Research On GA-based Anycast Routing Algorithm And Load-balanced Routing Protocol

Anycast is a new network service and a new feature of IPv6(Internet Protocol Version 6 for next generation network). It is designed to allow users to select and communicate with the"nearest"server in a group of content-equivalent servers. Anycast has numerous potential applications, but before that some problems need to be addressed. Until now, few researches have been done on anycast routing, especially on routing algorithm.In this paper, the up-to-date research on anycast is researched, and two anycast routing algorithms based on GA(Genetic Algorithm) are improved. In addition, a protocol for anycast routing is proposed. Firstly, double-populations and simulated annealing are introduced into the research on anycast routing algorithm and it indeed speeds algorithm's running rate. Secondly, the routing problem that refers to server load and reverse route's QoS constraints is researched. After defining the problem model, we resolve it with improved GA, and the simulating results are encouraging. When designing the protocol, two things are put in the first place. Firstly, the server load should be balanced, and so it should be considered into the selection of routes; secondly, the new protocol should be, if not fully, as compatible as possible to current standards and protocols.In order to verify the validity and convergence of the new algorithms, many simulations have been taken. The results show that both improved algorithms are efficient and convergent. Compared with other protocols, the proposed protocol is obviously maneuverable and valuable.