Upload
others
View
15
Download
0
Embed Size (px)
Citation preview
Embedded Database Performance Report
Actian Zen more than 6x faster than Couchbase
MCG Global Services Benchmark Results
September 2018
Key insights
This benchmark did a head-to-head comparison of Actian Zen Core and Couchbase Lite, both running on a Nokia 2, Android 7 handset (ARM-based) with 1GB DRAM
Actian Zen Core outperformed Couchbase Lite for Indexed and Non-Indexed data management by:
– more than 5x on inserts, up to 2x on deletes, and over 2x on updates
Actian Zen Core plus Actian Zen Edge was also faster as a client-server combination than Couchbase Lite plus Couchbase Sync Gateway plus Couchbase server, replicating IoT Device or Smartphone to gateway data sharing —the area where it tends to really matter in embedded Edge applications.
Checkout the Actian Zen performance advantage today!
Visit https://www.actian.com/zen
EmbeddedNoSQLDatabasePerformanceBenchmark
ProductProfileandEvaluation:ActianZenandCouchbaseByWilliamMcKnightandJakeDolezalMcKnightConsultingGroupSeptember2018Sponsoredby
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page2
ExecutiveOverview
Embeddeddatabasesarebuiltintosoftware,transparenttotheapplication’senduserandrequirelittleornoongoingmaintenance.Embeddeddatabasesaregrowinginubiquitywiththeriseofmobileapplicationsandinternetofthings(IoT)givinginnumerabledevicesrobustcapabilitiesviatheirownlocaldatabasemanagementsystem(DBMS).Developerscancreatesophisticatedapplicationsrightontheremotedevice.Fortheseuses,theembeddedarchitectureispreferredoverclient-serverapproacheswhichrelyondatabaseserversaccessedbyclientapplicationsviainterfaces.Today,tofullyharnessdatatogainacompetitiveadvantage,embeddeddatabasesneedahighlevelofperformancetoprovidereal-timeprocessingatscale.Alltheseweb,mobile,andIoTapplicationshavegeneratedanewsetoftechnologyrequirements.Embeddeddatabasearchitectureneedstobefarmoreagilethaneverbefore,andrequiresanapproachtoreal-timedatamanagementthatcanaccommodateunprecedentedlevelsofscale,speed,anddataflexibility.Sometimesrelationaldatabasesareunabletomeetthesenewrequirements,anddevelopersarethereforeturningtoNoSQLdatabasetechnology.NoSQLusecasesaboundwheretheneedforflexibleschemaorschema-lessdatawouldtripupconventional,relationaldatabases.Toquantifyembeddeddatabaseperformance,weconductedthisbenchmarkstudy,whichfocusesontheperformanceofmobileapplication-ready,NoSQL,embeddeddatabasesolutionsActianZenandCouchbase.Theintentofthebenchmark’sdesignwastorepresentasetofbasicdatabasetransactionsthatanorganizationdevelopingedgeapplicationsmightencounter.ThetestmethodologywasbasedonandlargelyfollowedtheBenchmarkofEmbeddedDatabaseson.NETconductedin2017byChristopheDiericxofrelationaldatabasetechnologies;however,ourownbenchmarkharnesswasdevelopedandadaptedtoaNoSQLusecase.WeconductedthebenchmarkonZenandCouchbaseLiteinstalledonthesameAndroiddevice,andZenandCouchbaseServerinstalledonaserver.Inourexperience,performanceisaveryimportantaspectofanembeddeddatabaseselection,butitisonlyoneaspectandmanyfactorsshouldbeconsidered.Overall,thebenchmarkresultswereinsightfulinrevealingthequeryexecutionperformanceofActianZenandCouchbaserevealingsomeofthedifferentiatorsinthetwoproducts.ActianZenEdgewasfasteracrosstheboardincludingtheareawhereittendstoreallymatterinembeddeddatabases—writespeed.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page3
