http://lattes.cnpq.br/4195265595859937; FERREIRA, Iury Gregory Melo.
Resumo:
In CloudComputingenvironments,especiallythoseusingtheinfrastructureasaservice model, theelasticitycharacteristicinresourceprovisioningcomeswiththeneedtomanage resources sothequalityofservicecancontinuetobeguaranteedtousersandalsoto maintain agoodperformanceoftheinfrastructure.Thisworkproposesheuristicsthat are abletoassistintheloadbalancingoftheserversinaCloudinfrastructure,proposing migrations toreducetheoverheadintheserversthatwereidentifiedasoverloaded,since with thepassageoftimethereisanaturalvariationintheamountofresourcesinuse.This variationinaconsequenceofremovaloradditionofapplicationsandevenoftheusageof verticalscalingtoimproveapplication’sperformance.Atoolwasimplementedtomake
use oftheheuristicalgorithmsandthustoaidintheexperimentsandtheirvalidation,the
metrics usedcomedirectlyfromheterogeneousserversoftheOpenStackCloudofthe
DistributedSystemsLaboratory.TheresultsshowthatinadditiontothedecreaseinCPU
consumption ofserversthatwereoverloaded,itisalsopossibletoimprovetheperformance
of theseserversinsomecases.