october/november 2010 | vol. 3 no. 6 virtualizationreview...

36
+ Q&A with Pano Logic CTO Aly Orady Review: vSphere 4.1 Packs an Incremental Punch How-To Guy: Configuring vSphere 4.1 Active Directory Authentication October/November 2010 | Vol. 3 | No. 6 VirtualizationReview.com Redmond is looking to streamline the application process via its Windows Azure platform, but it can’t ignore server virtualization or VDI if it wants to overtake VMware. Microsoft: The Great Server

Upload: others

Post on 09-Aug-2020

8 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

+Q&A with Pano Logic CTO Aly Orady

Review: vSphere 4.1 Packs an Incremental Punch

How-To Guy: Confi guring vSphere 4.1 Active Directory Authentication

October/November 2010 | Vol. 3 | No. 6

VirtualizationReview.com

Redmond is looking to streamline the application process via its Windows Azure platform, but it can’t ignore server virtualization or VDI if it wants to overtake VMware.

Microsoft: The Great Server

Page 2: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

Implement a merger in yourserver room.Consolidation. Businesses of all sizes are doing it every day with virtualization.

By concentrating the power of many servers into fewer physical ones,

virtualization optimizes your existing machines to save you on time, storage

space, and power and cooling costs. It’s a great business deal, and CDW is

just the business partner to help you get in on it. Our expert Virtualization

Assessment Team uses proven tools to determine the best virtualization

solution for your investment. And with our extensive experience, we can

implement our top-of-the-line solutions with as little downtime as possible.

From design to implementation, beginning to end, at CDW, we bring it all

together to get your servers back in business.

©2010 CDW LLC

Take our assessment to see if virtualization is right for you.

CDW.com/virtualization | 800.399.4CDW

Page 3: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

“Microsoft is trying to change the conversation away from virtualization as an infrastructure technology to virtualization as simply an enabler for application delivery. I think that takes them back to their strengths.”Dave Bartoletti, Senior Analyst, The Taneja Group

VirtualizationReview.com | Virtualization Review | October/November 2010 | 1

OCTOBER/NOVEMBER 2010 | VIRTUALIZATION REVIEW | VOL. 3, NO. 6

VISIT VIRTUALIZATIONREVIEW.COM

contentsRegardless of what an organization uses its virtual lab for, almost everyone has found that there are certain challenges associated with creating and maintaining a virtual lab.

14

2 Online TOC31 Ad/Edit Index

UPFRONT6 Review vSphere 4.1 Packs an Incremental Punch

10 Q&A: An Interview with Pano Logic CTO Aly Orady To Be or Not To Be a True Zero Client

COV

ER IL

LUST

RAT

ION

FR

OM

SH

UTT

ERST

OCK

25

COLUMNS4 Hoard: BRUCE HOARD Make the IT People Smile

29 How-To Guy: DAVID DAVIS

Confi guring the New vSphere 4.1 Active Directory Authentication

30 Everyday Virtualization: RICK VANOVER

Creating a Virtualization Service Catalog

32 Take 5: BRUCE HOARD Impressions of Microsoft

FEATURES14 Microsoft: The Great Server Redmond is looking to streamline the application process via its Windows Azure Platform, but it can’t ignore server virtualization or VDI if it wants to overtake VMware.

20 Back to the Mainframe: x86 Server Virtualization on Big Iron x86 virtual servers are put on notice.

25 Provisioning and Managing Virtual Labs Simplifi ed virtual lab management software allows even users to create their own testing environments.20

Page 4: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

Citrix, Cisco Partner on Desktop Virtualization After Citrix Systems Inc. announced it was acquiring lifecycle toolmaker VMLogix, the company disclosed a technology collaboration deal with Cisco Systems Inc.

The two companies have created a packaged solution that promises to help organizations deploy desktops more easily. The solution combines the Cisco “Unifi ed Computing System” with Citrix XenDesktop. Read the entire story online. VirtualizationReview.com/Citrix1010

Study Says Virtualization ‘Staggering’ IT DepartmentsA recent survey conducted by Refl ex Systems and fi lled out by 300 enterprise IT managers says IT shops are “staggering” under the heavy load of virtual machines and the applications running on them. Bruce Hoard discusses the survey results and how they can help you plan your virtualization strategy. VirtualizationReview.com/Hoard1010

2 | October/November 2010 | Virtualization Review | VirtualizationReview.com

what’s [email protected]

VIRTUALIZATIONREVIEW.COMOCTOBER/NOVEMBER 2010 ■ VOL. 3 ■ NO. 6

EDITORIAL STAFF

Vice President, Editorial Director Doug Barney

Editor in Chief Bruce Hoard

Managing Editor Wendy Gonchar

Associate Managing Editor Katrina Carrasco

Associate Managing Editor Marti Longworth

COLUMNISTS

David Davis

Rick Vanover

ART STAFF

Creative Director Scott Shultz

Art Director Brad Zerbel

ONLINE/DIGITAL MEDIA

Editor, VirtualizationReview.com Michael Domingo

Director, Online Media Becky Nagel

Site Administrator Shane Lee

Designer Rodrigo Muñoz

President Henry Allain

Vice President, Publishing Matt N. Morollo

Director, Marketing Michele Imgrund

Online Marketing Director Tracy S. Cook

President & Chief Executive Offi cer Neal Vitale

Senior Vice President & Chief Financial Offi cer Richard Vitale

Executive Vice President Michael J. Valenti

Senior Vice President, Abraham M. Langer Audience Development & Digital Media

Vice President, Christopher M. Coates Finance & Administration

Vice President, Erik A. Lindgren Information Technology & Application Developement

Vice President, Carmel McDonagh Attendee Marketing

Vice President, Event Operations David F. Myers

Chairman of the Board Jeff rey S. Klein

Reaching the Staff

Staff may be reached via e-mail, telephone, fax, or mail.A list of editors and contact information is also available

online at VirtualizationReview.com.

E-mail: To e-mail any member of the staff , please use thefollowing form: [email protected]

Framingham Offi ce (weekdays, 9:00 a.m. – 5:00 p.m. ET)Telephone 508-875-6644; Fax 508-875-6633

600 Worcester Road, Suite 204, Framingham, MA 01702

Irvine Offi ce (weekdays, 9:00 a.m. – 5:00 p.m. PT)Telephone 949-265-1520; Fax 949-265-1528

16261 Laguna Canyon Road, Suite 130, Irvine, CA 92618

Corporate Offi ce (weekdays, 8:30 a.m. – 5:30 p.m. PT)Telephone 818-814-5200; Fax 818-734-1522

9201 Oakdale Avenue, Suite 101, Chatsworth, CA 91311

The opinions expressed within the articles and other contents herein do not necessarily express those of the publisher.

Virtualization Housecleaning 101Everyday Virtualization blogger Rick Vanover provides a quick checklist on how to keep your system running at top speed—focusing on the human factor—including: • Reconfi gure drive arrays: If you want to reformat a drive array at a diff erent RAID level, the previous datastore step may create a good time to correct this.• Permission and role reconciliation: Check that the current roles, active administrators, permissions and group setup are as expected.Read the entire list here: VirtualizationReview.com/Vanover1010.

Read Editor in Chief Bruce Hoard’s blog at VirtualizationReview.com/ Hoard. Topics recently covered include SunGuard Availability Services, the recent purchase of AVIcode by Microsoft and the importance of user rights management. Stay on top of your game—check The Hoard Facts daily for fresh insights into the rapidly evolving virtualization market.

The Hoard Facts

VirtualizationReview.com • Redmondmag.com • RCPmag.com • RedDevNews.comVisualStudioMagazine.com • MCPmag.com • CertCities.com • TCPmag.com • ENTmag.com

RedmondEvents.com • ADTmag.com • ESJ.com

Page 5: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

irtualization off ers a unique value proposition, with the opportunity to reduce costs while also

increasing service levels. Nowhere is this more apparent than with data protection. Of course, you

have to have the right tools to turn the opportunity into reality. Veeam Software has made enhanced

data protection and disaster recovery a reality for more than 12,000 customers. And with Veeam vPower™, Veeam is taking backup to a whole new level, using virtualization to do things never before possible.

NEW Veeam Backup & Replication v5Veeam has developed new patent-pending technology in Veeam Backup & Replication™ v5 for VMware that lets you run a VM directly from a backup fi le. We call it vPower™, and with it you can:

• Restore a failed VM from backup in a matter of minutes• Recover individual items from any virtualized application • Verify the recoverability of every backup

Veeam vPower ensures that your backups are good, slashes recovery time, and eliminates the need for expensive backup options—making data protection better, faster and cheaper.

Instant VM RecoveryWith vPower, you can run a VM directly from a compressed and deduplicated backup fi le. This ground-breaking technology eliminates the need to extract the backup—you simply start the VM from the backup. So if a VM goes down, you can restart it on any host in a matter of minutes.

It’s like having a “temporary spare” for your VMs.

To restore full service, simply use Storage vMotion to migrate the VM to production storage. Or just replicate the VM from the backup fi le to a production datastore—with 2-in-1 backup and replication from Veeam, you can replicate any time.

U-AIR: Universal Application-Item RecoveryvPower also enables quick recovery of individual objects from any virtualized application. We call it UAIR™— universal application-item recovery—and it’s a brand-new solution to the age-old problem of what to do when users accidentally delete important e-mails or scripts incorrectly update records.

U-AIR addresses the limitations of existing object-level recovery methods. U-AIR is:

• Inexpensive: doesn’t require additional agents, backups or software licenses

• Universal: works with any virtualized application and the application’s native utilities and permissions

• Durable: not tied to application internals so is easy to maintain and works seamlessly with new application patches and releases

SureBackup Recovery Verifi cationvPower also makes it possible to test your backups to ensure you can actually recover. Veeam SureBackup automates the entire verification process using available resources in your existing production or test environment. So you can verify the recoverability of every backup of every virtual machine every time, without additional hardware or staff. It’s something no other backup solution offers, physical or virtual.

Everything You’d Expect—And MoreSince its first release, Veeam Backup & Replication has been the leader in data protection and disaster recovery for VMware, delivering the most innovation, value and customer satisfaction.

vPower extends the advantage that Veeam customers enjoy. With vPower, you reduce costs while increasing service levels. After all, isn’t that what virtualization is all about?

VRVPADVERTISEMENT

Data Protection in a Virtual World: Better, Faster, Cheaper

VIRTUALIZATION REVIEW VENDOR PROFILE

For more information about vPower and Veeam Backup &Replication v5, please visit www.veeam.com/vpower

RPORTOReliability

SureBackup InstantRestore SmartCDP

Recovery verification NEW!

Proven FastSCPTM engine

Advanced VSS integration

Instant VM recovery NEW!

U-AIRTM NEW!(universal application-item recovery)

Instant file-level recovery (Windows and Linux)

2-in-1: backup & replication

Replicate on-site for HA, off-site for DR

Fraction of the cost of traditional CDP

Veeam vPower : Virtualization-Powered Data Protection

TM TM TM

TMTM

0810vrm_VRVP.Veamm.final.indd 1 8/2/10 2:07 PM

Page 6: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

4 | October/November 2010 | Virtualization Review | VirtualizationReview.com

ED NOTE

Bruce Hoard

PHO

TO B

Y TERR

Y POW

ELL

HOARD

Make the IT People SmileIt could be said that Unidesk President and CEO Don Bulens is in a rut: starting up a new company, cashing out, starting up another company, cashing out … and so it goes. He’s done it fi ve times so far, if you count launching the Notes business out of Lotus.

