Cloud Radio Access Networks with Coded Caching (1512.02385v2)

Abstract: A cloud radio access network (C-RAN) is considered as a candidate to meet the expectations of higher data rate de- mands in wireless networks. In C-RAN, low energy base stations (BSs) are deployed over a small geography and are allowed to connect to the cloud via finite capacity backhaul links where the information is processed. A conventional C-RAN, however, requires high capacity backhaul links, since the requested files need to be transferred first from the cloud to the BS before conveying them to the users. One approach to overcome the limitations of the backhaul links is to introduce local storage caches at the BSs, in which the popular files are stored locally in order to reduce the load of the backhaul links. Furthermore, we utilize coded caching with the goal to minimize the total network cost, i.e., the transmit power and the cost associated with the backhaul links. The initial formulation of the optimization problem for this model is non-convex. We first reformulate and then convexify the problem through some relaxation techniques. In comparison to the uncoded caching at the BSs, our results highlight the benefits associated with coded caching and show that it decreases the backhaul cost.