ThisistheessentialperformancemetricforIoTdata.Withoutsynchronizationofdatatotheserver,ActianZenEdgeoutperformedCouchbaseby5xoninserts,7xonqueriesof10,000documentsonanindexedkey,6xonqueriesof5,000documentsonanon-indexedkey,2xondeletesof10,000documentsonanindexedkeyandondeletesof5,000documentsonanon-indexedkey,2xonupdatesontheindexedkeyand3xonnon-indexedkeyupdates.Withsynchronizationofdatatotheserver,ActianZenEdgeoutperformedCouchbaseby6xoninserts,2xondeletesof10,000documentsonanindexedkeyandondeletesof5,000documentsonanon-indexedkey,2xonupdatesontheindexedkeyand3xonnon-indexedkeyupdates.ActianZenisamatureplatformforembeddeddatabaseapplicationswithover30yearsofengineeringanddevelopmentbehindit.Featuresthatcontributedtoitsextremelyfastperformanceinclude,butarenotlimitedto,theBtrieveAPIandTurboWriteAccelerator.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page4
EmbeddedNoSQLDatabaseSelection
Organizationsthatutilizeapplication-ladensmartdevicesrelyonembeddeddatabaseplatformstoprocessedgedataathighspeedandbringitinwithconsistencytoharmonizeanecosystemofactivity.Volumesfordatathatcanbeutilizedattheedgeisrapidlyexpanding—placingsignificantperformancedemandsonembeddedarchitectures.Thus,akeydifferentiatoristhedepthbywhichadatabasemaintainsperformancetoscalewithsimplequeriesrepresentativeofrealworldusecasesofembeddeddatabases—SQLandNoSQLalike. Whileperformanceisveryimportant,itisnottheonlyconsideration.DeveloperschoosingNoSQLmustconsiderdataaccess,scalability,andavailability.BothCouchbaseandActianZenweredesignedto“setitandforgetit,”withlittle-to-noongoingdatabaseadministration.However,ActianZenwasengineeredpurposefullytoparedownanenterpriseNoSQLplatformtobeembeddedwithinOEMenvironments.Therefore,ActianZenhasfeaturesthatCouchbasedoesnot—includingauto-reconnectnetworking,automateddefragmentation,multi-usersupport,andconcurrentwritecapabilities.BothplatformsofferNoSQLsupport.ZenisnativelyNoSQLandisflexibleenoughtobedocument-basedorakey-valuestore.CouchbaseisJSONdocument-oriented.Additionally,ZenexclusivelyoffersthehighperformanceBtrieve2API(whichistestedinthisbenchmark.)TheBtrieve2APIsupportsNoSQLandnativedevelopmentsupportforJavaandC/C++baseddevicesandSWIGforPython,Perl,andPHP—inadditiontoitsSQLsupport.CouchbaseisexclusivelyNoSQL,andonlyofferssoftwaredevelopmentkitsformobiledevices,suchasiOS,Android,and.NET.Whilethesubjectofthisbenchmarkisembeddedapplications,ActianZenedgeispartoftheoverallZenfamilyofZenCore,ZenEnterprise,andZenReportingEngine.Whencombined,thissuiteofproductsenablesnotonlyembeddedapplications,butclient-server(withzeroETL)andclouddeploymentsaswell.Inaclient-serverconfiguration,bothActianZenandCouchbasecomewiththecapabilitytoautomaticallysynchronizeinrealtimebetweenaremotedeviceandaserver—withoutETL.Thiscapabilityiscriticalfortoday’sneedsanduses,becausethepotentialnumberofmobiledevicescouldeasilynumberinthethousands,andallthatinformationmayneedtofunnelintoacoredatabaseonaserver.WithCouchbase,youmustutilizeaSyncGatewayutilitytoserveastheproxybetweenCouchbaseLiteontheremoteclientandCouchbaseServer.ActianhasrealtimesynchronizationcapabilityofActianZenEdgetoCoreviatheBtrieveAPIwithoutanintermediary,whichcanallowyoutoachievescalewithsimplicity.Platformmaturityisalsoaconsideration.Couchbasewasinitiallyreleasedin2010.ActianZenwasinitiallydesignedasBtrieve(andlaterPSQL)andhasbeeninproductionwithmanymulti-nationalorganizationswithover30yearsofengineeringandenhancement.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page5
ThisreportsfocusesontheperformanceoftwoembeddedNoSQLdatabaseoptions.Itisimportanttogetintotherightembeddeddatabaseearlyinthedevelopmentcyclewhenthestakesarelesscritical.Oneisaspecialtyapproachwithenterprisesoftwareoptimizedfortheembeddedarchitecture,andthelatteranopensource,multi-purposedatabaseplatform.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page6
BenchmarkSetup
Thebenchmarkwasexecutedusingthefollowingsetup,environment,standards,andconfigurations.
