11g Data Warehousing

download 11g Data Warehousing

of 37

Transcript of 11g Data Warehousing

  • 8/18/2019 11g Data Warehousing

    1/37

    Oracle Data Warehouse Strategic Update

    Ray Roccaforte

  • 8/18/2019 11g Data Warehousing

    2/37

  • 8/18/2019 11g Data Warehousing

    3/37

    The following is intended to outline our general

    product direction. It is intended for information

    purposes only, and may not be incorporated into any

    contract. It is not a commitment to delier any

    material, code, or functionality, and should not berelied upon in ma!ing purchasing decisions.

    The deelopment, release, and timing of any

    features or functionality described for "racle#s

    products remains at the sole discretion of "racle.

  • 8/18/2019 11g Data Warehousing

    4/37

    !our Pillars of Oracle"s DW Strategy

    $ "racle as a %ystems Proider& Database 'achine

    $ 'a(imum )eerage of 'assie *ardware +apabilities

    $ mbedded Analytics

    $ "racle as a DW %olutions Proider&$ (adata Intelligent Warehouse for Industries

  • 8/18/2019 11g Data Warehousing

    5/37

    !our Pillars of Oracle"s DW Strategy

    "racle as a %ystems Proider& Database 'achine

    $ 'a(imum )eerage of 'assie *ardware +apabilities

    $ mbedded Analytics

    $ "racle as a DW %olutions Proider&$ (adata Intelligent Warehouse for Industries

  • 8/18/2019 11g Data Warehousing

    6/37

    -

    Sun Oracle Data#ase $achine

    $ Database 'achine eliminates the comple(ityof deploying database systems

    $ alanced configuration$ (treme performance and integrated analytics out of the bo(

    $ Deploys in days rather than months$ Pre/built, tested, standard, supportable configuration

    $ Includes *ardware Installation 0 %ite Planning

    $ %oftware +onfiguration

  • 8/18/2019 11g Data Warehousing

    7/37

    %&adata Hard'are (rchitecture

    Data#ase )rid

    $ 1 co*pute serers2345

    Storage )rid

    $ 36 storage serers 2745

    $ 388 T High Speed dis!,or 99- T High +apacity dis!

    $, -. P+I !lash

    $ Data *irrored acrossstorage ser/ers

    Scalea#le )rid of industry standard ser/ers for +o*pute and

    Storage

    Infini.and et'or1

    $ Redundant 0)#3s s'itches

    $ Unified ser/er 4 storage net

    : 7838 "racle +orporation5

  • 8/18/2019 11g Data Warehousing

    8/37

    e' %&adata 6278

    $ ;ull Rac! Database 'achine for large deployments

    $ +onsolidation, or )arge ")TP and DW wor!loads$ +omplements e(isting

  • 8/18/2019 11g Data Warehousing

    9/37

    !our Pillars of Oracle"s DW Strategy

    $ "racle as a %ystems Proider& Database 'achine

    'a(imum )eerage of 'assie *ardware +apabilities

    $ mbedded Analytics

    $ "racle as a DW %olutions Proider&$ (adata Intelligent Warehouse for Industries

  • 8/18/2019 11g Data Warehousing

    10/3738

    38

    %&adata Hy#rid +olu*nar +o*pression

    $ Data is grouped by columnand then compressed

    $ :uery $ode for data

    warehousing$ "ptimi@ed for speed$ 06 co*pression typical$ %cans improe proportionally

    $ (rchi/al $ode for

    infreuently accessed data$ "ptimi@ed to reduce space

    $ 3B< compression is typical

    $ 4p to B8< for some data

  • 8/18/2019 11g Data Warehousing

    11/37

    33

    $ %un "racle Database 'achine proides semiconductorcache hierarchy

    Sun Oracle Data#ase $achine;