Of course, as ruts go, this one has been pleasantly profi table, and the entrepreneurial Bulens looks just fi ne for a guy who’s been putting in beau coup hours on and off for more years than he probably cares to remember.

For the record, he was a major player in the start-ups of Radnet, a collaborative Web applications company; Trellix, a Web-publishing tools fi rm founded by Dan Bricklin; EqualLogic, a high-end enterprise storage company; and, most recently, Unidesk, which is in the business of desktop virtualization management. (By the way, Dell acquired EqualLogic for a cool $1.4 billion in 2008, so you do the math.)

Bulens talks of three challenges he’s faced each time he shifts into start-up mode. The fi rst is engineering a product that’s “good enough” to satisfy at least a subset of what might be a long-term audience. The second is establishing a go-to-market strategy that calibrates pricing-to-value, distribution, customer relationships and the proper balance between professional services and ease of use. Addressing the third challenge, Bulens says: “This is probably what becomes the most essential of all—a lifelong journey around making sure the team is comprised of the kinds of people who can help the company be successful, and who are passionate, purposeful and aligned with the vision of what the company is all about.”

When asked to recall mistakes he’s made along the way, he takes a moment to contemplate before declaring: “In all the companies I’ve been involved with, there have been painful decisions that we delayed or didn’t act upon where there were people who weren’t the right fi t—even if they were experts in their craft.”

The good news, he says, is after you “act on those misalignments,” there’s a tremendous boost of productivity and spirit among the ranks that has a measurable impact on how the company operates.

Bulens is turned on by the notion of simplifying a task that’s typically complex. Notes made it simple to develop collaborative applications, Trellix simplifi ed Web publishing, EqualLogic reduced the complex job of managing datacenter storage, and Unidesk is all about simplifying the “tedious and thorny” challenges of managing PCs in the new world of desktop virtualization—or, as Bulens puts it: “We’re capitalizing on the new platform capabilities of hypervisor-based desktops to try to make an industry.”

He also subscribes to a second notion, which is ensuring that the product experience and entire culture of the company are based on creating the best possible customer experience: “customer delight,” in Bulens’s words. He also wants to “make the IT people smile” at customer companies, because that creates “tremendous customer advocacy and passion.” That positive sentiment is evident in Unidesk’s 14 customers and in the dozens of customers who are in the middle of pilots.

Given the scale of the success of EqualLogic, it’s not surprising that Bulens cites it as his most-outstanding start-up experience—but he reveres those memories for more than just the money. He remembers that people thought the new team was “crazy” when they initially described their plans. “It takes special people to envision these things,” he says. “With EqualLogic, in a week we decided we wanted to work together because we liked each other.” VR

Bulens is turned on by the notion of simplifying a task that’s typically complex.

Page 7: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

800-754-3310www.aberdeeninc.com/vm006

STORAGE WITHOUT BOUNDARIES

Imagine what you can achieve with Aberdeen.

Aberdeen’s AberSAN Z-Series scalable storage platform brings the simplicity ofnetwork attached storage (NAS) to the SAN environment, by utilizing the innovativeZFS file system.

Featuring the high performance of the Intel Xeon processor 5600 series, theAberSAN Z-Series offers enterprise level benefits at entry level pricing, deliveringthe best bang for the buck.

Who gives you the best bang for the buck?

NetAppFAS 2050

EMCNS120

AberdeenAberSAN Z20

Deduplication ✓ ✓ ✓

Thin Provisioning ✓ ✓ ✓

VMware® Ready Certified ✓ ✓ ✓

Citrix® StorageLink™ Support ✓ ✓ ✓

Optional HA Clustering ✓ ✓ ✓

Optional Virtualized SAN ✓ ✓ ✓

Optional Synchronous Replication ✓ ✓ ✓

Optional Fibre Channel Target ✓ ✓ ✓

Unified Storage: NFS, CiFS, iSCSI ✓ ✓ ✓

Unlimited Snapshots x x ✓

Unlimited Array Size x x ✓

Hybrid Storage Pooling x x ✓

Integrated Search x x ✓

File System WAFL 64-bit UxFS 64-bit ZFS 128-bitRAID Level Support 4 and DP 5 and 6 5, 6 and Z

Maximum Storage Capacity 104TB 240TB Over 1,000TBWarranty 3 Years 3 Years 5 Years

Published Starting MSRP Not Available Not Available $8,995Above specific configurations obtained from the respective websites on Sept. 13, 2010. Intel, Intel Logo, Intel Inside, Intel Inside Logo, Pentium, Xeon, and Xeon Inside

are trademarks or registered trademarks of Intel Corporation or its subsidiaries in the United States and other countries. All trademarks are the property of their respectiveowners. All rights reserved. For terms and conditions, please see www.aberdeeninc.com/abpoly/abterms.htm. vm006

Page 8: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

6 | October/November 2010 | Virtualization Review | VirtualizationReview.com

news • trends • analysis

REVIEW

Like its vSphere 4.0 predecessor,which debuted with a slew of new features barely a year ago,

vSphere 4.1 includes a mix of new features and enhancements. These offerings are broken down evenly across many areas of a vSphere installation. Here are a few confi guration notes for the features that infrastructure administrators will fi nd most attractive:

vStorage APIs for Array Integration (VAAI): This feature set is now in its infancy, but has great potential for helping enterprise storage systems make easy work of selected vSphere I/O operations. The current list of hardware vendors that support VAAI is short, and within that short list the number of actual products shipping is even smaller. VAAI gives the decision maker an opportunity to deliver a storage solution that’s much more aligned

to the I/O needs of a virtualized environment by optimizing repetitive functions at the storage processor, rather than the hypervisor.

Storage I/O Control (SIOC): This new feature is designed to help address the No. 1 performance issue that virtualization administrators face: slow storage. While virtualization doesn’t fi x underperforming storage, SIOC allows administrators to apply share-based prioritization to storage resources. Effectively Distributed Resource Scheduling (DRS) for a datastore, SIOC enforces the I/O shares across all virtual machines (VMs) that access a datastore.

SIOC establishes a latency value for each datastore. The default value is 30 ms, which is rather generous. There’s no default latency value that will work for every environment, as everyone’s requirements vary. Within many virtualization circles, 20 ms latency is the threshold where administrators start to become concerned. When latency on the datastore exceeds the threshold value, SIOC assigns disk access based on the share value of the VMs on that datastore. Like other share functions, confi gurations can get complicated. A generally accepted practice is sticking with the default share values of high, normal or low when assigning resource allocations for VMs, and staying away from directly entering a shares value.

SIOC is different from other DRS-like features in that it fl ows from

the datastore to the VMs. SIOC is not based on the aggregated resources of the cluster, but instead on the individual datastore object. The silver lining is that individual datastores are confi gured for SIOC, so administrators can prioritize where prioritization is needed. Figure 1 shows the SIOC confi guration window.

SIOC only works with Virtual Machine File System (VMFS) volumes, confi rming a long-standing opinion that VMFS is the fi rst in line for new features. Network File System (NFS)-based storage systems currently do not support SIOC.

Network I/O Control (NIOC): Much like SIOC, the new NIOC feature allows advanced prioritization of network traffi c groups that are using a vSphere vNetwork Distributed Switch. The traffi c groups include FT VM traffi c, iSCSI traffi c, vMotion traffi c, management traffi c, NFS traffi c and guest VM traffi c. This categorization is important as Ethernet-based storage protocols (iSCSI and NFS) become more popular in virtualized infrastructures. NIOC can help protect investments by reducing port costs, as well as by giving the administrator granular control of the network interfaces. Network throughput is usually not one of the bottlenecks for most virtualization installations, but NIOC is one step ahead of the problem.

ESXi Host Authentication: Any admin will welcome a confi guration that delivers a unifi ed identity management solution. vSphere 4.1 introduces Active Directory-integrated authentication to ESXi hosts. This is similar to what XenServer 5.5 does, in that a computer account for the ESXi host is created in the AD domain. Most access to the hosts is performed

vSphere 4.1 Packs an Incremental PunchThe VMware update to vSphere boasts an amazing number of improvements and new features.Rick Vanover

Figure 1. The SIOC confi guration screen allows a latency congestion threshold value to be set for each datastore.

Page 9: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

How fast does a system have to be to capture a business opportunity?How do you cut costs without cutting results?How do you benchmark intelligence?

Page 10: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

8 | October/November 2010 | Virtualization Review | VirtualizationReview.com

Coming next issue: Readers Choice AwardsUpFront

via vCenter, but anything that reduces the frequency of root logging in will be welcome. This feature is also available on the free ESXi installation, now dubbed VMware vSphere Hypervisor.

Scaling Improvements: vSphere has increased its management capacity in a number of areas. For example, a high-availability cluster can now scale to the same limit as a DRS cluster. A smattering of concurrent activity increases have been added, including the number of concurrent vSphere client connections. Likely the most benefi cial to infrastructure administrators is the increase in number of concurrent vMotion operations per host. The supported limit is now eight current vMotion operations for 10 Gigabit networks, and four for 1 Gigabit networks. Figure 2 shows this in action.

These updates pose a minimal learning curve in the UI or day-to-day tasks that make up a typical installation. One new feature that the everyday administrator will appreciate is the DRS Resource Distribution

chart. This quick and easy chart shows how the workloads are balanced across servers for CPU and memory resources. Previously, administrators looked in a number of places to determine the cluster balance.

For highly consolidated workloads, vSphere 4.1 introduces an expanded category of functionality for its memory management technology suite in memory-compression technology. This is a meet-in-the-middle solution between hypervisor swapping and the more desirable technologies of transparent page sharing and balloon drivers. Memory compression won’t magically let you cram more workloads into existing infrastructure investments—a single-digit percentage gain in effi ciency is more realistic.

Be Sure to Course-CorrectOne of the fundamental principles of virtualization that makes the lives of administrators easier is the ability to change confi gurations with ease. Whether it’s a new server model functioning as host hardware, a new storage array or even the move from

ESX to ESXi, vSphere continues to be fl exible by making short work of the tasks through native functionality. The ESX vs. ESXi decision is now clear as vSphere 4.1 is the last release of the ESX hypervisor. ESXi will be the hypervisor for subsequent versions of vSphere.

The push to ESXi and x64 is the hidden homework behind all of these vSphere 4.1 features, but most of the changes won’t be as earth-shattering as they appear. In fact, many of the elements are made easy by tools provided with vSphere. The Data Migration Tool, for example, allows a vCenter Server confi guration to be moved from an x86 platform to an x64 vCenter Server with a SQL Express database running locally on the system.

The Effi ciency FactorThe tools are available with vSphere to push more out of your infrastructure investment. There are scores more features that simply won’t fi t in this quick review of vSphere 4.1. Effi ciency will be the distinguishing factor between one IT operation and another. Consider that

everyone is effectively purchasing the same equipment and software; the difference maker will be the effi ciency of confi guration and operations. Without mentioning cloud, it’s clear that an effi ciently confi gured and well-operating virtualized infrastructure is the gateway to greater things. VR

Rick Vanover, vExpert, VCP, MCITP, MCTS, MCSA, is a virtualization professional based in Columbus, Ohio. He’s an online columnist for VR and writes the Everyday Virtualization blog. You can reach him at [email protected].

REVIEW

Figure 2. vCenter will now manage four or eight simultaneous vMotion events per host.

Page 11: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

Smarter technology for a Smarter Planet:

It’s time to ask smarter questions.

1. Ba

sed o

n num

ber o

f day

s of p

erfor

man

ce le

ader

ship

for th

e TPC

-C, T

PC-H

10TB

, and

SAP 3

-Tier

SD be

nchm

arks

betw

een J

une 1

, 200

5, an

d Jun

e 1, 2

010.

