EatingtheITElephant:MovingfromGreenfield
DevelopmenttoBrownfield
byRichardHopkins;KevinJenkins
Publisher:IBMPress
PubDate:May06,2008
PrintISBN-10:0-13-713012-0
PrintISBN-13:978-0-13-713012-2
eTextISBN-10:0-13-714946-8
eTextISBN-13:978-0-13-714946-9
Pages:256
TableofContents|Index
Overview
APractical,Start-to-FinishApproachtoManaging,
Evolving,andTransformingLegacyITSystems
ForeveryITexecutive,manager,architect,programleader,
projectleader,andleadanalyst
"RichardandKevinintroduceustoarealitythat'soften
neglectedinourindustry:theproblemofevolvinglegacy
systems,adomaintheycall'Brownfielddevelopment.'The
authorsidentifytherootoftheproblemasthatofcomplexity,
andofferanapproachthatfocusesonthefundamentalsof
abstractionandefficientcommunicationtonibbleatthis
problemoftransformationbitbybit.Astheoldsayinggoes,the
wayyoueattheelephantisonebiteatatime.Richardand
Kevinbringustothetablewithknifeandforkandothertools,
andshowusawaytodevourthiselephantintheroom."
GradyBooch,IBMFellow,co-creatorofUML
"Mostorganizationsinthe21stcenturyhaveanexisting,
complexsystemslandscape.ItistimethattheITindustryface
uptotherealityofthesituationandtheneedfornew
developmentmethodsandtoolsthataddressit.Thisbook
describesanewapproachtothedevelopmentoffuture
systems:astructuredapproachthatrecognizesthechallenges
of'Brownfield'development,isbasedonengineeringprinciples,
andissupportedbyappropriatetooling."
ChrisWinter,CEngCITPFBCSFIET,IBMFellow,Memberofthe
IBMAcademyofTechnology
MostconventionalapproachestoITdevelopmentassumethat
you'rebuildingentirelynewsystems.Today,"Greenfield"
developmentisararity.Nearlyeveryprojectexistsinthe
contextofexisting,complexsystemlandscapes--oftenpoorly
documentedandpoorlyunderstood.Now,twoofIBM'smost
experiencedseniorarchitectsofferanewapproachthatisfully
optimizedfortheuniquerealitiesof"Brownfield"development.
RichardHopkinsandKevinJenkinsexplainwhyaccumulated
businessandITcomplexityistherootcauseoflarge-scale
projectfailureandshowhowtoovercomethatcomplexity"one
biteoftheelephantatatime."You'lllearnhowtomanage
everyphaseoftheBrownfieldproject,leveragingbreakthrough
collaboration,communication,andvisualizationtools--including
Web2.0,semanticsoftwareengineering,model-driven
developmentandarchitecture,andevenvirtualworlds.
Thisbookwillhelpyoureengineernewflexibilityandagilityinto
yourITenvironment…integratemoreeffectivelywithpartners…
prepareforemergingbusinesschallenges…improvesystem
reuseandvalue…reduceprojectfailurerates…meetany
businessorITchallengethatrequirestheevolutionor
transformationoflegacysystems.
·Systemcomplexity:understandit,andharnessit
Gobeyondthecomfortingillusionofyourhigh-level
architecturediagrams
·Howconventionaldevelopmenttechniquesactually
makethingsworse
Whytraditionaldecompositionandabstractiondon'twork-andwhattodoinstead
·ReliablyreengineeryourITinlinewithyourbusiness
priorities
Newwaystounderstand,communicate,visualize,
collaborate,andsolvecomplexITproblems
·Cuttheelephantdowntosize,onestepatatime
MasterallfourphasesofaBrownfieldproject:survey,
engineer,accept,anddeploy
EatingtheITElephant:MovingfromGreenfield
DevelopmenttoBrownfield
byRichardHopkins;KevinJenkins
Publisher:IBMPress
PubDate:May06,2008
PrintISBN-10:0-13-713012-0
PrintISBN-13:978-0-13-713012-2
eTextISBN-10:0-13-714946-8
eTextISBN-13:978-0-13-714946-9
Pages:256
TableofContents|Index
Copyright
ForewordbyGradyBooch
ForewordbyChrisWinter
Preface
Acknowledgments
AbouttheAuthors
PartI:IntroducingBrownfield
Chapter1.EatingElephantsIsDifficult
Today'sDeliveryMethods
WhyDoBigProjectsFail?