    !aster ra' scans

    %&adata dis1s

    00-. ra' dis1 capacity

    2 ).3sec

    %&adata S*art !lash +ache

    ,-. ra' capacity up to ,0-. user data

    ,0 ).3sec

    Data#ase DR($ +ache

    00). ra' capacity up to -. user data

    00 ).3sec

  • 8/18/2019 11g Data Warehousing

    12/37

    37

    In7$e*ory Parallel %&ecution

    $ Parallel uery processing across tables which are

    cached in memory

    $ *ow it wor!s&

    $ "racle determines appropriate tables for caching$ Tables are distributed across the buffer cache in all nodes

    $ During ueries, each node reads its local portion of the table

    $ ;ull parallel bandwidth accessing data in memory

    $ ?ot eery database obCects in a uery must be in memory to

    leerage in/memory access

    37

  • 8/18/2019 11g Data Warehousing

    13/37

    %&adata !lash +ache

    $ (adata has B T of ;lash& B- ;lash +ards aoid dis! controller bottlenec!s

    $  Automatically caches freuently/accessed hot# data in flash storage

    $ Intelligently manages ;lash& >ies speed of ;lash at cost of dis!

    $ (adata ;lash +ache achiees& O/er *illion IO3sec fro* S:= 8?@A Su#7*illisecond response ti*es 'ith ,0 ).3sec Buery throughput

    Infreuently

    4sed Data

    ;reuently4sed Data

  • 8/18/2019 11g Data Warehousing

    14/37

    3636

    Scanning fro* Dis1; %&adata S*art Scan

    $ (adata storage serers implement data

    intensie processing in storage$ Row filtering based on EwhereF predicate

    $ +olumn filtering

    $ Goin filtering

    $ Incremental bac!up filtering

    $ Scans on Hy#rid +olu*nar +o*pressed data

    $ Scans on encrypted data

    $ Data $ining *odel scoring

    $ 38( reduction in data sent to D serers

    is common

    $ ?o application changes needed$ Processing is automatic and transparent

    $ en if cell or dis! fails during a uery

  • 8/18/2019 11g Data Warehousing

    15/37

    !our Pillars of Oracle"s DW Strategy

    $ "racle as a %ystems Proider& Database 'achine

    $ 'a(imum )eerage of 'assie *ardware +apabilities

    mbedded Analytics

    $ "racle as a DW %olutions Proider&$ (adata Intelligent Warehouse for Industries

  • 8/18/2019 11g Data Warehousing

    16/37

    Oracle In7Data#ase (nalytics

    Oracle In7Data#ase (nalytics

    $  Analytics ealuated EcloseF to the data

    $ Integrated

    $ %ecure

    Oracle Data $ining; Unco/er 4 Predict

    $  Adanced Predictie

     Analytics

    $  Adanced Algorithms

    $ mbedded andIntegrated

    Oracle O=(P; (nalyCe 4 Su**ariCe

    $ %marter, ;aster I

    $ Improed Deeloper

    Productiity

    $ mbedded andIntegrated

  • 8/18/2019 11g Data Warehousing

    17/37

    In7Data#ase Data $ining%&adata S*art Scan

    $ 'odel scoring  

    performed

    in the (adata

    %torage %erer 

    $ Dramatic 27/B

  • 8/18/2019 11g Data Warehousing

    18/37

    %&adata Integration 'ith S(S

    $ %core %A% 'odels in (adata

    $ Import %A% models 2ia P'')5 to

    "racle Data 'ining

    $ )ogistic and 'ultiple Regressionmodels aailable today

    $ *igh/performance scoring in

    (adata

    $ %core directly against database tables

    $ 'odel scoring is e(ecuted in storage

    tier using natie "racle D' capabilities

  • 8/18/2019 11g Data Warehousing

    19/37

    Oracle O=(P

    $ Rapid Deelopment of rich dimensional analytics$ nhances the analytic content of usiness intelligence application

    $ (cellent uery performance for ad/hoc H unpredictable uery

    $ ;ast, incremental updates of data sets

    $ mbedded in the "racle database instance and storage.$ %afe, secure and manageable

