Re: [ws-rx] [NEW ISSUE] When to piggy-back RM headers

[Doug Davis <dug@us.ibm.com>]

Yup
-Doug

Anish Karmarkar <Anish.Karmarkar@oracle.com>

05/15/2006 09:47 PM

To	Mark Little <mark.little@jboss.com>
cc	Doug Davis/Raleigh/IBM@IBMUS, ws-rx@lists.oasis-open.org
Subject	Re: [ws-rx] [NEW ISSUE] When to piggy-back RM headers

I would image the same would apply for extensibility
points (?).

-Anish
--

Mark Little wrote:
> Doug, I just want to clarify the lack of reference to wsa:Metadata.
It's 
> much harder to quantify the relevance of the metadata to the EPR 
> comparison because it can change (become out-of-date, for example),
> whilst the rest of the endpoint is still valid. So I'm not advocating
> adding that to the equality definition you've got; I'd just like to
make 
> sure its absence is for the same reasons I'm thinking.
> 
> Mark.
> 
> 
> Doug Davis wrote:
>  >
>  > Based on some additional feedback I'd like to modify the
proposal 
> slightly:
>  >
>  > Proposal:
>  >
>  > Add after the first paragraph in section 3:
>  >
>  > Some RM header blocks may be added (or 'piggy-backed')
to messages 
> that happen to be targeted to the same endpoint to which those headers
> are to be sent, thus saving the overhead of an additional message
> exchange.  For the purpose of determining whether these soap
header 
> blocks may be added to a message, two Endpoint References are considered
> to be equal if the following are true:
>  > - The [address] properties of two endpoint references are
equal when 
> compared according to Section 6 of [RFC 2396]
>  > - They contain the same number of reference parameters
>  > - For each reference parameter in one endpoint reference
there exists 
> an equivalent reference parameter in the other.  One [reference
> parameter] is equivalent to another [reference parameter] if their
byte 
> streams per Exclusive XML canonicalization (with an empty "inclusives"
> list) are equal.  Note that this may result in incorrect answers
if 
> there are qnames in attribute or element content.
>  >
>  > (the 3rd bullet changed)
>  >
>  > Marc - when you update the issue list can you please update
the proposal?
>  >
>  > thanks
>  > -Doug
>  >
>  >
>  >
>  >
>  >
>  > Doug Davis/Raleigh/IBM@IBMUS
>  >
>  > 05/04/2006 12:18 AM
>  >      
>  > To
>  >       ws-rx@lists.oasis-open.org
>  > cc
>  >      
>  > Subject
>  >       [ws-rx] [NEW ISSUE] When to piggy-back
RM headers
>  >
>  >
>  >      
>  >
>  >
>  >
>  >
>  >
>  > Description:
>  > It is not clear when an implementation is allowed to piggy-back
RM 
> headers (acks, ackReq) in a message.  I suspect that most 
> implementations will simply compare the wsa:Address of the EPRs -
> however, since ref-p's are an integral part of EPRs they should really
> be included in the comparison.
>  >
>  > The latest WSA spec says:
>  >
>  > 2.3 Endpoint Reference Comparison
>  > This specification provides no concept of endpoint identity
and 
> therefore does not provide any mechanism to determine equality or
> inequality of EPRs and does not specify the consequences of their
> equality or inequality. However, note that it is possible for other
> specifications to provide a comparison function that is applicable
> within a limited scope.
>  >
>  > This proposal does just that - it proposes a comparison
function for 
> use just by RM for a very specific purpose.
>  >
>  > Target: core
>  >
>  > Type: design
>  >
>  > Proposal:
>  >
>  > Add after the first paragraph in section 3:
>  >
>  > Some RM header blocks may be added (or 'piggy-backed')
to messages 
> that happen to be targeted to the same endpoint to which those headers
> are to be sent, thus saving the overhead of an additional message
> exchange.  For the purpose of determining whether these soap
header 
> blocks may be added to a message, two Endpoint References are considered
> to be equal if the following are true:
>  > - The [address] properties of two endpoint references are
equal when 
> compared according to Section 6 of [RFC 2396]
>  > - They contain the same number of reference parameters
>  > - For each reference parameter in one endpoint reference
there exists 
> an equivalent reference parameter in the other.  One [reference
> parameter] is equivalent to another [reference parameter] if their
byte 
> streams per Exclusive XML cononicalization are equal.
>  >
>  > (some should recognize this from the submitted WSA spec)
>  >
>  > thanks
>  > -Doug
>