PROJECT TITLE :

Backpressure Delay Enhancement for Encounter-Based Mobile Networks While Sustaining Throughput Optimality

ABSTRACT:

Backpressure routing, in that packets are preferentially transmitted over links with high queue differentials, offers the promise of throughput-optimal operation for a wide selection of communication networks. But, when traffic load is low, backpressure ways suffer from long delays. This is of explicit concern in intermittent encounter-based mobile networks that are already delay-restricted due to the sparse and highly dynamic network connectivity. Whereas state-of-the-art mechanisms for such networks have proposed the employment of redundant transmissions to enhance delay, they do not work well when traffic load is high. In this paper we tend to propose backpressure with adaptive redundancy (BWAR), a novel hybrid approach that provides the most effective of each worlds. This approach is strong, distributed, and will not need any previous knowledge of network load conditions. We also present variants of BWAR that remove redundant packets via a timeout mechanism, which improve energy use. These algorithms are evaluated by mathematical analysis and by simulations of real traces of taxis in Beijing, China. The simulations ensure that BWAR outperforms ancient backpressure at low load, while outperforming encounter-routing schemes (Spray and Wait and Spray and Focus) at high load.