Avoiding Head-of-Line Blocking using an Enhanced Fairness Algorithm in a Resilient Packet Ring

IEEE is currently conducting an effort to standardize a full duplex, spatial reuse, ring network architecture, called the Resilient Packet Ring (IEEE P802.17). We discuss the relationship between spatial reuse and head of line blocking in such a ring. An algorithm that schedules packets individually based on destination address is outlined. We have written a model of the RPR architecture in the programming language Java, and simulated several scenarios in order to evaluate and compare our algorithm with a conventional ring fairness algorithm. Our performance evaluations show that it is possible to achieve significant higher throughput for traffic scenarios where head of line blocking traditionally has degraded performance. Where head of line blocking is not an issue, our algorithm behaves like the traditional one.