warm up for sap hana

Upload: suryya-kanta-adhikary

Post on 03-Apr-2018

214 views

Category:

Documents


0 download

TRANSCRIPT

  • 7/28/2019 Warm Up for Sap Hana

    1/19

    Chapter 5:

    Foundations for a New

    Enterprise Application

    Development Era

    Learning Map

    Founda'onsfora

    NewEnterprise

    Applica'on

    DevelopmentEra

    Founda'onsof

    DatabaseStorage

    Techniques

    TheFutureof

    Enterprise

    Compu'ng

    Advanced

    Database

    Storage

    Tech-

    niques

    In-Memory

    Database

    Operators

  • 7/28/2019 Warm Up for Sap Hana

    2/19

    Implications on Application

    Development

    Dr.-Ing. Jrgen Mller

    Simplified Application

    DevelopmentTradi&onal Column-oriented

    Applica&on

    cache

    Database

    cache

    Prebuiltaggregates

    Rawdata

    Nocachesneeded Noredundantdata/

    objects

    Nomaintenanceofindexesoraggregates

    Datamovementsareminimized

    111

  • 7/28/2019 Warm Up for Sap Hana

    3/19

    SAP ERP Financials onIn-Memory Technology

    In-memorycolumndatabaseforanenterprisesystem

    Combinedworkload(parallelOLTP/OLAPqueries) Leveragein-memorycapabili&esto Reduceamountofdata Aggregateon-the-fly Runanaly&c-stylequeries(toreplacematerializedviews) Executestoredprocedures

    UseCase:SAPERPFinancialssolu&on Postandchangedocuments Displayopenitems Rundunningjob Analy&calqueries,suchasbalancesheet

    112

    Current Financials

    Solutions

    113

  • 7/28/2019 Warm Up for Sap Hana

    4/19

    Onlybasetables,algorithms,andsomeindexes

    The TargetFinancials Solution

    114

    BKPFaccounting documents

    BSEG

    sum tables

    secondary indices

    dunning data

    change documents

    CDHDRCDPOS

    MHNKMHNDBSAD

    BSAKBSASBSIDBSIKBSIS

    LFC1

    KNC1

    GLT0

    In-Memory Financials on

    SAP ERP

    115

  • 7/28/2019 Warm Up for Sap Hana

    5/19

    BKPFaccounting documents

    BSEG

    116

    In-Memory Financials onSAP ERP (prototype)

    Feasibility of Financials on

    In-Memory Technology

    Modifica&onsonSAPFinancials Removedsecondaryindices,sumtablesandpre-calculatedandmaterialized

    tables

    Reducecodecomplexityandsimplifylocks InsertOnlytoenablehistory(changedocumentreplacement) Addedstoredprocedureswithbusinessfunc&onality

    Europeandivisionofaretailer ERP2005ECC6.0EhP3 5.5TBsystemdatabasesize Financials:

    23millionheaders/8GBinmainmemory 252millionitems/50GBinmainmemory

    (includinginvertedindicesforjoina_ributesandinsertonlyextension)117

  • 7/28/2019 Warm Up for Sap Hana

    6/19

    DBMS IMDBBKPF 8.7 GB 1.5 GBBSEG 255 GB 50 GB

    Secondary Indices 255 GB -Sum Tables 0.55 GB -Complete 519.25 GB 51.5 GB

    263.7 GB 51.5 GB

    Reduction by a Factor 10

    118

    Dunning Run

    Dunningrundeterminesallopenanddueinvoices Customerdefinedquerieson250Mrecords Currentsystem:20min Newlogic:1.5sec

    In-memorycolumnstore Parallelizedstoredprocedures SimplifiedFinancials

    119

  • 7/28/2019 Warm Up for Sap Hana

    7/19

    Why?

    Beingabletoperformthedunningruninsuchashort&melowersTCO

    Addmorefunc&onality! Runotherjobsinthemean&me!-inamul&-tenancycloud

    setuphardwaremustbeusedwisely

    120

    Bring Application Logic Closer

    to the Storage Layer

    Selectaccountstobedunned,foreach: SelectopenaccountitemsfromBSID,foreach:

    Calculateduedate Selectdunningprocedure,levelandarea

    CreateMHNKentries Createandwritedunningitemtables

    121

  • 7/28/2019 Warm Up for Sap Hana

    8/19

    Bring Application Logic Closerto the Storage Layer

    Selectaccountstobedunned,foreach: SelectopenaccountitemsfromBSID,foreach:

    Calculateduedate Selectdunningprocedure,levelandarea

    CreateMHNKentries Createandwritedunningitemtables

    1 SELECT

    10,000 SELECTs

    10,000 SELECTs

    31,000 Entries

    122

    Bring Application Logic Closer

    to the Storage Layer

    Selectaccountstobedunned,foreach: SelectopenaccountitemsfromBSID,foreach:

