QoS-Aware I/O Schedule for Virtual Machines in Cloud Computing Environment

Abstract

In the virtualization based cloud computing environment, multiple computers are allowed to run as virtual machines (VM) in a single physical computer. Since multiple VMs share the same network interface card (NIC). The I/O operation of one virtual machine is affected by others. The quality of service (QoS) of the network is hardly guaranteed. This paper presents credit based fairing scheduling algorithm for I/O virtualization (VCF), an efficient scheduling strategy with bandwidth guarantee that is implemented in the xen hyper visor layer. Our proposed scheme uses a credit based strategy to guarantee the reserved bandwidth of each VM. At the same time, VCF is aware of bandwidth requirement inside a VM using monitoring and prediction strategy. Thereby, VCF can also allocate the bandwidth according to the actual requirement of the VM, which can improve the bandwidth utilization. The proposed techniques are implemented on the xen virtual machine monitor and evaluated with micro-benchmarks on Linux operation system.