tabular mode and bism intro

Upload: adriana-oana

Post on 01-Jun-2018

230 views

Category:

Documents


0 download

TRANSCRIPT

  • 8/8/2019 Tabular Mode and BISM Intro

    1/22

    SQL Server Analysis Services

    Introduction to Tabular Mode and BISM

  • 8/8/2019 Tabular Mode and BISM Intro

    2/22

    Josh Fennessy

    • BI Architect – BlueGranite, Inc (http://www.blue-ranite.co!"

    http://www.blue-granite.com/http://www.blue-granite.com/

  • 8/8/2019 Tabular Mode and BISM Intro

    3/22

    Agenda

    •  Anal#$i$ Ser%ice$ – be&ore toda#

    • The BI Se!antic Model

    • Tabular Mode Architecture• 'e!on$tration

    • e%iew / )ue$tion$ / co!!ent$

  • 8/8/2019 Tabular Mode and BISM Intro

    4/22

      QL ERVER ANALY I

    ERVICE

     A brie& hi$tor#

  • 8/8/2019 Tabular Mode and BISM Intro

    5/22

    SSAS’ story

    • Ba$ed on *+A technolo# purcha$ed b# MST &ro! anora!aSo&tware in 0

    • *&&iciall# relea$ed in 1 a$ 2*+A Ser%ice$ in S3+ 4.5

    • ena!ed in S3+ 6555 to SSAS

    • Man# new &eature$ deli%ered in SSAS 6557

     – 'ata !inin

     – 8'M

  • 8/8/2019 Tabular Mode and BISM Intro

    6/22

    ANALY I ERVICE T!AY

    Broad adoption

    “Customers in the Magic Quadrant survey report that their Microsoft averagedeployment sizes are now larger than any other vendor in the survey in termsof users.” 

    “Use of OL! functionality "y Microsoft customers is more than dou"le that forthe rest of the survey respondents.” 

    #ource$ %artner Magic Quadrant for &' !latforms( )*++

    Highest rated infrastructure and development tools

    “Microsoft customers rate its &' platform infrastructure and development tools among the highest compared toother vendors( and a higher percentage of customers use them e,tensively.”

    #ource$ %artner Magic Quadrant for &' !latforms( )*++

    Large ecosystem

    -ide availa"ility of s/ills is among the top reasons customers select Microsoft over competing vendors.” 

    #ource$ %artner Magic Quadrant for &' !latforms( )*++

     

    http://www.gartner.com/technology/media-products/reprints/microsoft/vol2/article15/article15.htmlhttp://www.gartner.com/technology/media-products/reprints/microsoft/vol2/article15/article15.htmlhttp://www.gartner.com/technology/media-products/reprints/microsoft/vol2/article15/article15.htmlhttp://www.gartner.com/technology/media-products/reprints/microsoft/vol2/article15/article15.htmlhttp://www.gartner.com/technology/media-products/reprints/microsoft/vol2/article15/article15.htmlhttp://www.gartner.com/technology/media-products/reprints/microsoft/vol2/article15/article15.html

