Robustness study of MBAC algorithms

Measurement based admission control (MBAC) has long been recognized as an important method for providing stochastic service guarantees. The main work in the literature has mostly been on finding optimal MBAC algorithms that maximize network utilization. On the other hand, a critical issue for MBAC, which is the robustness of an MBAC algorithm in meeting the required quality of service (QoS) under changing traffic conditions, has long been overlooked and few results are available. In this paper, we investigate the robustness of MBAC in meeting the QoS target. Specifically we evaluate three representative MBAC algorithms and show their weaknesses and strengths through comparisons with an ideal admission controller.