For m

ore i

nfor

mati

on, s

ee ht

tp://

www.

tpc.o

rg an

d http

://ww

w.sa

p.co

m/so

lution

s/be

nchm

ark.

TPC,

TPC-

C an

d TPC

-H ar

e tra

dem

arks

of th

e TPC

. IBM

, the I

BM lo

go, ib

m.co

m, D

B2, P

ower

Sys

tems,

Smar

ter P

lanet

and t

he pl

anet

icon a

re tr

adem

arks

of In

terna

tiona

l Bus

iness

Mac

hines

Cor

p., re

gister

ed in

man

y jur

isdict

ions

world

wide

. Oth

er pr

oduc

t and

servi

ce na

mes

migh

t be t

rade

mar

ks of

IBM

or ot

her c

ompa

nies.

A cur

rent

list o

f IBM

trade

mar

ks is

avail

able

on th

e Web

at w

ww.ib

m.co

m/le

gal/c

opytr

ade.s

htm

l. © In

terna

tiona

l Bus

iness

Mac

hines

Corp

orati

on 20

10.

What exactly does a benchmark mean? For the last five years, IBM DB2® on Power Systems™ has ranked first on three of the industry’s leading performance benchmarks, longer than Oracle and Microsoft combined.1 But is that the best way to think about the possibilities of technology? What really matters isn’t some abstract measure of performance, it’s what companies actually do with that performance. For instance, Coca-Cola Bottling Company is using DB2 on Power to reduce licensing, maintenance and storage fees by $350,000. EuResist is using an integrated analytics solution to predict the most effective drug combinations for individuals with HIV, with 78% accuracy. And the Dubai Gold & Commodities Exchange is working with IBM Security Services to achieve system uptime of over 99.9%. On a smarter planet, these are the benchmarks that matter.

A smarter business is built on smarter software, systems and services. Let’s build a smarter planet. ibm.com/questions

Page 12: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

10 | October/November 2010 | Virtualization Review | VirtualizationReview.com

Go online for Rick Vanover’s Everyday Virtualization blogUpFront

V irtualization Review Editor in Chief Bruce Hoard recently interviewed Pano

Logic Inc. Chief Technology Offi cer Aly Orady about what is—and what isn’t—a true zero-client device.

VR: I think a lot of people are confused about what a true zero client is. Let’s start out by addressing that.Orady: A true zero client is an endpoint device that has no processor in it whatsoever, does not run an operating system, does not have software, does not have memory, does not require management, and cannot be infected with a virus. In other words, the point of zero-client computing is to move 100 percent of software into the datacenter. Any solution that has a processor, boots an operating system, boots some kind of fi rmware or requires any kind of management is not a true zero-client solution. To give you an example, there are a number of products on the market that claim to be zero clients that are really thin clients. The way that you can tell the difference is they always have a processor, and they generally boot some kind of operating system or some kind of software.

VR: What questions should prospective desktop virtualization users ask of vendors who claim to sell zero clients?Orady: Well, I think the best questions for them to ask are, “Does this device that you’re selling us boot any kind of software? Does it execute

any kind of code? Will I ever have to do a fi rmware update to the device? Can it be infected with a virus? Does it require management?” I think those are all really good questions.

VR: Can you describe the Pano Logic Zero client?Orady: We created a device that literally has no processor. It doesn’t run any code, it doesn’t have any fi rmware. If you talk about it at the most technical levels, it doesn’t have what’s typically called an instruction pointer, which is a thing that characterizes the processor. There are other technologies on the market that are really thin-client technologies that have been re-branded as zero client. Those technologies use some kind of an operating system or run some kind of software, and in that case, you

haven’t actually moved 100 percent of software into the datacenter, so that’s not zero-client computing, as far as we’re concerned.

Zero-client computing is about having nothing at the endpoint. Because if you have no processor at the endpoint, then you never have to manage the device, you never have to confi gure the device. The device cannot be infected with a virus. That’s really the point—eliminating management outside the datacenter. The second you have any kind of processor at the endpoint where you’re booting some kind of operating system, then you have a management burden. You have to update that fi rmware; you have to manage that fi rmware. There’s a potential of that software having memory leaks and performing poorly or getting infected with a virus or

Q&A

To Be or Not To Be a True Zero ClientPano Logic CTO Aly Orady draws a strict zero-client line. By Bruce Hoard

Pano Logic CTO Aly Orady

Page 13: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

Windows 7 Migration Simplified

Page 14: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

12 | October/November 2010 | Virtualization Review | VirtualizationReview.com

Coming next issue: Industry Q&AUpFront

any of the other burdens that come along with having software there.

VR: You claim to have a complete Virtual Desktop Infrastructure (VDI) system. How does that compare to products like VMware View or Citrix XenDesktop?Orady: There are a couple different ways of looking at that. We claim to have a complete solution in that we offer the virtual desktop solution end-to-end including the endpoint device, the protocol, the management system—the entire solution comes integrated together. It deploys very easily within 30 to 60 minutes and it just works. When you go to products like VMware View or XenDesktop, what you’re getting is the management systems that go in the back-end, but you’re not getting an endpoint solution. You typically have to go and get your own endpoint solution and integrate that, whether it be from a thin client vendor or a zero-client vendor such as Pano Logic. I guess the real question is, “How many different vendors am I going to have to work with in order to get my virtual desktop solution up and running?”

VR: How would you compare the Pano device to Xenith from Wyse or the new multi-protocol system from Teradici?Orady: Both Xenith from Wyse and the solution from Teradici are based on a client device that has a processor in it that boots an operating system or code. So, the Xenith system has a processor in it and it boots an operating system over the network. What they’ve done with that environment is remove the hard drive from their thin client, and they’re now streaming the operating system over the network to the thin client, but the thin client still has a processor in it. You need pretty

complex management systems in order to get all of this to work. You will have to worry about the fi rmware that gets booted onto those thin clients. You’re going to have to update that fi rmware from time to time. And even when you boot the fi rmware, once the thing has booted an operating system, well, it’s running an operating

system [and that means] it’s vulnerable to attacks just like any operating system is and is effectively a thin client at that point. The multi-protocol system from Teradici is the same way. It’s a device that has embedded fi rmware, and that fi rmware is used to connect back to the server. They happen to have put that fi rmware on a custom chip with a processor that’s not an x86 processor, but there’s a processor in there and it does run fi rmware. There are release notes that describe what their fi rmware does and, as a result, it requires management.

VR: I realize that every user is diff erent, but can you give me some indication of what the price-

point range is going to be for a Pano Logic system?Orady: It really depends on a number of factors. The easiest way of talking about it is, if a user is buying our solution a device at a time and has brought their own virtual infrastructure into the picture—they already have some form of hypervisor, they already have their own servers and storage—in that case, the entire solution end-to-end, from the Pano device to the protocols to the desktop provisioning and management systems in the datacenter for brokering and provisioning, works out to $319 per device, including all the software in the datacenter. If the user doesn’t have their own virtual infrastructure, then we’ve put together a bundle that they can purchase, which is a server pre-loaded with virtualization, pre-loaded with the Microsoft Windows licenses, pre-loaded with all the Pano Logic software. It’s a 50-user bundle or 50-desktop bundle that works out to $489 per desktop.

VR: Let’s talk a little bit about the company in a more general sense. Right now, what’s your current state of funding? Orady: Pano Logic has raised three rounds of venture funding. In total we’ve raised $44 million in backing, so we’re a very, very well-backed company. The most recent round of fi nancing came from the Mayfi eld Fund, which is basically one of the most prestigious funds in the valley.

VR: How many customers do you have at this point?Orady: We have about 350 customers, and we count customers as over 10 units. VR

Bruce Hoard ([email protected]) is the editor in chief of Virtualization Review magazine.

Q&A

“Zero-client computing is about having nothing at the endpoint. Because if you have no processor at the endpoint, then you never have to manage the device.”Aly Orady, Chief Technology Offi ce, Pano Logic Inc.

Page 16: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

COVER STORY | Microsoft

Microsoft wants you to know that the company is really bullish on virtualization. Redmond knows virtualization slashes IT expenses, rolls back the costs of deploying Windows and provides a

stepping stone to the cloud. That’s why Microsoft includes it with Windows Server 2008.

Microsoft also wants you to think about the benefi ts of virtualization—server consolidation, better business continuity, offl ine patching, provisioning and updates—and realize you can have those complicated capabilities

built into the cloud with the Windows Azure Platform, the Microsoft public cloud platform that lets you pay for what you use. Windows Azure allows users to develop, run and replicate their applications via Platform as a Service (PaaS) in Microsoft datacenters around the world, without having to virtualize them fi rst. Microsoft isn’t talking about server consolidation; it’s aiming for server elimination.

According to Dave Bartoletti, senior analyst at the Taneja Group, success in the cloud is all about apps, and Windows Azure is currently ahead of VMware Inc. in that regard.

IMA

GE FR

OM

SHU

TTERSTO

CK

Microsoft: The Great ServerRedmond is looking to streamline the application process via its Windows Azure Platform, but it can’t ignore server virtualization or VDI if it wants to overtake VMware.

14 | October/November 2010 | Virtualization Review | VirtualizationReview.com

By Bruce Hoard

Page 17: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

VirtualizationReview.com | Virtualization Review | October/November 2010 | 15

“I think Microsoft is trying to change the conversation away from virtualization as an infrastructure technology to virtualization as simply an enabler for application delivery,” says Bartoletti. “I think that takes them back to their strengths. Where Microsoft really owns us is from an application standpoint.”

In order to drive home his point, he notes that when they buy new PCs, many people have a discussion with themselves. Do they want to recreate the core set of Microsoft applications that they love, or would they prefer to go in a new direction with Google Apps and Gmail instead of Exchange? Or maybe it’s time to consider an infrastructure or PaaS from someone like Salesforce.com Inc.

“Microsoft wants to stem that discussion,” Bartoletti declares. “They want to make the standard Microsoft application stack easy to deliver from anywhere, and to look as cloud-enabled as possible.”

Successful Cloud CustomersDavid Greschler, Microsoft’s director of Virtualization Strategy, says Microsoft has already arrived in the cloud, claiming the company has more than 10,000 Windows Azure customers, including The Coca-Cola Company and OSI Restaurant Partners LLC (which runs Outback Steakhouse), who are scaling up and down dramatically. Mike Neil, general manager of Windows Server and Server Virtualization, elaborates on the Outback Steakhouse implementation by saying the company wanted to do a promotion to their customers that would get 500,000 of them to sign up in return for a blooming onion. Outback wanted to develop the application very quickly, so it did it in Windows Azure over a two-month period. The goal was to get people excited and in to their local Outback Steakhouse as soon as possible. In the end, the company burned through that half-million people in a little over two weeks, and it was able to scale its cloud capacity up and down to accommodate customer response. Outback just paid for what it needed, and the company didn’t over-provision its own hardware.

Neil further stresses the cost effi ciency of Windows Azure by comparing it to a more expensive alternative from a major managed services and cloud provider. If a customer goes to that provider to get a cloud-based virtual machine (VM), he says it’s the equivalent experience of having a physical machine somewhere off in somebody else’s datacenter. The customer still has to worry about what OS they’re going to run inside of it, so they still have to do the management, patching, maintenance and updates.

“In that case I’m not paying for the hardware as a capital expense,” Neil says. “I’m paying it as an operational expense, but I still have all the operational expense of managing it. With Azure, we provide you with an environment to run your application. We manage all of the underpinnings for you, and that signifi cantly reduces those costs for customers.”

