back-end disk I/O.

Symmetrix Optimizer technology is the mechanism used toimplement Virtual LUN migration for DMX-3 and DMX-4 arrays.Optimizer technology for analyzing SLV access patterns is acomponent of Fully Automated Storage Tiering (FAST) forSymmetrix VMAX arrays.

Symmetrix Optimizermigration

Starting with Enginuity 5772, Symmetrix Optimizer can performmigrations in addition to swaps. The key difference is that the usermust initiate the desired migration and there must be unallocateddisk space for the new location of the data. When a single device isbeing moved to unallocated space, there is no swap of locations fortwo devices. However the pre-migration location of the devicebecomes unallocated and can be used for a new purpose.

Migration moves devices from an old to a new location. The user onlydefines the destination as a group of similar disks or a single diskdrive. High-performance disks would be grouped separately fromhigh-capacity disks, making it easy to migrate devices to theappropriate tier of storage.



Migration contains the following features:

◆ One or more devices moved to one or more target disks.

◆ All mirrors of the devices are moved.

◆ Migrated volumes are distributed across specified recipientdrives by a sophisticated layout algorithm. Final placement of thevolumes is not known in advance.

◆ After completion of the migration, the original volumes aredeleted from their original locations.

Symmetrix Optimizer Virtual LUN technologyEMC Symmetrix DMX-3 and DMX-4 with Virtual LUN technologyintroduced in Enginuity 5772 enable data migration within an arraywithout host or application disruption. Virtual LUN technology is anenhancement to the Symmetrix Optimizer product that enablestransparent, nondisruptive data mobility among storage tiers withinthe same array with devices of the same RAID protection scheme.Virtual LUN technology is supported for both open system andmainframe platforms, and includes support for metavolumes.

Note: PowerPath Migration Enabler (PPME) TimeFinder/Clone support canbe used to nondisruptively migrate data while changing the RAID protectiontype in Symmetrix DMX systems as described in”PPME TimeFinder/Clone,”on page 112.

Virtual LUN technology offers two types of data movement:

◆ Migration

Migration provides users the ability to move data betweenhigh-performance disks and high-capacity disks, or to populatenewly added disk drives.

◆ Relocation

Relocation is used to support rollbacks, which can be invoked byusers to undo migrations that conflict with business rules. Forexample, an administrator might want to conform to a businessrule that a volume should in fact continue to reside in its originalhigh-priority location, invoking relocation rollback to put it backto its priority location.


94


Virtual LUNmanagement

The migration feature is based on Optimizer swap technology andshares the same requirements as the traditional swap environment.Dynamic Relocation Volumes (DRVs) must be available to provideadditional protection during the data movement process. Onlydevices with a matching sized DRV are eligible for Optimizer VirtualLUN control. RAID 6 devices are an exception and do not require aDRV since the two parity devices already provide double protection.

Data migration and relocation consume array resources anduser-definable limits can be set to avoid conflicts with productionapplications. The maximum number of volumes that can be movedeach day and the maximum simultaneous volumes to be movedwork together to limit resources that are used for the process. Apriority value can also be assigned to the copy process of eachparticipant volume. Together, these parameters affect how long amigration or relocation will take.

The Symmetrix configuration lock will remain in place while theVirtual LUN data movement is in progress. If other configurationmanagement functions are routinely scheduled, the Virtual LUNprocess will be limited by the maximum simultaneous volumes andmaximum number of volumes to ensure completion within atimeframe that would allow the configuration lock to be released forother activities.

Optimizer Virtual LUN functionality moves the logical volume as acomplete entity, including all mirrors of the device. Subsets of datathat reside on the logical volumes remain unknown to the process.Consequently, users are advised to understand data placement andassign related data to grouped logical volumes.

Virtual LUN migration is accomplished without applicationdowntime, providing the ability to meet service levels. After a LUN ismoved using Optimizer Virtual LUN technology, the LUN retains allof the characteristics of the original LUN, including its name and ID,so there is no need to bring the application offline.

Symmetrix Optimizer and its Virtual LUN component can bemanaged with Symmetrix Management Console (SMC) and theSolutions Enabler CLI command symoptmz.



Enhanced Virtual LUN technologyVirtual LUN technology, enhanced with Enginuity 5874 for theSymmetrix VMAX Series, enables transparent, nondisruptive datamobility among storage tiers within the same array and betweenRAID protection schemes without impacting local or remotereplication. Organizations can respond more easily to changingbusiness requirements when using tiered storage in the array. VirtualLUN technology is supported for both open system and mainframedevices, and includes support for metavolumes.

Virtual LUN technology offers two types of data movement:migration to unconfigured space and migration to configured space.In each case, the migration provides users the ability to move databetween high-performance drives and high-capacity drives, or topopulate newly added drives, with full inter-RAID flexibility.

RAID VirtualArchitecture

RAID Virtual Architecture (RVA) is the architecture employed inSymmetrix VMAX systems to implement RAID protection. RVAextends the design of RAID 6 used in the Symmetrix DMX-3 andDMX-4 storage arrays and, in so doing, provides a uniformimplementation of all RAID protection schemes – unprotected, RAID1, RAID 5, and RAID 6. The RVA design abstracts all RAID protectiontypes to a single mirror position for a given Symmetrix logicalvolume, thus freeing up device mirror positions for additionalfeatures, such as Virtual LUN migrations.

RAID Virtual Architecture temporarily allows for two distinct RAIDgroups to be associated with a device. This feature allows for aSymmetrix logical volume to be migrated from one RAID group toanother.

Virtual LUNtechnology

Virtual LUN technology enables transparent, nondisruptive datamobility for standard Symmetrix volumes between storage tiers andbetween RAID protection schemes. Virtual LUN can be used topopulate newly added drives or move devices betweenhigh-performance and high-capacity drives, thereby delivering tieredstorage capabilities within a single Symmetrix array. Migrations areperformed while providing constant data availability and protection.

Virtual LUN technology performs in-the-box tiered storage migrationby migrating data from one RAID group to another. The migration isfacilitated by the creation and movement of RAID groups throughonline configuration changes on the Symmetrix.


96


Starting with Enginuity 5875, Virtual LUN technology can alsoperform in-the-box tiered storage migration by rebinding a VP deviceto a target pool, migrating all allocated tracks to the target pool.

RAID Virtual Architecture allows, for the purposes of migration, twodistinct RAID groups, of different types or on different storage tiers,to be associated with a logical volume. In this way, Virtual LUNallows for the migration of data from one protection scheme toanother, for example RAID 1 to RAID 5, without interruption to thehost or application accessing data on the Symmetrix device.

Virtual LUN can be used to migrate standard Symmetrix devices andmetadevices of any emulation – FBA, CKD, and iSeries. Migrationscan be performed between all drive types includinghigh-performance enterprise EFD drives, Fibre Channel drives, andlarge capacity SATA drives.

Data can be migrated to either unconfigured or configured space. Inthe case of the migration to unconfigured space, a target RAID groupis created from the free pool in the array and migrated to. Whencomplete, the original RAID group is deleted and the storagereturned to the free pool.

When migrating to configured space, an existing, but unused,Symmetrix device is specified as a target and the source devicemigrated to the physical storage occupied by the target device. Thistarget device can either be specified by the user or automaticallyselected by the Symmetrix. Following the migration, the devicespecified as the target will occupy the storage originally associatedwith the device being migrated. An iVTOC is then performed on thetarget device leaving its original data inaccessible.

Note: In the case of migrating to configured space, the target device specifiedmust be of the same size and configuration as the source device beingmigrated. If migrating a metadevice, a metadevice of the same size andconfiguration must be specified, or a number of individual devices equal insize and quantity of the metamembers being migrated.

For the purposes of a migration, either the protection type or thestorage tier must change, or both. Symmetrix volumes of anyprotection can be migrated to any other supported RAID protectionwith one exception – protected devices may not be migrated to spacethat is unprotected.



Virtual LUN technology is fully interoperable with all otherSymmetrix replication technologies: SRDF, TimeFinder/Clone,TimeFinder/Snap, and Open Replicator.

Virtual LUN migrations can be managed via the SymmetrixManagement Console (SMC) graphical interface and the SolutionsEnabler CLI command symmigrate.

For more information reference the Best Practices for NondisruptiveTiering via EMC Symmetrix Virtual LUN Technical Note, available onhttps://support.emc.com.


98


Fully Automated Storage Tiering (FAST)EMC Symmetrix FAST technology, for standard provisioned,non-thin environments, is designed to automate the allocation andrelocation of application data across different performance tiers basedon the changes in the application performance requirements. FASThelps customers to maximize the benefits of tiered storage within thearray by prioritizing and optimizing application performance andreduce labor cost.

Symmetrix FAST proactively monitors volume workloads andautomatically moves the busier volumes to higher performing flashdrives, and the slower volumes to higher capacity, lowerperformance drives (SATA). The promotion/demotion activity isdriven by policies assigned to the FAST storage groups and isexecuted nondisruptively without affecting business continuity andavailability.

The management and operation of FAST are accomplished via theSymmetrix Management Console (SMC) graphical interface or theSolutions Enabler CLI commands symfast, symtier, symsg, andsymoptmz.

Management and operation of FAST is inherently different than othermigration strategies described in this book, because it is designed asan automated process. For more information see both FAST Theoryand Best Practices for Planning and Performance Technical Notes, andImplementing Fully Automated Storage Tiering (FAST) for EMCSymmetrix VMAX Series Arrays Technical Notes, available onhttps://support.emc.com.

FAST VPBeginning with Enginuity 5875, FAST VP provides automated tieredstorage for Virtually Provisioned (VP) devices. FAST VP offerssub-LUN data movement for VP devices providing increasedcapacity utilization and a reduction in time and complexity tomanage storage. Moving data at this much more granular level,allows FAST VP to be more responsive to changes in the productionworkload activity and to improve performance more efficientlyrequiring fewer EFDs in the system and placing more data on SATA.



Symmetrix FAST VP proactively monitors volume sub-LUNworkloads and automatically moves the busier extents to a VP poolof higher performing flash drives, and the more idle extents to a VPPool of higher capacity, lower performance SATA drives. Thepromotion/demotion activity is driven by policies assigned to theFAST storage groups and is executed nondisruptively withoutaffecting business continuity and availability. The management andoperation of FAST VP are accomplished through the same userinterfaces as for FAST for standard provisioned devices.

Operation of FAST VP is inherently different from other migrationstrategies described in this book, because it is designed as anautomated process. Implementing Fully Automated Storage Tiering withVirtual Pools (FAST VP) for EMC Symmetrix VMAX Series ArraysTechnical Notes, available on https://support.emc.com, provides moreinformation.

Fully Automated Storage Tiering (FAST) 99

100


Federated Tiered StorageThis section discusses EMC Federated Tiered Storage (FTS). Thefollowing topics are included:

◆ “Introduction to FTS” on page 100

◆ “FTS benefits” on page 101

◆ “FTS environment example” on page 101

◆ “FTS required components” on page 102

◆ “Modes of operation” on page 102

◆ “FTS and data migration” on page 104

◆ “Non-migration use cases” on page 104

Introduction to FTSFederated Tiered Storage (FTS) is a new feature of Enginuity 5876 thatallows supported, SAN-attached disk arrays to provide physical diskspace for a Symmetrix VMAX array. This permits the user to manage,monitor, migrate, and replicate data residing on Symmetrix VMAXand other supported arrays using familiar EMC software andEnginuity features. This includes data that already exists on externalarrays, as well as new storage that is being allocated.

FTS allows LUNs that exist on external arrays to be used as rawstorage space for the creation of Symmetrix devices in the same wayinternal Symmetrix physical drives are used. These devices arereferred to as eDisks, discussed further on page 102.

Data on the external LUNs can also be preserved and accessedthrough Symmetrix devices, allowing the use of SymmetrixEnginuity functionality (such as local replication, remote replication,storage tiering, data management, and data migration) with data thatresides on external arrays.

Note: For information on supported external storage arrays, refer to the EMCFederated Tiered Storage Simple Support Matrix available athttps://elabnavigator.EMC.com, Simple Support Matrix tab.



FTS benefitsFTS benefits include:

◆ Simplifies management of virtualized multi-vendor or EMCstorage by allowing heterogeneous arrays to be managed bySolutions Enabler and Unisphere for VMAX.

◆ Allows data mobility and migration between heterogeneousstorage arrays and between heterogeneous arrays and VMAX.

◆ Offers Virtual Provisioning benefits to external arrays.

◆ Allows VMAX enterprise replication technologies, such as SRDFand TimeFinder, to be used to replicate storage that exists on anexternal array.

◆ Extends the value of existing disk arrays by allowing them to beused as an additional storage tier.

FTS environment exampleFigure 27 shows a high-level view of an FTS environment.

Figure 27 High-level view of an FTS environment

Federated Tiered Storage 101

102


FTS required componentsFTS is implemented entirely in Enginuity and does not require anyadditional Symmetrix hardware. Connectivity with an external arrayis established through the same fibre optic SLICs currently used forconfiguring FAs and RFs. Instead of running FA or RF emulation,however, the processors run a new type of emulation, DX. DX andother required components are discussed in this section.

DX directors A new emulation, referred to as DX (for DA eXternal), has beendeveloped. DX adapts the traditional DA emulation model to act onexternal logical units as though they were physical drives. The factthat a DX is using external LUNs instead of a DA using internalLUNs is transparent to other director emulations and to theEnginuity infrastructure in general. With respect to most non-drive-specific Enginuity functions, a DX behaves the same as a DA.

eDisks An eDisk is a logical representation of an external LUN when it isadded into the VMAX configuration. The terms eDisk and externalspindle both refer to this external LUN once it has been placed in anexternal disk group and a virtual RAID group.

External disk group External disk groups are virtual disk groups created by the user tocontain eDisks. Exclusive disk group numbers for external diskgroups start at 512.

