Optimized Cloudlet Management in Edge Computing Environment

As mobile applications are becoming computational intensive, edge computing supported by Cloudlets is a promising and ongoing solution to overcome the obstacles of mobile devices capabilities and remote Cloud access limitations. Bringing the Cloud, closer to mobile user, at the network’s edge, is a plausible key solution to tackle high network latencies in a Wireless Metropolitan Area Network (WMAN). The inclusion of Cloud features at the access network can be accomplished by collocating an Access Point (AP) with a respective Cloudlet. However, this functionality induces a cost of additional energy consumption at the edge of the network, which may be an important issue for massive deployments not only for operational expenses, but also for environmental reasons. Thus, designing such an architecture should take into consideration not only technical requirements to support user Quality of Experience (QoE), but also power consumption expenses constraints imposed by operators. In this paper, we assume an Edge Computing environment and propose latency and energy aware heuristic policies, such as Active Cloudlet Selection and User Assignment policies to activate Cloudlets and route mobile users for service to the most appropriate active Cloudlet respectively. The performance of the proposed algorithms is evaluated through simulation experiments and results show that the proposed schemes are very efficient in terms of energy consumption, system’s latencies and QoS.