MFLWQ: A fair and adaptive queue management algorithm for scavenger service

Abstract

This paper presents a MFLWQ Algorithm for Scavenger Service, which is a non-elevated QoS technique proposed by the Internet2 project. The algorithm balances the bandwidth distribution among Scavenger Service flows by a modified Flow Random Early Detection (FRED) algorithm; estimates active Scavenger Service flow number by the variable nactive in FRED and accordingly tunes the bandwidth allocation between SS flows and BE flows in a logarithmical way to reach better performance and robustness. Simulations show that the algorithm provides more appropriate bottom bandwidth guarantee for Scavenger Service flows when protecting Best-Effort flows well and extends the applicability of Scavenger service to no-adaptive traffic like UDP. © Springer-Verlag Berlin Heidelberg 2005.