Neil adds that providing that sort of consistency, allowing the customer to develop an on-premises application, run it with a services provider, or run it in one of the Microsoft datacenters with Windows Azure, is the kind of customer experience Microsoft wants to provide.

The Battle RagesAt this point, Microsoft and VMware are fi erce competitors locked in a high-stakes battle. Microsoft is chipping away at VMware’s virtualization hegemony with a stable of products, including Hyper-V, Microsoft System Center, Microsoft Application Virtualization (or App-V), VDI Suites and Windows Azure. VMware is keeping pace with the new vSphere 4.1 and updated versions of ThinApp and View. VMware still has a commanding lead, but the company’s carefree days are behind it.

When I met with Greschler and Neil at VMworld, they took VMware to task for being too fi xated on, of all things, virtualization, saying: “They only have one trick, you know.” Neil notes, “Virtualization is the core of what they do. They have to sell a virtualization solution for them to be successful in their business. That’s where they’re driving customers, but it isn’t necessarily where customers want to go.”

Eager to drive home Microsoft’s competitive edge, Neil says that VMware shop CH2M HILL, a Fortune 500 engineering fi rm with offi ces around the world, is switching to Microsoft Hyper-V because the company fi nds it easier to maintain and manage in its environments. The company also says it’s less expensive than VMware, a charge that resonates with many VMware customers. When asked if CH2M HILL is ripping out its whole VMware infrastructure, Neil explains, “Yes, they’re switching over. It’s actually not that hard for a customer to switch from VMware. We run on the same hardware. In fact, we run on more hardware than they do, and our management solutions manage both VMware and Microsoft so they provide the single management tool for the customer who might be in transition. They can run and manage their VMware infrastructure that they have today, stand up their Hyper-V infrastructure, and move their workloads very easily across the two.”

No Winner in SightApprised of Microsoft’s depiction of VMware as a one-trick pony, Mark Margevicius, vice president and research director at Gartner Inc., says he sees value in both the Microsoft and VMware virtualization philosophies. He says Microsoft adds value by offering virtualization as a feature on a server product that still has a lot to offer even if the customer doesn’t use virtualization. Regarding VMware, he notes how customers have the fl exibility to add value components such as desktop virtualization and manageability to the baseline vSphere and ESX technologies.

Margevicius doesn’t see a competitive cloud advantage for either competitor, saying, “I don’t know how successful

Page 18: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

16 | October/November 2010 | Virtualization Review | VirtualizationReview.com

COVER STORY | Microsoft

either has been around the delivery of cloud-based services with their core products. I think both will struggle as cloud offerings evolve because their products, for the most part, were not designed with cloud in mind. They may have morphed to cloud offerings, but I don’t necessarily agree that either one is better situated at this point for wide-scale adoption in the cloud space.”

Dynamic Memory, not Memory OvercommitIn its position as Microsoft’s primary hypervisor, Hyper-V is on the line to perform against the best its competitors can muster, and toward that goal, the company partnered very closely with Intel Corp. and AMD Inc. on VT and V technologies, respectively, to make sure the chip makers’ hardware could handle virtualization. Microsoft then built Hyper-V on top of it. Going forward, Microsoft has continued to work with Intel and AMD on memory and I/O virtualization to drive down hypervisor overhead to the point where Hyper-V users can run large, compute-intensive, disk-intensive and network-intensive workloads.

“We have customers running SQL, Exchange and so on within those environments and they’re all very happy with the results they’re getting from a performance standpoint,” Neil says.

The lack of a memory overcommit capability became a sore point for Hyper-V, which had also been slow to receive live migration capabilities. Microsoft addressed the memory overcommit issue in Windows Server 2008 R2 SP1—which debuted in public beta this June, and was still in that position as of early October. The answer was dynamic memory, a technology Microsoft believes is less risky than memory overcommit, which VMware supports. Memory overcommit is a hypervisor feature that enables a VM to consume more memory space than a physical host has available.

With the dynamic memory capability of Hyper-V, guests participate with the hypervisor in managing memory across all of the VMs that are running. This is done by splitting up the physical memory to all the VMs, and moving it back and forth between VMs as necessary. This is done in a way that makes the guest OS running inside the VM aware of the dynamically shifting memory, so it’s never overcommitted.

“This way you get a lot of the same benefi ts from this sense of getting more virtual instances on the one physical box, getting the density much higher, but you get less risk from the standpoint of not overcommitting your resources, and you’re not getting into a situation where you’re going to see performance degradation and failure,” Neil says.

Bartoletti believes that despite advances with Hyper-V, VMware still has at least a fi ve-year, “pure technology” lead

on both Citrix XenServer and Hyper-V. According to him, the recent releases of both those products are just beginning to catch up with VMware in terms of raw performance, raw scalability and raw power in the datacenter. “That’s why I think we see so much more discussion from Microsoft about the desktop, and a much tighter integration with Citrix, who in many cases owns the desktop and owns application delivery to the desktop,” Bartoletti says.

Neil says Hyper-V has already proven its technical chops by competing against VMware in the still-lucrative server virtualization market, where VMware has reigned supreme for so long. “We’re slowing their growth, we’re taking their market share and we’re going faster in the market,” he declares.

Margevicius is less interested in technical comparisons between Hyper-V and vSphere than he is in determining which product is the best value. He says customers who are well entrenched with virtualization, have standardized on VMware and really push the boundaries of its feature set consider VMware to be a better product, if not necessarily the best value.

“However, for users who are just getting started and have minimal requirements, value is more important, and Hyper-V, being embedded within the server product, is clearly viewed by many customers who are not in virtualization yet as a very strong alternative,” he adds.

VDI: Composable Desktops, Folder RedirectionMicrosoft takes a measured approach to Virtual Desktop Infrastructure (VDI). As Greschler describes it, the space got defi ned by a lot of VDI hype that gained momentum when people said that it would be a great idea to take either Windows XP, Windows Vista or Windows 7, put it on a server and remote it to employee desktops. The idea sounded great in principal, Greschler notes, but it has a “huge set of headaches” associated with it. The fi rst problem is, users are just moving the problems they had on their desktops to their servers. The second problem is that they’re building out big server farms, which he compares to the VMware approach.

Greschler says the Microsoft approach to VDI—which relies on Microsoft System Center Confi guration Manager—involves breaking up the components of the desktop so they can be composed in real time wherever they’re needed. As he explains it, there are applications inside the OS that Microsoft virtualizes with App-V—its application virtualization software—enabling them to be delivered on-demand, without confl icts, via network streaming. “So they’re just data fi les delivered to your desktop, that fi re and run without impacting your OS,” he states.

Microsoft isn’t talking about server consolidation; it’s aiming for server elimination.

Page 20: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

18 | October/November 2010 | Virtualization Review | VirtualizationReview.com

COVER STORY | Microsoft

This is accomplished through the use of folder redirection, a component of Windows 7 that enables users to redirect their fi les to a server where they’re replicated and saved. This includes personal settings—such as personalization or user state virtualization—which may include things such as a user’s photo of children, toolbar settings, or choice of icons for their start menu or video screen. All of these profi les are replicated before they’re redirected to storage.

So what are the business benefi ts? Greschler says if your laptop is stolen, you can log back in, and based on the last time your data was replicated with folder redirection, you can literally restore the desktop down to the photos of your children and your toolbar settings.

VDI SuitesAccording to Sumit Dhawan, vice president for Citrix XenDesktop product marketing, “The Microsoft VDI Suites are built for enterprise customers looking to run VDI-based desktops that can be delivered to users remotely. However, it’s a platform and is only consumable for enterprise use when used with an enterprise solution, such as XenDesktop.”

The Standard and Premium VDI Suites enable customers to buy different components of the server and management infrastructure required to run VDI via a single license. The VDI Standard Suite is designed to enable organizations to deploy a basic VDI environment, and includes Hyper–V Server 2008 R2 and several System Center components, such as System Center Virtual Machine Manager, System Center Operations Manager 2007 R2, and System Center Confi guration Manager 2007 R2. Applications are dynamically delivered via App-V, and the Standard Suite also enables restricted use rights for Remote Desktop Services (RDS) to deliver VDI desktops to users. As Neil notes, “I think people lose sight of the fact that there are more customers using Terminal Services, now called RDS, than there are running VDI today.”

The VDI Premium Suite provides additional VDI fl exibility—RDS is unrestricted, for example—which allows users to deliver both session-based and VDI desktops. Also included is App-V for RDS, which Microsoft says enables organizations to dynamically deliver applications to the RDS server, reducing silos of RDS applications.

The Microsoft-Citrix RelationshipGiven the frequently cited close relationship between Microsoft and Citrix Systems Inc., it’s hardly surprising that Dhawan would be an expert in any Microsoft virtualization

products. Notes Greschler, “On the RDS space, we had a relationship with Citrix for 15 years before virtualization came on.” That relationship also encompasses the management side of things, where there’s a lot of integration between Citrix and System Center. For example, there are announced plans for System Center to manage XenServer in the future, and it already manages VMware. This fl exibility and interoperability with heterogeneous environments gives Microsoft an edge with customers who may want to run separate production, test and development environments.

Margevicius says System Center is designed to be an ecosystem of manageability around Microsoft products, adding, “Here in the management space, Microsoft gets very good results, and for the most part is viewed very

favorably by its customers. They like System Center and System Center does a nice job for them.”

End GameBartoletti thinks it all comes down to Microsoft and VMware, with Citrix somehow being folded into Microsoft because of Microsoft’s domination at the application layer. He foresees the emergence of a bifurcated market where VMware hangs on to its server virtualization business and ultimately lives or dies over the next three-to-four

years based on the evolving relationship between desktop virtualization and cloud computing. He sees the cloud as a way to attack the desktop, and if Microsoft or VMware really makes the cloud much more attractive, Bartoletti predicts there will be a major shift in what people expect from desktop virtualization.

“I think what we’re going to start seeing is who can package the real application stack, and this is where VMware’s behind Microsoft,” he says. “Packaging the application stack for desktop delivery is going to be key. I think what we’ll probably see is VMware tracking what’s going on in the smartphone space and trying to partner with a huge competitor of Microsoft’s, such as Google.”

Margevicius says VMware will have to fi nd value and differentiation beyond virtualization in order to continue its success. “If virtualization becomes commoditized and readily available within the OS platform that most people are going to deploy anyway, the value of something like VMware, which requires that a premium be paid, has to be defi ned and used by those customers,” he says. “So VMware has to continue to push in order to remain relevant.” VR

Bruce Hoard ([email protected]) is editor in chief of Virtualization Review.

According to Dave Bartoletti, senior analyst at the Taneja Group, success in the cloud is all about apps, and Windows Azure is currently ahead of VMware in that regard.

Page 21: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

Wiley, the Wiley logo, and the Sybex logo are registered trademarks of John Wiley & Sons, Inc. and/or its affiliates. All other trademarks are the property of their respective owners.

New VMware® Books from Sybex!

In-depth coverage to help you prepare for your exam, and the real world.

Visit www.sybex.com/go/VMware for more information.

And more! Don’t miss these VMware vSphere 4 must-haves from Sybex.

Untitled-2 1 7/12/10 10:39 AM

Page 22: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

20 | October/November 2010 | Virtualization Review | VirtualizationReview.com

FEATURE | Mainframe

For many IT planners, bombarded by the marketecture emanating from VMware Inc., Microsoft, Citrix Systems Inc., and a handful of smaller server virtualization

vendors over the past three years, hypervisor-based consolidation would seem to be just what the doctor ordered to cope with out-of-control server sprawl. Differences between products may still be debated, but the symptomatology delineated in vendor brochures and repeated by an echo chamber of the media, analyst community and blogosphere has achieved a general consensus.

