MICS: an efficient content space representation model for publish/subscribe systems

One of the main challenges faced by content-based publish/subscribe systems is handling large amount of dynamic subscriptions and publications in a multidimensional content space. To reduce subscription forwarding load and speed up content matching, subscription covering, subsumption and merging techniques have been proposed. In this paper we propose MICS, Multidimensional Indexing for Content Space that provides an efficient representation and processing model for large number of subscriptions and publications. MICS creates a one dimensional representation for publications and subscriptions using Hilbert space filling curve. Based on this representation, we propose novel content matching and subscription management (covering, subsumption and merging) algorithms. Our experimental evaluation indicates that the proposed approach significantly speeds up subscription management operations compared to the naive linear approach. Categories and Subject Descriptors C.4 [Performance of Systems...

• Type: Conference
• Year: 2009
• Where: DEBS
• Authors: Hojjat Jafarpour, Sharad Mehrotra, Nalini Venkatasubramanian, Mirko Montanari
• Keywords: Content Matching, Content Space, DEBS 2009, Distributed And Parallel Computing, Subscription