Optimising the multiplexing gain in Cloud RAN using the minimum spanning tree algorithm

Abstract

The cloud radio access network (Cloud-RAN) has been proposed as a cost efficient way of serving the increasing mobile traffic. The basic idea, supporting the cost efficiency, is the statistical multiplexing gain. Such a gain can be achieved when reusing the same processing power, realising BBU functions, for different, spatially distributed cells. Cells served by a single BBU pool need to be close to each other to support the mobility (handovers) and improve coordination. This article proposes the clustering algorithm, based on the minimum spanning tree (MSP). The algorithm allocates the geographically distributed sites onto a predefined set of BBU pools, keeping the clusters compact while achieving high multiplexing gain. We created, using the described algorithm, the BBU pools based on data traffic statistics from commercial network. We compared the multiplexing gain achieved with the maximum multiplexing gain achievable having given the daily profiles of analysed sites.