    $ enefits from and enhances (adata architecture$  Automatically gets full benefit of Database Machine Flash Cache

    $ ")AP uery IH" tends to be high olume, random reads

  • 8/18/2019 11g Data Warehousing

    20/37

    Oracle O=(P g Rich +alculations Deli/ered to S:=7.ased .I -ools

  • 8/18/2019 11g Data Warehousing

    21/37

    O=(P $D6 dri/er fro* Si*#a for $S %&cel

    : 788 "racle +orporation

    $D6 Parser $D6 Pro/ider %ngine

    S:= )enerator 

    S:= :uery of +u#e

    %&cel $D6 :uery

    $D6 Pro/ider forOracle O=(P

    %&cel 200

    (cel 7889

  • 8/18/2019 11g Data Warehousing

    22/37

    !our Pillars of Oracle"s DW Strategy

    $ "racle as a %ystems Proider& Database 'achine

    $ 'a(imum )eerage of 'assie *ardware +apabilities

    $ mbedded Analytics

    "racle as a DW %olutions Proider&$ (adata Intelligent Warehouse for Industries

  • 8/18/2019 11g Data Warehousing

    23/37

    $ Intelligence J Industry/specific data models

     J Pac!aged adanced analytics

    $ ... with (treme Performance

     J Improe uery performance 38/388(

    with (adata

    $ K at a )ower +ost

     J %implify your infrastructure

    $ K for ;ast Results

     J Gumpstart deelopment / delieralue uic!ly

     J  Automatically e(ploit "racle

    performance and analytic capabilities

    %&adata Intelligent Warehouse for Industries( +o*plete Data Warehouse Solution for Industries

  • 8/18/2019 11g Data Warehousing

    24/37

    %&adata Intelligent Warehouse Solution for IndustriesPac!aged (perience and Technology

    $ )eerage enterprise/wide data model forindustries

    $ >ain insights using prepac!agedadanced analytics

    .usiness Insight

    $ Improe uery performance 38 /388(

    $ >row solution to irtually any scale%&tre*ePerfor*ance

    $ Gumpstart deelopment and delier alueuic!ly

    $ )ower ris!s

    !ast-i*e7to7Ealue

  • 8/18/2019 11g Data Warehousing

    25/37

    %&adata Intelligent Warehouse for RetailPac1aged %&perience and -echnology

    +opyright : 7838, "racle and H or its affiliates. All rights resered.

    $ Improe uery performance38 /388(

    $ >row solution to irtually any scale

    %&tre*ePerfor*ance

    $ Gumpstart deelopment and delieralue uic!ly

    $ )ower ris!s

    !ast Results

    $ )eerage enterprise data model forretail

    $ >ain insights using prepac!agedadanced analytics

    .usinessInsight

    $ )eerage enterprise data model forretail

    $ >ain insights using prepac!agedadanced analytics

    .usinessInsight

  • 8/18/2019 11g Data Warehousing

    26/37

  • 8/18/2019 11g Data Warehousing

    27/37

    %&adata Intelligent Warehouse for Retail+onsolidated Eie' of the %ntire .usiness

    Oracle RetailData $odel

    $ POS

    $ Order $g*t$ Wor1force

    Scheduling

    Sell Side$ PO $g*t

    $ Warehouse$ Pricing$ +osting

    .uy Side

    $ HR$ !inance$ Sales !orecasting

    Inside

    $ Partners$ Suppliers$ Distri#utors

    Outside

    200 Industry7specific$easures 4 ?PIs

  • 8/18/2019 11g Data Warehousing

    28/37

    Interacti/e Dash#oards and Reportsmpowering nd 4sers

     Anticipate out/of/stoc! usingstatistical forecasts

    %&a*ple (nalytics; Retail

     Associate products using a mar!etbas!et analysis

     Analy@e store traffic and shopperconersion rates by time of day

    4nderstand effectieness of promotions by store

     Analy@e comparatie storeperformance oer time

    2F+opyright : 7838, "racle and H or its affiliates. All rights resered.

