An enhanced 1-hop clustering algorithm for Publish / Subscribe systems in AANETS

Content-based Publish / Subscribe communication paradigm offers a new approach to disseminate messages in the network, where the message content determines the recipients. Many applications used on AANETs, which are a subclass of VANETs, could be more efficient using this paradigm. Many Publish / Subscribe systems suitable for VANETs have been developed, however they are not efficient for some AANET applications. A promising approach is to build a Publish / Subscribe system over a cluster structure to reduce the control overhead and to offer a good scalability. However, the efficiency of this approach strongly depends on the performance of the clustering algorithm. The aim of this article is to propose a new clustering method, named CAPS, which will be the basis for a future content-based Publish / Subscribe system for AANETs. To validate our approach, a simulation model has been developed. Our algorithm has been compared to some other solutions in a modeled AANET context based on real air traffic traces. We show that CAPS gives better results than other solutions in terms of stability while maintaining at a low level the number of cluster groups.