Abstract:
Internet system is a collection of hosts, connected to each other. Communication among the hosts needs to have a protocol, which can be understood by each host in the Internet. IPv4 and the later version IPv6 protocols are designed for this purpose. Though IPv6 has begun to be deployed, IPv4 is still very widely used and it is impossible to expect a complete switch to IPv6 in the near future. It is therefore expected that IPv4 and IPv6 will co-exist for quite some time to come. Hence, an IPv4 host may need to communicate with an IPv6 host and vice-versa. This will necessitate translating an IPv6 packet to an IPv4 packet and vice-versa at network boundaries. The router sitting at the end of the network does translation of an IPv4 packet to an IPv6 packet and vice-versa at the network boundaries. We first designed and developed a mechanism in which a router whenever receives a packet from the IPv4 subnet intended for the IPv6 subnet, the router will translate the packet. The same thing happens when the router is forwarding a packet from the IPv6 network to the IPv4 network. To translate a packet from IPv4 to IPv6, each field in the IPv4 header is converted to a functionally equivalent field in the IPv6 header following the conversion scheme designed in the present work. Since ICMP is considered an integral part of IP, a mechanism for converting an ICMPv4 packet to an ICMPv6 packet and vice-versa has also been developed and implemented. This mechanism has been deployed on a Linux-based router and tested rigorously. The test results have been analyzed for correctness and estimated the time required for translating the packets.