DataPreparation
Anaimofthebenchmarkistosimulateatypicalreal-worldscenarioandusecaseforNoSQLembeddeddatabases.Inourbenchmark,wechoseasimple“schema”foranapplicationthatstorespeoples’contactinformationintheembeddeddatabase.Themodelconsistsofmultipledocumentsthatlooksimilartothefollowing:
{ "contact": { "id": 1, "lastname": "Rogers", "firstname": "Fred", "address": "381 Willinghelm Dr", "city": "Pittsburgh", "state": "PA", "zip": "15106", "country": "USA", "phone": "412-875-0921" } }
ThedatausedinthebenchmarkwasgeneratedrandomlyinrealtimebytheAndroidapplicationduringthebenchmarkexecution.Thekeyscity,state,andzipwereusedasselectioncriteriaintheSelect,Update,andDeletetests(describedbelow).Therefore,aparticularvaluewasrandomlyseededintothiskeyduringdatagenerationtoensuretherewouldbeenoughinstancesofthatvaluetoachievethedocumentcountsrequiredduringtheSelect,Update,andDeletetests.
Configuration
OurbenchmarkincludedtwodifferentembeddedRDBMS—ActianZenandCouchbase—installedonthesameAndroiddevice.Wealsotestedaconfigurationwithreal-timesynchronization.TheserverhadboththelatestversionsofActianZenCoreandCouchbaseServerinstalledonthesamemachine.Also,Couchbase’sSyncGatewaywasinstalledtoserveastheintermediarybetweenCouchbaseLiteandServer.Allcomponentsweredeployedonalocalareanetwork.
NoSQLDBMS
EmbeddedRDBMS ActianZen CouchbaseVersion 13.10.030 Server5.1.1Community
Lite2.1CommunitySyncGateway2.1Community
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page7
AndroidDevice
Hardware Nokia2TA-1035DSProcessor 1.3GHz64-bitquad-coreARMCortexA7RAM 1GB(8GBStorage)OS Android7.1.1Nougat
Server
Hardware LenovoThinkPadX1CarbonG620BS006UUSx64-basedPCProcessor [email protected] 8GBOS MicrosoftWindows10Enterprise10.0.16299
TestUseCases
Asaforementioned,thetestmethodologywasbasedonandlargelyfollowedtheBenchmarkofEmbeddedDatabaseson.NETconductedin2017byChristopheDiericx.ThetestinvolvessimpleusescasesofthemostbasicdatabaseCRUDoperations:selecting,updating,anddeletingrowsbasedonindexedandnon-indexedcolumns.Weconsideredotherbenchmarkframeworks,suchastheTransactionPerformanceCouncil(TPC).While,theirtestusecaseshavebeenappliedtoNoSQLtechnologiesinthepast,theyarenotveryapplicabletotypicalmobiledeviceapplications.MostIoTdevicesandmobileapplicationswillnotrequirethesophisticatedoperationsdemonstratedbythosebenchmarkframeworks.Therefore,weoptedforteststhatwoulddemonstraterawperformancethatcouldbefoundinmostembeddeddatabaseimplementations.BothplatformssupportarobustsetofNoSQLcapabilities.ForbothActianZenandCouchbaseLiteweusedthenativeAPIstoexecutethedatabasetransactionsinordertotestitsfunctionalityandperformance,ratherthanSQL(Zen)orN1QL(CouchbaseSQLforJSON).
UseCase1:OpenandCloseConnectionsinRapidSuccession
NOTE:WedidnotdothisrunfortheAndroiddevicesinceitisstandardpracticeformobiledeveloperstoopenadatabaseconnectionandleaveitopenwhiletheappisrunning.Also,wedidnotusethistestforthesynchronizationbenchmark,havingnoapplicableuseforthatworkload.