    Calculateduedate Selectdunningprocedure,levelandarea

    CreateMHNKentries Createandwritedunningitemtables

    123

    1 SELECT

    10,000 SELECTs

    10,000 SELECTs

    31,000 Entries

    Onesingle

    storedprocedure

  • 7/28/2019 Warm Up for Sap Hana

    9/19

    Selectaccountstobedunned,foreach: SelectopenaccountitemsfromBSID,foreach:

    Calculateduedate Selectdunningprocedure,levelandarea

    CreateMHNKentries Createandwritedunningitemtables

    1 SELECT

    10,000 SELECTs

    10,000 SELECTs

    31,000 Entries

    Onesingle

    storedprocedure

    Calculatedon-the-fly

    Bring Application Logic Closerto the Storage Layer

    124

    Results

    125

    Hardware: 4 CPUs x 6 Cores (Intel Dunnington), 256GB RAMCustomer Data: 250mio line items, 380k open, 200k due

    # Opera'on HANA2Version Variant2 Variant31 SelectOpenItems 0.63s 1.01s

    (incl.T047KNB5Join) 0.6s(incl.T047KNB5Join)2 Duedate,dunninglevel 27s deferredtoaggrega&on 0.5s3 Filter1(VerifyDunninglevels) 19s 1.1s 0.5s4 Filter2(CheckLastDunning) 15s 0.8s 0.4s5 GenerateMHNK(Aggregate) donein#1 1.2s donein#16 GenerateMHND(ExecuteFilters) donein#1 140ms donein#1

    Total 1Minute 3.0s(#3,#4exec.inparallel) 1.5s(#3,#4exec.inparallel)

    OriginalVersionneededabout20minutes

    Factor800xaccelera&onachieved

  • 7/28/2019 Warm Up for Sap Hana

    10/19

    Dunning Application

    126

    Dunning Application

    127

  • 7/28/2019 Warm Up for Sap Hana

    11/19

    Available-to-PromiseCheck

    CanIgetenoughquan&&esofarequestedproductonadesireddeliverydate?

    Goal:Analyzeandvalidatethepoten&alofin-memoryandhighlyparalleldataprocessingforAvailable-to-Promise(ATP)

    Challenges Dynamicaggrega&on Instantreschedulinginminutesvs.nightlybatchruns

    Real-&meandhistoricalanaly&cs Outcome

    Real-&meATPcheckswithoutmaterializedviews Ad-hocrescheduling Nomaterializedaggregates

    128

    In-Memory Available-to-Promise

    129

  • 7/28/2019 Warm Up for Sap Hana

    12/19

    Demand Planning

    Flexibleanalysisofdemandplanningdata

    Zoomingtochoosegranularity Filterbycertainproductsor

    customers

    Browsethrough&mespans Combina&onofloca&on-based

    geodatawithplanningdatainanin-memorydatabase

    Externalfactorssuchasthetemperature,orthelevelofcloudinesscanbeoverlaidtoincorporatetheminplanningdecisions

    130

    GORFID

    HANAforStreamingDataProcessing UseCase:In-MemoryRFIDDataManagement Evalua&onofSAPOER Prototypicalimplementa&onof:

    RFIDReadEventRepositoryonHANA DiscoveryServiceonHANA(10billion

    datarecordswithca.3secondsresponse&me) FrontendsforiPhone,iPad2

    KeyFindings: HANAissuitedforstreamingdata

    (usingbulkinserts)

    Analy&csonstreamingdataisnowpossible131

  • 7/28/2019 Warm Up for Sap Hana

    13/19

    Near Real-Time as aConcept

    Discovery Service

    Read Event

    Repositories

    Verification

    Services

    SAP HANA

    up to 8,000 read

    event notifications

    per second

    up to 2,000

    requests

    per second

    PA

    Bulk load every

    2-3 seconds:

    > 50,000 inserts/s

    132

    Learning Map

    Founda'onsfora

    NewEnterprise

    Applica'on

    DevelopmentEra

    Founda'onsof

    DatabaseStorage

    Techniques

    TheFutureof

    Enterprise

    Compu'ng

    Advanced

    Database

    Storage

    Tech-

    niques

    In-Memory

    Database

    Operators

  • 7/28/2019 Warm Up for Sap Hana

    14/19

    Views

    Dr.-Ing. Jrgen Mller

    Dynamic ViewsExcel SAP

    BusinessObjectsExplorer

    Any SoftwarePresentation Layer

    View View

    View View View...

    View

    View Layer(Calculations, Filter, ...)

