Analysis and Algorithms for Content-Based Event Matching

    Content-based  event  matching  is  an  important  problem  in  large-scale  event-based  publish/subscribe  systems.  However,  open  questions  remain  in  analysis  of  its  difficulty  and  evaluation  of  its  solutions.  This  paper  makes  a  few  contributions  toward  analysis,  evaluation  and  development  of  matching  algorithms.  First, based on a simplified yet generic model, we give  a  formal  proof  of  hardness  of  matching  problem  by  showing its equivalence to the notoriously hard Partial  Match  problem.  Second,  we  compare  two  major  existing matching approaches and show that countingbased algorithms are likely to be more computationally  expensive  than  tree-based  algorithms  in  this  model.  Third,  we  observe  an  important,  prevalent  characteristic  of  real-world  publish/subscribe  events,  and  develop  a  new  matching  algorit...

• Type: Conference
• Year: 2005
• Where: ICDCSW
• Authors: Satyen Kale, Elad Hazan, Fengyun Cao, Jaswinder Pal Singh
• Keywords: Development Of Matching , Distributed And Parallel Computing, Evaluation Of Matching , Hardness Of Matching , ICDCSW 2005