Brief Announcement: Distributed Self-organizing Event Space Partitioning for Content-Based Publish/Subscribe Systems

  • Roberto Beraldi
  • Adriano Cerocchi
  • Fabio Papale
  • Leonardo Querzoni
Part of the Lecture Notes in Computer Science book series (LNCS, volume 6976)

Context and Motivations

Publish/subscribe systems have commonly been divided in two large families on the basis of their event-selection model [2]: topic-based and content-based systems. The former trade reduced subscription expressiveness with simpler implementations and higher performance. Conversely, the latter allow to accurately map published data in a complex event schema on top of which expressive subscriptions can be defined, but incur the cost of more complex implementations that delivers reduced performance on large distributed settings. System developers are thus faced with a choice about which kind of system is best suited to the target application. A common solution to this dilemma lies in the event space partitioning [4] technique: the event schema is partitioned in a number of subspaces that are then statically mapped to topics. The partitioning must be globally known and subscribers are expected to subscribe those topics where subspaces that have a non-empty intersection with their content-based subscriptions have been mapped. Undesired events (false positives) can be filtered out at the receiver side. The event space partitioning granularity strongly affects the performance of such systems: if it is excessively coarse-grained too much resources are wasted to deliver false positives, while if it is too fine-grained the number of topics that will be generated, and that must be managed by the topic based system, could easily become huge. Current solutions [3] provide sub-optimal approximations that are calculated offline and then statically applied to the system.

Keywords

False Positive Receiver Side Event Space Target Application Complex Implementation 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

References

  1. 1.
    Beraldi, R., Cerocchi, A., Papale, F., Querzoni, L.: Distributed self-organizing event space partitioning for content-based publish/subscribe systems. Tech. rep., Università di Roma “La Sapienza” (2011), http://www.dis.uniroma1.it/~midlab/publications.php
  2. 2.
    Eugster, P., Felber, P., Guerraoui, R., Kermarrec, A.M.: The many faces of publish/subscribe. ACM Computing Surveys 35(2), 114–131 (2003)CrossRefGoogle Scholar
  3. 3.
    Opyrchal, L., Astley, M., Auerbach, J., Banavar, G., Strom, R., Sturman, D.: Exploiting IP multicast in content-based publish-subscribe systems. In: Coulson, G., Sventek, J. (eds.) Middleware 2000. LNCS, vol. 1795, pp. 185–207. Springer, Heidelberg (2000)CrossRefGoogle Scholar
  4. 4.
    Wang, Y., Qiu, L., Achlioptas, D., Das, G., Larson, P., Wang, H.J.: Subscription partitioning and routing in content-based publish/subscribe systems. Tech. rep., Microsoft Research (2002)Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2011

Authors and Affiliations

  • Roberto Beraldi
    • 1
  • Adriano Cerocchi
    • 1
  • Fabio Papale
    • 1
  • Leonardo Querzoni
    • 1
  1. 1.University of Rome “La Sapienza”RomeItaly

Personalised recommendations