Relocation analysis of stabilizing mac algorithms for large-scale mobile ad hoc networks (Extended abstract)

Throughput is a basic measure for communication efficiency. It is defined as the average fraction of time that the channel is employed for useful data propagation. This work considers the problem of analytically estimating the throughput of protocols for media access control (MAC) in mobile ad hoc networks (MANETs). The dynamic and difficult to predict nature of MANETs complicates the analysis of MAC algorithms. We use simple extensions to the interleaving model and evolving graphs, for defining the settings that model the location of mobile nodes. This work improves the understanding on impossibility results, the possible trade-offs and the analysis of fault-tolerant algorithms in MANETs. As the first result in the paper and as motivation for the ones that follow, we show that there is no efficient deterministic MAC algorithm for MANETs. Moreover, we prove a lower bound of the throughput in the radical settings of complete random relocation between every two steps of the algorithm. The lower bound matches the throughput of a strategy that is oblivious to the history of wireless broadcasts. Subsequently, we focus on the analysis of non-oblivious strategies and assume a bound on the rate by which mobile nodes relocate, i.e., randomly changing their neighborhoods. Our analysis is the first to demonstrate a novel throughput-related trade-off between oblivious and non-oblivious strategies of MAC algorithms that depends on the relocation rate of mobile nodes. We present a non-oblivious strategy that yields a randomized, fault-tolerant algorithm that can balance between the trade-offs. The studied algorithm is the first of is kind because it is a "stateful" one that quickly converges to a guaranteed throughput.