The WTE+ framework: automated construction and runtime adaptation of service mashups

  • Authors:

  • Anna Hristoskova;Bruno Volckaert;Filip Turck

  • Affiliations:

  • Department of Information Technology, Ghent University--IBBT, Ghent, Belgium 9050;Department of Information Technology, Ghent University--IBBT, Ghent, Belgium 9050;Department of Information Technology, Ghent University--IBBT, Ghent, Belgium 9050

  • Venue:
  • Automated Software Engineering
  • Year:
  • 2013

Quantified Score

Hi-index 0.00

Visualization

Abstract

An emerging software engineering methodology is the combination of functionality and content from existing software components into service mashups, creating greater value than the sum of the individual participating building blocks. For businesses, using catalogues of reusable services means agile development of new applications using open communication standards including the Simple Object Access Protocol (SOAP) for transmitting data, and the Web Services Description Language (WSDL) for defining services. The result is faster adaptation to the changing business environment creating value at reduced development time and cost while increasing revenues.This manuscript presents the developed service management framework and evaluation results of the WTE+ (Web2.0/Telco2.0/Enterprise2.0 with semantics) research project. Main goal is supporting application developers with the automatic construction and runtime reconfiguration of custom-made service mashups at a minimum performance cost without the need for constant IT intervention. Semantically enriched services are automatically combined into custom-made service mashups by the designed planning algorithms. These planning techniques are optimized with late binding and runtime adaptation to changing user-context taking fully into account the quality of service parameters of the available components which are provided at design time and updated at runtime. In addition, recovery mechanisms are provided in case of failed services and/or resources. During mashup execution, unavailable services are dynamically replaced by equivalent ones or alternative service mashups keeping in mind the current execution state.The developed planning algorithms are put through extensive performance and scalability experiments for a typical e-commerce scenario, in which e-shop services such as product payment and delivery are on-the-fly composed to an e-shop application. The results show that an automatic construction of a new application out of existing services can take up between 5 to 43 seconds for 500 services while runtime adaptation takes up to 5 seconds on average depending on the availability of equivalent services.