• DocumentCode
    239527
  • Title

    DEVS modeling of large scale Web Search Engines

  • Author

    Inostrosa-Psijas, Alonso ; Wainer, Gabriel ; Gil-Costa, Veronica ; Marin, Mario

  • Author_Institution
    Univ. de Santiago, Santiago, Chile
  • fYear
    2014
  • fDate
    7-10 Dec. 2014
  • Firstpage
    3060
  • Lastpage
    3071
  • Abstract
    Modeling large scale Web Search Engines (WSEs) is a complex task. It involves many issues such as representing user´s behavior, query traffic, several strategies and heuristics to improve query response time, etc. Typically, WSEs are composed of several services deployed in data centers, which must interact to get the best document results to user queries. Additionally, hardware specification like multithreading and network communications have to be taken into account. In this paper, we propose to model a servicebased WSE using the Discrete Event System Specification (DEVS) formalism, which is one of the most powerful methodologies for discrete event systems. We validate our proposed model against an actual MPI implementation of the WSE and a process oriented simulation. We evaluate the accuracy of the proposed model by evaluating metrics such as query throughput and we show that there is no relevant differences, just small fluctuations of less than 4%.
  • Keywords
    Internet; application program interfaces; discrete event simulation; discrete event systems; formal specification; message passing; multi-threading; query processing; search engines; DEVS formalism; DEVS modeling; MPI implementation; data center; discrete event system specification formalism; hardware specification; large scale Web search engines; multithreading; network communication; process oriented simulation; query response time; query throughput; query traffic; service based WSE; user behavior; Computational modeling; Engines; Indexes; Mathematical model; Program processors; Time factors; Web search;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Simulation Conference (WSC), 2014 Winter
  • Conference_Location
    Savanah, GA
  • Print_ISBN
    978-1-4799-7484-9
  • Type

    conf

  • DOI
    10.1109/WSC.2014.7020144
  • Filename
    7020144