The increasing popularity of cloud computing applications and the advances in virtualization software technologies (i.e., Virtual Machine deployment) have driven Data Center infrastructures toward a greater complexity and workload dynamicity. Thus, a unified control and management of computing and network resources would be required for assuring proper traffic performances in high volatile virtual machine deployments. This paper introduces a novel resource control platform for virtualized DC environments aimed at optimizing virtual machine placement on physical servers also considering traffic load across links in order to limit oversubscription-related problems. To this purpose OpenFlow statistics are elaborated for their distinguished features that well suit virtualized environments. Two novel algorithms have been conceived and compared by simulations for evaluating the effectiveness of proposed traffic-aware VM placement strategies.
Effective Resource Control Strategies using OpenFlow in Cloud Data Center
Martini B;
2013-01-01
Abstract
The increasing popularity of cloud computing applications and the advances in virtualization software technologies (i.e., Virtual Machine deployment) have driven Data Center infrastructures toward a greater complexity and workload dynamicity. Thus, a unified control and management of computing and network resources would be required for assuring proper traffic performances in high volatile virtual machine deployments. This paper introduces a novel resource control platform for virtualized DC environments aimed at optimizing virtual machine placement on physical servers also considering traffic load across links in order to limit oversubscription-related problems. To this purpose OpenFlow statistics are elaborated for their distinguished features that well suit virtualized environments. Two novel algorithms have been conceived and compared by simulations for evaluating the effectiveness of proposed traffic-aware VM placement strategies.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