EnvironmentalComplexity
BrownfieldSitesMustBeSurveyed
Endnotes
Chapter2.TheConfusionofTongues
IntroducingBrownfield
KeyCommunicationProblems
OvercomingCommunicationComplexity
Endnotes
Chapter3.Big-MouthedSuperheroRequired
Elephant-EatingStrategies
ConsumingtheEnvironment
ArchitectingtheElephantEater
TheElephantEaterinAction
TheBrownfieldBeliefs
Endnotes
Chapter4.TheTrunkRoadtotheBrain
AlternativeWallpapers
InvadingHilbertSpace
ArchitectureIstheSolution
BridgingtheBusiness/ITGap
Endnotes
Chapter5.TheMythicalMetaman
WhenYouEliminatetheImpossible
TheLongTailofBusinessServices
BusinessAttractorsforAttractiveBusinesses
TheDeathofBrownfield
Endnotes
PartII:TheElephantEater
Chapter6.AbstractionWorksOnlyinaPerfectWorld
ConsiderationsforanElephantEater
SystemsIntegrationandEngineeringTechniques
AbstractionIstheHeartofArchitecture
DoWeNeedaGrandUnifiedTool?
TheConnoisseur'sGuidetoEatingElephants
Endnotes
Chapter7.EvolutionoftheElephantEater
TheSourcesofBrownfield
Isn'tThisCASE?
Isn'tThisMDA?
Endnotes
Chapter8.BrownfieldDevelopment
AgileWaterfalls
TheBrownfieldDevelopmentApproach
Endnotes
Chapter9.InsidetheElephantEater
LookingInsidetheElephantEater
Step1:ParseViewandIdentifyPatterns
Step2:MergeViews
Step3:CreateTransforms
Step4:GenerateArtifacts
Steps5.1:TestArtifactsand5.1a:IdentifyGeneration
Faults
Step5.1b:AddandUpdateInformation
APortraitofanElephantEater
Endnotes
Chapter10.ElephantEateratWork
MakingtheMovetoBrownfield
TakingtheFirstStep
ABetterWaytoBuildInterfaces
ABetterWaytoBuildanEnterpriseServiceBus
TheEndofMiddleware?
EvolvingaDeployableEnterpriseArchitecture
Index
Copyright
Theauthorsandpublisherhavetakencareinthepreparationof
thisbook,butmakenoexpressedorimpliedwarrantyofany
kindandassumenoresponsibilityforerrorsoromissions.No
liabilityisassumedforincidentalorconsequentialdamagesin
connectionwithorarisingoutoftheuseoftheinformationor
programscontainedherein.
©Copyright2008byInternationalBusinessMachines
Corporation.Allrightsreserved.
NotetoU.S.GovernmentUsers:Documentationrelatedto
restrictedright.Use,duplication,ordisclosureissubjectto
restrictionssetforthinGSAADPScheduleContractwithIBM
Corporation.
IBMPressProgramManagers:TaraWoodman,ElliceUffer
Coverdesign:IBMCorporation
AssociatePublisher:GregWiegand
MarketingManager:KourtnayeSturgeon
Publicist:HeatherFox
AcquisitionsEditor:KatherineBull
DevelopmentEditors:KevinFerguson,GinnyBess
ManagingEditor:GinaKanouse
Designer:AlanClements
SeniorProjectEditor:LoriLyons
CopyEditor:KristaHansing
Indexer:LisaStumpf
Compositor:NonieRatcliff
Proofreader:AnneGoebel
ManufacturingBuyer:DanUhrig
PublishedbyPearsonplc
PublishingasIBMPress
IBMPressoffersexcellentdiscountsonthisbookwhenordered
inquantityforbulkpurchasesorspecialsales,whichmay
includeelectronicversionsand/orcustomcoversandcontent
particulartoyourbusiness,traininggoals,marketingfocus,and
brandinginterests.Formoreinformation,pleasecontact:
U.S.CorporateandGovernmentSales
1-800-382-3419
ForsalesoutsidetheU.S.,pleasecontact:
InternationalSales
Thefollowingtermsaretrademarksorregisteredtrademarksof
InternationalBusinessMachinesCorporationintheUnited
States,othercountries,orboth:IBM,theIBMlogo,IBMPress,
AD/Cycle,DB2,developerWorks,Rational,System360,Tivoli
andWebSphere.JavaandallJava-basedtrademarksare
trademarksofSunMicrosystems,Inc.intheUnitedStates,
othercountries,orboth.Microsoft,Windows,WindowsNT,and
theWindowslogoaretrademarksofMicrosoftCorporationin
theUnitedStates,othercountries,orboth.Othercompany,
product,orservicenamesmaybetrademarksorservicemarks
ofothers.
LibraryofCongressCataloging-in-PublicationData
Hopkins,Richard.
EatingtheITelephant:movingfromgreenfielddevelopmenttobrownfiel
HopkinsandKevinJenkins.
p.cm.
Includesindex.
ISBN0-13-7130120(pbk.:alk.paper)1.Informationtechnology.2.Business
enterprises—Planning.I.Jenkins,Kevin.II.Title.
