Optimal buffer partitioning for multiclass Markovian traffic sources

In this paper, we propose an algorithm for optimal buffer partitioning which requires the minimal channel capacity to satisfy the quality of service (QoS) requirements of input traffic. The traffic is generated by a Markov-modulated fluid source and has J (larger than 1) QoS classes at each state. The QoS is described by a packet loss probability requirement for each class. The buffer is partitioned with J-1 thresholds to provide the J loss priorities, thus the J classes of service. Traffic is admitted or rejected based on the buffer occupancy and its service class. We also present an approach for the buffer partitioning for heterogeneous Markov-modulated sources. Numerical results demonstrate that the proposed algorithm achieves a higher resource utilization efficiency than previously published results