Performance Analysis of Routing Protocols On IPv4 and IPv6 Addressing Networks

Abstract

With IPv4 addresses being exhausted, network engineers and researchers are encouraged to adopt IPv6. But before using the IPv6 network directly, engineers need to test their hardware and network performance under new conditions of IPv6 as it has an extended address, high complexity, overhead performance, and IPsec complications. As routing protocols play a crucial role in network performance, it leads to a network’s extended performance by finding the shortest path, good throughput, and lowest delay. As the specifications, viz. frame structure for IPv4 and IPv6 are entirely different, there are modified routing protocols specified for IPv6. Routing Information Protocol (RIP) and Routing Information Protocol Next Generation (RIPng) are distance vector routing protocols and use hop count as a cost. In this paper, we have used RIPng on the IPv6 addressing network and RIP on the IPv4 addressing network and then analyze and compare them on the basis of different performance parameters. For this comparison, three different applications – File Transfer Protocol (FTP), DB Query (DataBase), and electronic mail (e-mail) – are set on a network consisting of three different subnets, each having a diverse network topology. The performance parameters analyzed are global and object statistics, viz. ethernet delay, number of hops, applications response time, background traffic delay, traffic dropped, point-to-point links throughput, links utilization, and links queuing delay. The experimental results determine the strength of the routing protocols. Thus, the quantitative results give the option to choose the routing protocol according to the network scenarios. In terms of ethernet delay, traffic dropped, network convergence, and security, it is found that the RIPng_IPv6 network performs better than RIP_IPv4. RIPng_IPv6 has an ethernet delay of 2.9 milliseconds, traffic dropped of 0.29 packets/second, and network convergence of 17 seconds less than RIP_IPv4 values. However, the RIP_IPv4 network is scalable, uses less hop, and has 40 milliseconds of traffic delay, while RIPng_IPv6 has 0.40 seconds of traffic delay. RIP_IPv4 also has a better response time for all three applications, FTP as 100 milliseconds, DB as 40 milliseconds, and e-mail as 20 milliseconds which is much less than the values obtained for RIPng_IPv6 network. Therefore, according to the performance requirements, the network engineers/operators or researchers can use either the existing IPv4 network or a new IPv6 network to achieve the Quality of Service (QoS) target level.