Accelerating Digital Transformation: 10 Years of Software Center

447 Pages·21.618 MB·English

by  Jan Bosch, Jan Carlson, Helena Holmström Olsson, Kristian Sandahl, Miroslaw Staron

Jan Bosch · Jan Carlson 
Helena Holmström Olsson 
Kristian Sandahl 
Miroslaw Staron Editors
Accelerating
Digital
Transformation
10 Years of Software Center
Accelerating Digital Transformation
(cid:129) (cid:129) (cid:2)
Jan Bosch Jan Carlson Helena Holmstrom
(cid:129) (cid:129)
Olsson Kristian Sandahl Miroslaw Staron
Editors
Accelerating
Digital
Transformation
10 Years of Software Center
Editors
JanBosch JanCarlson
DepartmentofComputerScience DivisionofComputerScienceandNetworks
andEngineering Ma(cid:2)lardalenUniversity
ChalmersUniversityofTechnology Va(cid:2)stera˚s,Sweden
Gothenburg,Sweden
HelenaHolmstro(cid:2)mOlsson KristianSandahl
DepartmentofComputerScience DepartmentofComputerandInformation
Malmo(cid:2)University Science(IDA)
Malmo(cid:2),Sweden Linko(cid:2)pingUniversity
Linko(cid:2)ping,Sweden
MiroslawStaron
DepartmentofComputerScience
andEngineering
ChalmersUniversityofTechnology
Gothenburg,Sweden
ISBN978-3-031-10872-3 ISBN978-3-031-10873-0 (eBook)
https://doi.org/10.1007/978-3-031-10873-0
©TheEditor(s)(ifapplicable)andTheAuthor(s),underexclusivelicensetoSpringerNatureSwitzerland
AG2022
Thisworkissubjecttocopyright.AllrightsaresolelyandexclusivelylicensedbythePublisher,whether
thewhole orpartofthematerial is concerned, specifically therights oftranslation, reprinting, reuse
ofillustrations, recitation, broadcasting, reproduction onmicrofilmsorinanyotherphysicalway,and
transmissionorinformationstorageandretrieval,electronicadaptation,computersoftware,orbysimilar
ordissimilarmethodologynowknownorhereafterdeveloped.
Theuseofgeneraldescriptivenames,registerednames,trademarks,servicemarks,etc.inthispublication
doesnotimply,evenintheabsenceofaspecificstatement,thatsuchnamesareexemptfromtherelevant
protectivelawsandregulationsandthereforefreeforgeneraluse.
Thepublisher,theauthors,andtheeditorsaresafetoassumethattheadviceandinformationinthisbook
arebelievedtobetrueandaccurateatthedateofpublication.Neitherthepublishernortheauthorsor
theeditorsgiveawarranty,expressedorimplied,withrespecttothematerialcontainedhereinorforany
errorsoromissionsthatmayhavebeenmade.Thepublisherremainsneutralwithregardtojurisdictional
claimsinpublishedmapsandinstitutionalaffiliations.
ThisSpringerimprintispublishedbytheregisteredcompanySpringerNatureSwitzerlandAG
Theregisteredcompanyaddressis:Gewerbestrasse11,6330Cham,Switzerland
Foreword
Effectivecollaborationbetweenindustryandacademiaishard.Thisisthecasebe-
cause of a mismatch in incentives between researchers and industry practitioners
thatisquitesignificantandcausesnaturaltensionpoints.Industrypractitioners,in
theend,wanttoturnthefruitsofthecollaborationintobenefitsforthebusiness.Re-
searchersareinterestedinhigh-profilepublications.Industryoftenseekstoachieve
thesebenefitsasfastaspossiblewhereasresearchersoftenworkonatimelinethatis
muchmorerelaxed.Industrywantsaccesstoexpertsinthefieldwhereasresearchers
often use collaborationas an education and training contextfor PhD students and
earlycareerstaff.
Forcollaborationbetweenmultiplecompaniesanduniversitiestobesuccessful,
a carefullydesignedset of mechanismsneedsto be putin place to ensurethatthe
incentives for companies and researchers are as aligned as possible. Over the last
decade,SoftwareCenterhasevolvedasetofmechanismsthatallowthecollaboration
betweenthemembercompaniesandtheuniversitiestonotonlysurvivebutpositively
thrive. These mechanisms include the sprint model where all activities have a 6-
month sprint heartbeat, a mechanism where it is the companies, rather than the
researchers, that prioritize and fund the research projects, an active and engaged
steering committee, and several types of recurring events that allow for frequent
disseminationofresultsaswellasconstructiveengagementaroundtopicsofinterest.
