An efficient frame-based scheduling algorithm: credit round robin

Abstract

An efficient frame-based scheduling algorithm called credit round robin is proposed. Performance analysis of credit round robin shows that it provides a finer granularity bandwidth with lower latency and better fairness compared to deficit round robin (Shreedhar, M. and Varghese, G., IEEE/ACM Trans. on Networking, vol.6, no.5, p.611-24, 1998). with the same complexity. The major difference between deficit round robin and credit round robin is the frame size. The frame size of deficit round robin is a fixed value and precalculated in advance using the granularity of the worst case packet length and the reserved rate for all sessions. However, the frame size of credit round robin is variable and calculated using the head of line (HOL) packet length of a servicing session at the current time and its reserved rate. Since the packet length of the real-time traffic is not always of the maximum length, the frame size of the credit round robin is usually smaller than that of the deficit round robin. As a result, the performance of credit round robin is superior to deficit round robin.