1. NUMBER FIVE
OLAP(Online Analytical Processing) isthe technologybehindmanyBusinessIntelligence(BI)
applications.OLAPisapowerful technologyfordatadiscovery,includingcapabilitiesforlimitlessreport
viewing,complexanalyticalcalculations, andpredictive “whatif”scenario(budget,forecast) planning.
OLAPis an acronymfor Online Analytical Processing.OLAPperformsmultidimensional analysisof
businessdataandprovidesthe capabilityforcomplex calculations,trendanalysis,andsophisticated
data modeling.Itisthe foundationformanykindsof businessapplicationsforBusinessPerformance
Management,Planning,Budgeting,Forecasting,Financial Reporting,Analysis,SimulationModels,
KnowledgeDiscovery,andDataWarehouse Reporting.OLAPenablesend-userstoperformadhoc
analysisof datain multiple dimensions,therebyprovidingthe insightandunderstandingtheyneedfor
betterdecisionmaking.
OLAPsystemsvaryquite a lot,and theyhave generallybeendistinguishedbyalettertaggedontothe
frontof the acronym “OLAP,”forOn-Line Analytical Processing.MOLAPandROLAPhave classicallybeen
the most establishedtypes,andthe otherdistinctionsrepresentlittlemore thanthe marketing
programson the part of the vendorstodistinguishthemselves,forexample,SOLAPandDOLAP.Here,
we aim to give youanideaof whatthese distinctionshave meant.
Hybrid Transaction / Analytical Processing(HTAP)
Gartner coinedthe termHTAPin a paperin the beginningof 2014 to describe new in-memorydata
systemsthatdo bothonline transactionprocessing(OLTP) andonlineanalytical processing(OLAP).
HTAP reliesonnewerandmuchmore powerful,oftendistributed,processing:sometimesitinvolvesa
newhardware “appliance”,anditalmostalwaysrequiresanew software platform.Beyondthis,the key
pointseemstobe that all the technologyissitedinthe relational database. Andso,there’snomore
data replication,andnewtransactional informationbecomespartof ananalytical model inasfast a time
as is technologicallypossible.
Multidimensional OLAP(MOLAP) – Cube based.
MOLAP productsenable end-userstomodel dataina multidimensional environment,ratherthan
providingamultidimensional viewof relational data,asROLAPproductsdo (see nexttab).
The structure of a multidimensional model isnotaseriesof tables(asexistsinarelational database)but
whatis generallyreferredtoasa cube.Cubesmodeledinamultidimensional databaseextendthe
conceptassociatedwithspreadsheets:justasa cell ina spreadsheetrepresentsthe intersectionof two
dimensions(salesof productbyregion),acell ina cube representsthe intersectionof aninfinite number
of dimensionmembers(e.g.,Products,Customers,Regions,Months…nthdimension). Asina
spreadsheet,acell mightbe calculatedbyformulasinvolvingothercells.
2. Relational OLAP (ROLAP) –Star Schema based
ROLAPproducts(forRelational OLAP) are creditedwithbeingable todirectlyaccessdatastoredin
relational databases.The notionisthattheycan readilyretrievetransactionaldata,althoughthis
becomessuspectwhenverylarge datasetsare inplay,or if more complex calculationsare tobe
delivered,basedonthe transactional data.ROLAPproductsenable organizationstoleveragetheir
existinginvestmentsinRDBMS(relational database managementsystem)software.
ROLAPproductsaccess a relational database byusingSQL(structuredquerylanguage),whichisthe
standardlanguage thatis usedto define andmanipulate datain anRDBMS. Subsequentprocessingmay
occur in the RDBMS or withinamid-tierserver,whichacceptsrequestsfromclients,translatesthem
intoSQL statements,andpassesthemontothe RDBMS.
Structure of an Online Analytical Processing.
Each type of analysishasitsownadvantagesanddisadvantages.However,OLAPServicesprovides
MOLAP and ROLAPanalysisonthe same report,whichoffersmanydistinctbenefits,summarizedbelow:
Analyze reports at the “speedofthought” and manipulate themin real time.Using OLAPServices,you
can get fastresponse timesforreportsthatuse data directlyfromin-memoryIntelligentCubes,instead
of fromthe data warehouse.Youcancreate andanalyze new reportsinreal time throughinteractive
OLAPServicesmanipulations.
