Analyzing a distributed slot assignment protocol by Markov chains

Packet radio networks must be able to handle rapid changes in connectivity, e.g. for a continuous data exchange in intercar communication. In a concurrent slot assignment protocol this is achieved by random access to time synchronized frames and exchange of individual channel observation by each of the involved stations. An important question is how fast this synchronization tool works. The authors investigate its performance by analyzing the (random) first stable time in a Markov chain model. Numerical results for a variety of parameters are given.<<ETX>>