Fact 1: Hardware resource allocation and utilization ineffi ciency is rampant in contemporary x86-based distributed computing. The proliferation of the “one box-one application” model for software deployment has seen processor, bandwidth and capacity underutilized to embarrassing levels. The situation is getting more and more diffi cult to sustain or to justify in the current “do more with less” climate.

Fact 2: Maintaining a large complex of servers and their related storage infrastructure is nightmarish and costly. Tight coupling of software to hardware increases the potential for downtime, whether due to broken kit or OS/application software patching—the latter being a favorite whipping boy of both virtualization vendors and consumers. At a time when fewer workers carry the load once performed by many, dependency on automation is at an all-time high. As a consequence, downtime is no longer an option.

Fact 3: As hardware becomes more commoditized, it’s only natural—perhaps inevitable—for application and OS software functionality to be abstracted away from the underlying chips, circuit boards and feature cards of the hardware kit. Until that happens, the cost of even commodity hardware seems to accelerate rather than decelerate in response to Moore’s Law. This is intolerable in the current economy.

With these often repeated claims, the discussion around server virtualization has placed into sharp relief the endemic

x86 Server Virtualization on Big Iron

Back to theMainframe:

IMA

GE FR

OM

GETTY

x86 virtual servers are put on notice.By Jon William Toigo

Page 23: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

VirtualizationReview.com | Virtualization Review | October/November 2010 | 21

problems of distributed computing. At the same time, virtualization via x86 hypervisors has been posited as a cure-all for the costliness and waste associated with open systems. And it has done so using terms and language that even technology-challenged senior management can understand.

Business management has been told that divorcing hardware from software should cause the prices for computing gear to drop, an adjustment that more closely refl ects their increasingly commoditized status. The overall impact, vendors are telling the front offi ce, is that the cost of IT infrastructure should shrink: you don’t need as many servers to handle workload, nor as many administrators to maintain all of the hardware.

On its face, this campaign has appeared to be wildly successful. Though budgetary constraints are limiting the takeover of the entire distributed server world, hypervisor-based server virtualization has garnered tremendous mindshare. Virtualization has become in an extremely short time a popular buzz phrase, reaching the status of an “as-seen-on-TV” weight-loss product—you know, the kind preferred by four out of fi ve doctors.

Unfortunately, vendors have also encouraged “woo.” During an event in Orlando, Fla., last year, one VMware spokesperson declared that his company’s offering was “not just a product. We are a movement!” Notably missing from his recitation of

the many benefi ts that would accrue to server virtualization was a “Quack Miranda”—the fi ne print disclaimer about results varying widely and the lack of peer-reviewed research to validate claims.

What’s the Problem?While the virtualization advocates and the pundits have correctly identifi ed the symptoms of ineffi cient and costly distributed computing, they have at the same time largely ignored the bigger challenge: neither root causes nor treatment options have been rigorously analyzed or properly defi ned.

For example, resource allocation and utilization ineffi ciencies in Microsoft-based servers have been spelled out, but the fact that Linux-based machines have tended to present better resource-utilization effi ciency numbers than their Redmond-centric peers—and without the services of hypervisors—has not received much attention in all of the hype around hypervisors.

Also interesting is the fact that server effi ciency statistics have often been juxtaposed with mainframe effi ciency numbers—usually to provide some sort of comparative context. For example, storage allocation effi ciency in the mainframe world typically hovers at around 80 percent of optimal, while a typical Microsoft server scores in the 17 or 18 percentile. Without exploring the possible explanations for the difference, vendors were quick to prescribe server virtualization to fi x the problem.

The core explanation for the differences between the x86 server and mainframe effi ciency number, of course, has to do with the existence of OS-level resource management functionality that’s built into the mainframe OS. This is augmented by a set of de facto standards to which all third-party vendor gear must adhere. To plug its product into a mainframe, the third-party vendor must validate its conformance with IBM-dictated standards, including a set of management standards. By contrast, in the open systems world, management standards are a joke. Open standards have a tendency to be held hostage by three-letter vendors who have little interest in making it easier for the consumer to measure the relative performance of competing products. Management is a great leveler of marketecture: It enables consumers to unplug vendor A’s gear and substitute vendor B’s when B offers greater performance or ROI.

The mainframe comparison suggests that providing system-level management tools and meaningful cross-hardware management standards in the open systems world would likely elevate the resource allocation and utilization effi ciency numbers of servers—without requiring hypervisors to broker the I/O load at all. This point has been ignored by leading advocates of server virtualization, though it’s increasingly part of the “background chatter” around the subject as management software vendors such as Virtual

Page 24: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

22 | October/November 2010 | Virtualization Review | VirtualizationReview.com

FEATURE | Mainframe

Instruments and Tek-Tools (now aligned with SolarWinds) try to leverage the marketing around VMware to promote their “improved management equals greater resource effi ciency even in virtual servers” value proposition.

Coherent management is also central to the pitches of the budding crop of “mainframe mini-me’s.” Several vendors and vendor collectives are building what are essentially proprietary stovepipe platforms around a select set of software and hardware components. Proponents of this approach include Oracle Corp., with its recently acquired Sun Microsystems/STK server and storage hardware and its

hypervisor built in part on intellectual property acquired from Virtual Iron; Microsoft-HP-qLogic with Hyper-V hypervisor and HP management software; and VMware-Cisco-NetApp with the vSphere “OS” and the Cisco unifi ed network management wares.

Self-Serving De Facto StandardsPeel back the market speak proffered by each of these mainframe mini-me’s and it is evident that each one is seeking, among other things, to provide a set of de facto standards to ensure the interoperability and manageability of gear connected to their collective kit. The idea is to deliver on x86 technology the resource effi ciency, stability and resiliency of the mainframe.

The creation of proprietary hardware and software stacks featuring guest machine virtualization and unifi ed management is not new. Mainframe platforms have been delivering just such an architecture for at least 30 years. A key difference is that, within a mainframe, the failure of one guest machine does not cause all hosted apps to fail—the Achilles’ heel of many x86 virtual servers.

This partly explains IBM’s parental yawn when server virtualization fi rst came into vogue. If you ask them, IBMers will tell you that they invented virtualization, introducing the fi rst virtual machine (VM) OS in 1968. Twenty years later, they rolled out their vaunted Processor Resource/System Manager (PR/SM) to enable multiple VM environments to operate concurrently in the same physical hardware footprint. These logical partitions, or LPARs, are effi ciently allocated system resources, and they’re insulated from one another to prevent the results from becoming a Jenga game.

As early as 1998, the smart folks at Big Blue had seen the server consolidation trend coming. Their response was to invest in Linux and related open source development initiatives like Xen, a project aimed at the virtualization of PowerPC and x86 servers. At the same time, they worked on ways to consolidate x86 servers as virtual guests inside mainframe LPARs using the z/VM OS and Linux for System z.

As testimony to the fact that Big Blue’s idea of consolidating servers inside mainframe LPARs has not waned, you need only consider the recent announcement of SONAS—a scale-out NAS product that hooks into a Linux for System z environment like so much traditional block storage direct-access storage devices.

In early 2008, as the VMware buzz was gaining volume, IBM announced that it had a “better way” to virtualize and consolidate x86 sprawl—especially for those companies with a System z mainframe already deployed. Their announcements around the z10 mainframe touted an

ability to host up to 1,400 x86 servers in LPARs for about $400 each, well below the cost of the server hardware, software licenses and training requirements of VMware-style server virtualization. According to IBM spokespeople, a z10 mainframe can handle up to 60 LPARs and each partition (a virtual environment unto itself) can run the IBM z/VM OS capable of hosting many x86 guest machine images. How many servers you can elect to stand up in each LPAR requires an assessment of the guest machine’s workload processing requirements.

IBM caught quite a bit of heat for these claims. The vendors of x86 hypervisors thought they had dealt the death blow by citing the huge differences in acquisition price between mainframe hardware and servers. But that was not the case: On the heels of the z10 Enterprise Class mainframe and the alternative it presented to virtualization on x86 servers, IBM released “business class” versions of its big-iron platforms. These servers on steroids borrowed conceptually from mainframe architecture but delivered a better acquisition price point and support for a Linux OS. Business class versions of z Series mainframes, IBM said, could support the virtualization of up to 232 x86 servers—beating the best consolidation ratios available on open server platforms at the time.

It’s worth noting that, while neither the traditional enterprise mainframe or its business-class variant excelled at virtualizing Microsoft server OSes or applications (though third-party products enable these workloads to be ported into LPARs), IBM read the tea leaves correctly: The preponderance of servers being virtualized were fi le servers and low-traffi c Web servers, the latter typically running Linux and Apache. Corralling those workloads into mainframe LPARs was well within the realm of the possible.

Overselling Mainframe Benefits?What hurt the IBM proposition was the perceived overselling of mainframe benefi ts that supposedly extended to the VMs, but don’t. One of these is the IBM Parallel

If you ask them, IBMers will tell you that they invented virtualization, introducing the first virtual machine OS in 1968.

Page 25: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

www.techxtend.com1-800-330-9611 x7704

vReplicator

vRanger

vConverter

ShadowProtect

esXpress

Disk Director Suite

Disk Director Server

Lifeboat Distribution 9-10

SERV

ER &

WOR

KSTA

TION

ADMI

NIST

RATIO

N

BUSI

NESS

CON

TINUI

TY

& DI

SAST

ER R

ECOV

ERY

SECURITY

PERFORMANCEMONITORING & MANAGEMENT

STORAGE MANAGEMENT

SOFTWARE

DEVELOPMENT & TESTVIRTUALIZATION

PLATFORM

Traveller CPR

ShadowServerShadowSurfer

Reflex VSA

Veeam Monitor

Virtual Appliance Suite

Foglight

Vantage

vFoglight

vFoglight

Network Storage Server

Data Storage Server

PerfectDisk

SANsymphonySANmelodySANmaestroUpTempo

Backup &Recovery Server

Snap Deploy

World View products may also support virtualization platforms from

Citrix, Microsoft & VMware.

Check with your Lifeboat

Representative

AdminStudio

by Flexera

Orion Network Performance Monitor

StarWind Server

H O M E O F B AC K U P 2 . 0

Replay AppImage

2X LoadBalancer

2X ApplicationServer

2X LoadBalancer

2X ThinClientServer

2X VirtualDesktopServer

XP/VS Terminal Server

VDI-in-a-box

Veeam Configurator

Veeam Reporter

vWorkspace

VeeamBackup

VMware VirtualAppliance

Real Selection, Real Service, Real ValueBest of Breed Software & Services for Your Virtual World Needs

Contact us for a FREE on-site or web-meeting introduction to the business and technical implications of virtualization and cloud computing.

techxtend.com/VirtualWorldView

REGISTER TODAY! TechXtend.com/WebinarsView other virtualization webinars on-demand: techxtend.com/cloud

FREE VIRTUALIZATION WEBINAR SERIES: Build on VMware ESXi and vSphere for Centralized Management, Continuous Application Availability, and Maximum Operational Efficiency in Your Virtualized Datacenter.This webinar series is focused on the implementation and optimization of VMware-based virtualized server, storage,desktop and application environments. VMware’s next generation ESXi hypervisor enables single-server partitioning and forms the foundation for a virtualized datacenter. You can build upon this base virtualization layer with third party applications to obtain centralized management, continuous application availability, and maximum operational efficiency.

