software platform ecosystems - universitetet i oslo · • example: creditcard, vipps, cool...

SoftwarePlatformEcosystems

INF57502017

Unless noted,alltables,citations,andfigures are taken fromorare facsimiles from:Tiwana,Amrit.Platformecosystems:aligningarchitecture,governance,andstrategy.Newnes,2013.

Fromchapters 1,2and5

Contentsandlearning outcome of the lecture

• What platforms are,andtheir core components• Difference between software platforms andother typesof platforms• Drivers towards software platforms• Some important concepts• Some important principles• Important aspects of platform architecture• Platformlifecycles• Howdoes allthis relate toyour group assignments andthe DHIS2?

Why software platform ecosystems?

• Softwareplatform ecosystem «logics»increasingly plays amoredominantrole incompetition inadiversesets of markets• Competitionmigratingtorivalplatforms• potentmixofspecializedexpertisewiththediscipliningpowerofplatformmarketscanfosterinnovationatapacethatcantrumpeventhemightiestproductandservicebusiness,e.g.BlackberryvsAppleandGoogle;Cameraproducesvsmobilephones.

• Why inthe open source development course?

Maincomponents of asoftware platformiOS Android Nokia

Evolutionof platform ecosystems

• Architecture:Structure

• Governance:Process andrules

• Evolution:«…the interplay betweenitsirreversiblearchitectureandhowitisgoverned.”

FocusinTiwana (2013):software platforms:

• Platformswhere third partycomplementors add toplatform capabilities andfunctionality• Possibilities forhundreds orthousands of actors toadd functionality tothe sameecosystem

• Upstreamvaluechaintheplatformitself.Downstreamappdevelopers.Enduserscanuniquelymix-and-matchdownstreamcomplements– makingtheinnovationandadoptioninthedownstreamcentralforsuccessoffailure• Trueplatformsmustbeatleasttwo-sidedandspanatleasttwodistinctgroupsappdevelopersandend-usersthatinteractthroughtheplatform.• Mostsuccessfulplatformsbeganasstandaloneproductsorservices:iOS,Windows,Facebook,Amazon,eBay,Google,Firefox,Salesforce,andDropbox

• What does that imply andmean?

Driverstowards platformization

Moreabout drivers

• Deepening specialization:• Softwarecode grow larger andbecome morecomplex (morefunctionality)->morespecializationneeded forfurther growth.

• ->Morefocus needed forcompanies• ->Need forintegration of distributed knowledge fromothers• ->Moreeffort tocompete against successful platform owners

• Packetization:• Digitalization of anactivity,process,product orservice

• ->Enables transportation of information through the Internet – high speed,low cost – Removeslocationconstraints towork ->new possible businessmodels

• ->Deepening specialization• Example:globalradiology serviceinIndia(e.g.https://www.outsource2india.com/services/radiology.asp)

Moreabout drivers

• Softwareembedding:• Making software of businessprocesses andactivities

• Example:credit card,Vipps,coolphoto filters

• ->fromproducts toservices– clients toweb-based services,software based maps incars• ->physical – digitalboundary -• ->convergence across industries – gaming consoles andcameras into phones,Amazonkindle

• Internet of things:• Cheap sensorsonlineandnetworked

• Example:Sensorstomonitorpatients athome,door sensorstellingif you forget tolock your door

• ->Fromstock of datatostreams of data• ->Communication of contextual data• Examples:OneTeslacar tellingabout humpinthe road – allother cars get the information andadjust car configuration totake lessimpact when drivingthrough the sameplace.

• Optimalization of resources inahospital,dynamic prize regulations

Moreabout drivers

• Ubiquity:• Precence of Internet «everywhere»– lower prices – fasternetwork

• ->locationindependence of tasks andservices• ->networks of firms• ->crowdsourcing

• Example:Googlemaps traffic information

• Thecombination of the drivers• Pushing innovation ecosystems towards growing number of industries,like:

• mortgage,finance,drug development,software,automotive,healthcare,banking,foodservices,andenergy

Platformconcepts

• Platformlifecycle:

Lifecycle

Platformconcepts

• Platformproperties:

Properties

Platformconcepts

• Platformdynamics:

Platformguiding principles

• Platformstartup principles:

Platformguiding principles

• Platformdesignprinciples:

Platformguiding principles

• Platformevolution principles

Some key points• Thelifecycleofatechnologysolutionhasthree-dimensions:

• pre- orpost-dominantdesignstage(frommanytoone)• maturity trajectory (the S-curve)• proportionofthetotalprospectiveuserbaseadoption

• Multisidedness offers:same-sideandcross-sidenetworkeffects,lock-ins(coerciveandvalue-driven),prospectsofswallowingorbeswallowed

• Architecturesprovideblueprintformasscoordination.Conventionalcoordinationandcontrolmechanismscostlyandimplausibleinlargeecosystems

• Governancecanamplifyordiminishtheadvantagesofgoodarchitecture.Governanceandarchitecturemustbeco-designedandcoevolved

• Evolutionarypaceofaplatformisrelativetoitsrivals(theRedQueeneffect).• Emergentinnovationcanonlybefacilitated,notplannedbyaplatformowner.

