[NEW ISSUE] Inappropriate Fault destination

["Durand, Jacques R." <JDurand@us.fujitsu.com>]

Description

 

- The paragraph beginning Section 4 (Faults) suggests that faults that relate to the

processing of RM headers (and that relate to known sequences) SHOULD be sent to same

destination as <wsrm:SequenceAcknowledgement> messages.

But that does not make much sense for Faults of this type that originate in the RMS - such as

InvalidAcknowledgement, and that clearly result from an unexpected behavior of the RMD.

- The current text is not accurate in stating that “All other faults in the section

Relate to the processing of RM header blocks targeted….” , as some faults may not need

to originate in the processing of messages in order to

be generated , e.g. due to resource shortage (e.g. SequenceTerminated).

 

 

Justification

 

Recommending that some RMS-originating faults such as InvalidAcknowledgement, UnknownSequence etc.

go to the same destination as <wsrm:SequenceAcknowledgement> messages is simply not appropriate,

while it is the RMD that would benefit from taking knowledge of these.

In many cases, the destination of <wsrm:SequenceAcknowledgement> is the RMS itself.

The destination of such faults should either be left to an external agreement / policy, or recommended

to be the RM Destination.

 

 

Target: core

 

Proposal: (a + b)

 

(a)- Replace: "All other faults in this section relate to the processing of RM header blocks targeted at known Sequences"

with:

 "All other faults in this section relate to known Sequences", which is more accurate

(because some faults may be generated outside message processing, such as SequenceTerminated.)

 

(b)- Replace: "Entities that generate Sequence faults SHOULD send those faults

to the same [destination] as <wsrm:SequenceAcknowledgement> messages"

with:

"Entities that generate Sequence faults and that are on the destination end of the sequence SHOULD send those faults to the same [destination] as <wsrm:SequenceAcknowledgement> messages."