An efficient pub/sub protocol for WMNs based on Heuristic Markov Chain

We propose a novel Pub/sub protocol for Wireless Mesh Networks based on Heuristic Markov Chain, namely HMCPS, which can reliably and efficiently deliver events to mobile clients in a continuous manner. Using Heuristic Markov Chain, HMCPS predicts hot locations where a mobile client may stay with high probability, and then delivers interesting events to the hot locations in an efficient way. We also show that the delivering problem with minimum message cost is a NP-complete problem, namely the hot-location event delivery (HLD) problem. We have proposed one algorithm to obtain the spanning tree which minimizes the cost of message delivery in the HLD problem. The extensive simulation results show that HMCPS may adopt appropriate parameters to guarantee the reliability of event delivery with low cost. Compared with the pub/sub protocol using the paging algorithm, HMCPS can save the system message cost up to 80%.