CRoSS: A Combined Routing and Surrogate Selection Algorithm for Pervasive Service Offloading in Mobile Ad Hoc Environments

With the advantage of mobile devices and wireless communication technologies, service offloading, as an important branch of pervasive computing, attracts much research. In service offloading, by offloading part of computation in the mobile device to a powerful surrogate, the total service execution time can be reduced and mobile devices' battery lifetime can be greatly prolonged. In this paper, we focus on mobile ad hoc environments. Given a service's remote execution components, our goal is to find a surrogate and the path to the surrogate (i.e., the surrogate-path pair) so as to achieve a minimum surrogate execution time. Since the topology in mobile ad hoc environments can change frequently due to node mobility, we study the problem of combined routing and surrogate selection. The task is formulated as a combinatorial optimization problem and a combined routing and surrogate selection algorithm (CRoSS) is proposed. Based on cross-layer design principles, CRoSS enables the process of surrogate discovery and selection to be coalesced into the process of route discovery and maintenance of the underlying mobile ad hoc networks. It is analytically proofed that by utilizing the surrogate-path pair found by CRoSS the surrogate execution time is minimized. Simulation evaluations also show the constant efficiency of CRoSS.