Share IntelligentCube data securely.MicroStrategy’scentralizedmetadataandIntelligence Server
architecture allowsIntelligentCube datatobe sharedin a secure fashion.
Schedule IntelligentCube executionandmaintenance.To reduce stressonthe Intelligence Server,you
can schedule whenIntelligentCubesare executed.Thisallowsyoutotake advantage of Intelligence
Serverdowntime toexecute IntelligentCubeswithoutaffectingperformance foryourusercommunity.
3. You can alsoschedule whenIntelligentCubesare re-executedtosynchronize theirdatawithchangesto
the data in yourdata warehouse.
Drill from summary data to transaction-level details.Youcan drill frompredefinedreportstoconduct
advancedanalysisandtake full advantage of the IntelligentCube feature.Drillingisallowedwithinan
IntelligentCube forquick-responseMOLAPanalysis.Drillingcanalsobe enabledoutsideof anIntelligent
Cube for full ROLAPanalysis.
Use MicroStrategy Developer,Office,orWeb.UsingOLAPServices,youcanperformthe same
multidimensional analysiswhetheryouuse MicroStrategyDeveloper,Office,orWeb.
Apply securityrestrictionson users and objects.ReportingwithOLAPServicesandIntelligentCubes
adhere tothe same standardsof data access securityasthe restof your MicroStrategyproject.
Increase user self-service andproductivity.Since accessingIntelligentCubesforOLAPanalysisdoesnot
require runtime processingonthe datawarehouse andcanuse schedulestoreduce ITmanagement,
usershave increasedflexibilitytocreate andmodifytheirownreportstosuittheirunique work
environment.
Disadvantagesof OnlineAnalytical Systeminclude
1. Pre-modelingasa must. Regardingthe businessdata,the traditionalOLAPtoolsdonotallow
for the immediate analysiswithoutpre-modeling.These toolswithoutagoodOLAP engine
cannot convertthe data to a patternin whichbusinesspersonnel canoperate directly.
2. Great dependence onIT.Althoughbusinesspersonnelisthe intendeduserof OLAP,theywill
still have toworkwiththe IT prosbecause the traditional OLAPtoolsrequiresacomplex
modelingprocedureanditsusershave towrite a great numbersof codes/scripts/SQL.
3. Poor computation capability.The data computingreferstoa procedure of processingand
transformingdatathrougha seriesof specificstepstowardaconcrete goal.Data computingisa
basicfeature of OLAP.The traditional OLAPtoolsare of insufficientcomputational capabilities
and fewcomputational methodssuchasdrilling,slicing,rotation,andsimple column
computation.Thisisbecause theirarchitecturesare old,lackingthe innovation,andhardto
strike a balance betweenuser-friendlinessandflexibility.
4. Short of Interactive analysis ability.Data analysisis the mostimportantfeature of OLAP.Not
like the datacomputing,dataanalysisisan interactive procedure requiringthe perfectstep-by-
stepcomputational mechanism.Towardthe obscure goal,usersneedtoobserve the current
data and make the reasonable assumption,andthenverify/falsifythe assumptiontoultimately
achieve the rathercomplicatedbusinessanalysisgoal.Unfortunately,the model of traditional
OLAPtoolsis toooldto provide sofree a style of interactive analysis.
5. Short of Interactive analysis ability.Data analysisisthe mostimportantfeature of OLAP.Not
like the datacomputing,dataanalysisisan interactive procedure requiringthe perfectstep-by-
stepcomputational mechanism.Towardthe obscure goal,usersneedtoobserve the current
data and make the reasonable assumption,andthenverify/falsifythe assumptiontoultimately
4. achieve the rathercomplicatedbusinessanalysisgoal.Unfortunately,the model of traditional
OLAPtoolsis toooldto provide sofree a style of interactive analysis
6. Great potential risk. Traditional OLAPtoolshave ahuge potential riskdue tothe lackingof the
computationandlowinteractive analysisability,andthe implementationreliesonthe
cooperationwithITpros.The procedure andthe cycle are a bitlong.