  • 8/8/2019 Tabular Mode and BISM Intro

    7/22

    Analysis ServicesVision

    BI SemanticModel

    Build on the strengths

    and success ofAnalysis Services andexpand its reach to amuch broader userbase

    Embrace the relationaldata model – wellunderstood bydevelopers and I !ros

    Bring together therelational andmultidimensional

    models under asingle uni"ed BIplatform – best ofboth worlds#

    !rovide $exibility in theplatform to suit thediverse needs of BIapplications

    ANALY I ERVICE T"RR#

  • 8/8/2019 Tabular Mode and BISM Intro

    8/22

    $%siness Intelligence

    Se&antic "odel

  • 8/8/2019 Tabular Mode and BISM Intro

    9/22

    $I E"ANTIC "!EL

    One Model for all End User Experiences

    %lient oolsAnalytics& 'eports& Scorecards&

    (ashboards& %ustom Apps

    (ata Sources(atabases& )*B Applications& *(ata +eeds&

    Spreadsheets& ext +iles

    BI Semantic Model

    (ata model

    Business logicand ,ueries

    (ata access

     eam BI!ower!ivot for

    Share!oint

    !ersonal BI!ower!ivot for

    Excel

    *rgani-ational BIAnalysisServices

  • 8/8/2019 Tabular Mode and BISM Intro

    10/22

    $I Se&antic "odel#hat a'o%t e(isting Analysis Services a))lications*

    .ew applications

    .ew technology options

    /(enali0

    ExistingapplicationsBased on 1ni"ed (imensional

    Model

    ExistingapplicationsEvery 1(M becomes a BI

    Semantic Model

    ExistingapplicationsBased on 1ni"ed (imensional

    Model

    After 'M

  • 8/8/2019 Tabular Mode and BISM Intro

    11/22

    BI Semantic Model

    (ata model

    Business logicand ,ueries

    (ata access '*)A! M*)A! Verti!a,(irect2uery

    M(3 (A3

    Multi4dimensional

     abular

     hird4partyapplications

    'eportingServices

    Excel !ower!ivot

    (atabases )*B Applications +iles *(ata +eeds %loud Services

    Share!ointInsights

    $I " ARC+ITECT,RE

  • 8/8/2019 Tabular Mode and BISM Intro

    12/22

    $I " FEAT,RE

    • ich data !odelin

    capabilitie$

    • Sophi$ticated bu$ine$$

    loic u$in M'9 and 'A9

    • ine-rained $ecurit# –

    row/cell le%el

    • nterpri$e capabilitie$ –

    !ulti-lanuae and

    per$pecti%e$

    'ichness• ;ertia) &or hih

    per&or!ance, M*+A &or

    !i$$ion critical $cale

    • 'irect3uer# and *+A

    &or real-ti!e acce$$ to

    data $ource$• State-o&-the-art

    co!pre$$ion alorith!$

    • Scale$ to lare$t

    enterpri$e $er%er$

    Scalability• Multi-di!en$ional and tabular

    !odelin e

  • 8/8/2019 Tabular Mode and BISM Intro

    13/22

      CENARI- E.CEL VER ALE

    "!EL

    BI Semantic Model

    (ata model

    Business logicand ,ueries

    (ata access

    S2) Server (ynamics %'M

    End1ser

    Model (eveloper

    M(3 (A3

    Multi4

    dimensional abular

    Verti!a,

  • 8/8/2019 Tabular Mode and BISM Intro

    14/22

    #+AT !E $I " ! FR

    "E*3ui> ti!e? ic@ which one i$ a Tabular Model.

  • 8/8/2019 Tabular Mode and BISM Intro

    15/22

      A !ATA ACCE / TRA0E

    xVelocity

    In4memory column store5typical 67x compression

    Brute force memory scans5high performance by default5no tuning re,uired

    Basic paging support5 data

    volume mostly limited tophysical memory

    MOLAP

    (is8 based store5 typical 9xcompression

    (is8 scans with in4memorysubcube caching5 aggregationtuning re,uired

    Extensive paging support5

    data volumes can scale tomultiple terabytes

    DirectQuery

    !asses through (A3 ,ueries :calculations5 fully exploitsbac8end database capabilities

    .o support for M(3 ,ueries5no support for data sourcesother than S2) Server ;in

    (enali<

    OLAP

    !asses through fact tablere,uests5 not recommendedfor large dimension tables

    Supports most relational datasources5 no support foraggregations except S2)

    Server indexed views

  • 8/8/2019 Tabular Mode and BISM Intro

    16/22

    C, T" CALC,LATIN

    DA! Based on Excel formulas and

    relational concepts – easy toget started

    %omplex solutions re,uiresteeper learning curve –row="lter context& %alculate&etc>

    %alculated columns enablenew scenarios& however nonamed sets or calcmembers

    MD! Based on understanding of

    multidimensional concepts –higher initial learning curve

    %omplex solutions re,uiresteeper learning curve –%urrentMember& overwritesemantics& etc>

    Ideally suited for apps thatneed the power ofmultidimensionalcalculations – scopes&assignments& calc members

  • 8/8/2019 Tabular Mode and BISM Intro

    17/22

    +# +,L! I $,IL! "Y

    A L,TIN*

     wo Visual Studio ;BI(S< pro?ect types in (enali Multidimensional pro?ect – with M(3 and M*)A!='*)A!

     abular pro?ect – with (A3 and Verti!a,=(irect2uery

    Some %onsiderations

    • %ube write4bac8 needed@• !arent=%hild needed@• == +iscal %alendars• Excessive Many to Many• Extreme data volumes• )arge M( investment@• )arge 'AM footprint a

    negative@• +inancial models

    ;budgeting=forecasting<

    • 'eal4time ;(irect 2uery<• %ounting whatCs not present• Excel4based Modeling

    Attractive@• .on4relational data sources@

    • In4memory performancebene"t• )ower learning curve

    desirable@• Simpler models ;Sales& *)!

    transaction analysis& etc><

    "avors Multi#dim$MD! "avors %a&ular$DA!

  • 8/8/2019 Tabular Mode and BISM Intro

    18/22

    T+ER T+,0+T 1

    Multidimensional isnCt dead (A3 doesnCt address some common modeling

    re,uirements Vertipa, has more limited storage ;models must "t in'AM<

    Many simple data modeling tas8s are easier in (A3D manycomplex ones are easier in M(3

    As (A3=BISM evolves it will close the gap& but not for acouple years

    At 'M !ower View is a abular4only technology  his will probably force a decision to tabular in some

    scenarios Business Analytics is complex no matter what expression

    language is used (A3 isnCt a silver bullet& but it probably is easier to learn to

    implement basic=intermediate calculations than M(3 forthose new to *)A!

    Should I port my Multidimensional cube to abular during

    migration@ If calculations arenCt com lex and all necessar features

  • 8/8/2019 Tabular Mode and BISM Intro

    19/22

    T+ER T+,0+T 1

    !rocess for Multidimensional to abular migration Evaluate features in the gap

    Many4to4many ;can be done in calculations however<

    !arent=%hild %ube writebac8 %alculated members Etc>

    ow diFcult to rewrite calculations in (A3@ Is the data too large for abular mode@ ;terabytesG< Hill the server have enough 'AM@ Existing application impact@ (oes abular=(A3 solve unmet needs@

    Multi4select issues in calculations

    %ounting whatCs not there needs !erformance issues ;ad4hoc w=o aggregation issues<

  • 8/8/2019 Tabular Mode and BISM Intro

    20/22

    Demo

  • 8/8/2019 Tabular Mode and BISM Intro

    21/22

    REVIE#

    • BISM is designed to ma8e 1SE' experience

    smoother

    • %omplexity still exists in data modeling

    • Multi4dimensional is not gone

    • (A3 is still complex

  • 8/8/2019 Tabular Mode and BISM Intro

    22/22

    T+AN2 Y,3

    3ue$tion$ !ail !e - &enne$$#Cblue-ranite.co!