Optimal Choice of Random Variables in D-ITG Traffic Generating Tool using Evolutionary Algorithms

M Mosavi, F Farabi, S Karami
2015 Iranian Journal of Electrical & Electronic Engineering   unpublished
Impressive development of computer networks has been required precise evaluation of efficiency of these networks for users and especially internet service providers. Considering the extent of these networks, there are numerous factors affecting their performance and thoroughly investigation of these networks needs evaluation of the effective parameters by using suitable tools. There are several tools to measure network's performance which evaluate and analyze the parameters affecting the
more » ... ffecting the performance of the network. D-ITG traffic generator and measuring tool is one of the efficient tools in this field with significant advantages over other tools. One of D-ITG drawbacks is the need to determine input parameters by user in which the procedure of determining the input variables would have an important role on the results. So, introducing an automatic method to determine the input parameters considering the characteristics of the network to be tested would be a great improvement in the application of this tool. In this paper, an efficient method has been proposed to determine optimal input variables applying evolutionary algorithms. Then, automatic D-ITG tool operation would be studied. The results indicate that these algorithms effectively determine the optimal input variables which significantly improve the D-ITG application as the time cost of determining optimal DITG variables in automatic GA, ICA and ACO based methods has been improved up to 67.3 %, 69.7 % and 82.2 %, respectively. 1 Introduction1 Recently, the computer networks have been extremely complicated in software, protocols, equipment and etc [1] and lots of efforts have been made to understand the behavior of these networks [2]. There are many factors affect the network performance, including network congestion and load, network equipment and hardware, the amount of network users, the power of wireless network signal and the operating system used by users [3, 4]. In order to evaluate the networks performance there are several tools such as DSLProbe [5], Asymprobe [6], Spruce [7], Pathload [8], Netalyzy [9], Iperf, Netperf, IPtraffic, MGEN and D-ITG which measuring the parameters affecting the network performance. D-ITG has significant advantages over other traffic generators. D-ITG supports IPv4 and IPv6