T58.5.H692008
004.068—dc22
2008004231
Allrightsreserved.Thispublicationisprotectedbycopyright,
andpermissionmustbeobtainedfromthepublisherpriorto
anyprohibitedreproduction,storageinaretrievalsystem,or
transmissioninanyformorbyanymeans,electronic,
mechanical,photocopying,recording,orlikewise.For
informationregardingpermissions,writeto:
PearsonEducation,Inc
RightsandContractsDepartment
501BoylstonStreet,Suite900
Boston,MA02116
Fax(617)6713447
ISBN-13:978-013713012
TextprintedintheUnitedStatesonrecycledpaperatR.R.
DonnelleyinCrawfordsville,Indiana.
FirstprintingMay2008
Dedication
ToMam,Dad,Lyndsay,andthethreebestkidsinthe
world:Kirsty,Julia,andNicholas.Theywouldhave
preferredabookonrealelephants,butatleastthey
approveofthecover.
—R.H.
ToAnna,forsupportingmethroughthestressfultimesof
deliveringlargeprojects.
—K.J.
ForewordbyGradyBooch
Asimpleback-of-the-envelopecalculationsuggeststhat,
worldwide,weproduceabout33billionlinesofnewormodified
codeeveryyear.Cumulatively,thismeansthatsincethe1940s
and'50s(whenhigherorderprogramminglanguagesbeganto
gainsometraction),we'veproducedsomewherearoundone
trillionsourcelinesofcode.
Ontheonehand,thisvolumeofoutputsuggeststhatoursisan
incrediblyvibrantandinnovativeindustry.Ontheotherhand,
it'sahumblingthought,forthroughthosetrillionlinesofcode,
allhandcraftedbyindividualhumanlabor,we'vechangedthe
world.
Truthbetold,somenontrivialpercentageofthe33billionlines
yearlyisdeadonarrivalorsotransitorythatit'sthrownaway
quickly.Muchofthatcode,however,hasalongerhalf-life,and
evensomeofthatcodelivesafter10,20,oreven30ormore
years.Formanydevelopers,thecodetheywritetodaybecomes
tomorrow'slegacythattheirchildrenortheirchildren'schildren
maystareatsomeday,tryingtouseit,adaptit,evolveit,
askingthequestion,"Whattheheckwasthisdeveloper
thinking?"
Greenfielddevelopment,quitehonestly,isgreatfun,simply
becauseyougettostartwithacleanslateand,thus,arenot
burdenedbyanythingfromthepast.Forthemostpart,we
teachGreenfielddevelopmentinourschools;furthermore,
start-upcompanieslooksomuchmorenimblethantheirolder
counterpartsbecausetheydon'thavethemillstoneoflegacy
aroundtheirnecks.Woebeuntothestudentwhoentersthe
realworld(it'snotlikebeingattheuniversity,unlessyoumove
fromthecollegewombtothestart-upwombimmediately),and
woebeuntothestart-upcompanythatbeginstomatureinto
sustainabledevelopmentandsoonrealizesthatyoucan'tjust
startover.
RichardandKevinintroduceustoarealitythat'soften
neglectedinourindustry:theproblemofevolvinglegacy
systems,adomaintheycallBrownfielddevelopment.The
typicaleconomicallyinterestingsystemthesedaysis
continuouslyevolving(youcan'tshutitoff)andever-growing.
Theauthorsidentifytherootoftheproblemasthatof
complexity,andofferanapproachthatfocusesonthe
fundamentalsofabstractionandefficientcommunicationto
nibbleatthisproblemoftransformationbitbybit.Theirmodel
ofViews,Inventory,Transforms,andArtifactsoffersan
approachtoreasoningaboutandexecutingonthe
transformationofBrownfieldsystems.Theyproposea
Brownfieldlifecycleinvolvingsurveying,engineering,
acceptance,anddeploymentthatoffersameansofgoverning
thistransformation.
Astheoldsayinggoes,thewayyoueattheelephantisonebite
atatime.RichardandKevinbringustothetablewithknifeand
forkandothertools,andshowusawaytodevourthiselephant
intheroom.
GradyBooch
IBMFellow
January2008
ForewordbyChrisWinter
Ijoinedthecomputerindustryasacomputerprogrammer,
straightfromschool,in1969.Duringacareerthathasspanned
nearly40years,Ihaveworkedprimarilyintheareaof
applicationsdevelopmentandsystemsintegration.Iwrotemy
firstapplicationin1969;itwasaComputerAidedDesign(CAD)
graphicsapplicationforhardwareengineerstodesignPrinted
CircuitBoards.Thisapplicationgavetheboarddesigneratool
withthenecessaryphysicalrulesoftheelectroniccomponents
andhowtheycouldbeused.Intheearly1970s,Ideveloped
CADandotherapplicationstoassistbuildingarchitectsin
designinglargepublicbuildings,suchasschoolsandhospitals.