Also,SoftwareCenterhasbeenabletopredictandproactivelyaddresstopicsthat
havebecomeimportantforthemembercompaniesbeforetheneedatthecompanies
was there. At the start, the center predominantlyfocused on software engineering
challenges,andoverthelast10years,dataandartificialintelligenceaswellbusiness
modelsandproductmanagementchallengeshavealsobeenaddressed.
Itisthepartnersinthecenterthatfundthecenterwiththecompaniesproviding
cash and the universities in-kind funding. This allows the decision processes as
well as the collaboration between researchers and companies to be free from the
distractionsthatpubliclyfundedprojectstypicallybringtoacollaboration.Software
Centerhasa broadandmatureexperiencewith collaborationsbetweencompanies
and universities. These collaborations have long been a mainstay of research and
development–fromcreatingtheknowledgefoundationsforthenextgenerationof
v
vi Foreword
solutions,to servingas an extended“workbench”to solve short-term,incremental
problems.Thisprovidesa flowofnewlymintedtalentandallowsforinvestigation
oflong-termresearchchallenges.
ThisbookthatyouareholdingandforwhichIhavethepleasuretoprovidethe
forewordcapturessomeofthemostimpactfulandrelevantresearchworkconducted
overthelastdecade.However,itcantrulynotbeviewedasafairrepresentationas
hundredsofresearcharticleshavebeenpublishedbySoftwareCenter researchers.
Forafulloverview,IreferthereadertotheSoftwareCenterwebsite.
Over the last decade, I have had the honor to act as the chair for the Software
Centersteeringcommittee.SoftwareCenterhasbecomeanessentialinnovationand
researchpartner and to my understanding,the involvedpartnershavesignificantly
benefitedfromthecollaborations.Ithasbeenajoytocontributetobuildingupsuch
an effective collaborationbetween 17 companies and 5 universities, together with
alltheothersinvolvedinSoftwareCenter.Andalthoughthelastdecadewasgreat,I
expectevenmorefromthenextdecadeofSoftwareCenter!
ChairSoftwareCenterSteeringCommittee AndersCáspar
Stockholm,April2022
Preface
William Gibson, the famous science fiction writer, once stated that the future is
already here; it just is not evenly distributed. This basically sums up the state of
digitalizationanddigitaltransformationinsocietyinonesentence.Therearecom-
paniesthatwereborndigitalandthathaveincorporateddigitaltechnologiesatthe
verycoreoftheirbusiness,infusingbusinessmodels,waysofworking,andculture
withfastiterations,data-drivendecisionmaking,andmachinelearningmodels.
Most companies, however,were born in a differentage, and their initial differ-
entiationwasenabledbymechanicalengineering,manufacturing,orotherphysical
capabilities.Overtheyearsoftheirexistence,thesecompanieshavehadtoadoptnew
technologies,mindsets,waysofworking,andorganizationalapproachesinresponse
tothelatestinsightsandinnovationsinordertostaycompetitiveandtodrivegrowth.
Forinstance,manymechanicalcompanieshaveadopteddigitaldesigntoolsandad-
vancedsimulationapproachesto analyzethe propertiesof physicaldesignsbefore
the first instance of the design is created.This allows for a muchmore systematic
andstructuredexplorationofthedesignspace,resultinginbetteroutcomes.
Althoughthemajorityofcompanieswerenotborndigital,allofthesecompanies
havespenttheirentireexistencefocusingonincorporatingandembracingnewtech-
nologiesandinnovations.Thisraisesthequestionastowhydigitaltransformation
wouldbedifferentfromthecontinuousflowofchangethathasexistedfordecades,
ifnotcenturies.Ourresearchshowsthatthemainreasonisthescopeandmagnitude
of the change caused by digital transformation. It literally affects everything the
companyisconcernedwith,anditrepresentssuchasignificantparadigmshiftthat
itcausesstressandtensioninanyorganizationthatgoesthroughthetransformation
asmanyofthebeliefsheldbytheemployeesarefundamentallybrokenandneedto
bereplaced.
Although it is hard to structure a multifaceted transformationsuch as a digital
transformation,usingtheBAPOmodel[465]wecandiscussseveralaspects:
(cid:129) Business: One of thekey changesduringa digitaltransformationis the change
in business model. Traditional business models rooted in “atoms” tend to be
transactional in nature. This means that the user buys a product, uses it for a
vii
viii Preface
periodoftime,andthenreplacestheproductwithanewinstanceoftheproduct.
