Towards a network aware VM migration: Evaluating the cost of VM migration in cloud data centers

Host virtualization allows data centers to live migrate an entire Virtual Machine (VM) to support data center maintenance and workload balancing. Live VM Migration can consume nearly the entire bandwidth which impacts the performance of competing flows in the network. Knowing the cost of VM Migration propels data center admins to intelligently reserve minimum bandwidth required to ensure a network-aware VM migration. Recently, Remedy was proposed as a cost estimation model to calculate total traffic generated due to VM Migration. Unlike the previous approaches, Remedy makes it possible to incorporate network topology leading to a more intelligent allocation of network resources during VM migration. However, Remedy was evaluated within a simulated environment running on a single machine. In this paper, we empirically evaluate the performance of Remedy in an experimental GENI testbed characterized by wide-area network dynamics and realistic traffic scenarios.We deploy OpenFlow end to end QoS policies to reserve minimum bandwidths required for successful VM Migration. Preliminary results demonstrate that bandwidth reservation relieves the network of possible overloads during migration. We show that Remedy works best with link bandwidths of 1Gbps and above and pages dirty rates below 3000 pages/s. We present realistic scenarios that affect the accuracy of the cost estimation model. We conclude that link bandwidth, page dirty rate and user specified progress amount are the critical parameters in determining VM migration cost.