Capacity analysis for heterogeneous wireless mesh networks with Queueing Network Models

Gupta and Kumar [1] have shown that the throughput capacity which can be achieved is of order Θ(1/√(nlogn)) for wireless networks. In subsequent years, P.Li and Y.G. Fang thoroughly studied the same issue for both Regular and Random heterogeneous wireless networks [2]. In this paper, we derive the packet absorption probability of the multihop Regular Heterogeneous Wireless Networks (RHWN) based on [2]. With the derived absorption probability, we are able to obtain the achievable per-node throughput capacity and the end-to-end delay as well by applying the diffusion approximation method to the Queueing Network Model [3-4].We find that traffic pattern and network topology have great impact on end-to-end delay and throughput performance for RHWN. However, the most significant finding is the optimum number of helping nodes for RHWN to handle a very large number of user nodes. Certainly, such finding allows not only optimal use of existing resources, but also reduction of congestion in the network system.