Digitalcompaniestendtohavecontinuousbusinessmodels,suchassubscription-
orusage-basedmodels,thatautomaticallycausethecompanyandtheusertohave
acontinuousrelationshipaswell.Thiscontinuousrelationshipthenopensupfor
continuousimprovementsto the offeringwhich is initially a delighter but soon
becomesan expectation.Once it is an expectation,companiesthat do not offer
continuousimprovementsareatadistinctdisadvantage.
(cid:129) Architecture: The continuousbusiness model, especially the ability to contin-
uouslyimprovetheofferingtothecustomer,hassignificantimplicationsonthe
architecture and technology used in the product. Traditionally, many products
were built with deep integration between different subsystems to optimize for
billofmaterials.Theresultwasahighlyinterconnecteddesignthatexhibitshigh
coupling between the various parts, which complicates updating of individual
components and subsystems. This does not only affect upgradability but also
other quality attributes such as testability, as it is virtually impossible to test
individualcomponentsandsubsystemsindependently.Thisthenrequiresalltest-
ingtotakeplaceatthesystemlevel,whichiscomplex,expensive,andresource
consuming.
(cid:129) Process:Althoughmanycompanieshaveadoptedagilepracticesattheteamlevel,
itisinterestingtoobservethatbeyondtheteamlevel,manycompaniesstilllivein
awaterfallworldwherevariousactivitiesareperformedsequentiallyratherthanin
parallel.Astheseactivitiesareperformedoccasionallyandnotveryfrequently,
there is little pressure to automate, and consequently significant manual effort
is expended on what basically are repetitive tasks. Digital transformation and
the associated continuousimprovementmodelsrequiremuchhigherdegreesof
automationastasksthatwereperformedonceortwiceperyeararenowconducted
everyfewweeks.
(cid:129) Organization:Thechangestothebusinessmodel,architecture,andtechnology
as well as the process and ways of working also has significant impact on the
waycompaniesorganizethemselves.Ratherthanfunctionalunitsorganizedina
hierarchicalorganization,digitalizationrequirescross-functionalteamsoperating
inahighlyautonomousfashion,guidedbyquantitativeoutputmetricsthatthey
areaskedtoaccomplish,ratherthanbyfirstandsecondlevellinemanagersthat
tellpeoplewhattodo.
The impact of digital transformationsas described above only starts to scratch
the surfaceof the requiredchangesin organizations,butit providesan illustration
of the breadth and depth of the requiredchangesdue to digitalization. Everything
has to change and basically at the same time as a continuousbusiness model and
a static, immutable productdo not go well together. Similarly, agile practices and
yearlyreleasesdonotmixeither,andorganizationsdropagilewaysofworkingwhen
thereisnoneedtopushfunctionalityouttocustomers.Thecarefulsequencingand
buildingofcapabilitiesaswellasthetransitioningfromthetraditionaltonewways
ofworkingarechallengingandmanycompaniesstrugglewiththis.
Preface ix
Accelerating DigitalTransformation
Oneofthekeyquestionsthatmanyorganizationsaskthemselvesishowtheirdigital
transformationcouldbeaccelerated.Manyorganizationsfacecompetitivepressures
notonlyfromtheir peersin industry,butalso fromnewentrantswhoare digitally
bornandembracealltheadvantagesthatthisoffers.Althoughtheincumbentshave
thecustomerbaseandthenewentrantshavetoovercomethisinertia,thethreatof
beingdisruptedbypure digitalplayersis realin manyindustries.The bestway to
meetthethreatistoundothemainadvantagethatnewentrantshaveandtooperate
inthesamewaytowardcustomersasthenewentrants.
Althoughthismayseemlikeachangeontheinterfacebetweenyourorganization
and the customer, the effect is of course much more involved than just that. The
interface with the customer includes the business model as well as the way the
organization services its customers. As we discussed in the previous section, the
mainshiftinbusinessmodeltypicallyistomovefromatransactionaltoacontinuous
modelsothatthecustomerexperiencesacontinuouslyimprovingoffering.Especially
for embedded systems, the business model may involve a change in ownership of
thephysicalproductitself.Insteadofthecustomerowningtheproduct,thecustomer
hastheaccessandusebutisinfactnottheowneroftheproduct.Thesechangesgo
farbeyondtheinterfacebetweenthecompanyanditscustomers.
Itisclearthatchangeisneeded,anditneedstooccurrapidlyinordertoprotectthe
competitivepositionoftheorganization.Ourexperienceisthatallcompanieschange,
butthepaceofchangetendstoberelativelyslow.Organizationsareinherentlybuilt
to resist change, and accelerating the pace of change requires typically external
forces.
