An analytical model of proportional bandwidth adjustment for a two-tier hierarchical NEMO system

This paper presents an analytical model of proportional bandwidth adjustment (PBA) in a two-tier hierarchical NEMO system, composed of an access router (AR), mobile routers (MRs), and visiting mobile nodes (MNs). Depending on the location, an MN may use direct link to AR (Tier-0) or indirect link to AR via MR (Tier-1). To increase system utilization, sessions generated by high-speed MNs in one tier are allowed to use a certain amount of free bandwidth in another tier. Since Tier-1 traffic can be temporarily buffered at MR, to reduce the blocking probability of Tier-0, the bandwidth between Tier-0 and Tier-1 can be proportionally adjusted based on link quality ratio. In other words, more bandwidth is allocated to the tier with higher link quality. For the purpose of validation, we build a mathematical model with 6-D Markov chains to analyze the PBA performance in terms of session blocking probabilities and normalized goodput. Analytical results demonstrate that the proposed PBA can significantly reduce the session blocking probabilities and increase the overall normalized goodput when session arrival rates and link quality between Tier-0 and Tier-1 are very different.