Data parallel applications are being extensively deployed in cloud environmentsbecause of the possibility of dynamically provisioning storage and computation re-sources. To identify cost-effective solutions that satisfy the desired service levels,resource provisioning and scheduling play a critical role. Nevertheless, the unpre-dictable behavior of cloud performance makes the estimation of the resources actu-ally needed quite complex. In this paper we propose a provisioning and schedulingframework that explicitly tackles uncertainties and performance variability of thecloud infrastructure and of the workload. This framework allows cloud users to es-timate in advance, i.e., prior to the actual execution of the applications, the resourcesettings that cope with uncertainty. We formulate an optimization problem wherethe characteristics not perfectly known or affected by uncertain phenomena arerepresented as random variables modeled by the corresponding probability distri-butions. Provisioning and scheduling decisions – while optimizing various metrics,such as monetary leasing costs of cloud resources and application execution time –take fully account of uncertainties encountered in cloud environments. To test our framework, we consider data parallel applications characterized by a deadline con-straint and we investigate the impact of their characteristics and of the variabilityof the cloud infrastructure. The experiments show that the resource provisioningand scheduling plans identified by our approach nicely cope with uncertainties andensure that the application deadline is satisfied.
- Cloud computing
- Resource provisioning