Dynamic Bandwidth Allocation for OFDMA-PONs Using Hidden Markov Model

Accurate prediction of traffic conditions on orthogonal frequency division multiple access passive optical networks is important because of its vital role in network resource management and efficient bandwidth allocation. Given the dynamic and stochastic nature of network traffic, our proposed algorithm conducts a probabilistic approach by using the hidden Markov model (HMM). The HMM defines traffic states with two parameters: the mean and contrast of the bandwidth request observations. Simulation results demonstrate the performance comparison between with and without the prediction method in terms of throughput and end-to-end delay. As a result, the throughput improves 15% and the saturation offered load of the delay for the prediction and non-prediction is 0.8 and 0.7, respectively.