A new MANETs clustering algorithm based on nodes performances

A mobile ad hoc network (MANET) is a dynamic wireless network which can be formed without the need for any pre-existing infrastructure in which each node can act as a router. In comparison with fixed networks, some new security issues have arisen with the introduction of MANETs. Secure routing, in particular, is an important and complicated issue. Organizing mobile nodes into manageable clusters can limit the amount of secure routing information. Under a cluster structure, mobile nodes are managed by nodes called cluster heads. The clusterhead role is resource consuming since its always switched on and is responsible for the long-range transmission. If a node has this role, it would burn its resource quickly, and after it died, all its members would be headless. In this work, we propose a new clustering algorithm for ad hoc networks, where the clusters are formed around the most powerful nodes, i.e, the node that has the best material resources such as residual energy, free memory, processor speed and hard disk space is elected as cluster head. Our algorithm takes into account the node range by including in our calculation the node density. Therefore, we are sure that the cluster head is represented by the most powerful and the less mobile node in the network that can perform its roles in the best conditions. This solution will be used to manage cryptographic keys in an ad hoc network. Our algorithm gives major improvements regarding the number and the performance of elected cluster heads. Analysis and simulation results are used to show the performance of our algorithm, compared with other clustering algorithms in literature.