Opportunistic Peer-to-Peer Mobile Cloud Computing at the Tactical Edge

Mobile computing capabilities of the hand-held mobile devices are important for war fighters at the tactical edge to efficiently process their in-situ sensory data about the surrounding environment, so as to maintain situational awareness. Mobile cloud computing is generally applied to augment such capabilities of mobile devices, but has been limited to the interaction between the back-end cloud infrastructure, which availability, however, could be limited due to the dynamic network characteristics at the tactical edge. In this paper, we develop a novel theoretical framework to exploit the potential of peer mobile devices on mobile cloud computing, when they opportunistically move into the communication range of each other. Our proposed framework aims to adaptively configure the computational workload allocations among mobile devices, by considering both the energy consumption and timeliness of computational task execution. Probabilistic methods are adopted to quantitatively estimate the opportunistic network transmission delay and ensure that the computational results could be delivered back to the task initiator on time. Extensive trace-driven simulation results show that our proposed framework could significantly improve the mobile cloud computing performance, while reducing the energy consumption of such computing, compared to the existing schemes.