- From: Anish Karmarkar <Anish.Karmarkar@oracle.com>
- Date: Wed, 16 Feb 2005 11:35:17 -0800
- To: David Hull <dmh@tibco.com>
- CC: public-ws-async-tf@w3.org
David Hull wrote: > > Roberto Chinnici wrote: > >> >> * Description of the scenario >> >> In this use case, a client contacts a server using transport T1. The >> response is sent back asynchronously using a callback made over a >> potentially different transport T2. Even if the transports are the >> same, the request and response messages could use different connections, >> making this case different from #6. > > > I don't understand this statement. In case 6, if I send an HTTP > request, and the explicit reply-to address is also HTTP, then there is > one transport but two connections. On the other hand, if it's not HTTP, > then you have two different transports and (naturally) two connections. > Finally, if it's anonymous, you have one transport and one connection. > > As far as I can tell, case 6 subsumes case 3 and perhaps others. > > The usecases were created from the WSDL perspective. Usecase 3 is asynch req/response where there is a single WSDL operation and it is clear from the WSDL doc that the req/response is asynch Usecase 6 is again a single WSDL operation (req/res) but the WSDL does not explicitly say that it is asynch. It is the runtime ws-addr header that tells you whether it is synch or asynch (and this fact will be indicated in the WSDL). So I think there is a slight difference between the two usecases especially the way the operation will be expressed in WSDL. HTH. -Anish --
Received on Wednesday, 16 February 2005 19:35:54 UTC