Mobile Services Computation Offloading

Aiming to reduce task transmission latency, migration latency, and execution latency, new techniques are emerging to optimize applications, servers, and users jointly during tasks that are offloading to an edge server. In this chapter, we focus on several key issues about the offloading problem. Taking into account the unstable connectivity of mobile networks, we introduce a mobility model and a trade-off fault-tolerance strategy. This strategy is designed by a modified genetic algorithm to match the structure of this model. Besides, considering the partition of data stream application and edge servers cooperation, we discuss the offloading strategy with resource-limited mobile devices and computation-intensive services. A cross-edge computation offloading (CCO) framework for partitionable applications is proposed by the Lyapunov optimization. Edge servers’ clustering is also a major concern in the edge network environment to significantly improve the scheduling of network resources and satisfy the requirements of service subscribers. We designed a system that benefits from clustering edge servers based on the density of devices that connect to edge servers.