Page 26: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

FEATURE | Mainframe

Sysplex, a coupling facility that acts as shared memory for applications running in up to 32 z/OS images. In fact, Parallel Sysplex does not extend to z/VM or Linux for System z, IBM concedes. However, Sysplex-like functionality—involving the creation of a common timer in LPARs, whether located in one physical box or several—is supported. This is key for timestamp-sensitive functions, such as data mirroring, and for rollback operations used to recover databases to a certain state or point in time following an interruption event.

Additionally, Big Blue’s core systems-management functionality (System-Managed Storage [SMS], hierarchical storage management [HSM] and so on)—the biggest driver of resource utilization effi ciency—has never been fully ported to its Linux for System z. Systems-level management doesn’t extend out to the non-mainframe infrastructure elements that have routinely been deployed to serve x86 apps. Says IBM, “SMS and HSM management concepts” have been re-implemented into a new product called Scale Out File Service (SOFS) to enable connectivity to storage via NFS, CIFS, FTP and HTTP over Ethernet. IBM claims comparable resource effi ciencies can be realized using SOFS with fi le system-based workloads as accrued to transactional workloads using SMS. SOFS has been melded recently with SONAS, providing IBM with a new DASD capability in the form of clustered NAS that may prove its claims once and for all.

For connecting and managing other “non-traditional” storage behind a mainframe-hosted x86 apps, you may still need to deploy open systems-style management wares. This is a less-elegant approach than is the centralized resource management that comes with a traditional mainframe, but it is on par with current x86-based hypervisor solutions.

To IBM’s credit, the company is making signifi cant improvements in this space, leveraging technologies based on its own mainframe experience, as well as new technologies for managing heterogeneous x86 hypervisors, whether these are run from server platforms or are hosted together with x86 guest applications inside mainframe LPARs. This technology is already coming into its own as businesses use different hypervisors with different applications in the open systems world, creating in the process an additional management conundrum that goes by the name of virtual server sprawl.

[For more on the new IBM zEnterprise mainframe, go to VirtualizationReview.com/Toigo1010.—Ed.] VR

Jon W. Toigo is chairman of The Data Management Institute and CEO of data-management consulting and research fi rm Toigo Partners International. He is a contributing editor for ESJ.com and its Storage Strategies columnist. Toigo is the author of 14 books, including “Disaster Recovery Planning: Preparing for the Unthinkable” (Prentice Hall PTR, 2002) and “The Holy Grail of Network Storage Management” (Prentice Hall PTR, 2003).

Defining the ProblemThe point of this story is not to prescribe IBM mainframe-centric server virtualization and consolidation strategy as preferable to the approaches proff ered by the x86 hypervisor peddlers. At some point, real criteria must be formulated to help determine some way to defi ne an approach that is best suited to a given workload at a given company. In any balanced assessment, attention must be paid to at least three questions:

First: Is virtualization really needed at all to resolve the problem at hand? If a company is consolidating Web servers or fi le servers, other non-virtualization-based approaches exist: Common Web site managers like Plesk can be used to host and manage multiple Web servers on a single box, while distributed fi le systems providing a global namespace can be used to consolidate older fi lers. Hypervisors may be overkill with such workloads.

Second: What are the ramifi cations for server consolidation via various virtualization methods from an effi ciency, cost-of-ownership and manageability perspective? Mainframes may have a greater cost of initial acquisition than do commodity servers, but such a comparison may be specious. You need to factor in more than the acquisition cost and consider consolidation ratios, required staff and their training needs, downtime, software licenses, maintenance, energy consumption and a host of other factors that contribute to total cost of ownership. In many cases, you will not be surprised to see that the mainframe is actually quite a bit less expensive to own.

Also, play close attention to the highest common denominator issue in x86 virtualization. If you plan to virtualize numerous applications using x86 equipment and hypervisors and intend to optimize utilization by shifting workload from machine to machine, keep in mind that the application guest with the most demanding I/O requirements—the highest common denominator—tends to dictate the confi guration of each and every potential host platform. That means all servers may need to be “overconfi gured” to accommodate their potential use as a host for an I/O-intensive app. In some shops the resulting consolidated infrastructure actually costs quite a bit more than does the infrastructure that was replaced. Adding redundant NICs and HBAs to every server may impact your pricing model signifi cantly; conversely, re-using older servers may obviate that consolidation ratio you were seeking.

Additionally, you need to perform a clear-headed review of management pluses and minuses associated with each virtualization option. Virtualization does not mitigate the need for management; it increases it. Without a well-managed infrastructure, virtualization is like plastic surgery—it might make you prettier, but you may still have ugly children. An abstraction layer doesn’t change the DNA.

Last: How stable is the platform? Mainframe-based virtualization trumps x86 at present because the guests are insulated from each other. The best that the x86 virtualizers can off er is a complex failover strategy to recover guests on an alternate host if the stack collapses on the primary host. Simplicity should trump complexity.

Another point worth mentioning here: Some folks have actually bought into the woo that somehow propping applications and OS code in guest machines eliminates the fail-on-patch phenomenon we have all come to despise. You might want to revisit this thesis recalling what happened when VMware patched its ESX servers about 18 months back, causing many processes to stop unexpectedly in many client sites. Patching remains the bane of server administration. —J.W.T.

24 | October/November 2010 | Virtualization Review | VirtualizationReview.com

Page 27: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

FEATURE | Virtual Labs

Provisioning and Managing Virtual Labs

About 10 years ago, I walked away from my corporate job so that I could be a full-time, freelance technical

writer. As I made the transition, it quickly became apparent that I needed two separate networks—a production network and a lab network.

At the time, I already had a production network in place. I had a few desktops and laptops, a couple of domain controllers, a fi le server and a proxy server that provided my network with Internet access. While this might seem a little over the top for a home network, it’s important to realize that freelance writing is a business, and like any other business, it involves sending and receiving e-mails, maintaining certain fi nancial records and storing copies of various fi les.

While my production network met my business needs, it really wasn’t doing much in the way of helping me produce quality content. While I could use my production servers to capture certain screenshots or to get a feel for how certain tools and software worked, I was limited in what I could do safely.

To give you a better idea of what I mean, consider this: Recently I wrote

an article that discussed upgrading from Microsoft Offi ce SharePoint Server 2007 to SharePoint 2010. Without a separate lab, this article would have posed a major problem because I don’t run SharePoint on my production network.

But suppose I did have a SharePoint 2007 server on my production network—I may very well have wanted to upgrade it to SharePoint 2010. However, even if that had been the case, upgrading a production server just so I could write an article would’ve been foolish. Most of the editors I’ve worked with over the years tend to give fairly short deadlines. As such, the amount of planning required for a proper upgrade would likely have taken more time than the article’s deadline would’ve allowed, and rushing an upgrade isn’t wise.

Another reason why using my production network in my writing isn’t practical is because doing so could be a big security risk. Think about the types of information that could be exposed through a screenshot. Depending on what I’m writing about, a screenshot may reveal the IP addresses, SIDs, domain names or account names that I use internally.

Ultimately, I decided there was just no alternative to constructing a lab network. After doing so, my lab network made it a lot more practical for me to write about various enterprise networking technologies. However, it also presented a set of problems all its own.

For the most part, these problems were fi nancial in nature. Server virtualization had just been introduced and it really wasn’t practical yet, so I was forced to use physical hardware. As such, I spent an absolutely obscene amount of money on servers, network switches, cabling, software licenses and so on. Even after my lab was fully provisioned, there were ongoing expenses that I hadn’t thought about. It seemed that I was constantly spending money on server upgrades and repairs. Additionally, my electric bills began to resemble the national debt. Not only did the servers consume electricity, but they also forced my air conditioner to run a lot harder than it normally would have.

I continued to use physical servers in my lab until about three years ago. At that time, I decided server virtualization had fi nally matured to the point that it would be reliable enough to meet my

IMA

GE FR

OM

SHU

TTERSTO

CK

VirtualizationReview.com | Virtualization Review | October/November 2010 | 25

Simplifi ed virtual lab management software allows even users to create their own testing environments. By Brien Posey

Page 28: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

26 | October/November 2010 | Virtualization Review | VirtualizationReview.com

FEATURE | Virtual Labs

needs. I still use some physical servers in my lab, but the vast majority of my lab servers are virtual.

There’s no question that using server virtualization has made my life easier. I now spend far less money on server hardware, electricity and repairs than I did before I virtualized my lab. But the story doesn’t end there.

Creating a Virtual Lab: ChallengesLike me, there are many other organizations that have discovered having a virtual lab can be benefi cial. Such labs are used for everything from software development to testing patches prior to deploying them on the production network. Regardless of what an organization uses its virtual lab for, almost everyone has found

that there are certain challenges associated with creating and maintaining a virtual lab.

One of the biggest challenges is the amount of time and effort involved in confi guring a virtual lab for a specifi c task. For example, last year someone asked me to write about Microsoft Offi ce Communications Server (OCS) and how it integrates with Microsoft Exchange Unifi ed Messaging. I write about Unifi ed Messaging on a semi-regular basis, so I already had a Unifi ed Messaging lab deployment in place.

Unfortunately, OCS can’t be deployed seamlessly. The installation process makes a lot of changes to Active Directory. There are also certain changes that have to be made to the way that existing Unifi ed Messaging servers are confi gured. Therefore, I had a choice to make. I could either install OCS onto my existing lab network and have the network confi guration permanently

changed, or I could create a completely separate lab network.

In the end, I decided I needed to preserve my existing AD confi guration, so I created an entirely new set of virtual servers for my OCS deployment. In doing so, I had to create and manually provision a couple of domain controllers, a mailbox server, a Unifi ed Messaging server, a front-end OCS, a mediation server and some workstations. It took me the better part of a week to get the new infrastructure up and running and correctly confi gured. All of that work took place before I even started writing the article.

Of course, this is just one example of a situation in which I had to create a rather elaborate lab deployment so I could write an article about it. The reality is that I fi nd myself deploying

new lab servers on a regular basis. What made this particular situation so diffi cult was that there were so many different servers involved, and each of those servers had to be confi gured in a certain way.

VMLogix LabManagerThankfully, provisioning and managing virtual labs is becoming easier. Various software vendors have begun offering virtual lab management software that’s designed to make the process more effi cient. One of the leading providers of Web management software is a company called VMLogix Inc.

VMLogix, which is being acquired by Citrix Systems Inc., offers a virtual lab management application appropriately named LabManager. It’s important to note that LabManager is not intended to replace an organization’s virtualization platform, but rather augment it. LabManager works

alongside an organization’s existing virtualization platform, but remains hypervisor-agnostic. This means organizations that want to use VMLogix LabManager aren’t locked into using a specifi c virtualization product. It works equally well with VMware VSphere, Citrix XenServer and Microsoft Hyper-V.

Given all the virtualization platforms that I mentioned are mature and fairly comprehensive, you might be wondering what functionality LabManager adds, and how that functionality can make virtual lab management easier.

LabManager is designed to enable the process of building and deploying a virtual lab environment a lot quicker and easier. Interaction with it is done through a Web interface that acts as a front-end to your virtualization host. This interface provides users with templates that they can use to rapidly deploy lab servers.

Of course, most server virtualization products also allow administrators to perform rapid, template-based deployments of virtual machines (VMs). On my own network, for example, I’ve used Hyper-V to create lab machine templates. That way, whenever I need to deploy a new VM I don’t have to worry about starting completely from scratch, because the template already contains a pristine Windows installation.

VMLogix LabManager takes things a step further. While LabManager will certainly allow you to use a template to deploy a virtual server, you also have the option of using a template to deploy an entire lab confi guration. If you think back to my OCS lab that I described earlier, you’ll recall that building the lab required me to create 10 different VMs. Had I been using LabManager at the time, I could have created and confi gured all of the required virtual servers through a single administrative action. In doing so, the deployment time could’ve potentially been reduced from nearly a week to a matter of minutes.

