[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [wsrm] WSDL for WSRM ??
Paolo, You are not missing anything and they are all valid concerns. Infact, I've raised similar concerns when I was rallying for the need for 'Polling Pattern Ack.'. The specific issue I mentioned then was when a user defines (in WSDL) an operation as a pure one way operation, then they cannot use 'Response Pattern Ack' as it totally violates the user definition of it's operation. This is because 'Response Pattern Ack.' expects the response with a SOAP envelope on the same connection/ transaction thus violating the user's intention. BEA's proposal also dis-allows this combination. I believe that WS-RM MEP's shouldn't be defined in the WSDL and should be dealt by the RM processor transparently as long as it doesn't violate the user's intention (as in the above case). I made the same comment in the "WS-RM - WSDL" Sub-Task force discussions I've started yesterday. If you are interested, I can add you to that list too. -Sunil Paolo Romano wrote: > I think I am a bit confused about how the interaction between the rm processor > and the final web service should be described in wsdl. > > Let me try to explain why I am confused: In the ws-rm wsdl we will describe > (apart from the ws-rm header blocks) the supported MEPs. In WSDL 1.1 the choice > is only between one-way and request-response, since sollicit-response and > notification MEPs (operations in WSDL terminology) are not interoperably > supported (WS-I states they MUST not be used in practice). > > Suppose we choose to allow only one-way operations. This means that messages and > acknolowdgements will be exchanged asynchronously, carried by 2 separate HTTP > POSTs requests in opposite directions, and that the HTTP responses in both > directions will just carry the HTTP response code 202/200 without any SOAP > message. > > Now, WS-RM headers will be inserted in the header of (possibly already) existing > SOAP messages used to allow applications interaction. These services will > already have their own WSDL description, which defines the supported operations > (MEPs) by the service, choosing between one-way or request-response. > > Suppose the original "unreliable" web service has some operations which support > only the (synchronous) request response MEP. > The SOAP node including both the ws-rm layer and the final webservice at this > point disagrees with the WS-RM supported MEP. In other words, the ws-rm layer > will simply respond by sending the HTTP 202 code and closing the TCP connection. > Then the request will be forwarded to the above layer, which instead expects the > underlying connection to be still open, because it is designed to respond > synchronously. The service requestor as well will see its connection closed in > an unexpected way, that is, without any response from the remote application. > > There must be something that I am missing... > > How can in general the ws-rm layer and the above layers be coupled to guarantee > compatibility, since they both have distinct WSDL definitions of their own > supported MEPs. > > May be the sending application should use the MEP described in the WSDL of the > final receiving application, and specify it to the WS-RM sender processor (let's > say as a parameter). The WS-RM sender processor could insert the specific MEP > to be used in an apposite ws-rm header block which will determine the behavior > of the WS-RM receiver. If, for example, the one-way MEP is chosen then the > receiving WS-RM processor will send just a response code. If, on the other hand, > the req/response MEP is chosen then the receiving WS-RM processor will wait for > a response to be returned from the above layer (say the application)? > > I hope somebody can enlighten me... > > Paolo > > Scott Werden <scottw@wrq.com> ha detto: > > > > > Using two one-way MEPs for the exchange has the advantage that it will work > > for bindings to any transport. The problem is that the ACK message can never > > be bound to a port with a soap:address element since the adress comes from > > the original reliable message. I suppose we can leave the the port binding > > out for the ACK. > > > > Scott > > > > > -----Original Message----- > > > From: Tom Rutt [mailto:tom@coastin.com] > > > Sent: Wednesday, June 25, 2003 12:33 AM > > > To: Sunil Kunisetty > > > Cc: Mark D. Hansen; WSRM-TC (E-mail) > > > Subject: Re: [wsrm] WSDL for WSRM ?? > > > > > > > > > Sunil Kunisetty wrote: > > > > > > > Hi Mark, > > > > > > > > I mentioned that I'm going to look into annotating WSDL with > > > > RM operations, not MEPs though. If that's what you meant, I haven't > > > > had a chance to look into that yet. We could work together if > > > > you are interested too. > > > > > > > > Scott & Venkat also mentioned that they would chime in too. > > > > > > > > -Sunil > > > > > > > >"Mark D. Hansen" wrote: > > > > > > > > > > > > > > > >>I recall from the face-to-face meeting that a few of the > > > members of this committee were going to work on WSDL for the > > > message exchange patterns being defined. Has anything > > > happened with that? I've been working on it a little bit and > > > would be happy to help out as well. > > > >> > > > >>Mark Hansen > > > >>bus: (888) 360-7285 > > > >>fax: (914) 723-8671 > > > >>email: khookguy@yahoo.com > > > >> > > > >>You may leave a Technical Committee at any time by visiting > > http://www.oasis-open.org/apps/org/workgroup/wsrm/members/leave_workgroup.ph > > p > > >> > > >> > > > > > > > > >You may leave a Technical Committee at any time by visiting > > http://www.oasis-open.org/apps/org/workgroup/wsrm/members/leave_workgroup.ph > > p > > > > > > > > > > > One approach is to have oneway MEP for each direction of interchange. > > The headers > > are defined as parts, and are bound in the wsdl to the soap header. > > The body is a part which is bound to the soap body in the WSDL. > > > > This is what the BEA spec does for WSDL for Reliable Messaging. > > > > Tom Rutt > > > > -- > > ---------------------------------------------------- > > Tom Rutt email: tom@coastin.com; trutt@fsw.fujitsu.com > > Tel: +1 732 801 5744 Fax: +1 732 774 5133 > > > > > > > > > > > > > > You may leave a Technical Committee at any time by visiting > > http://www.oasis-open.org/apps/org/workgroup/wsrm/members/leave_workgroup.ph > > p > > > > You may leave a Technical Committee at any time by visiting > http://www.oasis-open.org/apps/org/workgroup/wsrm/members/leave_workgroup.php > > > > > > -- > Paolo Romano > > You may leave a Technical Committee at any time by visiting http://www.oasis-open.org/apps/org/workgroup/wsrm/members/leave_workgroup.php
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]