Thesesystemsassistedthearchitectsandcivilengineersinthe
designprocessofthebuilding;bycapturingthedesign,itwas
possibletoproduceallthenecessarydrawingstogetherwith
thebillsofmaterialsforthebuilding.
Intheintervening40years,Ihaveperformedavarietyof
differentroles,includingprogrammer,analyst,designer,
architect,projectmanager,andtroubleshooter.ThesystemsI
developedwereinabroadspectrumofindustries,including
manufacturing,banking,insurance,retail,utilities,andboth
localandfederalgovernment.Today,IamanIBMFellow[1]in
theIBMGlobalBusinessServicesdivisionandanactive
memberoftheIBMAcademyofTechnology.[2]Myprimary
responsibilityistotechnicallyshapeandensurethetechnical
healthoflargeandcomplexsystemsintegrationandstrategic
outsourcingprogramsandbids.IamaCharteredIT
Professional(CITP),aCharteredEngineer(CEng),aFellowof
theBritishComputerSociety(FBCS),[3]andaFellowofthe
InstitutionofEngineeringandTechnology(FIET).[4]
Lookingbacknowonwhatwetriedtoachievewiththedesign
andbuildofelectroniccircuitsandbuildingsintheearly1970s,
IamdisappointedandsomewhatdisillusionedbytheIT
industry'slackofsuccessinitsownadoptionofengineering-
basedmethodssupportedbycomputer-basedtoolstoarchitect,
design,build,integrate,andtestITsystems.Intoday'sworld,
itwouldbeinconceivabletodevelopacomplexsystemsuchas
theAirbus380withouttheengineeringdisciplinesandwithout
theengineeringtoolsprovidedbytheITindustry.TheIT
industryissignificantlylessmatureatadoptingengineering
techniquestodevelopitscomplexsystems.Itcannolonger
relyonrelativelyimmaturepracticesoftensupportedbyoffice
productivitytoolssuchaswordprocessors,presentationtools,
andspreadsheets.TheITindustryneedsabroaderadoptionof
trueengineering-basedtechniquessupportedbytoolsdesigned
forengineers.
Ithasbeenmypersonalexperienceinrecentyearsthatthe
overallcostandcomplexityofbuildingbespoke(custom)
applicationsorcustomizingCommercialOffTheShelf(COTS)
packageshasincreased—ashastherisk.Onfurther
investigation,itisapparentthatitisnotthebuildcostthathas
increased,buttheincreaseinthesizeandcomplexityofthe
integrationofsuchprojectsintothesystemslandscape.From
myownrecentexperience,theratioofeffortofnewbuildto
integrationis3:1.Foreverydollarspentonnewfunctionality,
thetotalcostisfourdollarstocutoverthisfunctioninto
production.Thiscostexcludesend-usertraining.Inan
environmentwherebothsizeandcomplexityofthesystems
landscapearecontinuallyincreasing,thereisaresulting
increaseinthecostsofmaintenance.Inaddition,organizations
areburdenedwithaneedtomeetincreasinglevelsof
legislationandregulation.Allofthisresultsinreducedbudgets
fornewdevelopmenttogetherwithdecreasingwindowsof
opportunitytodeploynewfunctionintheglobal24x7service
culture.ITinnovationisbeingstifled.Themethodsandtools
thatareinusetoday,albeitlimited,areinthemain,primarily
targetedatGreenfieldsystem'slandscapes.Therealityisthat
mostorganizationsinthetwenty-firstcenturyhaveanexisting,
complexsystemslandscape.WhenIrefertothesystems
landscape,ImeanboththebusinessanditsenablingIT
systems.TheseITsystems,inturn,arecomprisedof
applicationsandtheirdatadeployedonoftencomplexnetwork
andcomputerinfrastructure.Thedocumentationofsuch
systemsistypicallypooranditsongoingmaintenanceishighly
dependentonasmallnumberofknowledgeable"system
experts."[5]TheITindustryneedsamorestructuredapproach
tounderstandingthesesystemlandscapes.
Thisistherealityoftheworldinwhichtheauthorsofthisbook,
RichardHopkinsandKevinJenkins,andI,architect,design,and
implementnewsystemsforourclientsinexistingcomplex
systemslandscapes.ItistimethattheITindustryfaceupto
therealityofthesituationandtheneedfornewdevelopment
methodsandtoolsthataddresstheseissuesandtakeour
industryintothetwenty-firstcentury.
