• DocumentCode
    125461
  • Title

    Performance Functions for QoS Prediction in Web Service Composites

  • Author

    Volk, Florian ; Sokoli, Johanna ; Muhlhauser, Max

  • Author_Institution
    Telecooperation Lab., Tech. Univ. Darmstadt, Darmstadt, Germany
  • fYear
    2014
  • fDate
    June 27 2014-July 2 2014
  • Firstpage
    574
  • Lastpage
    581
  • Abstract
    We contribute performance functions to model the quality of service (QoS) of web services in horizontal composites. Performance functions are thereby used to formalize service level agreements (SLAs) to enable automated verification. Inter-Component rating is a model for gathering additional QoS information from web services that rate the output of their predecessor web services in composites. Furthermore, we contribute two approaches to verify if a web service adheres to its claimed QoS described with a performance function. For this verification, claimed performance is compared with reading points gathered from spot-testing the quality of the web service. The first approach we discuss is function interpolation, the second one applies methods from computational trust. The paper also discusses the advantages and drawbacks of both approaches as well as the influence of measuring errors.
  • Keywords
    Web services; interpolation; quality of service; trusted computing; QoS information; QoS prediction; SLA; Web service composition; Web service quality; computational trust; function interpolation; inter-component rating; performance function; performance functions; quality of service; service level agreements; Computational modeling; Customer satisfaction; Interpolation; Monitoring; Polynomials; Quality of service; Web services; Performance analysis; Prediction methods; Predictive models; Quality of service; web services;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Web Services (ICWS), 2014 IEEE International Conference on
  • Conference_Location
    Anchorage, AK
  • Print_ISBN
    978-1-4799-5053-9
  • Type

    conf

  • DOI
    10.1109/ICWS.2014.86
  • Filename
    6928946