On planning small cell backhaul networks

Abstract

Small cells are expected to play an important role in future mobile networks to meet the rapidly increasing capacity demand. However, this requires that cost efficient backhaul solutions are found. This is a very challenging task given the large number of small cells that must be connected. In this paper we propose a set of practical methods that mobile operators can use as tools when planning small cell backhaul networks. The tools can help operators find the best way to connect the base station nodes and which technology to use on each link. Two methods are proposed for planning fiber-only networks. Fiber only solutions are interesting since long term backhaul solutions are expected to be dominated by fiber links and can be used to find heuristic solutions to multi-technology problems by e.g. replacing some fiber links with radio links. Optimal fiber-only networks with a tree topology can be easily found by a minimum weight spanning tree algorithm. By introducing a new metric we show that more resilient ring topology based solutions can be found by solving a traveling salesman type of problem. For heterogeneous multi-technology backhaul networks we propose a Mixed Integer linear Programming (MIP) problem formulation that allows inclusion of a wide range of requirements and constraints. The solutions can be made robust against varying traffic conditions by specifying capacity demand for different scenarios and both point-to-point and point-to-multipoint type of technologies can be included in the optimization. To verify the methods an example case study is performed for a small city.