  • 8/18/2019 11g Data Warehousing

    29/37

    %&adata Intelligent Warehouse for +o**unications

    %ame idea, but specific to +ommunications industry

  • 8/18/2019 11g Data Warehousing

    30/37

    Interacti/e Dash#oards and Reportsmpowering nd 4sers

    Identify customers that are li!ely tochurn

    %&a*ple (nalyses; +o**unications

    'onitor dropped call rate oer time todetect failing networ! elements

    +ompare billed and actual +DRs to identifybilling errors

     Analy@e sales growth oer time byproduct and organi@ation

    G+opyright : 7838, "racle and H or its affiliates. All rights resered.

  • 8/18/2019 11g Data Warehousing

    31/37

    %&adata Intelligent WarehousePac1aged %&perience and -echnology

    +opyright : 7838, "racle and H or its affiliates. All rights resered.

    $ Improe uery performance38 /388(

    $ >row solution to irtually any scale

    %&tre*ePerfor*ance

    $ Gumpstart deelopment and delier

    alue uic!ly

    $ )ower ris!s

    !ast Results

    $ )eerage enterprise data model forIndustries

    $ >ain insights using prepac!agedadanced analytics

    .usinessInsight

  • 8/18/2019 11g Data Warehousing

    32/37

    %&adata Intelligent WarehousePac1aged %&perience and -echnology

    +opyright : 7838, "racle and H or its affiliates. All rights resered.

    $ Improe uery performance38 /388(

    $ >row solution to irtually any scale

    %&tre*ePerfor*ance

    $ Gumpstart deelopment and delier

    alue uic!ly

    $ )ower ris!s

    !ast Results

    $ )eerage enterprise data model forIndustries

    $ >ain insights using prepac!agedadanced analytics

    .usinessInsight

  • 8/18/2019 11g Data Warehousing

    33/37

    +usto* Solutions?ey I*ple*entation -as1s

    +hallenges;

    $ Design a data *odel that satisfies e&isting and future reBuire*ents

    $ %*ploy a di/erse s1ill set

    $ IntegrateA i*ple*entA ad*inister and tune disparate technologies

  • 8/18/2019 11g Data Warehousing

    34/37

    %&adata Intelligent Warehouse.est Practice I*ple*entation

    +o*plete !ast Results =o'er Ris1

    .enefits;

    $ UtiliCe an enterprise data *odel for retail

    $ =e/erage your e&isting Oracle e&pertise

    $ I*ple*ented using Oracle data 'arehousing #est practices

  • 8/18/2019 11g Data Warehousing

    35/37

    !ast ResultsRetail Perfor*ance $etrics and Insight out of the #o&

    %peed to Value, %implified Deployment with predictable cost

    .uild fro* Scratch 'ith

    .est of .reed (pproachOracle %&adata

     Intelligent Warehouse

    'ee1s or *onths*onths or years

     SiCing and +onfiguration

    Define $etrics 4 Dash#oards

     -raining 4 Roll7out

      Data Integration

      Data Integration

      SiCing and +onfiguration

     (nalysis and Design  Define $etrics 4 Dash#oards

      (nalysis and Design

     -raining 4 Roll7out

      -raining 4 Roll7out

  • 8/18/2019 11g Data Warehousing

    36/37

    %&adata Intelligent Warehouse Solution for IndustriesPac!aged (perience and Technology

    .usiness Insight

    %&tre*e Perfor*ance

    !ast -i*e7to7Ealue

  • 8/18/2019 11g Data Warehousing

    37/37

    Su**ary; Oracle"s DW Strategy

    $ "racle as a %ystems Proider& Database 'achine

    $ 'a(imum )eerage of 'assie *ardware +apabilities

    $ mbedded Analytics

    $ "racle as a DW %olutions Proider&$ (adata Intelligent Warehouse for Industries