OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.


Help: OASIS Mailing Lists Help | MarkMail Help

wsrm message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [wsrm] Duplicate message responses [was Re: [wsrm] aboutRequest-Response MEP]

Doug Bunting wrote:

> Sunil,
> If you believe it is impractical for receiver RMPs to cache consumer
> payloads until their associated incoming messages have expired, why do you
> believe it practical for sending RMPs to cache producer payloads (up to)
> the same amount of time?  The sending RMP may learn before the expiration
> time that it can clear a specific message from its cache but that is all
> that changes from one side to the other.

  The scenarios are different. In the request case, we only cache the  request payload for Message Order
  case and that too only for those messages that arrive out of order. Once we made the message "available"
  to the next layer, we reclaim the "request payload" and hold on to the metadata. Only in the worst case
  scenarios, such as when a previous message is lost for ever, we cache the request payload until that
  particular message expires.

 So "that" worst case scenario is made a regular scenario in the response case as per the proposal.
 And that too for all RM messages and not just for MO case. Also, most like message expire times
 are used optimistically and will have a large life span and thus throttling the server further.

 As I said  in the previous mail and also the mail in Feb., this is indeed the most appropriate solution,
 although may be far from practical.

> I do not believe it is practical to use a "reliable delivery" protocol that
> gives up on returning payloads after the first attempt.  I seem to remember
> we wanted to operate at least somewhat correctly across firewalls.

 Yes, we do want to work correctly across firewalls. But how is that related to
 duplicate elimination?


> thanx,
>         doug
> On 09-Jun-04 21:56, Sunil Kunisetty wrote:
> > Doug Bunting wrote:
> ...
> >   But this solution will result in practical nightmare. Caching 'payload
> > response'
> >   until the original message expires is very impractical.
> ...
> >   -Sunil

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]