UseCase2:InsertPerformance
Mobiledevicesandtheirapplicationswillundoubtedlyneedexcellentinsertperformance.Thismaythesinglemostimportantmetricformanyusecases.Forexample,consideranIoTdeviceisasensortakingreadingsatregularintervals.Inthecaseofreal-timeorrapidsensorreadings,insertperformanceiscritical.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page8
Test2 Insert25,000documents
NOTE:Atthebeginningofthetest,thedatabasecontainsanemptydatabase.TheInserttestprovidedthetestdatafortheremainingbenchmarks.
UseCase3:SelectPerformance
Certainly,wemustconsiderbothplatforms’abilitytoretrievedata.Ourtestcasesinvolveselectingbulkdocuments,ratherthansingledocumentsviaauniqueidentifier.Thefirstvariationofthetestfiltersonanindexedkey(state).Thesecondtestselectsfewerdocuments,butfiltersonakeythatdoesnothaveanindex(zip).
Test3a Select10,000documentsonanindexedkeyTest3b Select5,000documentsonanon-indexedkey
NOTE:Wedidnotusethistestforthesynchronizationbenchmark,sinceselectingdocumentsbythemselveswouldnotconstitutethecompleteworkload.
UseCase4:UpdatePerformance
WealsotestedtheperformanceofbulkdocumentupdatesusingthesameselectiontestcriteriaasTest3.Ourtestcasesinvolveselectingbulkdocumentsandupdatingasinglekey-value.Thefirstvariationofthetestfiltersonanindexedkey(state)andupdateszip.Thesecondtestselectsfewerdocuments,butfiltersonakeythatdoesnothaveanindex(zip)andupdatesstate.
Test4a Update10,000documentsonanindexedkeyTest4b Update5,000documentsonanon-indexedkey
UseCase5:DeletePerformance
Wealsotestedtheperformanceofbulkdocumentdeletes—again,usingthesameselectiontestcriteriaasTest3.Ourtestcasesinvolveselectingbulkdocumentsanddeletingthem.Thefirstvariationofthetestfiltersonanindexedkey(state)anddeletesthosedocuments.Thesecondtestselectsfewerdocuments,butfiltersonakeythatdoesnothaveanindex(zip)anddeletesthedocuments.
Test5a Delete10,000documentsonanindexedkeyTest5b Delete5,000documentsonanon-indexedkey
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page9
BenchmarkResults
ThefollowingfiguresdisplaytheaveragetimeelapsedforeachdatabasetransactionforbothActianZenandCouchbase.Eachtestwasexecuted5timesandthemedianvaluewasused.
LocalOnly(NoSynchronization)
Test2:Insert25,000documents
Belowaretheaveragetimes(inmicroseconds)ittooktoinsertacompletedocumentofrandomly-generateddataintotheContactsdatabaseonActianZenandCouchbase.
Thistestrevealedthefirstmajorperformancedifferentiator.ActianZen’saveragetimetoinsertasingledocument(takingtheaverageofall25,000inserts)was4.6timesfasterthanCouchbaseinserts.
Test3a:Select10,000documentsonanindexedkey
Belowaretheaveragetimesperdocument(inmicroseconds)ittooktobulkselectrecordsfromtheContactsdatabaseapplyingafilteronanindexedkeyforbothActianZenandCouchbase.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page10
Bothplatformsrespondedveryquickly.Couchbase’sfetchrateperdocument(takingtheaverageofall10,000documents)was7timesthatofActianZen’s.
Test3b:Select5,000documentsonanon-indexedkey
Belowaretheaveragetimesperdocument(inmicroseconds)ittooktobulkselectrecordsfromtheContactsdatabaseapplyingafilteronanon-indexedkeyforbothActianZenandCouchbase.
Againbothplatformsrespondedveryquickly.Couchbase’sfetchrateperdocument(takingtheaverageofall10,000documents)was5.8timesthatofActianZen’s.
Test4a:Update10,000documentsonanindexedkey
Belowaretheaveragetimes(inmicroseconds)ittooktoupdateasinglekeyintheContactsdatabaseapplyingafilteronanindexedkeyforbothActianZenandCouchbase.