Note: External spindles and internal physical spindles cannot be mixed in adisk group.

Virtual RAID group An unprotected, virtual RAID group is created for each eDisk thatgets added to the system. The RAID group is virtual because eDisksare not locally protected by the VMAX array. eDisks rely on theprotection provided by the external array.

Modes of operationFTS has two modes of operation, depending on whether the externalLUN is used as raw storage space or has data that must be preservedand accessed through a VMAX device. Each is discussed further inthis section:

◆ “Encapsulation” on page 103◆ “External Provisioning” on page 103



Encapsulation Encapsulation allows the user to preserve existing data on externalLUNs and access it through Symmetrix volumes. These devices arecalled encapsulated devices.

In the case of migration from an external array into a VMAX array,encapsulation is the mode of operation primarily involved.

Encapsulation offers two different options:

◆ Standard

With standard encapsulation, the external spindle is created andadded to the specified external disk group and unprotected RAIDgroup. Symmetrix devices are created at the same time, allowingaccess to data that has been preserved on the external LUN.

◆ Virtual Provisioning

Virtual Provisioning encapsulation and standard encapsulationshare the fact that the external spindle is created and added to thespecified external disk group and to an unprotected RAID group.

However, they differ in that with Virtual Provisioningencapsulation, data devices (TDATs) are then created and addedto a specified thin pool. Fully, non-persistently allocated thindevices (TDEVs) are also created and bound to the pool. Extentsare allocated to the external LUN through the TDAT.

External Provisioning When using FTS to configure an external LUN, Enginuity creates aneDisk and adds it to the specified external disk group. External diskgroups are separate from disk groups containing internal physicaldisks. These start at disk group number 512.

Because RAID protection is provided by the external array, eDisks areadded to unprotected virtual RAID groups. Symmetrix devices canthen be created from the external disk group to present to users.

IMPORTANT

External provisioning should only be used with external LUNs thatcontain no data or unwanted data. These devices are VTOCed aspart of the eDisk configuration process. Any data residing on theLUN, prior to adding it as an eDisk, will be lost.

Virtual Provisioning (VP) can be configured using externalprovisioning by creating data devices (TDATs) using an external diskgroup. Other than the fact that the TDATs are created on eDisks, theprocess for configuring VP is the same.


104


FTS and data migrationFTS can be used for migration of data between VMAX arrays andexternal storage as part of a tiering or asset-management strategy.

Virtual LUN Technology (VLUN) and FAST VP are fully supportedwith FTS, allowing both features to be used for non-disruptive datamigration and mobility between storage tiers on VMAX Familyarrays and any array configured as external storage.

Since the disks are virtual and exist in an external (and possiblythird-party) disk array into which the VMAX Series array has novisibility, this FTS tier is considered the lowest performing storagetier in a FAST VP policy.

Aside from using FAST VP for automatic tiering in a virtualprovisioning FTS environment, a typical migration scenario would bewhere an external LUN has data on it that the user wishes toencapsulate so that the data can be preserved and then migrated tothe VMAX. This can be done easily and simply using SymmetrixVirtual LUN Technology.

Data can also be moved freely between internal and external storageusing Symmetrix replication technologies, such as SRDF andTimeFinder.

Non-migration use casesConsider the following non-migration use cases:

◆ FTS continues the use of external disk arrays while takingadvantage of most VMAX Enginuity features.

FTS allows organizations to continue to use existing disk arraysas additional storage capacity. Along with this, the data (whetherit is existing data on the external array or new data), can bemanaged, controlled, and monitored in the same way as nativeSymmetrix data.

Almost all of the features supported on Symmetrix devices usinginternal storage are also supported with FTS. Features likeQuality of Service (QoS), FAST VP, and Virtual Provisioning,among many others, are available to be used with externalstorage.



◆ FTS protects data on external arrays using Symmetrix local andremote replication technologies.

Local and remote replication technologies, such asTimeFinder/Clone, TimeFinder/Snap, SRDF, and OpenReplicator are all supported with Symmetrix devices using eDiskstorage. For example, if the goal is to use SRDF to replicate databetween a DMX and a CLARiiON CX4, FTS allows it to beaccomplished.

◆ FTS provides a pool of extra VMAX storage in a VirtualProvisioning environment.

Because of the ease of migration using FAST VP between theVMAX Series array and any external array configured in an FTSenvironment, an external array could conceivably be used as atemporary repository for data in case of a shortage of availablephysical disk space in the VMAX Series array.

For example, a VMAX Series array with all SATA drives couldprovide spillover for a number of other Symmetrix arrays usingoversubscribed thin pools. As production VMAX Series arrays’thin pools reach a set threshold, a percentage of the least activeallocated capacity could be pushed to the FTS tier. Whenadditional storage is added to the production VMAX Series array,and the thin pool or pools are expanded, the data pushed to theexternal array could be pulled back to the production VMAXSeries array.


106



8

This chapter introduces several management software applicationsthat provide for control and monitoring of Symmetrix featuresdescribed in Chapter 4, “EMC Open Migrator/LM,” throughChapter 6, “EMC Open Replicator for Symmetrix.” It also detailsattributes of the PPME product. The topics are:

◆ Introduction ...................................................................................... 108◆ PowerPath Migration Enabler (PPME)......................................... 109◆ Solutions Enabler ............................................................................. 114◆ Symmetrix Management Console (SMC) ..................................... 118◆ ControlCenter ................................................................................... 121

PPME andManagement Software

for Symmetrix Arrays

PPME and Management Software for Symmetrix Arrays 107

108

PPME and Management Software for Symmetrix Arrays

IntroductionThis chapter includes software that runs on a host, not in the storagearray, SAN, or appliance. “Full selection model summary” on page 26step 4, which identifies potential data migration solutions needs, andstep 6, comparing and evaluating potential migration solutions,require more detail than is provided in “PowerPath MigrationEnabler (PPME)” on page 40, which can be used to manage EMCOpen Replicator for Symmetrix detailed in Chapter 6, “EMC OpenReplicator for Symmetrix.”

Additionally, this chapter includes more details on the multipleproducts that can be used to manage Symmetrix arrays and themigration products and features detailed in Chapter 2, “Foundationand Migration Products,” through Chapter 7, “Symmetrix SoftwareUsed for Migration.”

Most migrations will be scripted and therefore would use SolutionsEnabler Command Line Interface (CLI). There are also two GraphicalUser Interface (GUI) management products:

◆ Symmetrix Management Console (SMC)

◆ EMC ControlCenter.



PowerPath Migration Enabler (PPME)PowerPath Migration Enabler (PPME) is a host-based migrationproduct that migrates data between storage systems. PPME takesadvantage of PowerPath technology and works in conjunction withanother underlying technology, such as Open Replicator,TimeFinder/Clone, or Host Copy to actually migrate the data.

PPME provides a host-based solution with virtually no impact to hostresources by utilizing array-based or SAN-based replication (exceptwhen using Host Copy). PPME benefits data migrations in threesignificant ways:

◆ Greatly reduces or eliminates application disruption due to themigration

◆ Reduces migration risks

◆ Simplifies migration operations

PowerPath Migration Enabler is independent of PowerPathMultipathing technology and does not require the use of PowerPathfor multipathing.

Benefits of using PPMEAs discussed in “Data migration” on page 24, redirecting theapplication(s) to point to the migrated data in its new location willrequire an application outage unless this is done transparently to theapplication. PPME enables this type of transparent operation so thecutover to the migrated data does not require an application outage.Depending on the host type and the use of pseudo- or native-nameddevices, this complete elimination may not always be possible.Additionally, if PowerPath 5.x is not already installed on the host, aplanned application outage must occur for the reboot necessary toinstall or upgrade PowerPath.

Even if PPME cannot entirely eliminate application outages, it greatlyminimizes them and reduces data migration risk. For example, theinterruption for installing PowerPath 5.x can be scheduled to takeplace during normal maintenance windows prior to the actualmigration process. Complex migrations will almost always requirecertain setup activities for the migration, like updating HBA drivers,to be conducted during scheduled maintenance windows when thehost will need to be rebooted. There is a great difference between this

PowerPath Migration Enabler (PPME) 109

110


type of small activity as part of a maintenance window and morerisky procedures that have to be conducted when PPME is not used.One example of a risky procedure not needed when PPME is used, isthe potentially catastrophic cutover outage where a machine is shutdown, a few configuration changes are made, and hopefully themachine comes back up without issue. With PPME, the cutover taskis fully verified before being performed and can sometimes beconducted fully online. I/O redirection allows administrators topreview deployment without committing to it. With PPME, in allcases, the data remains accessible to host applications during themigration itself.

Eliminating or even just reducing application downtime during amigration greatly simplifies the planning for large-scale migrations.Migration window flexibility is important to the administratorsthemselves as this simplifies the migration planning process. They donot have to rely on others to follow their directions, nor do they needto work off their shift when migrations need to be done. The pressureto correctly complete critical and complex migration tasks duringoutages or off hours is reduced or eliminated.

PPME greatly simplifies migration operations, hiding thecomplexities of underlying migration products and integrating withthe host. This simplification is even more important in providing acommon interface across heterogeneous hosts, eliminating the hostspecific knowledge required to perform key migration tasks. Thesimplicity that PPME brings to migration operations may even allowless skilled and less expensive staff to execute the work.

At the present time, PPME works in conjunction with fourunderlying technologies, Open Replicator for Symmetrix,TimeFinder/Clone, and Host Copy.

Note: Users should upgrade to the latest version in order to get the latestfeatures and performance improvements.



PPME with Open Replicator technical overviewThe PPME powermig command is used to interface with OpenReplicator. PPME keeps source and target devices synchronized bycloning writes on the host. Open Replicator Hot Pull does the bulkdata copying between the arrays through the SAN. Currently, onlythe hot pull operation is supported.

Source arrays can be: Symmetrix 8000 series through all DMX andSymmetrix VMAX series, VNX series, or CLARiiON CX series. Targetarrays can be any Symmetrix VMAX or DMX array with theminimum microcode level to support Open Replicator. Figure 28,illustrates an example of PPME operations with pseudo-nameddevices and Open Replicator. PPME swaps the pseudo-nameddevices in the middle so that emcpower25c points to the target deviceinstead of the source device without any change required in theapplication.

Figure 28 PPME operation with pseudo-named devices and Open Replicator

Application

emcpower25c

Data RAID 1

Application

Data RAID 1

Data RAID 5

PowerPath Filter Driver

emcpower25c emcpower37c

SAN

Configure target

pseudo devices

Application

emcpower25c

Data RAID 5

Complete migration

and remove old storage

OR Copy

ICO-IMG-000452


112


PPME TimeFinder/ClonePPME TimeFinder/Clone technology refers to the migration of databetween devices within a Symmetrix system usingTimeFinder/Clone. There are two significant use cases for thistechnology.◆ First, PPME can be used to nondisruptively migrate data from a

thick to a thin (virtually provisioned) Symmetrix device on bothSymmetrix VMAX and DMX arrays.

◆ Second, PPME can be used to nondisruptively migrate data whilechanging the RAID protection type in Symmetrix DMX systems.

PPME TimeFinder/Clone can be used when the source and target areof equal size, or when the target is larger than the source.

Note: See”Enhanced Virtual LUN technology,” on page 95 for how theSymmetrix VMAX can nondisruptively change the RAID protection typewithout the need for PPME TimeFinder/Clone.

PPME Host CopyPPME works in conjunction with the host operating system tomigrate data from the specified source logical unit to the targetlogical unit. A Host Copy migration does not use or require a directconnection between the arrays containing the source and targetlogical units. Host Copy can be used to migrate plain text data, or itcan be used to migrate data to/from an encrypted logical unit.Hostcopy migrations consume host resources similar to OpenMigrator/LM, and PPME provides parameters that controlhost-resource usage, along with operations to pause and resume ahostcopy migration. PPME Host Copy can migrate data at the LUNlevel, while Open Migrator/LM can migrate data at any level in theI/O stack.More information on PPME Host Copy can be found in the EMCPowerPath Migration Enabler Host Copy White Paper, located onhttps://support.emc.com.



powermig summaryTo help illustrate both what PPME can do and the simplicity of itscommand set, all 15 of the powermig actions are listed below:

◆ abort — Aborts the migration and returns the source to the stateit was in before the migration began.

◆ cleanup — Cleans up data on the target logical unit to preventconfusion to the application or operating system caused by twological units with identical data.

◆ commit — Commits the migration by making the target logicalunit the object for reads and writes.

◆ getHandle — Retrieves the migration handle for a migration inprogress.

◆ help — Provides a description and usage for the powermigcommands and options.

◆ info — Displays information about the specified migrationhandle or about all active migrations.

◆ query — Displays the current state of the migration.

◆ recover — Recovers to the migration state you were transitioningto when the interruption occurred.

◆ selectSource — Designates the source logical unit as the recipientof all I/O.

◆ selectTarget — Designates the target as the recipient of all I/O.

◆ setup — Defines the source and target logical units to use in themigration, and specifies the underlying technology to use inconjunction with PowerPath Migration Enabler.

◆ sync — Synchronizes the target and source logical units.

◆ throttle — Sets the speed of synchronization in progress.

◆ undoRedirect — Turns off the redirection of I/O from the sourceto the target logical unit. This command applies only when thesource uses a native device name.

◆ version — Displays the version of PowerPath Migration Enableryou are running. The version is the same as PowerPath.

In contrast, the Solutions Enabler symrcopy command used only tocontrol Open Replicator has 15 actions alone:


114


◆ create, remove, activate, recreate, restore, rename, set mode, setpace, terminate, set ceiling, list, list ceiling, query, verify, andexport.

Note: The EMC PowerPath Migration Enabler User Guide contains greaterdetail. The EMC PowerPath Migration Enabler Release Notes includes specificversion functionality for all supported host platforms. Both are available onhttps://support.emc.com.

