ws-rx message
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]
Subject: New questions about MSFT's MC Proposal
- From: Doug Davis <dug@us.ibm.com>
- To: ws-rx@lists.oasis-open.org
- Date: Thu, 16 Nov 2006 05:56:00 -0500
Finally did a quick scan of the latest
MC proposal from Marc - it has the following changes to the CS/CSR
MEP:
/wsrm:CreateSequence/wsrm:MakeConnectionTo
The RM Source SHOULD include this element
in any CreateSequence message it sends when it anticipates MakeConnection
will be used to retrieve Sequence Traffic and Sequence Lifecycle Messages,
e.g. when sent to an anonymous endpoint. This element is of type wsa:EndpointReferenceType
(as specified by WS-Addressing). It specifies the endpoint reference to
which messages containing MakeConnection header blocks related to the created
Sequence are to be sent. The RM Source MAY use this EPR, including reference
parameters, to relate MakeConnection requests to specific sequences.
...
/wsrm:CreateSequenceResponse/wsrm:Accept/wsrm:MakeConnectionTo
The RM Source SHOULD include this element
in any CreateSequence message it sends when it anticipates MakeConnection
will be used to retrieve Sequence Traffic and Sequence Lifecycle Messages,
e.g. when offered sequence was from an anonymous endpoint. This element
is of type wsa:EndpointReferenceType (as specified by WS-Addressing). It
specifies the endpoint reference to which messages containing MakeConnection
header blocks related to the created Sequence are to be sent. The RM Source
MAY use this EPR, including reference parameters, to relate MakeConnection
requests to specific sequences.
- In the CSR stuff there appears to
be typos since it talks about the RMS including a MCTo in a CS when were
talking about the CSR.
- This text talks about the RMS including
a MCTo when it "anticipates" a MC will be used - how can it know
this? There is nothing in the CS that would indicate it - use of
wsa:Anonymous is not enough for the CSR side and on the CS side there's
nothing to look at since its the first message being sent. This seems
to require quite a bit of out-of-band knowledge especially in the case
of a endpoint talking to both anon and non-anon clients at the same time.
- Following on the previous bullet,
it really seems like this proposal will require a MCTo EPR to be generated
and supported in _all_ situations since there is no indication of when
it may or may not be needed.
- This proposal still has the problem
that an rogue MC to the service endpoint will cause the generation of unused
Sequences since it has no way of knowing whether or not its rogue. So
saying it MAY return a MCRefused fault does no good since there's no way
for the recevier of the MC to know when to fault.
- re:MessagePending/MCTo EPR - how
does the sender of this header know that the recevier of this header is
the right one? If the wsa:To is anonymous, and there are lots of
messages going out to lots of different anon endpoints, how does the right
MP/MCTo header get placed on the right outgoing message? I've heard
people talk about cases where the RM logic is on a gateway type of machine
- so how does the gateway know which message is which when it may not have
access to the application logic/data ? All it sees is some anon message
going out.
- And I believe my previous comments/questions
still apply [1].
thanks
-Doug
[1] http://www.oasis-open.org/apps/org/workgroup/ws-rx/email/archives/200611/msg00011.html
[Date Prev]
| [Thread Prev]
| [Thread Next]
| [Date Next]
--
[Date Index]
| [Thread Index]
| [List Home]