Research Article
A Conqueror Traffic Based Ranking Metrics Heuristic for Assigning Synthetically Generated Traffic Matrices to a Topology
@INPROCEEDINGS{10.1109/COMSWA.2007.382611, author={ Saman Taghavi Zargar and Mohammad Hossein Yaghmae}, title={A Conqueror Traffic Based Ranking Metrics Heuristic for Assigning Synthetically Generated Traffic Matrices to a Topology}, proceedings={2nd International IEEE Conference on Communication System Software and Middleware}, publisher={IEEE}, proceedings_a={COMSWARE}, year={2007}, month={7}, keywords={conqueror traffic based traffic matrices internet traffic matrices placement problem heuristics ranking metrics heuristic synthetic traffic matrices}, doi={10.1109/COMSWA.2007.382611} }- Saman Taghavi Zargar
Mohammad Hossein Yaghmae
Year: 2007
A Conqueror Traffic Based Ranking Metrics Heuristic for Assigning Synthetically Generated Traffic Matrices to a Topology
COMSWARE
IEEE
DOI: 10.1109/COMSWA.2007.382611
Abstract
In this paper we introduced the synthetically generated traffic assignment problem and we proposed conqueror traffic based ranking metrics heuristic method for solving this problem. We modified previous ranking metrics heuristic method and we tried to assign the synthetically generated traffic to a topology based on traffic type. In our work, transmission control protocol (TCP) and user datagram protocol (UDP), as two major traffic types of today's Internet, are considered to generate special traffic matrices for them. By generating special traffic matrices for each type of traffic, traffic matrices may better describe the network conditions and consider the traffic type constraints. We made a comparison between our proposed algorithm and the two previous algorithms, called Load Minimization Solution (LMS) and ranking metrics heuristic (RMH), using simulation and results are presented. Simulation results indicate that our proposed heuristic method performs better in all cases which conqueror traffic (TCP & UDP) exists and correctly identifies more flow rates than the two other algorithms. Moreover, results show that, it is better to use LMS in the case of no conqueror traffic exists. The total absolute error rates between the amount of originally assigned flow rates and the rates that the RMH, LMS, and our conqueror traffic based RMH are assigned is another comparison we made which, proves that our heuristic method assigns the rates much closer to the original rates.