Solutions EnablerAn EMC Solutions Enabler installation provides the host withSYMAPI, CLARAPI, and STORAPI shared libraries for use bySolutions Enabler applications, and the Symmetrix Command LineInterface (SYMCLI) for use by Storage Administrators and SystemsEngineers.

SYMCLI is a specialized library of UNIX-formatted commands thatcan be invoked one at a time. It supports single command line entriesand scripts to map and perform control operations on devices anddata objects. It also monitors device configuration and status ofdevices that make up the storage environment. The target storageenvironments are typically Symmetrix, but can be VNX orCLARiiON when you have a license and work with the mappingSRM component.

Solutions Enabler tasks can be summarized as follows, grouped bythe product guide the tasks are documented in:

Symmetrix array overall base management◆ Access Control, User Authorization, and Authentication

◆ Basic Monitoring of storage arrays, devices, and components

◆ Storage Devices and Device Grouping

◆ Statistics

◆ Gatekeepers

◆ Change Tracker

Business continuance, data replication/availability◆ Data Replication, Mirroring, Clones, Snap Copies

◆ Remote Site Mirroring or Backup

◆ Disaster Recovery



◆ Auto-provisioning, Device Masking and Array Controls:

◆ Configuration Change

◆ Fully Automated Storage Tiering (FAST)

◆ Device Pooling and Masking

◆ Path Isolation

◆ Quality of Service

◆ Performance Optimization

◆ Double Checksum

Storage Resource Management (SRM)Host Application View, Data Mapping within:

◆ Databases

◆ File Systems

◆ Logical Volume Managers

◆ Data Objects (extents)

◆ Statistics

A selection of SYMCLI commands likely to appear in a migrationscript or in support of a migration are:

◆ symapierr — Used to translate SYMAPI error code numbers intoSYMAPI error messages.

◆ symbcv — Perform BCV support operations on Symmetrix BCVdevices.

◆ symdev — Perform operations on a device given the device'sSymmetrix name.

◆ symdg — Perform operations on a device group (dg).

◆ syminq — Issues a SCSI Inquiry command on one or all devices.Can be used to obtain WWNs for Open Replicator from multiplestorage arrays.

◆ symlabel — Perform label support operations on a Symmetrixdevice. Can be used to relabel a replica of a device so two devicesvisible to the same host have different labels.

◆ symld — Perform operations on a device in a device group (dg).

◆ sympd — Perform operations on a device given the device'sphysical name (host pathname).

Solutions Enabler 115

116


◆ symreturn — Used for supplying return codes in pre-action andpost-action script files.

◆ symaccess — Administer Symmetrix Access Logix.

◆ symcg — Perform operations on an composite group (cg).

◆ symchksum — Administer checksum checks when an Oracledatabase writes data files on Symmetrix devices.

◆ symclone — Perform Clone control operations on a device groupor on a device within the device group (TimeFinder/Clone).

◆ symconfigure — Perform modifications on the Symmetrixconfiguration.

◆ symconnect — Setup or Modify Symmetrix Connection Securityfunctionality.

◆ symfast — Administer Symmetrix FAST (Fully AutomatedStorage Tiering) policies, associations, and the FAST Controller.

◆ symmask — Setup or Modify Symmetrix Device Maskingfunctionality.

◆ symmaskdb — Backup, Restore, Initialize or Show the contentsof the device masking database.

◆ symmigrate — Migrate the physical disk space associated with aSymmetrix device to a different data protection scheme, or todisks with different performance characteristics.

◆ symmir — Perform BCV control operations on a device group oron a device within the device group (TimeFinder/Mirror).

◆ symoptmz — Perform Symmetrix Optimizer control operations,including Virtual LUN migration.

◆ symqos — Perform Quality of Service operations on SymmetrixDevices, including tuning replication processing.

◆ symrcopy — Perform Symmetrix Rcopy control operations ondevices in a device file (Open Replicator).

◆ symrdf — Perform RDF control operations on a device group oron a device within the device group (SRDF).

◆ symrecover — Perform automated SRDF session recoveryoperations.

◆ symreplicate — Perform automated, consistent replication ofdata given a pre-configured SRDF/TimeFinder setup(SRDF/AR).



◆ symsnap — Perform Symmetrix Snap control operations on adevice group or on devices in a device file (TimeFinder/Snap).

◆ symstar — Perform SRDF STAR management operations(SRDF/Star).

◆ symtier — Create and manage storage tiers within a Symmetrix.

Solutions Enabler tasks can be run from any host which has at leastone Symmetrix device visible to it, which is called a gatekeeper device.Solutions Enabler commands are sent in-line through the device I/Opath to the Symmetrix and can execute any of the above tasks, onlyrestricted by security limitations configured by the storageadministrators. Most Solutions Enabler tasks can also be executedremotely from any host that can connect remotely to a host with agatekeeper device, with additional configurable security limitations.

Solutions Enabler 117

118


Symmetrix Management Console (SMC)As a small and easy-to-implement, browser-based graphicalinterface, SMC can be hosted on a small Windows, Linux, or SunOSserver and is accessible through a client Web browser from nearlyanywhere in the world.

SMC is also available as a virtual appliance for ESX V3.5 (and later)Servers in a VMware environment.

Starting with Enginuity 5874, SMC is available on the Symmetrixservice processor as part of EMC’s management integration. Becauseof this architecture, SMC is an ideal graphical complement to CLIusage and even to full EMC ControlCenter for performing basicdevice management of a Symmetrix system.

SMC is also an exceptional product for those comfortable withSYMCLI, but who might be looking for an easier graphical control oftheir Symmetrix. It is easy and quick to install. Similar to SolutionsEnabler, the SMC Server can be run from any host which has at leastone Symmetrix device visible to it. The SMC Server can also beconfigured to remotely connect to another Solutions Enabler hostwhich has at least one Symmetrix device visible to it.

With SMC, users can manage operations from device creation tovirtual provisioning to replication configuration and monitoring.And as needs change and grow, higher-level, storage resourcemanagement capabilities can easily be added using the EMCControlCenter family.

SMC’s browser interface enables users to quickly and efficientlyperform device management activities on your Symmetrix system.SMC can perform initial system discovery and configuration,including the ability to run discovery requests. The discoveryrequests will identify standard items such as arrays, logical devices,and physical drives. SMC performs traditional Symmetrixconfiguration activities like device creation, configuration, and basicvolume masking. SMC’s GUI interface is an ideal method toprovision the target environment. SMC also can be used to move andmigrate data with Symmetrix Virtual LUN technology.

Figure 29 on page 119 shows page 3 of the Optimizer Data MigrationWizard: scheduling the data move. The start of the Virtual LUNprocess can be scheduled in this SMC window, or it can startaccording to policies defined in the Optimizer program. In thisexample, an explicit start time is specified.



Figure 29 Optimizer Data Migration Wizard page 3

SMC provides dependable local and remote replication management.In today's business environments where information availability isparamount, the use of local and remote replication is extensive. Insome cases, replication activities are scheduled and run in a routinefashion day after day. However, many times ad hoc requests aremade, such as the creation and subsequent access to TimeFinderbusiness continuity volumes (BCVs).

The BCVs might be used for restore operations, application testingand development, or as part of a one-time migration process. All ofthese replication requests can be serviced quickly and easily withSMC. SMC supports all features of local and remote replication andOpen Replicator sessions. In addition, users can performmanagement tasks using the GUI, and then use the Command-LineGenerator to create text that is used to repeat the task via the CLI inthe future.

ICO-IMG-000453

Symmetrix Management Console (SMC) 119

120


SMC provides robust information and system security. Simple andefficient management of data is important, but it is only valuablewhen multiple levels of secure access to specific hosts and users isprovided. SMC provides role-based user authentication and detailedauditing to protect critical information from accidental or maliciousactivity. SMC also enables users to manage Symmetrix access controlsthat authenticate and restrict host control of Symmetrix devices. Inaddition, SMC supports iSCSI CHAP (Challenge HandshakeAuthentication Protocol) secrets and RADIUS (RemoteAuthentication Dial-In Service) server settings to provide additionalauthentication and authorization capabilities.

SMC also simplifies day-to-day Symmetrix managementresponsibilities by providing real time monitoring of Symmetrixsystems, including a comprehensive set of alerts, including physicaldrive status, that can be easily viewed and acted upon, and providescomprehensive audit records of host actions taken on the Symmetrixsystem.



ControlCenterEMC ControlCenter is used to simplify and automate key tasks, suchas discovery, monitoring, reporting, planning, and provisioning foreven the largest, most complex storage environments. ControlCentercan manage the storage, fabric, and hosts. It provides a unified viewfor multiple arrays, a single-pane-on-glass interface. Whenperforming migrations, ControlCenter’s GUI interface is an idealmethod to both provision the target environment and reconfigure theSAN.

ControlCenter is a family of products with additionally licensedfeatures. Ordering ControlCenter includes both Solutions Enablerand Symmetrix Management Console. The key SymmetrixManagement piece is the ControlCenter Symmetrix Manager. VNXseries, VNXe, CLARiiON, Celerra, and EMC Centera managementtools can be launched from within the ControlCenter framework.

The ControlCenter infrastructure includes one or more dedicatedserver hosts and agents that reside on managed hosts. With thisinfrastructure, ControlCenter is able to scale to support very largeenterprises.

ControlCenter Symmetrix ManagerKey Symmetrix Manager tasks and benefits include:

◆ Discover, monitor, and configure Symmetrix storage from a singleconsole-and automate key system management and replicationtasks.

◆ Integrated end-to-end Storage Resource Management (SRM).

◆ By automating key system management and replication tasks,you can decrease complexity, save time, and reduce overallmanagement costs.

ControlCenter 121

122


ControlCenter family add-on licensesBrief descriptions of add-on license functionality:

◆ Automated Resource Manager — Meet and exceed service-levelcommitments by fulfilling requests for storage at the applicationlevel-quickly and easily.

◆ Performance Manager — Identify and solve performance issuesthroughout the information infrastructure. Get a comprehensiveview of performance across the environment-and zero in onproblems quickly.

◆ SAN Advisor™ — Simulate changes to the SAN beforeimplementing them-avoiding problems before they happen.Proactively minimize errors and downtime.

◆ SAN Manager — Streamline and centralize management of theentire multi-vendor SAN. Manage better and faster-all from oneeasy-to-use interface.

◆ StorageScope — Get integrated asset and use reports across themulti-vendor storage infrastructure. Gain the insights needed toimprove asset utilization, simplify management, and plan for thefuture.

◆ StorageScope File Level Reporter — Extend monitoring andreporting to the file level. Know more about storage activity todetermine where to move or delete files and reclaim capacity.



Figure 30 shows using Topology View to verify host connectivity tostorage devices.

Figure 30 EMC ControlCenter topology view

ControlCenter 123

124



9

This chapter identifies multiple business issues that may affect thechoice of a migration solution. The topics are:

◆ Introduction ...................................................................................... 126◆ Budgetary factors ............................................................................. 126◆ Human resources ............................................................................. 127◆ Verification and validation ............................................................. 127◆ Application interruption................................................................. 128◆ Other factors ..................................................................................... 128◆ Transformational migrations.......................................................... 129

Business Factors

Business Factors 125

126

Business Factors

IntroductionNow, having completed all of the detail on the existing and targetenvironments and the potential data migration solutions, it is time tomove on to step 5 in “Full selection model summary” on page 26,identifying business factors that limit potential data migrationsolutions. Some of these factors have been alluded to in earlierchapters, but they are presented here at this point, just as they appearin this order in the model, narrowing the potential solutions downfrom the wider consideration before coming to this step.

Budgetary factorsAn obvious business factor is the financial budget. Key elements ofthe budget that will affect the choice of data migration solution arethe capital, project, services, and training budgets.

◆ The capital budget will limit potential options for the targetenvironment.

◆ The project budget will limit the potential options for hiredservices and temporary resources for conducting the migration.

◆ The services budget will limit the maintenance costs associatedwith the target environment.

◆ The training budget needs to be considered if training is neededto adequately support the data migration itself and the targetenvironment.

If the data migration process is going to repeated frequently, ratherthan a one time migration, budgets beyond this single project need tobe considered. It may be more economical to purchase a productwhich can be used for multiple future migrations than utilize a tool aspart of a single-use contract.

Other budgets should also be considered when the migrationprovides an opportunity to transform the IT infrastructure; moreinformation can be found in “Transformational migrations” onpage 129.


Business Factors

Human resourcesThe existing support infrastructure elements, including humanresources and service contracts, will also affect the choice of the datamigration solution.

Key human resource issues include internal staff ability as well as theexisting product knowledge of that staff. Lack of product knowledgefor the target environment or the data migration solution itself meansthat there must be sufficient training budget, and perhaps equallylimiting, time available to take training.

Existing service contracts may already include certain key supportfor the data migration, or updated/new service contracts may benecessary to support both the migration and the target environment.Outside human resources available through a service contract maysubstitute for training internal human resources.

Verification and validationAnother important set of business factor relates to the timeline andstandard verification procedures.

In order to implement a chosen data migration solution, it may benecessary to upgrade or install selected hardware and software.

Usually there will be some sort of verification procedure to validatethat these changes have not adversely affected productionapplications. Depending on the rigorousness of these validationprocedures, a solution that has fewer verification requirements mightbe preferred. Additionally, there may be validation requirementsassociated with the data migration itself, adding requirements forhost access for testing the data in the new location before switchingproduction to the new location.

Human resources 127

128

Business Factors

Application interruptionPotential data migration solutions will also vary in the ways in whichthe production application needs to be interrupted. The range ofeffect on the production application varies from zero interruption tocomplete shutdown of the application from the start of the migrationuntil it is completed and verified. Business requirements willdetermine when and for how long it is permissible to interrupt theproduction application, and therefore limit the choice of datamigration solution. It is important to remember that disruptiveinstallation and upgrade processes must be considered as potentiallyinterrupting the application in addition to the actual data migrationand cutover.