There’s no question that using server virtualization has made my life easier. I now spend far less money on server hardware, electricity and repairs than I did before I virtualized my lab.

Page 29: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

VirtualizationReview.com | Virtualization Review | October/November 2010 | 27

Such rapid deployments are possible because LabManager comes with a confi guration editor that allows you to defi ne exactly what you need in your virtual lab by using a collection of modular components. Virtual lab confi gurations consist of VM templates, network settings and policies, as shown in Figure 1. Ultimately, users can provision environments by simply selecting the desired confi guration and clicking on a Deploy button.

Of course, simplifying the provisioning process would be pointless if the lab management software wasn’t fl exible enough to meet an organization’s ever-changing needs. LabManager allows administrators to customize a server’s confi guration before the server is even deployed. Suppose, for example, that you needed to create a SQL server as a part of your virtual lab. There’s no need to create a Windows Server and then install SQL Server on top of it. Instead, you can simply tell LabManager to install SQL onto the new server. Upon doing so, LabManager will handle the provisioning process without the administrator having to perform any manual confi guration.

The same Web interface that allows applications to be added to virtual servers allows for other confi guration

changes as well. For example, you can change a server’s environment variables or add an additional NIC to the server, all before the server is even created.

Network IsolationEarlier I mentioned that I’ve created some templates of pristine Windows deployments that I can use whenever I need to create a new VM. The problem with creating such templates is that there are some Windows components that must be unique. Each server on a Windows network requires its own IP address, MAC address, NetBIOS name and SID. This simple fact can make cloning virtual servers a bit problematic.

Normally, if you want to clone a Windows deployment you must run the Sysprep utility against the virtual

server. Sysprep removes anything that uniquely identifi es the VM so that it won’t confl ict with other servers on your network. Even though Sysprep sounds like an ideal solution, it really isn’t.

One of the things that Sysprep removes is the server’s computer name. Because the computer name is required in order for a server to be a domain member, cloned servers must be manually joined to a domain. Furthermore, many server applications require the server to be a domain member. This means you can’t install such applications ahead of time.

In many ways, a Sysprep image is a lot like a brand-new Windows deployment (although some types of customizations can be included in the image). In fact, the fi rst time you boot a server that was cloned from a Sysprep image, you’re required to run a mini Windows Setup.

My point here is that cloning VMs in the traditional manner is ineffi cient. However, the issues that affect the cloning process don’t come into play when LabManager is being used. LabManager uses network isolation to ensure that VMs don’t confl ict with each other, even if they’re running completely identical confi gurations.

User ManagementBecause lab management software makes virtual lab deployment so much easier, some organizations are fi nding that virtual labs can be benefi cial to

Figure 1. LabManager allows you to deploy multiple servers simultaneously.

Figure 2. The administrator has the ability to approve user requests.

Page 30: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

28 | October/November 2010 | Virtualization Review | VirtualizationReview.com

FEATURE | Virtual Labs

departments other than IT. For instance, a company’s sales department might use a virtual lab as a way of demonstrating a product to a client. Likewise, organizations may choose to use virtual labs for user training or to train the help-desk staff. There’s really no limit to what a virtual lab can be used for.

Fortunately, VMLogix realized that once users begin to discover all of the potential uses for virtual labs, the IT department could be overwhelmed by user requests for virtual labs. Therefore, the company equipped LabManager with a self-service portal that users can use to create their own virtual lab confi gurations, though

whether or not users will be allowed to deploy their own virtual labs is ultimately up to the administrator (see Figure 2, p. 27). VMLogix LabManager allows administrators to apply a set of permissions to each confi guration so that end users will only be able to deploy virtual lab confi gurations for which they’ve been authorized.

Combating VM SprawlAlthough the primary purpose of LabManager is to facilitate the rapid deployment of lab environments, it’s also designed to solve another problem that’s common in lab environments—VM sprawl. This is an

important consideration because, without a mechanism in place to prevent sprawl, the excessive creation of virtual servers can quickly deplete the available hardware resources. This is especially true if an organization chooses to allow end users to provision and deploy their own virtual labs.

As I explained earlier, I tend to deploy additional lab servers every time I have to write about something new. Because creating a lab deployment without the aid of lab management software can involve so much work, I usually don’t delete the newly created VMs when I fi nish the article. After all, I never know when an

editor might have a question or might need a screen capture, and I really don’t want to have to completely rebuild the entire lab if that happens. Besides, I never know when I might be asked to write about a topic that would require a similar confi guration.

Last summer I began running out of disk space on one of my host servers, so I decided to see if there were any VMs that I could get rid of. As I looked through the server’s contents, I realized that I had dozens of unnecessary VMs. I couldn’t even remember what some of the VMs had been used for. Being that I operate a one-man shop, you can only imagine

how quickly lab servers could get out of hand in a larger organization.

LabManager helps with VM sprawl by retaining ownership of any virtual labs that are created. When a user creates a virtual lab, he doesn’t own the lab servers. Instead, LabManager leases the virtual lab to the user for a fi nite amount of time. When the lease nears its expiration date, LabManager alerts the user, and gives him the option of extending the lease if the virtual lab is still being used. If the user allows the lease to expire, then LabManager will automatically delete the virtual lab.

Administrators also have the option of imposing quotas on end users as a way of preventing them from creating an excessive number of lab servers. At any time, administrators can run a quota report similar to the one shown in Figure 3. That way, they can see how many VMs each user is using, as well as the amount of host server resources being consumed.

Not Just for IT ProsBecause of their cost and complexity, virtual labs were once used solely by IT professionals and software developers. Today, though, virtual lab management software has simplifi ed the creation of virtual labs to the point that even end users are able to create them. This has allowed virtual labs to be used for numerous creative purposes such as sales demos, product training and even help-desk functions. Perhaps the greatest benefi t to virtual lab management software, however, is that it greatly reduces the time and effort required in building virtual lab environments. VR

Brien Posey ([email protected]), MCSE, is a Microsoft MVP and the co-founder and vice president of research at Relevant Technologies, which specializes in IT best practices and information security. Posey is the author of many technical articles for online publications, and has also written numerous white papers.

Figure 3. Administrators can run a virtual lab quota report.

Page 31: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

VirtualizationReview.com | Virtualization Review | October/November 2010 | 29

HOW-TO

By David Davis

ILLUSTR

ATION

BY M

AX

GREG

OR

GUY

Confi guring the New vSphere 4.1 Active Directory AuthenticationWith the release of vSphere 4.1, VMware offi cially announced that this will be the last release that offers the company’s full ESX Server hypervisor with a console operating system (COS). In the future, the hypervisor used by all vSphere installations will be the thin ESXi Server. In order to make this more palatable for VMware admins, VMware has made a number of improvements to ESXi around offering greater command-line options, offi cially supporting a thin console and allowing it to join a Windows Domain as a member server. This last new feature is offi cially called “Directory Services.”

Before You BeginIf you were about to add a new Windows server to your AD Domain, you’d need to double check a few things. These are the same prerequisites for adding an ESX Server to your AD Domain. They are:

■ AD Domain Controller & DNS Server preconfi gured■ ESX/ESXi has DNS preconfi gured■ ESX Server can resolve its Fully Qualifi ed Domain Name (FQDN) ■ Network Time Protocol (NTP) Time Sync on ESX/ESXi

vSphere 4.1 Benefi tsOnce Directory Services is confi gured, you’ll be able to use your Windows AD username and password to administer an ESXi or ESX server using the vSphere client (directly to the ESX/ESXi Server), SSH, direct console user interface (DCUI) with ESXi, and local console on ESX Server or Tech Support Mode (TSM) on an ESXi server. In the past, the administration of ESX and ESXi servers using these methods would call for the use of the root username and password. Due to the lack of knowledge in confi guring other local Linux-like accounts, all too commonly the root account was generally used and the credentials could be too widely distributed to the server admin staff.

With these ESX/ESXi servers being able to join Windows AD, the account that would be used to log in is the user’s real Windows account. The benefi ts to this are: The user logging into the ESX/ESXi server knows these credentials and they’re his own, not a shared root account; no local user accounts need to be set up on the ESX/ESXi servers; and the actual Windows user account will be recorded in security logs on the server, which can be used for auditing.

Additionally, with the ESX/ESXi servers being members of the Windows domain, they can now be included in Windows-based inventory tools like other servers in the datacenter.

Confi guring AD AuthenticationOnce the prerequisites are met, the actual confi guration of Directory Services is simple, as it’s similar to joining a Windows PC to a domain. To make the ESX or ESXi server a member server in Windows AD, use the vSphere client, go to the Confi guration tab for the ESX/ESXi server, and click on Authentication Services under Software, then click on Properties. In the window that appears, add your Windows domain and the administrator credentials to join the domain.

Once joined, go to your AD domain controller and add a Windows group called “ESX Admins.” This will be your new point of control for AD. It’s here that you’ll add or remove Windows users or groups that will be authorized to administer the virtual infrastructure. From here, you can start using Windows AD user accounts that are members of this special group to perform local management on the ESX/ESXi servers.

Quick and EasyUsing the new vSphere 4.1 Directory Services to join an ESX or ESXi server to a Windows domain is quick and easy. While it’s a straightforward feature, it’s also a long-awaited feature that I applaud VMware for adding. I believe this new feature will make ESXi much more appealing to Windows admins by allowing them to use the same credentials they already use and further solidifying the place of vSphere in the enterprise datacenter. VR

David Davis is director of infrastructure at Train Signal Inc., which offers video training for IT pros. His certifi cations include vExpert, VCP, CISSP and CCIE #9369. He’s authored hundreds of articles and nine different training courses at Train Signal.

Page 32: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

30 | October/November 2010 | Virtualization Review | VirtualizationReview.com

EVERYDAY

By Rick Vanover

ILLUSTR

ATION

BY M

AX

GREG

OR

VIRTUALIZATION

Creating a Virtualization Service CatalogAll too often, organizations function without a Service Level Agreement (SLA) in the face of stringent expectations. Does this sound familiar? The infrastructure administrator should be one step ahead of those expectations with documented materials. One approach for the collection of virtualization technologies is providing a service catalog that documents what is available and at what level. The service catalog should dispel any myths of the private cloud. How many times have we heard something to the effect that simply because a server is a virtual machine (VM), off-site disaster recovery is built-in? The fact is, virtualization technologies are frequently blurred with cloud technologies.

As a style guide, infrastructure administrators can check out the Storage Networking Industry Association (SNIA) Tiered Services and Classifi cation presentation. While written specifi cally for storage, this can also be a format guide for all key categories of virtualized infrastructures. With this in mind, virtualization service catalogs should include all areas of virtualized infrastructures in a simple menu format. Here are some categories to consider:

Guest operating systems: This feature outlines which guest OSes are offered in virtual environments. This can help users keep up with new releases as well as unsupported offerings such as Windows 2000, which has fallen off of the Microsoft extended support list.

Data protection solutions: If standard backups are available, workload-protection solutions or full site-recovery solutions can include their respective recovery time objectives and recovery point objectives.

Workload provisioning: There’s an entire set of challenges around cost allocation and chargebacks, but the fact is that most organizations don’t charge back. A virtualization service catalog should document what types of workloads are available free of infrastructure charges. Many organizations have rules such as “500GB or less is free” in regards to enterprise storage. Virtualization revolves around advance capacity in terms of storage, processor and memory, implying there’s always headroom. Effective service

catalogs should document that headroom for organic server growth. As an alternative, when a request comes in above this amount, there should be a requirement to purchase infrastructure that will satisfy the request.