Thisonewasaclosetest.Theaveragetimetoupdateasinglekey(takingtheaverageofall10,000updates)wasonly1.9timesfasterthanCouchbaseupdates.
Test4b:Update5,000documentsonanon-indexedkey
Belowaretheaveragetimes(inmicroseconds)ittooktoupdateasinglekeyintheContactsdatabaseapplyingafilteronanon-indexedkeyforbothActianZenandCouchbase.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page11
Thistesthadsimilarresultsastest4a.ActianZen’saveragetimetoupdateasinglekey(takingtheaverageofall5,000updates)was2.9timesfasterthanCouchbaseupdatesusingthesamefilter.
Test5a:Delete10,000documentsonanindexedkey
Belowaretheaveragetimes(inmicroseconds)ittooktodeleteadocumentintheContactsdatabaseapplyingafilteronanindexedkeyforbothActianZenandCouchbase.
Bothwereveryfast.Theaveragetimetodeletearow(takingtheaverageofall10,000deletes)forActianZenwassimply2timesfasterthanCouchbasedeletes.
Test5b:Delete5,000documentsonanon-indexedkey
Belowaretheaveragetimes(inmicroseconds)ittooktodeleteadocumentintheContactsdatabaseapplyingafilteronanon-indexedkeyforbothActianZenandCouchbase.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page12
Deletingdocumentsonanon-indexedkeyproducedresultsconsistentwithbefore.ActianZen’saveragetimetodeleteadocument(takingtheaverageofall5,000deletes)was1.88timesfasterthanCouchbaseupdatesusingthesamefilter.
Synchronization
Test2:Insert25,000documentsandsync
Belowaretheaveragetimes(inmicroseconds)ittooktoinsertacompletedocumentofrandomly-generateddataintotheContactsdatabaseontheActianZenandCouchbasedatabases.
Thistestrevealedthefirstmajorperformancedifferentiator.ActianZen’saveragetimetoinsertasingledocument(takingtheaverageofall25,000inserts)was5.9timesfasterthanCouchbaseinserts.
Test4a:Update10,000documentsonanindexedkeyandsync
Belowaretheaveragetimes(inmicroseconds)ittooktoupdateasinglekeyintheContactsdatabaseapplyingafilteronanindexedkeyforbothActianZenandCouchbase.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page13
ActianZen’saveragetimetoupdateasinglekey(takingtheaverageofall10,000updates)was2timesfasterthanCouchbaseupdates.
Test4b:Update5,000documentsonanon-indexedkeyandsync
Belowaretheaveragetimes(inmicroseconds)ittooktoupdateasinglekeyintheContactsdatabaseapplyingafilteronanon-indexedkeyforbothActianZenandCouchbase.
Thistesthadsimilarresultsastest4a.ActianZen’saveragetimetoupdateasinglekey(takingtheaverageofall5,000updates)was3.3timesfasterthanCouchbaseupdatesusingthesamefilter.
Test5a:Delete10,000documentsonanindexedkeyandsync
Belowaretheaveragetimes(inmicroseconds)ittooktodeleteadocumentintheContactsdatabaseapplyingafilteronanindexedkeyforbothActianZenandCouchbase.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page14
Bothwereveryfast.Itsaveragetimetodeletearow(takingtheaverageofall10,000deletes)wasalittlelessthanhalfthetimeCouchbasetook.
Test5b:Delete5,000documentsonanon-indexedkeyandsync
Belowaretheaveragetimes(inmicroseconds)ittooktodeleteadocumentintheContactsdatabaseapplyingafilteronanon-indexedkeyforbothActianZenandCouchbase.