It is not just the interruption of the primary application that may be aconcern to the business. Critical applications typically have both localand remote data replication in place to limit the cost of any downtimeshould a problem occur. However, the target environment mayinitially have a copy of the data, but may not have all of the replicassynchronized. Depending on the size of the data, the time it takes forthe replication synchronization exposes the business to a largerwindow than may be acceptable for a potential downtime eventwithout the mitigating replicas in place.

Other factorsThe business factors described so far in this chapter is not exhaustive.For example, many customers have defined managementframeworks that require certain elements for new applications to fitin the environment. Depending on company policies, this may limitthe choice of data migration solution. The key point is to take the timeto try to expose these factors early in the data migration planningprocess in order to avoid more difficult to correct changes later.

In real environments, business factors are often more complexbecause lines of business may have competing priorities that do notalign.


Business Factors

Transformational migrationsA particular data migration does not exist in a vacuum. Onemigration may be part of what will be a series of data migrations,which therefore may justify implementing a more sophisticated datamigration solution that will provide for future continuousmigrations.

Data migrations can be looked at narrowly, and sometimes for simplemigrations this is very appropriate. A simple, single relocationsolution is applicable to migrations that are:

◆ Limited in scale, single open systems environments

◆ Simple replication scenarios

◆ Tolerant of risk and or outages

◆ Data movement from like to like

◆ Limited or no IT infrastructure optimization

A simple migration is tactical in nature, but sometimes this path ofleast resistance may secure short-term gains at long-term expenses.The migration may end in the inefficient use of devices, incorrectconfiguration of replication groups, or undersized volumes that willnot support future application growth. One way of looking atInformation Technology (IT) is to see it as perpetually in one of threephases:

◆ Getting ready to move to something new

◆ Moving to something new

◆ Recovering from moving to something new

Migrations provide the opportunity to transform the ITinfrastructure.

A simple relocation has small scope and little opportunity fortransformation. Consolidating servers and relocation increases thescope and opportunity for transformation. When consolidating boththe storage arrays and storage network and adding in tiering, there isa big jump in scope and transformational opportunity. Consolidatingand tiering entire data centers offers the largest scope andopportunity.

Figure 31 on page 130 illustrates that the increasing opportunity fortransformation results in increasing the scope of data migration.

Transformational migrations 129

130

Business Factors

Figure 31 Transformational data migrations

As the complexity increases with more types of applications, moreinterrelationships between data, and more scale, the stakes are raisedand there is less tolerance for outages due to the large impact of anoutage. Due to the increased service level requirements withincreasing complexity, it is usually necessary to contract professionalservices to ensure a migration that moves the data with zero defects,minimal disruption to the business, and within the plannedtimeframe.

Migration products that have added value as permanentinfrastructure solutions include:

◆ EMC VPLEX

◆ EMC RecoverPoint

◆ EMC Rainfinity

◆ EMC PPME

◆ EMC FTS


10

This chapter will bring together the earlier chapters with examples ofchoosing data migration solutions for Symmetrix storage systems.The topics are:

◆ Use cases choosing a migration solution for Symmetrix ........... 132◆ Open Replicator for Symmetrix/PPME solution use case......... 136◆ Symmetrix Remote Data Facility (SRDF) solution use case....... 141◆ Virtual LUN solution use case........................................................ 147◆ Federated Live Migration (FLM) solution use case .................... 151◆ EMC Open Migrator/LM solution use case ................................ 156

Applying the Model

Applying the Model 131

132

Applying the Model

Use cases choosing a migration solution for SymmetrixThe previous chapters of this TechBook have described the myriadfactors that influence the choice of a migration solution and details ofEMC migration solutions for Symmetrix. In order to illustrate usingthis information to determine and validate migration solutions, thefollowing four use cases are presented:

◆ Open Replicator for Symmetrix/PPME Solution

◆ Symmetrix Remote Data Facility (SRDF) Solution

◆ Enhanced Virtual LUN Solution

◆ EMC Open Migrator/LM Use Solution

Table 1 summarizes when to use the five solutions highlighted in thesimple model and Open Migrator/LM as a host-based solution.

Table 1 Five solutions summary (page 1 of 2)

Virtual LUNFederated LiveMigration SRDF

OpenReplicator/PPME Open Migrator/LM

Runs on Array: SymmetrixVMAX or DMX

Array: SymmetrixVMAX

Array: Symmetrix Array: SymmetrixVMAX or DMX

Host: Windows,UNIX, or Linux

Availability Online: applicationand data availablethroughoutmigration

Online: applicationand data availablethroughoutmigration

Online: applicationand data availableuntil planned cutover

Online or offline: nocutover outagerequired when usingPPME and PP 5.x

Online: applicationand data availableuntil plannedcutover

OS Support All Symmetrix-supported hostoperating systems

Pre-qualified stacksfor Windows, UNIX,Linux

All Symmetrix-supported hostoperating systems

All DMX-supportedopen systemsoperating systems

Windows, UNIX,Linux

Functionality • CLI and GUI• Tiering• RAID Protection

(SymmetrixVMAX only)

• Schedulingcontrol

• CLI and GUI• Smaller to larger

volumes• Hot pull• Auto-failback


volumes• Many flavors /S,

/A, DM, AR, Star,Concurrent,Cascaded, EDP& Four-Site


volumes• Push or Pull,

“Live” or BCV

• CLI(UNIX/Linux),GUI (Windows)

• Smaller tolarger volumes

• Robust andtunable


Applying the Model

For each use case, both the simple selection model and the morecomplex selection model introduced in Chapter 1, “Introduction,”will be applied. Both models are repeated below.

Systemto SystemConnectivity

• Within aSymmetrix array

• Symmetrix DMXto VMAX

• Check eLab forupdatedpre-qualifiedstacks

• Symmetrixto/fromSymmetrix only(May requireupgrade or onlysupportSRDF/DM)

• Fibre Channel,GigE, andESCON

• SymmetrixVMAX or DMXto/from anyqualified array:

• CLARiiON,HP/IBM/HDS/Sun

• Receive datafrom multiplesources

• Fibre ChannelSAN and FC/IP

• Any-to-any• Source and

target must bevisible to host

• SCSI/FibreChannel

When To Use

• Applicationavailability/transparent

• Large amountsof data

• Tiering within asingle array

• Maintain sourcedevice size

• Tunable

• Applicationstays onlinewithout hostmigrationsoftware

• Consolidationsand technologyrefreshes

• Tunable

• When SRDFalready in use orbeing added fordisaster recovery

• Large amounts ofdata

• Fast andhomogeneous

• Smaller to largervolumes

• Tunable• Need DR

immediately inplace at target

• Application muststay online(PPME)

• Large amounts ofdata

• Fast andheterogeneous

• Smaller to largervolumes

• Consolidatingmultiple arrays

• Tunable

• Quick andsimple

• Small amountsof data

• Available serverCPU cycles

• Smaller tolarger volumes

• Migrating fromone volumemanager toanother

• Need controland validation

Table 1 Five solutions summary (page 2 of 2)

Virtual LUNFederated LiveMigration SRDF

OpenReplicator/PPME Open Migrator/LM

Use cases choosing a migration solution for Symmetrix 133

134

Applying the Model

Simple selection modelTo review, the five simple selection model steps discussed in “Simpleselection model” on page 22, are:

A. If the migration is for changing the tier or data protection type ofdata within a Symmetrix VMAX array, use Virtual LUNTechnology (this can also be used for DMX arrays, except whenalso changing the protection type where PowerPath MigrationEnabler with TimeFinder/Clone can be used).

B. If the migration is a consolidation or technology refresh of entirestorage arrays that need to be transparent to the host application,use Federated Live Migration (FLM) for pre-qualified stacks.

C. If the migration is from Symmetrix-to-Symmetrix storage arrays,use SRDF.

D. If the migration involves heterogeneous storage arrays (includingsame vendor but not compatible for B), then use EMC OpenReplicator for Symmetrix.

E. If the above four choices cannot be used, then use a host-basedtool, for example, EMC PowerPath Migration Enabler with HostCopy, or EMC Open Migrator/LM.

Full selection modelTo review, the six full selection model steps discussed in “Fullselection model summary” on page 26, are:

1. Define the reasons for the data migration, clearly notingmandatory and optional objectives.

2. Inventory the existing environment identifying storage elementsthat must participate with the chosen data migration solution,and resources available to support the migration itself.

3. Inventory the target environment identifying storage elementsthat must participate with the chosen data migration solution,and resources available to support the migration itself. This stepshould include scalability considerations to ensure the targetenvironment is not obsolete too soon. Additionally, the solutionmight include adding in semi-permanent infrastructure tosupport future data migrations.


Applying the Model

4. Identify potential data migration solutions that can successfullymove the data from the existing environment to the targetenvironment.

5. Identify business factors that limit potential data migrationsolutions due to budget, human resources, and applicationoutage and verification requirements.

6. Compare and evaluate the potential data migration solutionsincluding the criteria identified in steps 1–5.

Use case information flowFirst, the use case introductory paragraph in very brief form includes:

• The reasons for the data migration (full model step 1)

• Key components of the existing environment inventory(partial step 2)

• Key components of the target environment inventory (partialstep 3)

• Key business factors that help determine the data migrationsolution (partial step 5)

Second, the simple selection model is presented with brief analysis.

Third, the full selection model is presented with brief analysis:

2 and 3. Existing and Target environment detail.

5. Key business factors that help determine the datamigration solution, presented out of order, becausesteps 4 and 6 are combined in this briefrepresentation.

4 and 6. Details of the potential migration solutions withanalysis grouped by the location where the solutionresides.

6. Alternate solution summary, briefly stating whycertain potential solutions were not selected in thiscase.

Use cases choosing a migration solution for Symmetrix 135

136

Applying the Model

Open Replicator for Symmetrix/PPME solution use caseThe customer is migrating data for a technical refresh and aconsolidation of a number of older systems into a single SymmetrixVMAX. The older systems consist of two Symmetrix DMX 1000s andfour CLARiiON CX600 arrays. The primary applications to bemigrated reside on Sun, HP and Windows hosts. EMC PowerPath 5.3is already installed on the hosts. The application cannot tolerate anydowntime, some slowdown if absolutely necessary at selected times,might be tolerable.

Simple selection modelBriefly apply the simple selection model:

A. The migration involves multiple arrays, eliminating EMC VirtualLUN technology as a potential solution.

B. All arrays are connected by a SAN, and the target is an EMCSymmetrix VMAX array that supports Federated Live Migration(FLM). However current FLM support could only migrate fromthe DMX arrays, but not the CLARiiON arrays, therefore apartial, but not a full solution.

C. Some of the migration is from older EMC storage arrays to newerEMC arrays, but also migration from older CLARiiON arrays isrequired. Therefore, EMC SRDF can only offer a partial, but not afull solution. Additionally, SRDF can copy the datanondisruptively, but a host interruption would still be required topoint to the data in its new location.

D. All arrays are connected by a SAN, and the target is an EMCSymmetrix VMAX array that supports EMC Open Replicator forSymmetrix. Would need to use in concert with PowerPathMigration Enabler (PPME) for zero application downtime.

E. Choice D seems effective, so it is not necessary to use a host-basedtool.

The next step is to add some of the complexity that may validate thissolution selection or require the selection of an alternate solution.


Applying the Model

Existing and target environmentsApply steps 2–3 listed in “Full selection model” on page 134:

◆ Existing DMX 1000s

• Running Enginuity 5671

• FA directors connect to host through the SAN, all 16 FC portsin use, very small margin, if any, of capacity remains if alsoused for migration

◆ Existing CLARiiON CX600s

• Maximum Fibre Channel ports in use, very small margin, ifany, of capacity remains if also used for migration

◆ Existing SAN switches:

• Some existing spare ports available for new connections

• No more expansion capacity (new blades) available foradditional ports

◆ Application Hosts:

• Fibre Channel HBAs for connection through SAN to Storage

• HBA firmware is compatible with new Symmetrix VMAX

• PowerPath 5.3 is installed, which means that the use ofPowerPath Migration Enabler (PPME) or Federated LiveMigration (FLM) is possible for zero host downtime

• Little excess CPU capacity or I/O bandwidth available formigration use

◆ New Symmetrix VMAX:

• Available FA director ports exceed current application need,plenty left for independent migration activity

• Available FA bandwidth exceeds existing application needs,more than sufficient for planned response time improvements

• Running Enginuity 5874 (would need to upgrade to 5875 forFLM support)

• RDF Migration supported back to 5671

• Open Replicator supported for up to 1024 simultaneoussessions

Open Replicator for Symmetrix/PPME solution use case 137

138

Applying the Model

◆ New SAN switches:

• Will gain higher bandwidth (4 Gb/s) and larger number ofport connections to Symmetrix VMAX

• Compatible with Symmetrix VMAX

• Will interoperate with older existing SAN switches

Business factors“Full selection model” on page 134, step 5:

◆ Capital budget sufficient for new Symmetrix VMAX and SANswitches

◆ Project budget includes migration software and contractedservices costs

◆ Wary of complex migration mixing strategies, prefer singlemigration solution, with single validation methodology

◆ Wish to expand LUN sizes to allow for application growth as partof the migration

◆ Application not tolerable of downtime

Symmetrix VMAX migration software“Full selection model” on page 134, steps 4 and 6:

◆ Federated Live Migration (FLM):

• Upgrade to Enginuity 5875 and Solutions Enabler 7.2 requiredfor VMAX FLM support

• FLM pre-qualified stack applies for DMX to VMAX change

• FLM does not currently support migrating from CLARiiONarrays, so no single solution as desired

