A finite user model for stability analysis of Announced Retransmission Random Access (ARRA) systems is presented.This analytical approach augments earlier work on ARRA in which an infinite population Poisson model was used to demonstrate capacities of 0.53 and 0.6 for Basic and Extended ARRA respectively. Using the finite user model described here,the stability properties of ARRA systems are investigated.Considering first fixed (nonadaptive) retransmission policies,it is demonstrated that like ALOHA and CSMA,ARRA systems can exhibit bistable behavior and saturation.The model is useful for determining the maximum permissible retransmission probability for stable operation. The higher expected delay associated with an unconditionally stable nonadaptive system can be avoided by adaptively varying retransmission parameters with system backlog. Optimal retransmission control (ORC) policies of this type are derived and evaluated.The results confirm the viability of stable,high capacity ARRA systems as an attractivternative to other comparable random access protocols such as slotted ALOHA and Tree algorithm CRAs.
展开▼
