• DocumentCode
    2832246
  • Title

    Modeling Load on a Publish/Subscribe System

  • Author

    Singhera, Zafar U.

  • fYear
    2008
  • fDate
    1-5 Sept. 2008
  • Firstpage
    706
  • Lastpage
    710
  • Abstract
    Message oriented middleware (MOM) systems have been using publish/subscribe mechanism for asynchronous messaging among backend servers. Latest developments in modern Web-based clients have extended the use of publish/subscribe mechanism from backend servers to front-end clients. Some of the recently proposed protocols, like Bayeux, use topic-based publish/subscribe mechanism to exchange messages between clients and servers. This paper shares our experience in designing a load generation framework for benchmarking of an HTTP publish/subscribe server. It analyzes topic-based publish/subscribe domain by identifying major domain entities. It documents issues and challenges in load generation and benchmarking of such systems. It presents a model to represent a topic-based publish/subscribe system. We also share our experience in using the proposed model for load-generation and benchmarking of an HTTP/Bayeux based publish/subscribe server implementation. Some implementation details about our benchmarking framework, including an XML schema to represent a set of workloads, are shared. The paper is concluded by sharing shortcomings in the proposed model, and our plans on extending this work to other asynchronous and synchronous messaging paradigms.
  • Keywords
    Internet; XML; hypermedia; middleware; Bayeux protocol; HTTP publish-subscribe server; Web-based clients; XML schema; asynchronous messaging; load generation framework; message oriented middleware; topic-based publish-subscribe system; Databases; Expert systems; Load modeling; Message systems; Message-oriented middleware; Protocols; Subscriptions; Usability; Web server; XML; Computer Performance; Load Modeling;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Database and Expert Systems Application, 2008. DEXA '08. 19th International Workshop on
  • Conference_Location
    Turin
  • ISSN
    1529-4188
  • Print_ISBN
    978-0-7695-3299-8
  • Type

    conf

  • DOI
    10.1109/DEXA.2008.121
  • Filename
    4624801