◆ Symmetrix Remote Data Facility (SRDF):

• SRDF/DM (Data Mobility) supported for 5671 to 5874;cutover before data is fully copied is not possible without fullSRDF support

• Application downtime required for cutover to target deviceson the VMAX

• Cannot be used for migrating CLARiiON arrays either, so nosingle solution as desired


Applying the Model

◆ EMC Open Replicator for Symmetrix:

• Sufficient FA ports and bandwidth for fast migration

• Migration sets fit within 1024 session limit

• Support for migration to larger targets meets business factor

• Hot pull with donor update:

– Utilizes sufficient Symmetrix VMAX bandwidth instead oflimited old array bandwidth for the migration

– Provides recovery mechanism should some problem occurduring the migration

EMC host migration and migration management software“Full selection model” on page 134, steps 4 and 6:

◆ PowerPath Migration Enabler (PPME):

• Enables zero interruption to the application providing amethod to switch the application target transparently to theapplication

• Includes a CLI interface for scripting the migration usingpowermig

• Migration Enabler mirrors I/O to keep the source and targetLUNs synchronized throughout the migration process

• Changes the ASCII name of the target logical unit's disk labelto one that matches the logical unit type of the target array asneed for Solaris

• Includes powerformat to safely update disk-labelinformation, preserve partition definitions and data, and makenewly available disk capacity (larger target) available for use

• Installs of PPME can also be done without applicationinterruption because PowerPath 5.3 is already installed

• Supports CLARiiON and Symmetrix source arrays and aSymmetrix VMAX target array

◆ Solutions Enabler 7.1:

• Includes a CLI interface for verifying the Open Replicatorsetup configuration using symsan

• Includes a CLI interface for scripting the Open Replicatormigration using the symrcopy command, though powermigwill mostly be used in this case

Open Replicator for Symmetrix/PPME solution use case 139

140

Applying the Model

◆ Symmetrix Management Console

• Includes a GUI interface for symsan and Open Replicatorfunctionality, may be used for setup, monitoring ortroubleshooting

◆ EMC Ionix ControlCenter SAN Manager

• Provides consistent generic interface for reconfiguring andmanaging the SAN

Alternate solution summary“Full selection model” on page 134, step 6:

Multiple host migration solutionsEvaluate potential host migration solutions:

◆ Meets criteria of ability to move data from heterogeneous oldarrays to new Symmetrix VMAX

◆ Requires host resource capacity for the migration that is notavailable

◆ PowerPath Migration Enabler (PPME) hostcopy solution avoidsapplication interruption.

◆ EMC Open Migrator/LM install on Windows would require atleast one short application interruption.


Applying the Model

Symmetrix Remote Data Facility (SRDF) solution use caseThe customer is migrating the entire data center from its originallocation to a new location. The old location will continue to functionas a remote backup center for the new primary data center. ExistingSymmetrix DMX-4 arrays will remain in the backup center while theapplications will be consolidated on new VMware ESX servers andnew Symmetrix VMAX arrays. Old HDS array data used for ancillaryapplications is to be removed from the data center; data needs to bemigrated to new Symmetrix VMAX arrays. Since host servers andancillary application support are also moving to the new data center,it is acceptable to interrupt the applications for the changeover, butthe entire changeover, must be completed within the scheduled fourhour window.


A. The migration involves multiple arrays, eliminating EMC VirtualLUN technology as a potential solution. Since the move includessome consolidation, Virtual LUN technology can be used at alater time for moving applications to the appropriate tiers in theSymmetrix VMAX arrays if adjustment is necessary. FullyAutomated Storage Tiering (FAST) could also be used tocontinuously analyze the utilization (busy rate) of Symmetrixarray devices, and automatically move data volumes betweentiers to fine tune performance and reduce costs.

B. Although the principle migration is from EMC SymmetrixDMX-4 to Symmetrix VMAX arrays, seemingly fitting currentFLM support, the plan is for new servers. Therefore, FLM’snondisruptive nature having the host see both the source andtarget devices is not a fit.

C. The principle migration is from EMC Symmetrix DMX-4 toSymmetrix VMAX arrays, implying that SRDF is likely the bestsolution. However, there is a plan to also migrate some ancillaryapplications from HDS storage to the new Symmetrix VMAXarrays. Additionally, SRDF will be continued to be used after themigration for remote backup.

Symmetrix Remote Data Facility (SRDF) solution use case 141

142

Applying the Model

D. All arrays are connected by a SAN in the old data center, and anew SAN will be configured in the new data center. It was notplanned to interconnect the SANs due to distance limitations, butthe plan includes leased telecommunication services to supportremote mirroring using SRDF. EMC Open Replicator forSymmetrix can be used to move data from the HDS arrays to theEMC Symmetrix arrays, but may need to stage the data to thelocal DMX arrays first, unless the SAN is interconnected betweenthe sites.

E. Choice C seems effective, so not necessary to use a host-basedtool.

The simple selection model identifies both SRDF and EMC OpenReplicator as potential solutions, but neither alone can offer acomplete solution. Next step is to consider the complexity more fullyand determine the best solution for this particular case.

Existing and target environments“Full selection model” on page 134, steps 2–3:◆ Existing DMX-4s:


• Currently using 4 FA director boards

• Room to expand with 2 additional director boards to allow forconfiguration of SRDF RA ports

• Open Replicator:

– Supported for up to 1024 simultaneous sessions– Supports HDS array model as remote, and DMX as control

◆ HDS array• SAN connected storage

◆ Existing SAN switches:• Only a couple of existing spare ports, desired to keep as spare

ports for FA connections to application hosts

◆ Existing Hosts:• Fibre Channel HBAs for connection through SAN to Storage

• HBA firmware compatibility with new Symmetrix VMAX notan issue as new consolidated servers using VMware will bepositioned in the new data center


Applying the Model


◆ Symmetrix VMAX:• Available FA director ports exceed current application need,

plenty left for independent migration activity if necessary


• Available RA director ports sufficient to meet remotereplication fault tolerance and performance requirements

• Running Enginuity 5874:

– Full RDF connectivity supported back to 5773– Open Replicator supported for up to 1024 simultaneous

sessions• New telecommunication links needed to support remote

SRDF mirroring

• Distance between sites is > 200km


◆ Capital budget sufficient for outfitting entire new datacenter

◆ Project budget includes licenses for new software and contractedservices costs (EMC Open Replicator for Symmetrix not includedin projected budget)

◆ Major concern is for flawless cutover within the four hourchangeover window

◆ Plan to expand LVM VGs by adding additional LVs to provide forapplication data growth



• Full SRDF supported for 5874 to 5773

• Can use SRDF adaptive copy mode for initial transfer prior tocutover


144

Applying the Model

• Can use dynamic swap to switch the R1/R2 relationship atcutover

• Can use SRDF/A for the long distance remote mirroring fordisaster recovery

• Cannot be used for migrating HDS arrays directly, but can beused for remotely mirroring data first migrated to the remoteDMX-4s


• HDS arrays are remote and no plan to interconnect the SANs

Symmetrix DMX-4 migration software“Full selection model” on page 134, steps 4 and 6:

◆ Symmetrix Remote Data Facility (SRDF), Full SRDF supported for5773 to 5874 (more detail listed under SRDF in “SymmetrixVMAX migration software” on page 143)


• With the addition of two additional director boards for SRDF,also gain 2 more FA ports to support migrated applicationsand bandwidth for fast migration

• Migration sets fit within the 1024 session limit

• Movement of ancillary applications will not occur on the samecutover date

• Will likely use hot pull with donor update for old siteapplication migration to Symmetrix, before moving theapplication to new datacenter

• See Open Replicator use case above for more information



• Includes a CLI interface, symrdf command to support:

– Dynamic SRDF configuration– Setting adaptive copy mode– Dynamic Swap


Applying the Model

– Setting asynchronous mode and monitoring SRDF/A• Includes a CLI interface for verifying the Open Replicator

setup configuration using the symsan

• Includes a CLI interface for scripting the Open Replicatormigration using the symrcopy command

• Includes a CLI interface symconfigure to provision the newSymmetrix VMAX arrays

◆ Symmetrix Management Console:

• Includes a GUI interface for symsan and Open Replicatorfunctionality, may be used for setup, monitoring ortroubleshooting

• Includes a GUI interface for SRDF functionality, may be usedfor configuration, swap, and planned ongoing monitoring ofSRDF/A

• Includes a GUI interface to provision new Symmetrix VMAXarrays

◆ ControlCenter Symmetrix and SAN Managers:

• Will be implemented in the new data center for single-paneview of the entire infrastructure including the remoteSymmetrix arrays at the backup site

• Can be used to provision the new Symmetrix VMAX arraysand configure the new SAN


EMC Open ReplicatorEvaluate potential EMC Open Replicator solution:

◆ Not in the current project budget

◆ Meets criteria of ability to move data from HDS arrays to DMX-4

◆ Meets criteria of ability to move data from DMX-4 arrays to newSymmetrix VMAX arrays

• Requires SAN connectivity just for the migration which is notpart of the customer long-term plan

◆ So not a good fit to be the sole migration solution


146

Applying the Model


◆ Meets criteria of ability to move data from HDS arrays to DMX-4s

◆ Requires host resource capacity for the migration that will impactcurrent performance

◆ Requires heavy load on application network resources to movethe data between the data centers

◆ So in general a poor fit to be the main migration solution


Applying the Model

Virtual LUN solution use caseIn this use case, the customer wants to migrate a large amount of datawithin a Symmetrix array from a higher to lower tier as part of theirInformation Lifecycle Management (ILM) strategy. In theirSymmetrix VMAX, the source data resides on Tier-1 Fibre Channel15k rpm RAID-1 protected disks and the desire is to move the data toTier-2 SATA-II RAID-6 protected disks. The application cannottolerate any downtime.


A. The migration is within a Symmetrix VMAX array indicating thatVirtual LUN technology may provide the best migration solution.The customer may want to consider implementation of FullyAutomated Storage Tiering (FAST) to automate a dynamic tieringpolicy based on user defined policies.

B. Federated Live Migration (FLM) cannot be used within a singleSymmetrix array.

C. SRDF cannot be used within a single Symmetrix array.

D. EMC Open Replicator for Symmetrix cannot be used within asingle Symmetrix array.

E. Choice A seems effective, so it is not necessary to use a host-basedtool.


Existing and target environmentsApply “Full selection model” on page 134, steps 2–3

◆ Existing:

• Running Enginuity 5874• Data on Tier-1 Fibre Channel 15k rpm RAID-1 protected disks• Space available to add additional Tier-2 disk capacity

◆ Application hosts

Virtual LUN solution use case 147

148

Applying the Model

◆ PowerPath version 5.3 is installed, meaning that the use ofPowerPath Migration Enabler (PPME) is possible for zero hostdowntime

◆ Little excess host CPU capacity or I/O bandwidth available formigration use

◆ Target data location:

• Tier-2 1 TB SATA-II drives added to the configuration of theexisting Symmetrix VMAX


1. Capital budget sufficient for new Tier-2 disk storage.

2. Project budget includes a Virtual LUN license but no contractedservices costs.

3. Current allocated space for the application is nearly fully utilized;additional storage is needed very soon.

4. ILM strategy: new Tier-2 storage must be lower cost and due toaging of the data, will be accessed less often and therefore lowerperformance is acceptable.

5. Application cannot tolerate downtime.

6. Local replication (TimeFinder/Clone) already in use to create acopy used for data mining operations.


◆ Virtual LUN technology

• Fully transparent, nondisruptive data migration betweenstorage tiers and between RAID protection schemes

◆ TimeFinder/Clone

• Available for local data replication within a single Symmetrixarray

• Nondisruptive data copying between storage tiers andbetween RAID protection schemes


Applying the Model

• No provision for nondisruptive application redirection to datain its new location without PPME

• Reuse of existing data mining local replication copy wouldrequire full synchronization


◆ Solutions Enabler 7.1:• Includes a CLI interface symmigrate command for scripting

the Virtual LUN migration• Includes a CLI interface symconfigure command to

provision the new Tier-2 storage◆ Symmetrix Management Console:

• Includes a GUI interface for Virtual LUN migrationfunctionality, and may be used for setup, monitoring, ortroubleshooting

• Includes a GUI interface to provision new Tier-2 storage◆ PowerPath Migration Enabler (PPME):

• Supports TimeFinder/Clone as the underlying technology tocopy data between tiers within the Symmetrix array

• Enables zero interruption to the application, providing amethod to switch the application target transparently to theapplication

• Includes a CLI interface for scripting the migration usingpowermig

• Migration Enabler mirrors I/O to keep the source and targetLUNs synchronized throughout the migration process

• Installation of PPME can also be done without applicationinterruption because PowerPath 5.3 is already installed

Alternate solution summary“Full selection model” on page 134, step 6.

Virtual LUN solution use case 149

150

Applying the Model

PPME TimeFinder/CloneEvaluate a potential PPME TimeFinder/Clone solution:

◆ Meets criteria of lowering costs, migrating from Tier-1 to Tier-2data, and changing the RAID protection scheme within theSymmetrix array

◆ Meets criteria for zero application interruption◆ Greater performance impact than a Virtual LUN solution◆ Requires interruption and full synchronization of an existing local

TimeFinder/Clone replica


◆ Meets criteria of lowering costs, migrating from Tier-1 to Tier-2data, and changing the RAID protection scheme within theSymmetrix array


◆ Only PPME hostcopy meets the criteria for zero applicationinterruption


Applying the Model

Federated Live Migration (FLM) solution use caseThe customer is migrating data for a technical refresh and aconsolidation of two Symmetrix DMX 1000s and two SymmetrixDMX 2000s into a single Symmetrix VMAX. The primary applicationsto be migrated reside on Windows 2003, Linux RHEL 4 U2, and AIX5.3 hosts. PowerPath 4.5 is already installed on the hosts. Theapplication cannot tolerate any downtime. Some slowdown, ifabsolutely necessary at selected times, might be tolerable.


