Secure Mobile Edge Computing via a Drone-Enabled FSO-Based Heterogeneous Network

A secure mobile edge computing (MEC) framework, which leverages free space optics (FSO) to provision backhauling from drone-mounted base stations (DBSs) to a macro base station (MBS), is proposed. DBSs can work as computing nodes to provide computing services to user equipments (UEs) as well as relays to forward the tasks to the MBS at the same time. Both DBSs and MBS are provisioned with servers. UEs can offload their tasks to the MBS directly, DBSs directly, or the MBS via a DBS. The DBSs are to be placed at the optimal locations to provide computing services and uplink communications for the ground UEs. We then formulate the joinT bandwidth and computation rEsource assignment, DBS traNsmisSion power contrOl, UE power contRol, UE association and DBS placement (TENSOR) problem to jointly maximize the average secrecy rate and minimize the task completion time. Since TENSOR is a mixed-integer nonlinear problem, we decompose it into four sub-problems: resource (bandwidth and computation) assignment and DBS transmission power control problem, UE power control problem, UE association problem, and DBS placement problem. Successive convex approximation and an iterative algorithm are leveraged to solve the four sub-problems separately. In each iteration, we use the output of the first three as the input for the fourth. The performance of our algorithm is superior to the greedy algorithm, block search placement algorithm, and equally shared resource allocation algorithm upon which the average secrecy rate is improved by at least 19%.