Modeling and Trade-off for Mobile Communication, Computing and Caching Networks

TL;DR

This paper investigates how caching and computing at mobile devices in a multi-user MEC system can be jointly optimized to reduce wireless bandwidth consumption under latency constraints, using advanced optimization techniques.

Contribution

It formulates a joint caching and computing policy optimization problem, proves its NP-hardness, and proposes an effective solution with analytical insights into resource impacts.

Findings

Joint caching and computing reduce bandwidth significantly.

Optimal policies depend on resource availability and system parameters.

Analytical results reveal resource influence on bandwidth savings.

Abstract

Computation task service delivery in a computing-enabled and caching-aided multi-user mobile edge computing (MEC) system is studied in this paper, where a MEC server can deliver the input or output datas of tasks to mobile devices over a wireless multicast channel. The computing-enabled and caching-aided mobile devices are able to store the input or output datas of some tasks, and also compute some tasks locally, reducing the wireless bandwidth consumption. The corresponding framework of this system is established, and under the latency constraint, we jointly optimize the caching and computing policy at mobile devices to minimize the required transmission bandwidth. The joint policy optimization problem is shown to be NP-hard, and based on equivalent transformation and exact penalization of the problem, a stationary point is obtained via concave convex procedure (CCCP). Moreover, in a symmetric scenario, gains offered by this approach are derived to analytically understand the influences of caching and computing resources at mobile devices, multicast transmission, the number of mobile devices, as well as the number of tasks on the transmission bandwidth. Our results indicate that exploiting the computing and caching resources at mobile devices can provide significant bandwidth savings.