A. The migration involves multiple arrays, eliminating EMC VirtualLUN technology as a potential solution.

B. All arrays are connected by a SAN, and the target is an EMCSymmetrix VMAX array that supports Federated Live Migration(FLM) and all stacks appear to fit the pre-qualified conditions.The requirement for a nondisruptive migration also seems tomake FLM an excellent fit.

C. The migration is from older EMC storage arrays to newer EMCarrays. Although, SRDF can copy the data nondisruptively, a hostinterruption would be required to point to the data in its newlocation.

D. All arrays are connected by a SAN, and the target is an EMCSymmetrix VMAX array that supports EMC Open Replicator forSymmetrix. Would need to use in concert with PowerPathMigration Enabler (PPME) for zero application downtime.However, PPME requires PowerPath 5.x, so an application outagewould be required to upgrade PowerPath on the hosts.

E. Choice B seems effective, so it is not necessary to use a host-basedtool.


Federated Live Migration (FLM) solution use case 151

152

Applying the Model

Existing and target environmentsApply “Full selection model” on page 134, steps 2–3:

◆ Existing DMX 1000s:


• FA directors connect to host through the SAN. All 16 FC portsare in use, and a very small margin, if any, of capacity remainsif also used for migration

◆ Existing SAN switches:

• Some existing spare ports available for new connections

• Empty slots available for capacity expansion if needed


• Fibre Channel HBAs for connection through SAN to Storage

• HBA firmware is compatible with new Symmetrix VMAX

• PowerPath 4.5 is installed, which means that the use ofFederated Live Migration (FLM) is possible for zero hostdowntime. Use of PowerPath Migration Enabler (PPME)would require a short outage for the installation of PowerPath5.x, which is a requirement for PPME.

• Conform to FLM pre-qualified stacks


◆ New Symmetrix VMAX:

• Available FA director ports exceed current application need,plenty left for independent migration activity


• Running Enginuity 5875 including support for FLM

• FLM/Open Replicator supported for up to 1024 simultaneoussessions

• RDF Migration supported back to 5671

◆ SAN switches:

• Compatible with Symmetrix VMAX


Applying the Model


◆ Capital budget sufficient for new Symmetrix VMAX

◆ Wary of complex migration mixing strategies, prefer singlemigration solution, with single validation methodology

◆ Application not tolerable of downtime

◆ End of year processing is pushing for a fast migration withlimited time for extensive planning, also with no scheduledmaintenance windows until after year end


◆ Federated Live Migration (FLM):

• Upgrade to Solutions Enabler 7.2 required for Enginuity 5875and VMAX FLM support can be performed on a control hostwithout any requirement for a maintenance window onapplication hosts

• FLM pre-qualified stack applies for all of the plannedmigration, mitigating the need for remediation duringmaintenance windows, allowing for a quicker start to theactual migration

• Sufficient FA ports and bandwidth for fast migration

• Migration sets fit within 1024 session limit

• Hot pull with donor update:

– Utilizes sufficient Symmetrix VMAX bandwidth instead oflimited old array bandwidth for the migration

– Provides recovery mechanism should some problem occurduring the migration

• Completely nondisruptive, not even requiring any migrationoperations to be run on the application hosts, except a SCSIbus scan to discover the new VMAX device paths


• SRDF/DM (Data Mobility) supported for 5671 to 5875,cutover before data is fully copied not possible without fullSRDF support


154

Applying the Model

• Application downtime required for cutover to target deviceson the VMAX


• Fits with mostly the same criteria as using Open Replicatortechnology for FLM

• Not completely nondisruptive, because of the need to upgradePowerPath to 5.x in order to install PPME


◆ PowerPath Migration Enabler (PPME):

• Could be run to enable a nondisruptive Open Replicatormigration, after a short application interruption to upgrade toPowerPath 5.x enabling PPME installation


• Only required on a control host for an FLM migration

• Includes a CLI interface for verifying the FLM setupconfiguration using symsan

• Includes a CLI interface for scripting the FLM migration

◆ Symmetrix Management Console

• Includes a GUI interface for symsan and FLM functionality,may be used for setup, monitoring, or troubleshooting

◆ ControlCenter SAN Manager

• Provides consistent generic interface for reconfiguring andmanaging the SAN


Applying the Model

Alternate solution summary“Full selection model” on page 134, model step 6:

SRDFEvaluate a potential EMC SRDF solution:

◆ Meets criteria of ability to move data from old DMX arrays to newSymmetrix VMAX

◆ Fails to meet criteria for zero application interruption due to needto redirect applications to use new VMAX devices

Open Replicator with PPMEEvaluate a potential EMC Open Replicator with PPME solution:


◆ Fails to meet criteria for zero application interruption due to needto upgrade PowerPath version to enable installation of PPME




◆ PowerPath Migration Enabler (PPME) Host Copy solution avoidsapplication interruption during migration cutover, but wouldhave short interruption for required PowerPath upgrade to 5.x

◆ EMC Open Migrator/LM install on Windows would require atleast one short application interruption


156

Applying the Model

EMC Open Migrator/LM solution use caseIn this use case, there is a show-stopping condition. As part of thismigration the customer is introducing a replacement Logical VolumeManager (LVM). In the simple model, this rules out any device basedcopy method including Virtual LUN technology, SRDF, OpenReplicator, and PPME Host Copy because they copy at a level belowthe LVM, and in this case it is necessary to copy at the LVM level orabove.

Host-based migration lowest common denominatorWhen it comes to choosing a migration methodology, host-based datamigration technologies are often considered the lowest commondenominator. If circumstances rule out the possibility of usingSAN/Array-based migration tools, one will often fall back to ahost-based data migration. This could involve either block-level copytechnologies, such as EMC Open Migrator, PPME Host Copy, AcronisTrue Image, and Host LVM Mirroring, or file system-leveltechnologies, such as EMCopy, rsync, and Robocopy. Althoughhost-based migrations will work in almost all cases, performing datamigrations at the host layer typically requires more manual effort,takes more time, and has a higher risk of human error thanalternative methods that operate at the LUN/SAN/Array layer.

There are many factors that can rule out the possibility for an Arrayor SAN-based data migration, such as:

◆ Migrating from internal disks, JBOD, or SCSI-attached externalarrays, sources that do not have the potential for SANconnectivity.

◆ Migrating from SAN-based arrays that do not interoperate or arenot supported with SAN Copy, Open Replicator, RecoverPoint, Iand so on.

◆ Migrating to target LUNs that are smaller than source LUNs(shrinking targets).

◆ Customer does not want to purchase a product such as VPLEX, orRecoverPoint for a one-time migration.


Applying the Model

◆ Customer will not allow EMC to temporarily install RecoverPointfor a one-time migration, for example, a secure governmentfacility may not permit EMC to remove hardware from a site onceit has been installed.

◆ Lack of SAN connectivity, for example, not enough free ports onthe Arrays or the SAN to dedicate to replication.

When one encounters these limiting factors, it is often necessary touse the lowest common denominator.


A. Not applicable.

B. Not applicable.

C. Not applicable.

D. Not applicable.

E. The above three choices cannot be used, therefore a host-basedtool like EMC Open Migrator/LM must be used.

Existing and target environments“Full selection model” on page 134, steps 2–3:

◆ Storage Array details do not matter, only requirement is that themigration host must see both the source and target devices

◆ SAN switch details do not matter, only requirement is that themigration host must see both the source and target devices


• HP-UX v11.11

• hfs

• HP LVM

• PowerPath


◆ New Host Environment:

• Switching to Veritas VxVM LVM

EMC Open Migrator/LM solution use case 157

158

Applying the Model


◆ Very small migration budget as part of upgrade.

◆ Large concern about migration affect on production application,wants ability to limit the migration, especially during busy times.

◆ Can tolerate a lengthy migration period.

◆ Individual applications can tolerate short interruptions duringscheduled maintenance windows.

EMC host migration software“Full selection model” on page 134, steps 4 and 6:

◆ EMC Open Migrator/LM for UNIX:

• Price does not exceed budget.

• Includes CLI interface for scripting the migration usingstormigrate and sessions for treating sets of source/targetvolumes as a single entity.

• Includes user tunable migration rate while active.

• Installed, operated, and uninstalled without system reboot.

• Enables data migration with a single application disruption,taken at the end of the migration task to redirect theapplication to the new target.

• Once synchronization completes, mirroring continues tomaintain the source and target data images in sync, until thedisruption can be scheduled at the next maintenance window.


Multiple host and LVM migration tools that are LVM aware couldmeet the basic data movement requirement, but often lack theresiliency and sophistication of Open Migrator/LM includingrecovery across reboots and tunable migration.


Glossary

This glossary contains terms related to disk storage subsystems.Many of these terms are used in this manual.

Aadministrator A person responsible for administrative tasks such as access

authorization and content management. Administrators can alsogrant levels of authority to users.

agent A software entity that runs on endpoints and provides managementcapability for other hardware or software. An example is an SNMPagent. An agent has the ability to spawn other processes.

allocate To assign a resource for use in performing a specific task.

allocated storage The space that is allocated to volumes, but not assigned.

arbitrated loop A Fibre Channel interconnection technology that allows up to 126participating node ports and one participating fabric port tocommunicate. See also ”Fibre Channel” and “loop topology.”

audit To review and examine the activities of a data processing systemmainly to test the adequacy and effectiveness of procedures for datasecurity and data accuracy.

authority The right to access objects, resources, or functions.

authorizationchecking

The action of determining whether a user is permitted access to aRACF-protected resource.


160

Glossary

automated operations Automated procedures to replace or simplify actions of operators inboth systems and network operations.

Bbackup A copy of computer data that is used to recreate data that has been

lost, mislaid, corrupted, or erased. The act of creating a copy ofcomputer data that can be used to recreate data that has been lost,mislaid, corrupted or erased.

back-up The process of creating a copy of data to ensure against accidentalloss.

bandwidth A measure of the data transfer rate of a transmission channel.

bridge Facilitates communication with LANs, SANs, and networks withdissimilar protocols.

Business ContinuityPlanning

An enterprise wide planning process which creates detailedprocedures to be used in the case of a disaster. Business ContinuityPlans take into account processes, people, facilities, systems, andexternal elements.

Ccache A random access electronic storage in selected storage controls used

to retain frequently used data for faster access by the channel.

CAS See ”content- addressable storage (CAS).”

channel director The component in the Symmetrix subsystem that interfaces betweenthe host channels and data storage. It transfers data between thechannel and cache.

CIFS See ”Common Internet File System (CIFS).”

CKD Count Key Data, a data recording format employing self-definingrecord formats in which each record is represented by a count areathat identifies the record and specifies its format, an optional key areathat may be used to identify the data area contents, and a data areathat contains the user data for the record. CKD can also refer to a setof channel commands that are accepted by a device that employs theCKD recording format.


Glossary

CLI See ”Command Line Interface (CLI).”

client A function that requests services from a server, and makes themavailable to the user. A term used in an environment to identify amachine that uses the resources of the network. See also”client-server.”

client authentication The verification of a client in secure communications where theidentity of a server or browser (client) with whom you want tocommunicate is discovered. A sender's authenticity is demonstratedby the digital certificate issued to the sender.

client-server In TCP/IP, the model of interaction in distributed data processing inwhich a program at one site sends a request to a program at anothersite and awaits a response. The requesting program is called a client;the answering program is called a server.

client-serverrelationship

Any process that provides resources to other processes on a networkis a server. Any process that employs these resources is a client. Amachine can run client and server processes at the same time.

Command LineInterface (CLI)

A mechanism for interacting with a computer operating system orsoftware by typing commands to perform a given task, referred to as“entering” a command: the system waits for the user to conclude thesubmitting of the text command by pressing the Enter key. Acommand line interpreter then receives, analyses, and launches therequested command. Upon completion, the command usually returnsoutput to the user in the form of text lines on the CLI.

Common Internet FileSystem (CIFS)

An open cross-platform mechanism for client systems to request fileservices from server systems over a network. It is based on the SMBprotocol widely used by PCs and workstations running a widevariety of operating systems.

content-addressable storage

(CAS)

A mechanism for storing information that can be retrieved based onits content, not its storage location. It is typically used for high-speedstorage and retrieval of fixed content, such as documents stored forcompliance with government regulations.


162

Glossary

connection In TCP/IP, the path between two protocol applications that providesreliable data stream delivery service. In Internet communications, aconnection extends from a TCP application on one system to a TCPapplication on another system.

consistency group A consistency group is a user-defined group of devices that can spanmultiple Symmetrix systems and, if needed, provide consistencyprotection. Consistency means that the devices within the group actin unison to preserve dependent-write consistency of a database thatmay be distributed across multiple Symmetrix systems or multipleRDF groups within a single Symmetrix.

consistent copy A copy of data entity (for example, a logical volume) that contains thecontents of the entire data entity from a single instant in time.

console A user interface to a server. That part of a computer used forcommunication between the operator or user and the computer.

count-key-data (CKD) A data recording format employing self-defining record formats inwhich each record is represented by a count area that identifies therecord and specifies its format, an optional key area that may be usedto identify the data area contents, and a data area that contains theuser data for the record. CKD can also refer to a set of channelcommands that are accepted by a device that employs the CKDrecording format.

Ddaemon A program that runs unattended to perform a standard service.

DAS See ”direct-attached storage (DAS)” and compare with “networkattached storage.”

data availability Access to any and all user data by the application.

data integrity The condition that exists as long as accidental or intentionaldestruction, alteration, or loss of data does not occur.

data migration The one time movement of data from source to target, where the datawill subsequently only be accessed at the target.

default A value, attribute, or option that is assumed when no alternative isspecified by the user.


