In Carmesh, we are looking at different networking solutions for delivering services to cars, mainly in an urban enviroment. One connectivity solution is wireless mesh networks. One critical aspect of Wireless Mesh Networks (as well as other networks) is routing. That is the focus of this post.
A simple google search returns almost 60.000 scientific papers treating this aspect of Wireless Mesh Networks alone. Clearly, it is a hot topic which stimulates new ideas and activity.
Most of the work done in this area aims to improve the overall network capacity.
A simple taxonomy would split the routing mechanisms in two: proactive and reactive routing protocols. Reactive (on-demand) routing protocols include AODV, DSR, and proactive routing protocols include OLSR.
Proactive routing protocols are characterised by the fact that each node maintains local information in their routing table information on how to reach any other node in the network. To cope with a possible change in the network topology, nodes update their tables periodically.
Reactive routing, on the other hand, searches for a route to a destination node only when needed. Hence, the network is flooded with packets only when a source node has data to send. This type of routing works well in the case of ad-hoc networks, where mesh nodes are mobile and hence the network link topology changes quite dynamically. Consequently, it would be useless for each node to keep a local view of the network.
In Carmesh, we envision a network with static mesh nodes providing connectivity to mobile clients: proactive routing protocols are most suitable. In this type of network, links created between mesh nodes have a longer life. In this way, whichever way clients are roaming in the city and re-connecting to various access points, there will always be a link between the source of the data and the access point they are connected to. In the context of Carmesh, we are working on evaluating the performance of these routing protocols for the case of vehicular clients.