Animportantfirststepinresolvingthisistoprovideaname
thatdescribesboththeproblemanditssolution.Inthesearch
foraname,theauthorshaveturnedtothebuildingindustry
wherenewbuildingsareincreasinglybeingdevelopedon
Brownfield[6]sites.Thisisanalogoustothemajorityoftoday's
newsystemsthatarebeingdevelopedonBrownfieldsystems
landscapes;itismyexperiencethatmorethan90percentof
newdevelopmentisdeployedintoaBrownfieldenvironment.
Thechallengesarenotrestrictedtojustthetransformationof
legacysystems,butwiththeintegrationintotheBrownfield
systemslandscapeitself.
Thisbookdescribesanewapproachtothedevelopmentof
futuresystems.Itisastructuredapproachthatrecognizes
thesechallenges,itisbasedonengineeringprinciples,anditis
supportedbyappropriatetooling.Itisspecificallydesignedto
solvethechallengesofBrownfielddevelopment.
ChrisWinter
CEngCITPFBCSFIET,IBMFellow
MemberoftheIBMAcademyofTechnology
ForewordEndnotes
1. "IBMAppointsSixNewFellowsWhoExplorethe
BoundariesofTechnology." />May2007.
2. IBMAcademy. />3. BritishComputerSociety. />4. TheInstitutionofEngineeringandTechnology.
/>5. Lindeque,P."WhydolargeITprogrammesfail?"
/>Index=390,September2006.
6. BrownfieldisdescribedbytheNational
AssociationofRealtors®as"Theredevelopment
ofexistingurban,suburban,andruralproperties
alreadyservedbyinfrastructureincluding
'brownfields'sites,thatareormaybe
contaminated,stimulatesgrowthandimprovesa
community'seconomicvitality.Developmentin
existingneighborhoodsisanapproachtogrowth
thatcanbecost-effectivewhileproviding
residentswithacloserproximitytojobs,public
services,andamenities."
Preface
Withineverybusiness,thereisadesireforrapidchangeto
meetcustomerdemands.Suchchangesusuallyinvolve
changingsupportingITsystems.Whenlargebusinesschanges
arerequired,theaccompanyingITchangestendtobe
significant,too.However,alltoooften,thesebigprojectshit
problems,runoverbudget,aredelayed,orsimplyget
cancelled.Evenin2006,65%ofITprojectsfailedononeof
thesecounts.[1]Largeprojectshaveanevenpoorersuccess
rate.Suchoddsareveryworryingwhenthestakesarevery
high.Thisbookidentifiesthefundamentalissuesattheheartof
theITindustry'scurrentapproachesandprovidesanewway
forward.Allpeopleinvolvedinlarge-scalebusinessandIT
changeshouldreadthisbook.
TheDaytheElephantWasBorn
TheITindustryhasmanykeydates,buttheintroductionin
1964ofIBM'snew-generationmainframe,calledthe
System/360,markedthestartofanewera.Untilthatpoint,
buyinganewbusinesscomputermeantrewritingyourexisting
software.TheSystem/360changedallthatwiththe
introductionofafamilyofcompatiblecomputersandassociated
devices:Aprogramthatranononewouldrunonany.The
industryfollowedsuitwithequivalentproducts,andthenature
ofITchangedinonefellswoop.
ITinvestmentscouldnowbeeasilypreserved.Theprograms
thatranontheSystem/360stillrunonIBM'smainframe
platformstoday.
Thiswasanimperceptiblechangeatfirst,butitwasahugely
significantmilestone.Atthispoint,ITcomplexitystarted
accumulatingwithintheenterprise.Systemsgrewwiththe
business.Thousandsofperson-yearsoftime,effort,andmoney
flowedintotheseITsystems.Theygotcomplex.Theybecame
elephants.
Inthemeantime,ITfashionscameandwent.Overtheyears,
theoriginalstructuredprogramshavebeenaugmentedby
object-orientedprogramming,wrappedbycomponent-based
development,andadvertisedbyServiceOrientedArchitecture
(SOA).Eachofthesemovementshashaditsownstrategyfor
dealingwiththecomplexity,butnoneeverreallytookitto
heart.
Today'sITsystemsaresocomplexthattheysimplydefy
everydaycomprehension,spillingoutofourmindsaswetryto
getourheadsaroundthem.Responsibilityformaintainingthem
issplitamongavarietyofskilledgroupsandmyriadproducts
andprogramsthatcoexisttosupportthefunctionsofthe
enterprise.TodealwiththisHydra,wedrawhigh-level
architecturediagramsthatcomfortusbymakingthingslook
simple.Thesediagramsareanillusion,atrick,afacade.They
are,atbest,approximationsforeasyconsumptionandhighlevelcommunication.Atworst,theyinstillfalseoptimismabout
ourabilitytomakechangestothatcomplexity.
