[MOBY-l] Discovery protocols review

[Phillip Lord]

I thought that you might also be interested in the following....

http://twiki.mygrid.info/twiki/bin/view/Technologies/OntologicalDescription

http://twiki.mygrid.info/twiki/pub/Technologies/OntologicalDescription/nesc2.pdf

which describe the basic ideas that I have, for doing semantic service
discovery within mygrid. The architecture that we have allows us to
split service discovery from the registration (although does not
mandate this), which allows service discovery by arbitrary metadata,
even if the service registry does not understand this metadata. Also
with a small modification it allows third party metadata (ie not
provided by the service provider directly). 

I would be interested in opinions. 

Cheers

Phil

>>>>> "Ken" == Ken Steube <steube at sdsc.edu> writes:

  Ken> Hello, here is my review of discovery protocols...looking
  Ken> forward to seeing everybody this weekend!




  Ken> MOBY-related Protocols: Discovery

  Ken> Web service discovery mechanisms are created to provide a means
  Ken> to locate web services that are of interest to your research or
  Ken> business.  More specifically, three types of lookups are used
  Ken> when locating web services:

  Ken> 	white pages organization's name and contact info yellow pages
  Ken> 	industrial category and geographical location green pages how
  Ken> 	to use a service

  Ken> Included below is a summary of the features, strengths and
  Ken> weaknesses of MOBY-Central, UDDI, LDAP and ebXML.  I added
  Ken> CORBA Trader to the list to contrast an older binary protocol
  Ken> method to the newer text-based protocol methods.

  Ken> MOBY-Central
  Ken> 	Name: Model Organism Bring Your own Reference:
  Ken> 	http://biomoby.org Reference:
  Ken> 	http://biomoby.org/wilkinson_links_2002.pdf Purpose: an
  Ken> 	architecture for the discovery and distribution
  Ken> 		of biological data through web services
  Ken> 	Provides: white pages, yellow pages, green pages Strengths:
  Ken> 	object-driven registry query system with object and
  Ken> 		service ontologies, can easily adapt to new data
  Ken> types,
  Ken> 		we have our own data types so it's easy to have output
  Ken> 		of one service be input to another
  Ken> 	Weaknesses: not industry standard, relatively few developers
  Ken> and users