Adaptive live VM migration over a WAN: modeling and implementation

Abstract

Recent advances in virtualization technology enable high mobility of virtual machines and resource provisioning at the data-center level. To streamline the migration process, various migration strategies have been proposed for VM live migration over a local-area network (LAN). The most common solution uses memory pre-copying and assumes the storage is shared on the LAN. While applied to a wide-area network (WAN), the VM live migration algorithms need a new design philosophy to address the challenges of long latency, limited bandwidth, unstable network conditions and the movement of storage. This paper proposes a three-phase fractional hybrid pre-copy and post-copy solution for both memory and storage to achieve highly adaptive migration over a WAN. In this hybrid solution, we selectively migrate an important fraction of memory and storage in the pre-copy and freeze-and-copy phase, while the rest (non-critical data set) is migrated during post-copying. We propose a new metric called performance restoration agility, which considers both the downtime and the VM speed degradation during the post-copy phase, to evaluate the migration process. We also develop a profiling framework and a novel probabilistic prediction model to adaptively find a predictably optimal combination of the memory and storage fractions to migrate. This model-based hybrid solution is implemented on Xen and evaluated in an emulated WAN environment. Experimental results show that our solution wins over all others in adaptiveness for various applications over a WAN, while retaining the responsiveness of post-copy algorithms.