Predicted Sum: A Robust Measurement-Based Admission Control with Online Traffic Prediction
Matyas Egyhazy, Yao Liang · IEEE Communications Letters · 2007
We present a new measurement-based admission control (MBAC) scheme with online traffic prediction, referred to as predicted sum (PS). A simple adaptive online traffic predictor based on normalized least mean squares (NLMS) algorithm is employed. Our scheme has two unique merits: (1) robustness to traffic characteristics and network environment, and (2) easy and accurate control for the provisioning of network utilization and quality-of-service. The simulation results demonstrate its significant performance enhancement over the well-known MBAC measured sum (MS) scheme