In ATM networks, the available bit rate (ABR) service enables data sources to efficiently utilize network resources without adversely affecting the performance of non-ABR, guaranteed traffic such as video. Most research to date on ABR performance has not used realistic assumptions. Unrealistic assumptions such as persistent data sources, an absence of non-ABR traffic, and inaccurate TCP models may lead to misleading results. We address the above shortcomings of previous research. In a comparison of two major ABR schemes, the rate based Explicit Rate Indication for Congestion Avoidance (ERICA) and the credit based Quantum Flow Control (QFC), we find that the performance may be significantly better with QFC than previously thought when bursty traffic sources and bursty interfering traffic is considered.
展开▼
