Content-Centric Transmission Design in Fog Radio Access Network With Partition-Based Caching

This paper presents a joint optimization of caching and multicast beamforming for a content-centric transmission in a fog radio access network (F-RAN). Users requesting the same file are grouped together and served using multicast transmission. Each edge node (EN) has a local cache and can pre-store contents via partition-based caching. To improve cache performance and reduce fronthaul cost, the maximum distance separable (MDS) code is applied to the file coding. We consider the mixed timescale joint optimization. In short-term optimization, our goal is to optimize the beamforming vector to minimize the total network consumption under the quality-of-service (QoS) constraint. Meanwhile, semi-definite relaxation (SDR) and convex-concave procedure (CCP) algorithms are introduced to solve short-term problem. In long-term optimization, based on the channel state information (CSI) and user requests, we formulate an integer programming problem of minimizing the expectation of the total network consumption under the constraint of cache capacity and the parameter of MDS code. Furthermore, the considered long-term problem is decomposed and a heuristic method is proposed by replacing the objective function with the expectation of the file splitting number. After transforming the integer variables, the subproblem is transformed into the difference of convex (DC) programs and effectively solved using the CCP algorithm. Numerical results show that the proposed content-centric transmission can strike a better balance between power consumption and fronthaul consumption than existing schemes.