Backpressure routing for the backhaul in sparse small cell deployments

This paper presents Backpressure for sparse small cell deployments (BS), a transport network layer solution aimed to carry backhaul traffic under sparse topologies. BS leverages queue backlog and geographic information to take forwarding decisions on a per-packet basis, without incurring into excessive path lengths and queueing latencies. Simulations were carried out in ns-3 under diverse traffic loads. Our scheme shows similar performance to that of an idealized zero-overhead shortest path routing protocol (ISPA) under light traffic conditions, while outperforming a state of the art geographic routing scheme (i.e., GPSR). Under more severe traffic conditions, BS shows better performance than ISPA and GPSR. In particular, BS attains a reduction of up to 85% in latency regarding GPSR. Overall, this shows how a (quasi-)stateless low overhead scheme can scale to be able to serve the realistic backhaul topologies of interest in future massive small cell deployments.