• Spontaneouslyarisefromtheselfishpursuitofself-interestbyindividualecosystemparticipants.• Chicken-or-eggproblemandthepenguinproblemtogetofftheground- unattractiveforeithersidetojoinunlessthereisacriticalmassontheotherside.Uncertaintyaboutwhetherotherswilljointheplatformecosystemcanstallinitialadoption,creatingthepenguinproblem

• Balanceautonomywithintegration(theseesawproblem)separablebutre-integratable (HumptyDumpty).Organizedtomirrorthearchitectureandthe“microarchitecture”(themirroringprinciple).

Platformarchitectures

• Thearchitecture enable (ornot)participation among potential andactual third partyinnovators• Thirdpartyinnovators mustbeable andmotivated toparticipate• Abilitythrough architecture• Motivation through governance

• Mainarchitecture parts(andtheir interconnectedness)• Platformcore• Platforminterfaces• «Apps»

Managing complexity

• What iscomplexity?• Afunction of the number of parts,typesof parts,andnumber andtypesofconnections between the parts.• Structural (difficult todescribe)• Behaviorally (difficult tocontrol andpredict)

• Toohigh complexity will leadtoatleast• Incomprehensibility• Gridlock

• ->lossof predictable outputfrominput– ripple effects• ->co-innovation risk(80%x80%x80%=51%)– need toreduce dependencies atthe rightplace

Managing complexity

• Inaplatform ecosystem with numerous actors,complexity mustbecontrolled somehow toreduce riskof gridlocks,unpredictable rippleeffects andco-innovation problems

• ->Architecture• Balancing between control andautonomy• Keeping transaction costs andcoordination cost aslow aspossible

Architecturesolutions toorchestrate

• Partitioning (modularization)– core <->apps - degrees• Creating «autonomous»subsystems

• Tocognitively manageable parts• Blackboxing

• Visibleinformation:what they doandhow tointeract with them• Hidden information:how they work

• Systemsintegration• Developmentactivities coordination between platform owner andapp developers

• Managing dependencies• Minimizing need forcoordination

• Apps mustbeintegrated tothe platform toenable value toend-users• Platform– app integration – uneven development,platform changes – ongoing effort• App – app integration

Architecturesolutions

• Relatively stablecore• Platformarchitecture• Visiblepart:Shared sets of assetsthrough defined interfaces• Hidden:inner functions of theplatform core tomakeinterfaceswork andbehave asthey do

• Dynamicsandvariability inapps->innovation• Microarchitecture

App architecture (microarchitecture)

Possible partitioning of layers

Many possibilities forpartitioning the app

Many possibilities forpartitioning the app

App architecture choices haveconsequences

• Hard,or,impossibletomaximize allpositiveconsequences;always trade-offs between partitioning inside the platform andacross the Internet• Early architecture choices are hardtochange later

• ->creating path dependencies inarchitectures• Some characteristics showupimmediately:

• speed,security,reliability,scalability,testability,andusability• Someatlaterstages:

• maintainability,extensibility,evolvability,andthecapacitytomutateandenvelopadjacentappmarketsegments

• Developersneedknowledgeaboutwhichtypesofapparchitecturesgiveswhichtypesoftrade-offsandadvantages• ->design,notexperiencetoolate

Platformarchitecture

• Inpractice,irreversible• ->havetostick with early choices andtheir consequences

• Desirableproperties• Simple;defined interfaces,functionlity etc.• Resilient;notbreaking the ecosystem upon app failure• Maintainable;minimizing consequences of local changes• Evolvable;balancing between stability/control of interfaces andautonomy ofinnovation

• But also here,trade-offs.

Moreon modularization andamount ofmodularity• Monolithic versusmodular• Noteither or– rather acontinuum between the two extremes,wheremostliesinbetween

• Some important aspects:• Division of work among several organizations/actors

• Emergent properties• Dependencies among modules isrestricted todefined interfaces• Need tobecompliant only tointerface specifications• Possible performance sacrifications

Balancing needs andimplications

Balancing needs andimplications

What isin,what isout?

• High-reusability functionality• Generic functionality• Stablefunctionality• Interfacesintegralpartsof the platform

• Highuncertainty functionality – out

• But also in:• Forattractiveness• Expectation fromend-users

Theinterfaces

• Standardization• Stability• Versatility• flexibility instandards• highly dependentfunctionality stays inthe platform

• Openness• who can participate

DHIS2asaplatform ecosystem?

• Howdoyour developed apps relate toplatform architectures asdescribed?• Dothe architectural choices inyour app (together with the DHIS2)imply anything forfurther development andevolvement of your app,andinrelation tothe DHIS2core• Dependencies – loose coupling• Modularization• Usage of APIs• Placement of functionality andlayers

Platformvs application vs Informationinfrastructure