Such"fluffycloud"diagramscannothidegenuinecomplexity
forever.Toachieveyourbusinessgoalsandchangethose
systems,youmustunderstand,communicate,andharnessthe
realcomplexity.Noonecanunderstandthewholebeast,so
vastamountsofwell-coordinatedteamworkandunambiguous
communicationarerequiredtocompletesuchtasks.This
combinationofhighlevelsofcomplexityandtheneedforclear
communicationofthatcomplexityamonghundredsof
individualsdestroysbigprojects.
DoINeedtoMovefromGreenfieldto
Brownfield?
ITsystemsaregenerallynotimplementedonGreenfieldsany
more.Theaccumulatedcomplexitysince1964meansthatthe
environmentformostbigITprojectsisoneofimmense
challenge,entangledinanalmostuncountablenumberof
environmentalconstraints.
Thisistheunderlyingreasonforthedemiseofmostlarge-scale
ITprojects.Only30%oflargeITprojectssucceed.
Bigprojectsareusuallyexecutedon"contaminated"sites,
whereyouneedtobecarefulofwhereandhowyoubuild;a
changeinoneplacecanripplethroughtoothersystemsin
unexpectedways.Suchsitesaremorebrownthangreen,and
theITindustryneedstoadoptaBrownfield-orientedapproach
toaddressthemsuccessfully.
ThisbookintroducessuchaBrownfieldapproachandexplains
whycurrentmethodsarestillessentiallyGreenfield.Itis
specificallywrittenforpeoplewhowanttochangetheir
businessandknowthattheycandoitonlybybuildingonwhat
hasgonebefore.Ifanyofthefollowingistrue,thisbookisfor
you:
YouareaCIO,CTO,ITdirector,projectexecutive,project
director,chiefarchitect,orleadanalystwhois
contemplatingasignificantchangeinyourITlandscape.
YoucannotaffordtoreplaceyourwholeITlandscape.
Yoursystemstalktoafairnumberofsystemsoutsideyour
directcontrol.
YouwouldliketoreengineeryourexistingITenvironment
sothatitwillremainflexibleforthefuture.
Youaredeeplyunhappywiththecurrentfailureratesof
largeITprojects.
YouarecontemplatingsendingasignificantpartofyourIT
developmentandtestingworkoff-shore.
EatingtheITElephantwaswrittenbytwofull-timeExecutiveIT
ArchitectsfromIBMwhocanandhavetickedeverysingleone
ofthoseboxesonanumberofoccasions.Wehavebeen
accountableforthetechnicaldirectionandday-to-day
implementationofsomeofthelargestsystemsintegrationand
reengineeringprojectsthatIBMhasundertaken.Webelieve
stronglythatexistingGreenfielddevelopmentapproachesare
anincreasinglypoormeansofaddressingtoday'sbusiness
problemsthroughITsolutioning.Tobeblunt,wehavea
numberofyearsofhard-wonexperience,andwehavegrown
tiredoftherecurringproblemsofITdelivery.Inrecentyears,
wehavedeliberatelysoughtadifferentapproach;thefollowing
pagesdetailthefruitsofourlaborsandthatofourcolleagues.
Hereticswemightbe,butpragmatistswearealso,and,hand
onheart,wecansaythattheinsightweshareherehas
significantlyacceleratedandsimplifiedanumberofrecentIBM
engagements.
Wedon'tthinkthehighfailurerateofmajorITprojectsisdoing
ourindustryanyfavorsandwouldliketopopularizethe
approachthathasserveduswell.Ifwecanhelpmitigatethe
impactoftheunavoidablycomplexITenvironmentandknock
downsomebigprojectcommunicationbarriers,webelievethat
successratewillimprove.
AReader'sDigest
Thisbookisnotatechnicalmanualnoracookbook;itdoesnot
containasinglelineofcode,andwehavetriedtominimizethe
useoftechnicaldiagramsandjargon.Thisisabookabout
changingthewayweapproachlargeandcomplexbusinessand
ITreengineeringprojects.
Tomakethebookasaccessibletoasmanypeopleaspossible,
wehavesplititintotwoparts.
PartIisforallreaders.Initially,itdefineswhatiswrongwith
large-scaleITprojectsanddeterminestherootcauseoffailure
(seeChapters1,"EatingElephantsIsDifficult,"and2,"The
ConfusionofTongues").Theheartofthebook(Chapters3,
"Big-MouthedSuperheroRequired,"and4,"TheTrunkRoadto
theBrain")concentrateondefininganalternativesolution—an
ElephantEater—andtheBrownfieldapproachthatgoeswithit.
InChapter5,"TheMythicalMetaman,"welookatthenew
speciesofbusinessesthatemergeasaresult.
