On live migration and routing integration for delay-sensitive cloud services in wireless mesh networks

Virtual machine live migration is a promising solution to provide low end-to-end delay for delay-sensitive applications hosted in data centers. This enables a cloud operator to minimize end-to-end delay by placing virtual machines to the nearest data centers from users. However, a network operator does routing with a different objective, e.g., minimize network resource usages. Since the both operators' objectives are contradictory to each other, we previously proposed to integrate live migration and routing in backbone networks to resolve the contradiction. However, the dominant factor of end-to-end delay for cloud services is a congestion delay at access networks, which was not taken into account in our previous work. In this paper, we apply the integration into wireless access networks prone to suffer from large end-to-end delay and validate its effectiveness in such networks by formulating it as a solvable mixed integer non-liner program. Our new integration model incorporates congestion delays due to wireless links and therefore enables cloud operators to maintain small end-to-end delays for cloud applications. The numerical results indicate that integration of live migration and routing is successful at providing low end-to-end delay to delay-sensitive applications by resolving the contradiction.