Delay analysis of a newly proposed tree based collision resolution algorithm with known multiplicity feedback

Abstract

This thesis investigates the performance of two frame-based MAC protocols namely framed slotted Aloha and the tree algorithms with different types of feedback information; binary, ternary, and known multiplicity. Four fundamental mechanisms for resolving collision are introduced as basic building blocks for the construction of a wide range of random access MAC protocols. The proposed analytical evaluation has shown that the use of feedback information, if used efficiently, plays a vital role in delay performance improvement. The achievable delay performance is shown to be highly dependent upon how the feedback information is used in the contention resolution. For known multiplicity, the maximum achievable MST of 0.533 is obtained by our proposed random access protocol that is derived by the combination of splitting mechanism, adaptive frame size, slot-skipping type II, and non-uniform access probability. We also present a preliminary study of a generic model that serves multi-class nodes with different quality of service requirements using slotted aloha, designed specifically for reservation-based MAC protocol with slotted Aloha that allows us to develop a variety of prioritization schemes, whereby nodes can be prioritized through reservation success rates, while aiming to maintain high efficiency of channel utilization.