PartIIexplainsthetechnicalandpracticalaspectsofBrownfield
forsomeonewhomightwanttoimplementsuchanapproach.It
startsbyanalyzingexistingElephantEatingtechniques(see
Chapter6,"AbstractionWorksOnlyinaPerfectWorld")and
explainswhyBrownfieldisdifferent(seeChapter7,"Evolution
oftheElephantEater").InChapters8,"Brownfield
Development,"and9,"InsidetheElephantEater,"welook
insidetheElephantEaterandatsomeofthenewtechnologies
thathavebeenusedtoimplementit.Thebookconcludesby
explaininghowtheBrownfieldapproachcanbeimplementedon
aprojectandthebenefitsitcanbring(seeChapter10,
"ElephantEateratWork").
Forthosewhotakethekeymessagesonboard,awealthof
technicalinformationhasalreadybeenpublishedthatwill
enableanyorganizationtoadoptthecoretechnologiesthatwe
haveused(orequivalentones)toimplementBrownfieldintheir
ownway(seethe"Endnotes"sectionsofChapters8and9).We
hopethatenablingbusinessandITchangeviaanewproject
approach,nottechnology,isattheheartofthisbook.
PartI:IntroducingBrownfield
Chapter1,"EatingElephantsIsDifficult,"introducesthe
metaphorthatperformingalargeITprojectcanbecompared
toeatinganelephant.Itlooksatwhybigprojectsfailand
providesbestpracticesonhowtoovercomesomeofthe
commonreasonsforfailure.
Chapter2,"TheConfusionofTongues,"explainswhythis
accumulatedITcomplexityistherootcauseoffailure,focusing
onthehumancommunicationproblemsitcreates.Itgoesonto
specificallyexaminethe"greatdivide"betweenbusinessandIT
thatcompoundstheproblem.
Chapter3,"Big-MouthedSuperheroRequired,"introducesthe
coreconceptsofBrownfield.ItlooksathowBrownfieldcanbe
implementedtocreateanefficientElephantEater.
Chapter4,"TheTrunkRoadtotheBrain":WedespairatIT
professionals'inabilitytocommunicateaseffectivelyand
efficientlyasthoseinothersimilarprofessions(suchasreal
architects).Chapter4describeshowtheBrownfieldapproach
combinedwiththeVITAarchitectureopensupnewformsof
communication,remotecollaboration,andvisualizationof
complexITproblems.
Chapter5,"TheMythicalMetaman":Thefirstpartofthebook
concludeswithanexaminationofthelikelyimpactof
Brownfield.Itforecastsanewbreedofbusinessesthatare
infinitelymorecustomerfocusedandagilethantoday'sand
explainshowsuchbusinessesmightcomeintobeing.
PartII:TheElephantEater
Chapter6,"AbstractionWorksOnlyinaPerfectWorld":This
moretechnicalhalfofthebookopensbydefiningthe
characteristicsofanElephantEater.Itconsidersexisting
"ElephantEating"approachesandnotesthattheytendto
compoundprojectdifficultiesviatheirextensiveuseof
decompositionandabstraction.
Chapter7,"EvolutionoftheElephantEater,"looksat
Brownfield'stechnicalandprojectroots,andexplainsitskey
differencesfrompreviousideas.Itendswithsomelikely
scenariosandreal-lifeprojectexamplesforwhichBrownfield
hasbeenorcouldbeespeciallybeneficial.
Chapter8,"BrownfieldDevelopment,"introduceshowthe
Brownfielddevelopmentapproachcanbedeployedonaproject.
ItshowshowtostrikeanewbalancebetweenAgile-and
Waterfall-baseddevelopmenttechniquesandprovidessomeof
thebestelementsofeach.Italsodescribesthecorephasesof
Survey,Engineer,Accept,andDeploy,andstatesthebenefitsof
theapproach.
Chapter9,"InsidetheElephantEater":IfChapter8described
whathappensonaBrownfieldproject,Chapter9explainshow
ithappens.Thischapterlooksinsidetheworkingsofan
ElephantEaterandexplainshowiteatstheelephant.The
chapteralsoservesasaneasy-to-readintroductiontothenew
semantictechnologiesthatunderpinWeb2.0andthesemantic
web.
Chapter10,"ElephantEateratWork":Thebookconcludeswith
alookatthepracticalapplicationsoftheElephantEaterand
howitcanhelpsolvesomeoftoday'smostdifficultIT
problems.Thischapterincludesasummaryofthekeybenefits
oftheBrownfieldapproach.
WalkingtheBrownfields
Wehopethatyouwillenjoyreadingthisbookasmuchaswe
enjoyedwritingit.Ifyou'dliketoseemore,gotothewebsite
www.elephanteaters.org.Additionally,ifyouwouldliketosee
moreofthedynamicnatureofBrownfield,therearetwo
exhibitionsinSecondLife.OneSecondLifesiteisdedicatedto
thebook[Cypa30,180,302].TheotherSecondLifesiteis
dedicatedtotheuseofBrownfieldwithinIBMat[IBM1140,
150,60].Welookforwardtomeetingyouthere.
