Discovery:Implementation

Discovery Component Implementation

The group has finished the setup of the basic infrastructure, which is a discovery framework within the WSMX framework. This framework is able to manage different discovery components capable of processing different kinds of requests. Additionally together with the WSMT Developer Tools group a GUI for editing goals and Web services and a GUI for Web Service discovery have been completed.

WSMX Discovery

The WSMX discovery component contains discovery engines for the different discovery approaches as described in Discovery:

• Keyword-based discovery engine
• Lightweight DL discovery engine
• Lightweight Rule discovery engine
• Heavyweight Rule discovery engine
• Heavyweight "extended" (not supported anymore)

Discovery framework

The discovery framework within the WSMX discovery component does:

• Align the different discovery approaches in one framework
  • discovery using single engines
  • discovery using some/all engines

• Validate Web services and goals, based on the discovery engine targeted to be used
  • different service descriptions depending on targeted engine
  • modeler needs to be aware of description differences
  • validator supports the modeler in writing valid service descriptions

• Provide discovery component configuration means
  • added to the goal‘s nonfunctional properties
  • discovery engine to use
  • seeked matches (exact, plugin, subsume, ...)

Discovery component download

Download a WSMX binary distribution from sourceforge or get the latest nightly build from http://www.wsmx.org/downloads.html.

In particular you will need the core platform, and optionally other components that you want to use or be used by the Discovery component.

Tutorial

The short Discovery Tutorial can be used to test the functioning of the component and also to get familiar with it.

Keyword-based Discovery

• Within WSMX a keyword based engine has been developed that operates over the non-functional properties of WSML files. This engine is part of the standard WSMX distribution.

• Additionally seekda - a spin off of the University of Innsbruck - operates a keyword-based search on publicly available Web services: http://seekda.com/ . The keywords are in this case not taken from any semantic service description, but from the WSDL[1] description.

Service Location

Service Location / Service Crawling research is done in tight collaboration with seekda. The implemenation is tight into the seekda Web Service search engine and can be directly used at http://seekda.com .