    Persistency Layer

    (Main Memory)

    Object Hierarchy

    Node Tables Node TablesNode Tables Node Tables

    View Other DB

    LogicalLog

    i

    i i i i

    DBPersistence

    Store

    Restart

    Write CompleteObjects

    Recovery

    135

  • 7/28/2019 Warm Up for Sap Hana

    15/19

    Learning Map

    Founda'onsfora

    NewEnterprise

    Applica'on

    DevelopmentEra

    Founda'onsof

    DatabaseStorage

    Techniques

    TheFutureof

    Enterprise

    Compu'ng

    Advanced

    Database

    Storage

    Tech-

    niques

    In-Memory

    Database

    Operators

    Bypass Solution

    Dr.-Ing. Jrgen Mller

  • 7/28/2019 Warm Up for Sap Hana

    16/19

    Bypass Solution Allows a SmoothTransition to In-Memory Technology

    Millionsofoldun-op&mizedlinesofcodeatthecustomerssiteTransi&onrequired

    Row-storereplacement Part-for-partreplacementwithbypass Transformrow-storetocolumn-storeonthefly Changeofapplica&oncode

    140

    OLAPEngine

    Traditional DBw/ Cubes

    Bypass Solution for Transition

    ERP

    Traditional DB

    ETL

    Traditional BITodays System

    141

    BIA

    SAP

    Excel

    .

    .

    .

    BusinessObjects

  • 7/28/2019 Warm Up for Sap Hana

    17/19

    OLAPEngine

    ETL

    Traditional DBIMDB

    SSD

    Bypass Solution for Transition

    Traditional BITodays System with IMDB

    ERP

    Traditional DBw/ Cubes

    BIA

    SAP

    Excel

    .

    .

    .

    BusinessObjects

    STEP 1: Install and run the in-memory database in parallel

    NewApplications

    OLAPEngineETL

    Traditional DBIMDB

    SSD

    Bypass Solution for Transition

    Traditional BITodays System with IMDB

    ERP

    Traditional DBw/ Cubes

    BIA

    SAP

    Excel

    .

    .

    .

    BusinessObjects

    STEP 1: Install and run the in-memory database in parallel

    STEP 2: Generate business value from the first day on

  • 7/28/2019 Warm Up for Sap Hana

    18/19

    NewApplications

    OLAPEngine

    EL

    Traditional DBIMDB

    SSD

    Bypass Solution for Transition

    Traditional BI with IMDBTodays System with IMDB

    ERP

    IMDBw/o cubes

    BIA

    SAP

    Excel

    .

    .

    .

    BusinessObjects

    STEP 1: Install and run the in-memory database in parallel

    STEP 2: Generate business value from the first day on

    STEP 3: Re-create traditional BI w/o cubes in IMDB

    NewApplications

    OLAPEngineEL

    IMDB

    SSD

    Bypass Solution for Transition

    ERP

    IMDBw/o Cubes

    BIA

    SAP

    Excel

    .

    .

    .

    BusinessObjects

    SAP

    BusinessObjects

    Excel

    BI 2.0

    .

    .

    .

    OLAP

    Engine

    Traditional DB

    Traditional BI with IMDBTodays System with IMDB

    STEP 1: Install and run the in-memory database in parallel

    STEP 2: Generate business value from the first day on

    STEP 3: Re-create traditional BI w/o cubes in IMDB

    STEP 4: Introduce next-gen BI running in parallel without materialized views

  • 7/28/2019 Warm Up for Sap Hana

    19/19

    NewApplications

    IMDB

    SSD

    Bypass Solution for Transition

    ERP

    SAP

    BusinessObjects

    Excel

    BI 2.0

    .

    .

    .

    OLAP

    Engine

    Traditional DB

    Todays System with IMDB

    STEP 1: Install and run the in-memory database in parallel

    STEP 2: Generate business value from the first day on

    STEP 3: Re-create traditional BI w/o cubes in IMDB

    STEP 4: Introduce next-gen BI running in parallel without materialized views

    STEP 5: Eliminate all the traditional BI, virtualize all in-memory BI, using non-materialized views

    NewApplications

    IMDB

    SSD

    OLTP & OLAP

    Bypass Solution for Transition

    ERP FutureReleases

    SAP

    BusinessObjects

    Excel

    BI 2.0

    .

    .

    .

    OLAP

    Engine

    Future System with IMDB

    STEP 1: Install and run the in-memory database in parallel

    STEP 2: Generate business value from the first day on

    STEP 3: Re-create traditional BI w/o cubes in IMDB

    STEP 4: Introduce next-gen BI running in parallel without materialized views

    STEP 5: Eliminate all the traditional BI, virtualize all in-memory BI, using non-materialized views

    STEP 6: Run ERP and BI on IMDB