Deletingdocumentsonanon-indexedkeyproducedresultsconsistentwithbefore.Couchbase’saveragetimetodeleteadocument(takingtheaverageofall5,000deletes)wasmorethandoublethatofActianZenupdatesusingthesamefilter.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page15
Conclusion
Withoutsynchronizationofdatatotheserver,ActianZenEdgeoutperformedCouchbaseby5xoninserts,7xonqueriesof10,000documentsonanindexedkey,6xonqueriesof5,000documentsonanon-indexedkey,2xondeletesof10,000documentsonanindexedkey,2xondeletesof5,000documentsonanon-indexedkey,2xonupdatesontheindexedkeyand3xonnon-indexedkeyupdates.Withsynchronizationofdatatotheserver,ActianZenEdgeoutperformedCouchbaseby6xoninserts,2xondeletesof10,000documentsonanindexedkey,2xondeletesof5,000documentsonanon-indexedkey,2xonupdatesontheindexedkeyand3xonnon-indexedkeyupdates.ActianZenoutperformedCouchbaseinallofthefundamentaldatabaseoperations.ThesetestedoperationsunderlienearlyalloperationsthatoccuronanembeddeddatabaseforanIoTormobileimplementation,soitisunlikelymorecomplexoperationswouldhaveadifferentresult.ActianZenisamatureplatformforembeddeddatabaseapplicationswithover30yearsofengineeringanddevelopmentbehindit.TheBtrieve2APIhadclearperformanceadvantageswithouttheoverheadofCouchbase.Also,Zen’sTurboWriteAcceleratorcouldalsoshedlightintoitsperformanceadvantages.Sinceitcostsmuchlesstocontinuewritingthantostopandrestart,contiguouswritesaresignificantlyfasterthannon-contiguouswrites.TheTurboWriteAccelerator(TWA)pre-allocatesopenslotswithinthephysicalfilesothatmultiplepagescanbewrittenasasinglecoalescedpage—improvingI/Operformanceandreducingtheoverheadofinteractionwiththeoperatingsystem.Theresultoftheapplicationofthemethodologytothearchitecture,bothexplainedhereinandreplicable,showamarked,andsometimesastonishing,performanceadvantagetoActianZen.Thisisespeciallytrueintheimportantwriteoperationsinsert,updateanddelete.Overall,ActianZenisanexcellentchoiceforIoTormobilecompaniesneedinghighperformanceandascalableembeddeddatabase.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page16
AboutMcKnightConsultingGroup
WilliamMcKnightisPresidentofMcKnightConsultingGroup(MCG)(http://www.mcknightcg.com).Heisaninternationallyrecognizedauthorityininformationmanagement.HisconsultingworkhasincludedmanyoftheGlobal2000andnumerousmidmarketcompanies.Histeamshavewonseveralbestpracticecompetitionsfortheirimplementationsandmanyofhisclientshavegonepublicwiththeirsuccessstories.Hisstrategiesformtheinformationmanagementplanforleadingcompaniesinvariousindustries.JakeDolezalhastwodecadesofexperienceintheInformationManagementfieldwithexpertiseinbusinessintelligence,analytics,datawarehousing,statistics,datamodelingandintegration,datavisualization,masterdatamanagement,anddataquality.Jakehasexperienceacrossabroadarrayofindustries,including:healthcare,education,government,manufacturing,engineering,hospitality,andgaming.HehasadoctorateininformationmanagementfromSyracuseUniversity.MCGservicesspanstrategy,implementation,andtrainingforturninginformationintotheassetitneedstobeforyourorganization.Westrategize,designanddeployinthedisciplinesofMasterDataManagement,BigDataStrategy,DataWarehousing,AnalyticDatabasesandBusinessIntelligence.
McKnightConsultingGroup EmbeddedDatabaseBenchmark
©McKnightConsultingGroup2018 http://www.mcknightcg.com Page17
AboutActian
Actian,thehybriddatamanagement,analyticsandintegrationcompany,deliversdataasacompetitiveadvantagetothousandsofcustomersworldwide.ThroughthedeploymentofinnovativehybriddatatechnologiesandsolutionsActianensuresthatbusinesscriticalsystemscantransactandintegrateattheirverybest–onpremise,inthecloudorboth.FormoreinformationaboutActianVectorandtheentireActianportfolioofhybriddatamanagement,analyticsandintegrationsolutionson-premiseorinthecloud.FindoutmoreaboutActianVectorforsingleserversandforHadoopclusters,orgetlinkstodownloadsforon-premisedeploymentorcloudinstances.