This paper evaluates the architecture of a wireless mesh network. This paper mainly evaluates the network with unplanned node placement and multi hop routing compared to a planned single hop network. The network implemented is the Roofnet 802.11b mesh network. The paper first explains the architecture and routing protocol of the Roofnet mesh network, which was tested in the MIT campus. The nodes were distributed in a unplanned but not completely random fashion. The routing protocol used was Srcr, which tries to find the highest throughput route between any pair of Roofnet nodes.
The evaluation method measures several different parameters to observe the effects of different values on the throughput of the network. From the results, we can see that the less hops means higher throughput. They also take into account the effect of link quality and distance. The observation was that as distance incrase, the throughput decreases, which makes complete sense because link quality should degrade and signal strength also degrades with distance. As density increases, the throughput also increases. This is a combined effect of lower hop count, but also interference. However, it offers a wider choice of high quality links.
The paper also compares multi-hop to a single hop network, which shows that multi-hop performs better, however the assumption is that the gateways are set and all connections are forced to operate in single hop. One could imagine a combination of planned and random multi-hop access to obtain the best performance results.
Subscribe to:
Post Comments (Atom)
No comments:
Post a Comment