Performance objectives: There’s nothing more frustrating than trying to troubleshoot a symptom such as a vague report indicating a system is slow. It’s also a good idea to spell out disk latency for a LUN (or datastore). If infrastructure administrators provide the performance characteristics up front, there are no surprises later. There may be a revelation when it’s found that specifi c applications are consuming more than the established baseline,

and at that point the options are either to move them or add additional resources to mitigate the performance increase. Either way, simply having a baseline number for datastore latency ahead of time is a good idea.

Cost model: While most organizations don’t have a formal chargeback process, a cost model for unit costs should be defi ned. Obvious categories include guest OS licenses, agent software costs for in-guest packages and anything else on a per-OS basis.

Access and security models: All VMs are not created equal, nor are they accessed in the same way. Virtualization tools allow advanced role creation and assignment, but can become complicated. The infrastructure should provide a menu of what access is available and how.

The objective of virtualization service catalogs is to provide infrastructure menus to internal customers. This can look a lot like the SNIA resource—or it can look like a resource within your own organization that delivers a catalog of services. Above all else, you can’t satisfy your customers’ expectations if you don’t know what the current run rate is. VR

Rick Vanover ([email protected]), vExpert, VCP, MCITP, MCTS, MCSA, is an IT Infrastructure Manager for Alliance Data in Columbus, Ohio. He’s a 12-year IT veteran and online columnist for VR. Follow him on Twitter @RickVanover.

A virtualization service catalog should document what types of workloads are available free of infrastructure charges.

Page 33: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

VirtualizationReview.com | Virtualization Review | October/November 2010 | 31

ADVERTISING SALESINDEX

ID STATEMENTVirtualization Review magazine (ISSN 1941-2843) is published bi-monthly by 1105 Media, Inc., 9201 Oakdale Avenue, Ste. 101, Chatsworth, CA 91311. Complimentary subscriptions are sent to qualifying subscribers. Annual subscription rates payable in U.S. funds for non-qualifi ed subscribers are: U.S. $39.95, International $64.95.

SUBSCRIPTION INQUIRIES, BACK ISSUE REQUESTS, AND ADDRESS CHANGES: Mail to: Virtualization Review magazine, 9201 Oakdale Avenue, Suite 101, Chatsworth, CA 91311, Email [email protected] or call (818) 814-5353, fax number (818) 936-0267.

POSTMASTER: Send address changes to Virtualization Review magazine, 9201 Oakdale Avenue, Suite 101, Chatsworth, CA 91311. Canada Publications Mail Agreement No: 40612608. Return Undeliverable Canadian Addresses to 9121 Oakdale Avenue, Suite 101, Chatsworth, CA 91311.

COPYRIGHT STATEMENT© Copyright 2010 by 1105 Media, Inc. All rights reserved. Printed in the U.S.A. Reproductions in whole or part prohibited except by written permission. Mail requests to “Permissions Editor,” c/o Virtualization Review magazine, 2600 El Camino Real, Ste 300, San Mateo, CA 94403.

LEGAL DISCLAIMERThe information in this magazine has not undergone any formal testing by 1105 Media, Inc. and is distributed

without any warranty expressed or implied. Implementation or use of any information contained herein is the reader’s sole responsibility. While the information has been reviewed for accuracy, there is no guarantee that the same or similar results may be achieved in all environments. Technical inaccuracies may result from printing errors and/or new developments in the industry.

CORPORATE ADDRESSCorporate Headquarters: 1105 Media9201 Oakdale Ave. Ste 101, Chatsworth, CA 91311www.1105media.com

MEDIA KITSMedia Kits: Direct your Media Kit requests to Matt Morollo, VP Publishing, 508-532-1418 (phone), 508-875-6622 (fax), [email protected]

REPRINTSReprints: For single article reprints (in minimum quantities of 250-500), e-prints, plaques and posters contact:PARS InternationalPhone: 212-221-9595E-mail: [email protected]/QuickQuote.asp

LIST RENTALThis publication’s subscriber list, as well as other lists from 1105 Media, Inc., is available for rental. For more information, please contact our list manager, Merit Direct. Phone: 914-368-1000; E-mail: [email protected]; Web: www.meritdirect.com.

Matt Morollo VP, Publishing508-532-1418 phone508-875-6622 [email protected]

Al TianoNational Sales Manager818-814-5227 phone818-734-1529 [email protected]

Danna VedderMicrosoft Account Manager253-514-8015 phone775-514-0350 [email protected]

SALES STAFF

Tanya EgenolfAdvertising Sales Associate760-722-5494 phone760-722-5495 [email protected]

PRODUCTION

Jenny Hernandez-AsandasDirector, Print Production

Teresa AntonioProduction [email protected] phone818-734-1528 fax

Advertising IndexAdvertiser Page URLAberdeen 5 www.aberdeeninc.comCDW Corp. C2 www.cdw.comCitrix Systems, Inc. C4 www.citrix.comIBM Corporation 7 www.ibm.comIBM Corporation 9 www.ibm.comIBM Corporation C3 www.ibm.comQuest Software 17 www.quest.comRectiphy Corporation 13 www.rectiphy.comSymantec Corporation 11 www.symantec.comTechXtend 23 www.techxtend.comVeeam Software 3 www.veeam.com

Wiley Publishing 19 www.wiley.com

Editorial IndexCompany Page URLAdvanced Micro Devices Inc. 16 www.amd.comCisco Systems Inc. 22 www.cisco.comCitrix Systems Inc. 8, 12, 16, 20, 26, 32 www.citrix.comGoogle Inc. 15 www.google.comHewlett-Packard Co. 22 www.hp.comIBM Corp. 21 www.ibm.comIntel Corp. 16 www.intel.comNetApp 22 www.netapp.comOracle Corp. 22 www.oracle.comPano Logic Inc. 10 www.panologic.comSalesforce.com Inc. 15 www.salesforce.comTeradici Corp. 12 www.teradici.comVMLogix Inc. 26 www.vmlogix.comVMware Inc. 6, 12, 14, 20, 26, 29, 32 www.vmware.comWyse Technology Inc. 12 www.wyse.com

SUBSCRIPTION INQUIRIES, BACK ISSUE REQUESTS, AND ADDRESS CHANGES: Mail to: Virtualization Review magazine, 9201 Oakdale Avenue, Suite 101, Chatsworth, CA 91311, e-mail [email protected], call 818-814-5353 or fax to 818-936-0267.

Page 34: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

32 | October/November 2010 | Virtualization Review | VirtualizationReview.com

IMA

GE FRO

M SH

UT

TERSTO

CK

TAKE FIVE 5 TIPS AND TRICKS TO TAKE WITH YOU : : BY BRUCE HOARD, EDITOR IN CHIEF

Impressions of Microsoft

Let me know what you think about Microsoft’s virtualization and cloud offerings: [email protected].

Microsoft is defi nitely gunning for VMware. You can’t talk for long with

anyone on the Microsoft virtualization

team before the comments and

comparisons with VMware start popping

up. This whole us-versus-them attitude

was made crystal clear during VMworld

when Microsoft took out a full page ad

in the form of an open letter to VMware

customers in USA Today . It claimed

VMware is trying to lock customers

into three-year license agreements for

virtualization projects, which would not

be in customers’ best interests because

VMware “cannot provide you with the

breadth of technology, fl exibility or scale

that you’ll need to build a complete cloud

computing environment.”

TAKE

1

Microsoft is surprisingly low-key on its Virtual Desktop Infrastructure (VDI) Suites, which the company describes as a “collection of technology to help organizations deploy the basic infrastructure for VDI.” Redmond says that Hyper-V Server 2008 R2 is the virtualization platform of choice for both the standard

and premium versions of VDI Suites, and that management of the virtual infrastructure is handled by various components of System Center. Applications are handled by App-V because the Microsoft Desktop Optimization Pack is also included in the Standard Suite. Microsoft says the suites are well-suited for small to midsize businesses, but at the same Redmond says there’s no problem scaling. It all seems rather murky for Microsoft customers wanting to know more about VDI.

TAKE

2

Redmond talks a good game when it

comes to cloud, saying the company has

learned a lot from running such gigantic

cloud services as MSN (600 million

unique users), Bing (more than 4 billion

search queries monthly), 500 million-plus

active Windows Live IDs and 20 million users of the

Xbox Live gaming service. Throw in more than 500

government entities and 10,000 Windows Azure

users in 40 countries, and Microsoft has a pretty

good cloud argument to make.

TAKE

3

Ouch. Back in May, David Greschler, director of Microsoft

virtualization strategy, told members of the Microsoft Partner

Network about the availability of a “profi tability monitoring

tool” in the form of an interactive spreadsheet designed to

depict the advantages of adding a Microsoft virtualization

practice to their business. What would these Microsoft partners

fi nd if they dutifully used this tool? According to Greschler,

“As much as we’d like to tell you otherwise, if you plug your

numbers in to look at your work on a project-by-project basis,

the numbers could favor VMware.”

TAKE

5

Bob Muglia, president of the Microsoft Server and Tools Division, which reigns over the company’s virtualization/cloud operations, says all the right things about Microsoft’s future plans, noting his company’s goal is to offer customers everything they require to move into the cloud

“on their own terms.” This translates to: “Don’t worry—we won’t mess with your legacy applications.”

4TAKE

By now, everybody knows that Microsoft is making a big push into the virtualization/cloud computing market via Windows Server 2008 R2, the free Hyper-V and a little help from Citrix, who’s more than happy to share its technology and customers with Redmond. So it shouldn’t come as a surprise to anyone that Redmond’s building some serious momentum as more and more Microsoft customers become familiar with its virtualization and cloud wares.

Page 35: October/November 2010 | Vol. 3 No. 6 VirtualizationReview ...pdf.1105media.com/VRMmag/2010/701060191/OctNov.pdf · 12,000 customers. And with Veeam vPower™, Veeam is taking backup

Building the engines of a Smarter Planet:

How midsize businesses can collaborate better on a global scale.As a midsize company, you can now collaborate in new ways to drive innovation, with IBM LotusLive™. It’s an integrated collaboration tool that allows you to focus on growing your business rather than on managing your applications. No additional hardware or software. Just a browser. It’s a cloud-based service that delivers e-mail, plus online meetings, social networking and more, in one security-rich environment for internal and external collaboration. A powerful tool for midsize companies, leading to better decisions and smarter businesses. Consider the benefits:

1 Prices subject to change and valid in the U.S. only. Actual costs will vary depending on individual customer confi gurations and environment. IBM, the IBM logo, ibm.com, LotusLive, Smarter Planet and the planet icon are trademarks of International Business Machines Corp., registered in many jurisdictions worldwide. Other product and service names might be trademarks of IBM or other companies. A current list of IBM trademarks is available on the Web at www.ibm.com/legal/copytrade.shtml. © International Business Machines Corporation 2010.

IBM LotusLive is priced with midsize companies in mind. Collaboration offerings starting at

$3per user per month.1

View the IBM LotusLive demonstration at ibm.com/engines/lotuslive

1 Save time and money. The power of the cloud means there are no upfront deployment costs. IBM does the work. You focus on your business.

2 Improve productivity. With all the collaboration tools rolled into one, millions of users can see improvements in their businesses.

3 Keep your exchanges more secure. Security-rich infrastructure from IBM helpsprotect your data.

To learn more about services like LotusLive, connect with a Business Partner orget set up now, call 1-877-IBM-ACCESS or visit ibm.com/engines/lotuslive

Midsize businesses are the engines of a Smarter Planet.