A New Matching Structure and Interval Division on Content Based Publish/Subscribe System

In the content based publish/subscribe system, the efficiency of matching algorithm impact the integral performance directly. In order to achieve efficient matching and low forwarding, there must be a good matching structure. Most content-based publish/subscribe systems adopt tree matching structure, becausetree matching structure is easy to construct and maintain. At present, there are high efficient matching and high forwarding problem in tree matching structure. In this paper, we propose a new tree matching structure, named QuTREE. This tree structure can convert the predicate matching problem into thequery problem, and devide this interval, reduce the depth of the interval tree effectively, improve the matching efficiency. Most interval division of matching tree structure adopt dichotomy. It is not only simple implementation but also can avoid overloading the nodes, forwarding of dichotomy is overhead. Aiming at this problem, this paper propose a subscription interval division algorithm based on the golden section method, the algorithm can effectively avoid overloading of nodes, balance forwarding overhead and maintenance cost.