Glossary

delayed fast write There is no room in cache for the data presented by the writeoperation.

dependent-writeconsistency

A data state where data integrity is guaranteed by dependent-writeI/Os embedded in application logic. Database management systemsare good examples of applications that utilize the dependent-writeconsistency strategy. Database management systems must deviseprotection against abnormal termination in order to successfullyrecover from one. The most common technique used is to guaranteethat a dependent write cannot be issued until a predecessor write hascompleted. Typically the dependent write is a data or index writewhile the predecessor write is a write to the log. Because the write tothe log must be completed prior to issuing the dependent write, theapplication thread is synchronous to the log write (that is, it waits forthat write to complete prior to continuing). The result of this kind ofstrategy is a dependent-write consistent database.

dependent-write I/O An I/O that cannot be issued until a related predecessor I/O hascompleted. Most applications, and in particular databasemanagement systems (DBMS), have embedded dependent-writelogic to ensure data integrity in the event of a failure in the host orserver processor, software, storage subsystem, or if an environmentalpower failure occurs. See also ”dependent-write consistency.”

destage The process of writing data from cache to a disk device.

device type The general name for a kind of device; for example, standard, BCV,VDEV, or Clone.

direct-attachedstorage (DAS)

Computer storage that is directly attached to one computer or serverand is not, without special support, directly accessible to other ones.

director The component in the Symmetrix subsystem that allows theSymmetrix to transfer data between the host channels and diskdevices.

directory (1) A type of file containing the names and controlling informationfor other files or other directories. Directories can containsubdirectories, which can contain subdirectories of their own. (2) Afile that contains directory entries. No two directory entries in thesame directory can have the same name. (POSIX.1). (3) A file thatpoints to files and to other directories.


164

Glossary

disaster recovery The process of restoring a previous copy of the data and applyinglogs or other necessary processes to that copy to bring it to a knownpoint of consistency.

disaster restart The process of restarting dependent write-consistent copies of dataand applications, using the implicit application of DBMS recoverylogs during DBMS initialization to bring the data and application to atransactional point of consistency. If a database is shut downnormally, the process of getting to a point of consistency duringrestart requires minimal work. If the database abnormally terminates,the restart process takes longer depending on the number and size ofin-flight transactions at the time of termination. An image of thedatabase is created using the EMC consistency technology while thedatabase is running, and is done without conditioning the database.The database, is in a dependent-write consistent data state, which issimilar to that created by a local power failure. This is also known asa DBMS restartable image. The restart of this image transforms it to atransactionally-consistent data state by completing committedtransactions and rolling back uncommitted transactions during thenormal database initialization process.

disk director The component in the Symmetrix subsystem that interfaces betweencache and the disk devices.

EE_Port An E_Port is an inter-switch expansion port that connects to the

E_Port of another Fibre Channel switch, in order to build a largerswitched fabric.

enterprise network A geographically dispersed network under the backing of oneorganization.

Enterprise SystemsConnection (ESCON)

A set of products and services that provides a dynamically connectedenvironment using optical cables as a transmission medium.

ESCON Enterprise Systems Connection architecture; a set of IBM and vendorproducts that connect mainframe computers with each other andwith attached storage, locally attached workstations, and otherdevices using optical fiber technology and dynamically modifiableswitches called ESCON directors.


Glossary

FF_Port A fabric port that is not loop-capable. It is used to connect an N_Port

to a switch.

fabric Fibre Channel employs a fabric to connect devices. A fabric can be assimple as a single cable connecting two devices. The term is oftenused to describe a more complex network utilizing hubs, switches,and gateways.

fast write In Symmetrix, a write operation at cache speed that does not requireimmediate transfer of data to disk. The data is written directly tocache and is available for later destaging.

FBA Fixed Block Architecture, disk device data storage format usingfixed-size data blocks.

FC See ”Fibre Channel.”

FCIP See ”Fibre Channel over IP.”

FCP See ”Fibre Channel protocol.”

FCS See ”Fibre Channel standard.”

fiber optic The medium and the technology associated with the transmission ofinformation along a glass or plastic wire or fiber.

Fibre Channel A technology for transmitting data between computer devices at adata rate of up to 4 Gbps. It is especially suited for connectingcomputer servers to shared storage devices and for interconnectingstorage controllers and drives.

Fibre ChannelArbitrated Loop

A reference to the FC-AL standard, a shared gigabit media for up to127 nodes, one of which can be attached to a switch fabric. See also”arbitrated loop” and “loop topology.”

Fibre Channel over IP Fibre Channel over IP is defined as a tunneling protocol forconnecting geographically distributed Fibre Channel SANstransparently over IP networks.

Fibre Channelprotocol

The serial SCSI command protocol used on Fibre Channel networks.


166

Glossary

Fibre Channelstandard

An ANSI standard for a computer peripheral interface. The I/Ointerface defines a protocol for communication over a serial interfacethat configures attached units to a communication fabric. Refer toANSI X3.230-199x.

FICON An I/O interface based on the Fibre Channel architecture. In this newinterface, the ESCON protocols have been mapped to the FC-4 layer,that is, the Upper Level Protocol layer, of the Fibre Channel Protocol.It is used in the S/390 and z/Series environments.

file system An individual file system on a host. This is the smallest unit that canbe monitored and extended. Policy values defined at this leveloverride those that might be defined at higher levels.

FL_Port A fabric port that is loop capable. It is used to connect NL_Ports tothe switch in a loop configuration.

flash drive A storage device that uses flash memory rather than conventionalspinning platters to store data. Flash drives tend to physically imitateconventional hard drives in size, shape, and interface so that theymay act as a replacement for hard drives. Note that with nothingbeing mechanically driven in a flash drive, the name is actually amisnomer. The motivation to call it a “drive” comes from the fact thatit is serving the purpose of a part that has traditionally beenmechanically driven.

Ggatekeeper A small logical volume on a Symmetrix storage subsystem used to

pass commands from a host to the Symmetrix storage subsystem.Gatekeeper devices are configured on standard Symmetrix disks.

gateway In the SAN environment, a gateway connects two or more differentremote SANs with each other.

A gateway can also be a server on which a gateway component runs.

gateway node A node that is an interface between networks.

Gigabit Ethernet Technologies for transmitting Ethernet frames at a rate of a gigabitper second, as defined by the IEEE 802.3-2005 standard.

gigabyte (GB) 109 bytes.


Glossary

Graphical UserInterface (GUI)

A type of user interface which allows people to interact with acomputer and computer-controlled devices. It presents graphicalicons, used in conjunction with text, labels, or text navigation to fullyrepresent the information and actions available to a user. But insteadof offering only text menus, or requiring typed commands, theactions are usually performed through direct manipulation of thegraphical elements.

GUI See ”Command Line Interface (CLI).”

Hhardware Physical equipment, as opposed to the computer program or method

of use; for example, mechanical, magnetic, electrical, or electronicdevices. See also ”software.”

hardware zoning The members of a hardware zone are based on the physical ports onthe fabric switch. Zoning can be implemented in the followingconfigurations: one to one, one to many, and many to many.

HBA See ”host bus adapter.”

head and diskassembly (HDA)

A field replaceable unit in the Symmetrix subsystem containing thedisk and actuator.

highly parallel Refers to multiple systems operating in parallel, each of which canhave multiple processors.

host Any system that has at least one Internet address associated with it. Ahost with multiple network interfaces can have multiple Internetaddresses associated with it. This is also referred to as a server.

host bus adapter A Fibre Channel HBA connection that allows a workstation to attachto the SAN network.

host not ready In this state, the volume responds not ready to the host for all readand write operations to that volume.

hub A Fibre Channel device that connects up to 126 nodes into a logicalloop. All connected nodes share the bandwidth of this one logicalloop. Hubs automatically recognize an active node and insert thenode into the loop. A node that fails or is powered off is automaticallyremoved from the loop.


168

Glossary

hypervolume A user-defined storage device allocated within a Symmetrix physicaldisk.

hypervolumeextension

The ability to define more than one logical volume on a singlephysical disk device making use of its full formatted capacity. Theselogical volumes are user-selectable in size. The minimum volume sizeis one cylinder and the maximum size depends on the disk devicecapacity and the emulation mode selected.

II/O device An addressable input/output unit, such as a disk device.

iFCP See ”Internet Fibre Channel protocol (IFCP).”

I/O device An addressable input/output unit, such as a disk device.

Internet Fibre Channelprotocol (IFCP)

The Internet Fibre Channel Protocol specification defines iFCP as agateway-to-gateway protocol for the implementation of a FibreChannel fabric in which TCP/IP switching and routing elementsreplace Fibre Channel components.

internet protocol(IP)

A protocol used to route data from its source to its destination in anInternet environment.

internet SCSI(iSCSI)

Internet SCSI encapsulates SCSI commands into TCP packets;therefore enabling the transport of I/O block data over IP networks.

iVTOC Instant Volume Table of Contents (VTOC) is a method to initialize avolume as containing all zeros by marking it, delaying any actualwriting of zeros until actually used.

JJBOD Just a Bunch Of Disks. A disk group configured without the disk

redundancy of the RAID arrangement. When configured as JBOD,each disk in the disk group is a rank in itself.

KK Kilobyte, 1024 bytes.


Glossary

Llocal volumes Volumes that reside on an Symmetrix system but do not participate in

SRDF activity.

locality of reference Locality of reference in SRDF/A environments improves theefficiency of the SRDF network links. Even if there are multiple dataupdates (repeated writes) in the same cycle, the systems send the dataacross the SRDF links only once.

locally not ready If the local primary SRDF volume fails, the host continues torecognize that volume as available for read/write operations as allreads and writes continue uninterrupted with the secondary (target,R2) volume in that remotely mirrored pair.

logical unit number(LUN)

A volume identifier that is unique among all storage servers. TheLUN is synonymous with a physical disk drive or a SCSI device. Fordisk subsystems such as the IBM Enterprise Storage Server, a LUN isa logical disk drive (a unit of storage on the SAN which is availablefor assignment or unassignment to a host server). The LUNs areprovided by the storage devices attached to the SAN.

logical volume A user-defined storage device.

long miss Requested data is not in cache and is not in the process of beingfetched.

longitude redundancycode (LRC)

Exclusive OR (XOR) of the accumulated bytes in the data record.

loop topology The available bandwidth is shared with all the nodes connected to theloop. If a node fails or is not powered on, the loop is out of operation.This can be corrected using a hub. A hub opens the loop when a newnode is connected and closes it when a node disconnects. See ”FibreChannel Arbitrated Loop” and ”arbitrated loop.”

LUN masking Allows or blocks access to the storage devices on the SAN. Intelligentdisk subsystems provide this kind of masking.

LUN See ”logical unit number (LUN).”

MMB Megabyte, 106 bytes.


170

Glossary

media The disk surface on which data is stored.

microprocessor A processor implemented on one or a small number of chips.

mirrored pair A logical volume with all data recorded twice, once on each of twodifferent physical devices.

mirroring The Symmetrix maintains two identical copies of a designatedvolume on separate disks. Each volume automatically updatesduring a write operation. If one disk device fails, Symmetrixautomatically uses the other disk device.

mirroring (RAID 1) The highest level of performance and availability for allmission-critical and business-critical applications by maintaining aduplicate copy of a volume on two disk drives.

multiprocessing The simultaneous execution of two or more computer programs orsequences of instructions. See also ”parallel processing.”

multiprocessor (MP) A CPC that can be physically partitioned to form two operatingprocessor complexes.

multisessionconsistency (MSC)

mode

Beginning with Enginuity 5x71 for mainframe and open systems,SRDF/A is supported in configurations where there are multipleprimary Symmetrix systems or multiple primary Symmetrix SRDFgroups connected to multiple secondary Symmetrix systems ormultiple secondary Symmetrix SRDF groups. This is referred to asSRDF/A Multi-Session Consistency or SRDF/A MSC. SRDF/A MSCconfigurations can also support mixed open systems and mainframedata controlled within the same SRDF/A MSC session.

NN_Port A node port. A Fibre Channel-defined hardware entity at the end of a

link which provides the mechanisms necessary to transportinformation units to or from another node.

NAS See ”network attached storage.”

network attachedstorage

A NAS device is attached to a TCP/IP-based network (LAN orWAN), and accessed using CIFS and NFS-specialized I/O protocolsfor file access and file sharing.


Glossary

Network File System(NFS)

The Network File System (NFS) is a client/server application that letsa computer user view and optionally store and update file on aremote computer as though they were on the user's own computer.

network topology A physical arrangement of nodes and interconnectingcommunication links in networks based on application requirementsand geographical distribution of users.

NL_Port A node loop port. A node port that supports arbitrated loop devices.

Oopen system A system whose characteristics comply with standards made

available throughout the industry, and therefore can be connected toother systems that comply with the same standards.

operating system (OS) Software that controls the execution of programs and that mayprovide services such as resource allocation, scheduling,input/output control, and data management. Although operatingsystems are predominantly software, partial hardwareimplementations are possible.

orphan data Data that occurs between the last, safe backup for a recovery systemand the time when the application system experiences a disaster. Thisdata is lost when either the application system becomes available foruse, or when the recovery system is used in place of the applicationsystem.

