Efficient composite event detection based on DHT protocol

Current communication systems restrict user affordances in terms of expressiveness and interaction. Complex Event Processing (CEP) have gained much importance to overcome these shortcomings. It allows users to interact with powerful expressiveness when defining logical and temporal patterns of exchanged events. This expressiveness allows cleaning network traffic by eliminating the routing of useless events and consequently reducing the overall consumed energy. However, handling the expressiveness of desired events, filtering coming events still challenging features, especially with the growth of data size encapsulated as events in the network. In this work, we rely on Publish/Subscribe system based on Distributed Hash Tables (DHT), which offers intrinsically flexible event routing, scalability and load balancing in order to manage distributed composite events routing. For efficient event filtering, we propose a three dimensional indexing hash space named CECube as a smart data structure for rapid CEP over DHT. The CECube indexes firstly composite subscriptions, then basing on a simple binary search, it serves as publications filter and helps making the right decision for what events should be aggregated and forwarded to the adequate subscribers. The performance of our solution is evaluated using Free Pastry simulator. The results demonstrate firstly that our approach is efficient in terms of filtering process and that the average number of routing nodes is decreased. Secondly, we prove the superiority of our approach as compared to another existing work.