Endnotes
1. TheinitialCHAOSreportfromStandishGroupin1994reporteda16%
projects.Thissuccessratehasgenerallyincreasedovertheintervening
Groupreported35%ofITprojectsbeingontimeandwithinbudget,an
requirements.Theonlyblipinthatrecordappearedin2004,whenfailu
StandishGroupexplainedthatin2004thereweremorebigprojects—t
becausetheyareoftenforcedtoabandoniterativedevelopmenttechni
In2007,arivalreporttoCHAOSbySauer,Gemino,andHornerReichl
foundthatmorethan65%ofITprojectssucceeded,butitfoundnosu
than200person-years.Thisbooklooksspecificallyatthoselargeproje
Hayes,F.ChaosIsBack.
www.computerworld.com/managementtopics/management/project/sto
Krigsman,M.RearrangingtheDeckChairs:ITProjectFailures
/>
Rubinstein,D.StandishGroupReport.www.sdtimes.com/article/story-2
Sauer,C.,A.Gemino,andB.HornerReigh."TheImpactandSizeandV
Performance."CommunicationsoftheACM50no.11(November2007)
Acknowledgments
Thisbookwouldnothavebeenpossiblewithoutthesupport
andunderstandingofourfamilies(whodon'tseeusenoughat
thebestoftimes).Inaddition,wewouldliketoacknowledge,in
alphabeticalorder,thecontributionofthefollowingpeopleor
companies:
BobLojek—FortakingModelDrivenArchitecturetothe
nextlevelandforinventingthefirstplausibleattemptata
softwareBabelFish.
ChrisWinter—AtrulyinspiringIBMFellowwhobelieves
moreincommunityandthecapabilityofpeoplethan
anyoneelsewe'vemetintheindustry.Chriswasthekey
driverandtechnicalsponsorbehindthiswork.
ChristianHance—ChristianisaprojectexecutiveinIBM
whohadenoughfaithinhisarchitectstoenablethemto
createthefirstiterationofBrownfield.Wearehugely
gratefulthathealsoagreedtoreviewthisbook.
CúramSoftware—Forintroducingustothepracticalities
ofModelDrivenArchitectureinthefirstplace.
FredBrooks—Whocallshisseminalwork,TheMythical
ManMonth,abibleofsoftwareengineeringbecause
everyonehasheardofit,somehavereadit,butveryfew
followit.Wehopeyoudoallthreeandhopethisisnearer
toatestamentthantosomeapocrypha.We'dliketothank
Fredforthequote,"Brownfieldismuch,muchharderthan
Greenfield,whethersoftwareorhouseremodeling."
IanHughes—Probablybetterknowngloballyas
"ePredatorpotato,"IanintroducedustoSecondLifeand,
hopefully,putonemorenailintothecoffinofPowerPointbasedarchitectures(seeChapter4).
IanScott—ExecutiveITarchitectsreallyshouldn'tcode,
butIantookthebravestepofinheritingthepatented
approachandcodethatproducedthepicturesinChapter4.
Thecapabilitiesofthiscodearenowworld-leading,thanks
toIan.
IBM—Thisbookproposesanumberofradicalideas;that
IBMwouldendorsethepublicationofsuchabookshows
Gerstner'sDancingElephantvisionliveson.
MandyChessell—Forcrystallizingthearchitecturein
everyone'sheads,patentingit,andunwittinglyinventing
theacronymVITA.Oh,andtheslides—veryniceslides.
JohnTait—Johnmadeusthinkaboutthetimelineofthis
book.Whydidourgut-feelplacetheoriginsoftheproblem
35yearsago?Whenwethoughtaboutit,wealreadyknew
theanswer,butwehadcompletelymissedthequestion.
KatherineBull—Wealwayshadenthusiasm,impetus,and
drivefortheideasinthisbook,butKatherineaddedthe
sametothepublishingprocess.
KevinFerguson—Kevinsavedusgoodnessknowshow
manydays,thankstohisthorougheditandreviewofour
text.NoonewillsuspectwecomefromCanada.
PhilTetlow—Philisthemostenthusiasticproponentofall
thingssemanticandinherentlycomplex(includingWeb
Science)andhassupportedthiseffortthroughout.
Ourclients—Lastbutnotleast.Wecan'tpossiblyname
you,asthelawyerswouldundoubtedlygetexcited,but,
withoutadoubt,thethingthatkeepsusmotivatedand
innovativeistryingtosolvetheproblemsofourclients,for
theircustomers.WehopeBrownfieldwillassistinsolving
someofthoseintransigentproblemsyouface.