Pparallel processing The simultaneous processing of units of work by many servers. The

units of work can be either transactions or subdivisions of large unitsof work (batch). See also ”highly parallel.”

password A unique string of characters known to a computer system and to auser, who must specify the character string to gain access to a systemand to the information stored within it.

permanent link loss If SRDF/A experiences a permanent link loss, it drops all of thedevices on the link to not ready state. This results in all data in theactive and inactive primary Symmetrix cycles (capture and transmitdelta sets) being changed from write pending for the remote mirror toowed to the remote mirror. In addition, any new write I/Os on the


172

Glossary

primary Symmetrix system result in tracks being marked owed to theremote mirror. All of these tracks are owed to the secondarySymmetrix once the links are restored.

point of consistency A point in time to which data can be restored and recovered orrestarted and maintain integrity for all data and applications.

port An endpoint for communication between applications, generallyreferring to a logical connection. A port provides queues for sendingand receiving data. Each port has a port number for identification.When the port number is combined with an Internet address, it iscalled a socket address.

port zoning In Fibre Channel environments, the grouping together of multipleports to form a virtual private storage network. Ports that aremembers of a group or zone can communicate with each other butare isolated from ports in other zones. See also ”LUN masking” and”zoning.”

primary SRDF volumes A volume that contains production data that is mirrored in a differentSymmetrix system. Primary volumes are also referred to as source orR1 volumes. Updates to a primary volume are automaticallymirrored to a secondary volume in the remote Symmetrix system. Seealso ”secondary SRDF volumes.”

printer A device that writes output data from a system on paper or othermedia.

protocol The set of rules governing the operation of functional units of acommunication system if communication is to take place. Protocolscan determine low-level details of machine-to-machine interfaces,such as the order in which bits from a byte are sent. They can alsodetermine high-level exchanges between application programs, suchas file transfer.

RRAID Redundant array of inexpensive or independent disks. A method of

configuring multiple disk drives in a storage subsystem for highavailability and high performance.

RAID 0 A protection method where data is striped across several disks toincrease performance. Unless combined with RAID 1, does notnatively provide protection from data loss due to drive failure. See


Glossary

also ”RAID 10.”

RAID 1 A protection method that provides the highest level of performanceand availability for all mission-critical and business-criticalapplications by maintaining a duplicate copy of a volume on two diskdrives. See also ”mirroring.”

RAID 5 A protection method that provides high performance with automaticstriping across hypervolumes. Lost hypervolumes are regeneratedfrom remaining members. RAID 5 is configured in (3+1) and (7+1)groups. RAID 5 technology stripes data and distributes parity blocksacross all the disk drives in the RAID group.

RAID 6 A protection method that supports the ability to rebuild data in theevent that two drives within the RAID group fail.

RAID 10 A protection method that combines RAID 1 and RAID 0; used inmainframe environments.

read access Permission to read information.

read hit Data requested by the read operation is in cache.

read miss Data requested by the read operation is not in cache.

ready volume A state indicating the volume is available for read/write operations.

recovery The process of rebuilding data after it has been damaged ordestroyed, often by using a backup copy of the data or by reapplyingtransactions recorded in a log.

recovery system A system used in place of a primary application system that is nolonger available for use. Data from the application system must beavailable for use on the recovery system. This is usuallyaccomplished through backup and recovery techniques, or throughvarious DASD copying techniques, such as remote copy.

remote operations Operation of remote sites from a host system.

remotely not ready If primary SRDF volumes are remotely not ready, write updates willnot propagate to the secondary volumes. Changes to the primaryvolumes are marked invalid as owed to the secondary volumes.


174

Glossary

reserve capacityenhancement

SRDF/A Reserve Capacity enhances the ability of SRDF/A tomaintain an operational state when encountering network resourceconstraints that would have previously suspended SRDF/Aoperations. With SRDF/A Reserve Capacity functions enabled,additional resource allocation can be applied to address temporaryworkload peaks, periods of network congestion, or even transientnetwork outages.

restore A process that reinstates a prior copy of the data.

resynchronization A track image copy from the primary volume to the secondaryvolume of only the tracks which have changed since the volume waslast in duplex mode.

rolling disaster A series of events that lead up to a complete disaster. For example,the loss of a communication link occurs prior to a site failure. Mostdisasters are rolling disasters; their duration may be onlymilliseconds or up to hours.

SSAN See ”storage area network.”

SATA See ”Serial Advanced Technology Attachment (SATA).”

scrubbing The process of reading, checking the error correction bits, and writingcorrected data back to the source.

SCSI adapter Card in the Symmetrix subsystem that provides the physical interfacebetween the disk director and the disk devices.

secondary SRDFvolumes

Volumes that contains a mirrored copy of data from a primary SRDFvolume. Secondary volumes are also referred to as target or R2volumes. See also ”primary SRDF volumes.”

semi-synchronousmode

Used mainly for an extended distance SRDF solution,semi-synchronous mode allows the primary and secondary volumesto be out of synchronization by one write I/O operation. Data mustbe successfully stored in the Symmetrix system containing theprimary volume before an acknowledgement is sent to the local host.This mode is not supported for FICON or at Enginuity 5772 and later.


Glossary

Serial AdvancedTechnology

Attachment (SATA)

A computer bus primarily designed for transfer of data between acomputer and mass storage devices such as hard disk drives.

server A program running on a mainframe, workstation, or file server thatprovides shared services. This is also referred to as a host.

shared storage Storage within a storage facility that is configured such that multiplehomogeneous or divergent hosts can concurrently access the storage.The storage has a uniform appearance to all hosts. The host programsthat access the storage must have a common model for theinformation on a storage device. Programs must be designed tohandle the effects of concurrent access.

short miss Requested data is not in cache, but is in the process of being fetched.

simple networkmanagement

protocol (SNMP)

A protocol designed to give a user the capability to remotely managea computer network by polling and setting terminal values andmonitoring network events.

small computersystem interface

(SCSI)

An ANSI standard for a logical interface to computer peripherals andfor a computer peripheral interface. The interface utilizes a SCSIlogical protocol over an I/O interface that configures attached targetsand initiators in a multi-drop bus topology.

SNMP agent An implementation of a network management application which isresident on a managed system. Each node that is to be monitored ormanaged by an SNMP manager in a TCP/IP network, must have anSNMP agent resident. The agent receives requests to either retrieve ormodify management information by referencing MIB objects. MIBobjects are referenced by the agent whenever a valid request from anSNMP manager is received. See also ”simple network managementprotocol (SNMP).”

SNMP manager A managing system that executes a managing application or suite ofapplications. These applications depend on MIB objects forinformation that resides on the managed system.

SNMP trap A message that is originated by an agent application to alert amanaging application of the occurrence of an event.

SNMP See ”simple network management protocol (SNMP).”


176

Glossary

software zoning Is implemented within the Simple Name Server (SNS) running insidethe fabric switch. When using software zoning, the members of thezone can be defined with a node WWN, port WWN, or physical portnumber. Usually the zoning software also allows you to createsymbolic names for the zone members and for the zones themselves.

software (1) All or part of the programs, procedures, rules, and associateddocumentation of a data processing system. (2) A set of programs,procedures, and, possibly, associated documentation concerned withthe operation of a data processing system. For example, compilers,library routines, manuals, circuit diagrams. See also ”hardware.”

SQL Structure Query Language.

SRDF group SRDF groups define relationships between Symmetrix systems. AnSRDF group is a set of SRDF director port connections configured tocommunicate with another set of SRDF director ports in anotherSymmetrix system. Logical volumes (devices) are assigned to SRDFgroups.

SRDF link One end-to-end SRDF connection between a given pair of Symmetrixsystems.

stage The process of writing data from a disk device to cache.

storage administrator A person in the data processing center who is responsible fordefining, implementing, and maintaining storage managementpolicies.

storage area network A managed, high-speed network that enables any-to-anyinterconnection of heterogeneous servers and storage systems.

storage control unit The component in the Symmetrix system that connects Symmetrix tothe host channels. It performs channel commands and communicateswith the disk directors and cache. See also ”CAS.”

string A series of connected disk devices sharing the same disk director.

switch A component with multiple entry and exit points or ports thatprovide dynamic connection between any two of these points.


Glossary

switch topology A switch allows multiple concurrent connections between nodes.There can be two types of switches; circuit switches and frameswitches. Circuit switches establish a dedicated connection betweentwo nodes.

Frame switches route frames between nodes and establish theconnection only when needed. A switch can handle all protocols.

TTCP See ”transmission control protocol.”

TCP/IP Transmission Control Protocol/Internet Protocol.

topology An interconnection scheme that allows multiple Fibre Channel portsto communicate. For example, point-to-point, arbitrated loop, andswitched fabric are all Fibre Channel topologies.

transaction A unit of work performed by one or more transaction programs,involving a specific set of input data and initiating a specific processor job.

transactionalconsistency

Transactional consistency is a DBMS state where all in-flighttransactions are either completed or rolled back.

transmission controlprotocol

A communications protocol used in the Internet and in any networkthat follows the Internet Engineering Task Force (IETF) standards forInternetwork protocol. TCP provides a reliable host-to-host protocolbetween hosts in packet-switched communications networks and ininterconnected systems of such networks. It uses the Internet Protocol(IP) as the underlying protocol.

Vvirtual storage (1) The storage space that can be regarded as addressable main

storage by the user of a computer system in which virtual addressesare mapped into real addresses. The size of virtual storage is limitedby the addressing scheme of the computer system and by the amountof auxiliary storage available, not by the actual number of mainstorage locations. (2) An addressing scheme that allows external diskstorage to appear as main storage.


178

Glossary

virtualization A technique for hiding the physical characteristics of computingresources from the way in which another function interacts withthose resources.

volume A general term referring to a storage device. In the Symmetrixsubsystem, a volume corresponds to single disk device.

Wwait state Synonymous with waiting time.

waiting time (1) The condition of a task that depends on one or more events inorder to enter the ready condition. (2) The condition of a processingunit when all operations are suspended.

WAN Wide area network.

wave divisionmultiplexing

WDM allows the simultaneous transmission of a number of datastreams over the same physical fiber cable, each using a differentoptical wavelength. WDM receives incoming optical signals frommany sources (Fibre Channel, IP, ESCON, FICON) which it convertsto electrical signals, then assigns them a specific wavelength (orlambdas) of light and retransmits them on that wavelength. Thismethod relies on the large number of wavelengths available withinthe light spectrum. Coarse WDM (CWDM) and Dense WDM(DWDM) are based on the same methodology as WDM enablingmore data streams over the same physical fiber.

WDM See ”wave division multiplexing.”

web browser A software application which enables a user to display and interactwith text, images, videos, music and other information typicallylocated on a Web page at a website on the World Wide Web or a localarea network. Web browsers communicate with Web serversprimarily using HTTP (hypertext transfer protocol) to submitinformation to Web servers as well as fetch Web pages from them.

world wide name A unique number assigned to Fibre Channel devices (including hostsand adapter ports). It is analogous to a MAC address on a networkcard.

write hit There is room in cache for the data presented by the write operation.


Glossary

write miss There is no room in cache for the data presented by the writeoperation.

WWN See ”web browser.”

Zz/OS A widely-used operating system for the IBM zSeries mainframe

computers that uses 64-bit real storage.

zoning In Fibre Channel environments, zoning allows for finer segmentationof the switched fabric. Zoning can be used to instigate a barrierbetween different environments. Ports that are members of a zone cancommunicate with each other but are isolated from ports in otherzones. Zoning can be implemented in two ways: hardware zoningand software zoning. See also ”hardware zoning” and “softwarezoning.”


180

Glossary


Index

Aallocation

relation to capacity 36application

I/O redirection with PPME 109storage redirection 40

Bbusiness factors

in full selection model 26

Ccapacity

excess and constrained 36

Ddata growth

reason for moving data 30data migration

business factors 126Open Migrator scenario 48Open Replicator scenario 77PPME with Open Replicator scenario 111SRDF scenario 68transformational 129Virtual LUN scenario 91

EEMC ControlCenter

add-on license functionality 122detail 121

existing environmentin full selection model 26, 36

FFAST

see Fully Automated Storage TieringFederated Live Migration

description 81nondisruptive migration 82

Federated Tiered Storage 22, 100full selection model

defined 26iterative process 26relationship to chapters 26

Fully Automated Storage Tieringdetail 98in simple model 22in use case 147Symmetrix Optimizer 92

Hhost

information to collect 40migration management software 108

host-based migrationlowest common denominator 156

II/O stack 38Information Lifecycle Management (ILM)

reason for moving data 33Virtual LUN Technology as an enabler 91


182

Index

Invistaexploration needed 140, 155

Mmigration

project steps 20migration transparency

virtualization 24

OOpen Migrator

detail 48in simple selection model 22summary table 132use case 156

Open Replicatorcapability summary 72data migration scenario 77definitions 73in simple selection model 22Interactions with SRDF 76summary table 132use case 136with PPME 111

Pperformance improvements

reason for moving data 31Symmetrix Optimizer 31

PowerPath Migration Enabler (PPME)detail 109use case 136, 151with Open Replicator overview 111

Rreasons for data migration

in full selection model 26reasons for moving data

eight categories 30

Ssimple selection model

defined 22Solutions Enabler

detail 114SRDF

adaptive copy mode 60data migration scenario 68device migration 69Dual-Site SRDF Migration Service 70family product definitions 57Four-site SRDF data migration 69in simple selection model 22Interactions with Open Replicator 76summary table 132use case 141

storage arrayinformation needed 43

storage networkinformation to collect 42

storage resource management mapping 39Symmetrix

Virtual Provisioning 84Symmetrix Management Console (SMC)

detail 118Symmetrix Manager

key tasks 121Symmetrix Optimizer

balancing logical disk placement 31Virtual LUN technology 93

Ttarget environment

in full selection model 26, 36technical refresh

reason for moving data 30thin devices

and replication products 85definition 85

tiered storageinformation per tier 44reason for moving data 32

TimeFinderdata migration scenarios 89family product definitions 87thick to thin migration 89

transformational migration 129


Index

Uuse case

information flow 135utilization

rate 36

VVirtual LUN

summary table 132

use case 147Virtual LUN technology 93

in simple selection model 22, 134Virtual Provisioning

space reclamation after migration 86Virtual Provisioning 84virtualization

broad definition 24method for migration transparency 24


184

Index


choosing a data migration solution for emc symmetrix · pdf filechoosing a data migration...

Documents