A Treap-Based Congestion Control Model for M/G/k Queue Network

Abstract

This study presents a Treap-based Congestion Control Model (Tb-CCM) for ensuring stability in M/G/k queue network. In this model, packets of varying sizes of data were transmitted from N sources over time into the network, consequently resulting in traffic. Since the model has one server at inception, it could not process more than one packet at a time. Packets arriving in the network whenever the server is busy were arranged as nodes in a tree. These packets arrive in a Poisson distribution process of α(k) = (λk/k!)e−λ and the general distribution process for packet’s service time is an arbitrary probability density function fy(γ). The model combines the features of max- heap and binary search trees which are manipulated to manage and transmit nodes to the server based on node rotations. The algorithm of the proposed model and that of Random Early Detection with Reconfigurable Maximum Dropping Probability (RRMDP) were implemented using Optimized Network Evaluation Tool (OPNET) 14.5 simulator. The performance of both methods in network throughput, latency, average queue size management and queuing delay were compared. Simulation results showed a considerable improvement in the performance of the proposed model in network throughput, latency, queuing delay as well as average queue size. Consequently, it was concluded that the proposed model is more effective in the management of M/G/k queue networks.