Leveraging Active-Guided Evolutionary Games for Adaptive and Stable Deployment of DVFS-Aware Cloud Applications

This paper proposes and evaluates a multiobjective evolutionary game theoretic framework for adaptive and stable application deployment in clouds that support dynamic voltage and frequency scaling (DVFS) for CPUs. The proposed algorithm, called AGEGT, aids cloud operators to adapt the resource allocation to applications and their locations according to the operational conditions in a cloud (e.g., workload and resource availability) with respect to multiple conflicting objectives such as response time, resource utilization and power consumption. In AGEGT, evolutionary multiobjective games are performed on application deployment strategies (i.e., solution candidates) with an aid of guided local search. AGEGT theoretically guarantees that each application performs an evolutionarily stable deployment strategy, which is an equilibrium solution under given operational conditions. Simulation results verify this theoretical analysis; applications seek equilibria to perform adaptive and evolutionarily stable deployment strategies. AGEGT allows applications to successfully leverage DVFS to balance their response