Adaptive Batch Resolution Algorithm with Deferred Feedback for practical CSMA Wireless Networks

A batch is a group of nodes that have to transmit a single packet each to a common receiver in the shortest time. Most of existing batch resolution algorithms assume immediate feedback and generally neglect the feedback time, being considered much shorter than the packet transmission time. This conjecture, however, fails to apply in many practical high-rate wireless systems, with the consequence that the classical performance analysis of batch resolution algorithms may result overoptimistic. In this report we propose and analyze a batch resolution algorithm for CSMA wireless networks that waives the immediate feedback approach in favor of a deferred feedback method, which shall reduce the overhead costs. The scheme, named Adaptive Batch Resolution Algorithm with Deferred Feedback (ABRADE), is obtained by merging a batch size estimate module with a framed ALOHA access scheme, whose frame length is dynamically adapted to the residual batch size in order to minimize the overall batch resolution time. ABRADE’s adaptation strategy is designed under the assumption of known batch size. To remove this constraint, we propose a batch size estimation module that, coupled with ABRADE, gives rise to ABRADE, a complete batch resolution algorithm that works even with no a priori information on the batch size. We then extend ABRADE with a batch size estimation module, thus obtaining a complete batch resolution algorithm, called ABRADE, that works even with no a priori information on the batch size. ABRADE is compared against the batch resolution algorithms based on the immediate feedback paradigm. Results confirm that, in practical CSMA systems, ABRADE outperforms the algorithms based on the immediate feedback paradigm, both in case of partial and no a priori knowledge of the batch multiplicity.