53
The final twochanges in thedatacenter
skill set identifiedby theDCUGsurvey
reveal theneed forgreaterbusiness
andvendormanagement skills. The first
maybepartlydrivenby thegrowing
numberofdatacenters thataredirectly
connected to revenuegenerationand
aremore integral tobusiness
operations.Executives responsible for
those facilitiesneed tostaycurrenton
technologywhilealsounderstanding—
andanticipating—businessdemands
andobjectives.
Businessskillswill alsogrow in
importance for those inorganizations
thatdeploy internal cloudsandcharge
back ITservices, effectively
transforming InformationTechnology
fromaservice that supports the
business toabusiness thatdelivers
services.
Vendorandpartnermanagement skills
areexpected togrow in importanceas
organizations relymoreoncloud
services. Insteadof justproviding the
technologysystemsonwhich the
businessdepends, vendorswill
increasinglybehosting theapplications
andstoringdatacritical today-to-day
businessoperations.Problems resulting
frompoorvendor selectionand
managementwill haveamore
immediateanddirect impacton
businessoperations.
STRATEGIESFORADAPTING
In lightof thecurrentoperational issues
manyorganizations faceand the
pressingneed toevolve towardmore
holisticmanagement, thechallenge for
datacentermanagersseemsdaunting.
But therearestrategies for findinga
betterbalancebetween thedemands
of todayand theneedsof the future:
1.Explore thepotentialofnewdata
center infrastructuremanagement
(DCIM)platforms.
As thedatacenterbecamemore
complex, theneed toconsolidateand
analyzedataacrosssystems increased;
however, that samecomplexitymade it
harder toachieve real-timevisibility
acrosssystems.Previous-generation
management systemssimplydidn’t
have thescaleor sophistication tobring
together thenecessarydataand
convert it intoameaningful viewof
real-timeoperations.
Current-generationsystemshave
overcome thischallenge through the
useofdedicatedappliancescapableof
consolidatingdata fromacrosssystems
and theenvironment toprovidea
meaningful viewof real-time
operations. Thisdevelopment
essentiallybreaks through thecurrent
system-levelmanagementceiling that
haspreventeddatacentermanagers
fromseeing thebigpicture.
2.Breakdownorganizational silos
Withaholistic, real-timeviewofdata
centeroperations, itbecomeseasier to
breakdownorganizational silosbased
ondatacenter systems. System-level
expertisewill remaincritical to
successfuldatacenteroperationsbut it
ismostvaluablewhensystem-level
expertsshare thesameviewof
operationsandwork together.
Geographicsiloscanalso impede
optimization.Moreorganizationsare
seekingglobal servicepartners thatcan
deliverconsistent supportandservices
anywhere in theworld.
3. Supplement internal skillsasnecessary
Basedon thechanges to thedata
center skill set identifiedby theDCUG,
thedatacentermanagerof the future
willbeexpected tooperatemore likea
“general contractor,”managing to the
bigpicture, collaboratingacross the
businessandactivelymanaginga
networkofpartners thatdeliver
specificskillsandcapabilities.
PROCESS:ENHANCING
OPERATIONALEFFICIENCY
Process inefficiencies, oftenoverlooked
in the race todeploynew technology,
can roban ITorganizationof itsagility,
diverthuman resources fromstrategic
pursuitsand introducevulnerabilities
that lead todowntime. Typically, these
inefficienciescanbe traced tooneof
fourcauses:
1.Overdependenceonmanual
processes
With themonitoring technologies
available today, itno longermakes
sense tohavedatacenterpersonnel
walking the floor tomonitorequipment
statusor take inventory.Datacenter
management teamshave toshed the
“cobbler’schildren”mentality inwhich
theyaresooccupiedsupporting
servicedelivery that they lag the restof
thebusiness inadopting technology to
automateprocesses.
2. Informationsilos
Whenoperatingdata is fragmented
across theorganization, personnel have
toeitherchasedown information from
multiplesourcesormakedecisions
withouta full understandingof the
impacton interdependent systems,
potentiallycreatingnewproblems to
beaddressed.
3. Insufficient information
What’sworse thanhaving tochasedata
isnothavinganyway toacquire it. This
forcesdecision-making tobe
conductedbasedon instinct rather
thandata.
4.Poorlydefinedprocesses
Insomecasesadequateprocesseshave
notbeendefinedordocumented. This
isparticularly true for tasks thatoccur
less frequently, suchascommissioning
or service.
Theseprocess inefficienciesmanifest
themselves ineveryphaseof thedata
center lifecycle, fromplanningand
commissioning toongoing
management.
PLANNINGANDCOMMISSIONING
Nearly70%ofearlyequipment failures
canbe traced todesign, installationor
startupdeficiencies. Inefficiencies in
theplanningphaseofanew facilityor
build-outnotonly increase the riskof
failurebutalsocreate futureproblems
whenadditional capacity is required.
Thesecan result from insufficient
information regarding theneedsof the
businessandbestpracticesand
technologiesavailable tomeet those
needs.Oneexample isplanningdata
center spacebeforedesigncriteria
havebeen finalized.Aproper
understandingofdatacenter
infrastructure technologiesand the
densities thatcansafelybeachieved
canhaveasignificant impactondata
centercostsandgrowthplans.Best
practice infrastructuredesigns thatcan
be tailored tobusiness requirements,
alongwithuseof standardized
technologies,mitigatemanyof the risks
associatedwithplanningdatacenter
infrastructure.
Commissioningcanalso introduce
problems. Impropercoordinationand
calibrationofprotectivedevices,wiring
errors, designerrorsandother issues
canall affectequipmentperformance.
Thesecangenerallybeavoided
througha thoroughandsystematic
commissioningprocess. This isa time
whenaprofessional service
organizationwith theability to test
systemcomponents togetherprior to
installationandconducton-site
inspectionand testingprior tostartup
candeliver significantvalue.
Commissioningandstartup testingalso
providesbaseline information thatcan
beused toevaluate futuremaintenance
decisions.