The history of Software Center starts with the above realization by key R&D
leaders at four large software development organizations in Gothenburg,Sweden,
i.e.,Ericsson,SaabSurveillance,VolvoCarCorporation,andABVolvo.Theselead-
ersrequestedfromthepresidentofChalmersUniversityofTechnologyatthattime
that the university start a center, in part funded by these companies, to help them
collaborateinbuildingsoftwaredevelopmentexpertiseintheirrespectiveorganiza-
tions.SoftwareCenterstartedin2011withthefouraforementionedcompanies,with
GothenburgUniversityandChalmersUniversityofTechnologyasfoundingmem-
bers.Overthelastdecade,thecenterhasgrownto17companiesand5universities,
andthescopeofresearchhasevolvedandexpandeddramatically.
Initially,theprimaryfocusofthecompanieswastheadoptionofagilesoftware
development practices, but over the years, the research has expanded to include
workonsoftwarearchitecture,metrics,customerdata,businessecosystems,artificial
intelligence,businessmodels,digitalization,etc.Asyouwillreadintheremainder
ofthebook,thereisa multitudeoftopicsaddressedinthe center—allinresponse
to the digital transformation that the companies are increasingly involved in and
pursuing.
The companiesuse SoftwareCenter notonlyforthe researchthatinspiresspe-
cific changesand improvementsat the company.The other primarymechanismis
thatcompaniesworkonexchangingknowledgebetweeneachother,withoutthein-
x Preface
volvementoftheresearchersattheuniversities.Overtheyears,wehavedeveloped
a variety of mechanisms to support the knowledge exchange, research, and adop-
tionforthecentermembers,includingourreportingworkshopthatrunstwice per
year,ourYouTubechannel,industrialPhDcandidates,weeklybrownbagseminars,
amonthlynewsletter,andourwebsitetomentionafew.
This Book
Thisbookaimstocelebratethe10-yearanniversaryofSoftwareCenterbypresenting
someofthemostimpactfulandrelevantpublicationsthattheparticipantsinthecenter
havecontributedoverthelastdecade.
ThebookisorganizedintothefivethemesthatresearchinSoftwareCenteriswo-
venaround,i.e.,ContinuousDelivery,ContinuousArchitecture,Metrics,Customer
DataandEcosystem-DrivenDevelopment,and,finally,AIEngineering.
ThefocusoftheContinuousDeliverythemeistohelpcompaniestocontinuously
build high-quality products with the right degree of automation. The researchers
of the theme work with improving large-scale utilization of CI/CD (Continuous
Integration/ContinuousDelivery)inareassuchasoptimizingautomatedtestingand
improvingthequalityofthetestcasesandthetestenvironment.Workisalsodone
in finding the role of different manual testing activities and formal verification in
CI/CD.MostcompaniesalreadyhaveaCI/CDtoolchaininplacethatperformsits
individualfunctionswell,buttogetanoverviewoftheprocess,forinstancetoassess
theconfidenceabouttheproductquality,isnottrivial.Oneapproachinvestigatedin
thethemeistousevisualizationtechniquesasameansforfulfillingtheinformation
needs. The Continuous Architecture theme addresses challenges that arise when
balancing the need for architectural quality and more agile ways of working with
shorter development cycles. We have, for example, investigated how architectural
technicaldebtcanbeidentified,managed,andprioritizedandhowcompaniescan
keepdevelopmentartifactsconsistentasthesystemevolves,suchascodeandmodels
at different levels of abstraction. The research has also addressed how to extend
agileprinciplesfrompuresoftwaretofitthedevelopmentofmechatronicproducts,
includingmoreincrementalprocessesforsafetyandsecurityassurance.
TheMetricsthemestudiesandprovidesinsighttounderstand,monitor,andim-
provesoftwareprocesses,products,andorganizations.Itaddresseschallengesrelated
to finding the rightmetrics for the rightorganization (e.g., quantifyingproductiv-
ity during organizationaltransformations),right product(e.g., finding the optimal
featuresetforarelease),andprocess(e.g.,findingtherightmetricstomonitorthe
evolution,ordecay,ofsoftwarearchitectures).Theresearchersinthemetricstheme
utilizethenewestmethodsfrommachinelearning,artificialintelligence,andmining
softwarerepositoriestooptimizetestselection,identifycyber-securityrisks,support
agiletransformations,andsupportautomotivesoftwaredevelopment.
The fourth theme, Customer Data and Ecosystem-Driven Development, helps
companiesmakesense of the vastamountsof data thatare continuouslycollected