That's still not good enough. We do not define the actual timeout that the
MSH uses to determine when to retry. That timeout could be longer or
shorter than RetryInterval. We also do not say whether RetryInterval is
measured from the sending of the message or the expiration of the MSH
timeout. I recommend defining the MSH timeout and adding it to the CPA.
In addition, the send time is really unknown since it depends on properties
of the network as well as MSH processing times, which are system dependent.
Regards,
Marty
*************************************************************************************
Martin W. Sachs
IBM T. J. Watson Research Center
P. O. B. 704
Yorktown Hts, NY 10598
914-784-7287; IBM tie line 863-7287
Notes address: Martin W Sachs/Watson/IBM
Internet address: mwsachs @ us.ibm.com
*************************************************************************************
"Burdett, David" <david.burdett@commerceone.com> on 09/12/2001 02:55:26 PM
To: "'David Fischer'" <david@drummondgroup.com>, ebXML Msg
<ebxml-msg@lists.oasis-open.org>
cc:
Subject: RE: T2 Clarify TimeToLive
Almost. I think that really it should be ...
TimeToLive > Sum Of [SendTime + (Retries * RetryInterval)] over
all hops + some percentage
As you cannot always know how many hops there will be nor how fast each hop
would be, you really should make TimeToLive a lot more than the figure
above. In practice, I think that feedback from actual use is likely to be
the most effective way to work out the values to use.
David
-----Original Message-----
From: David Fischer [mailto:david@drummondgroup.com]
Sent: Wednesday, September 12, 2001 8:03 AM
To: ebXML Msg
Subject: RE: T2 Clarify TimeToLive
>> I think adding this to the spec as a non-normative note is a good idea.
Very Good. I mostly realized the potential for problems when I realized
this
could cause a recalculation of the signature. I guess the answer is:
TimeToLive > SendTime + (Retries * RetryInterval)
David Fischer
Drummond Group.
-----Original Message-----
From: Burdett, David [mailto:david.burdett@commerceone.com]
Sent: Tuesday, September 11, 2001 11:50 PM
To: 'Arvola Chan'; David Fischer; ebXML Msg
Subject: RE: T2 Clarify TimeToLive
Currently TimeToLive is defined in section 8.4.6.4 as ...
The TimeToLive element indicates the time by which a message should be
delivered to and processed by the To Party.
... and in section 10.2.3 as ...
The TimeToLive value indicates the time by which a message should be
delivered to and processed by the To Party.
What I think you are suggesting is that we provide guidance on how it
should
be set. There are two factors which can affect this:
1. Setting it to a value long enough so that it is longer that the time
required to do all the retries on all the hops between the From and To
Parties. By definition this is not precisely computable.
2. Setting it to a value that makes sense in a business context, e.g. if
this PO gets there in a day then it is fine.
I think adding this to the spec as a non-normative note is a good idea.
David
-----Original Message-----
From: Arvola Chan [mailto:arvola@tibco.com]
Sent: Tuesday, September 11, 2001 5:12 PM
To: David Fischer; ebXML Msg
Subject: Re: T2 Clarify TimeToLive
David:
I agree with you that we need to clarify in the spec how TimeToLive
should be set for a given message. For reliably delivered messages,
I don't think we want to reset the the TimeToLive on every retry. It just
incurs unnecessary signature recomputation overhead. I would suggest
setting
TimeToLive to current time (for the initial attempt) + (retry + 1) *
retryInterval. This must also be less than the receiver MSH's persist
duration.
For messages delivered by BestEffort, the TimeToLive can either be
omitted, or perhaps set to the current time plus the
RequestingBusinessActivity's or RespondingBusinessActivity's
timeToReceiveAcknowledgment.
Cheers,
-Arvola
-----Original Message-----
From: David Fischer <david@drummondgroup.com>
To: ebXML Msg <ebxml-msg@lists.oasis-open.org>
Date: Tuesday, September 11, 2001 4:52 PM
Subject: T2 Clarify TimeToLive
I don't know if this is a problem or not. TimeToLive is a time, not an
interval, and should probably be set to current time plus 1/2 RetryInterval
(half of the round trip from sender to receiver back to sender). Won't
this
have to be recalculated each retry? Won't this invalidate the signature?
Maybe I'm getting the correct value of TimeToLive wrong? What in the world
should this be? Maybe this is more along the lines of "This PO is valid
for
30
days" so this value should be set to current time plus 30 days? In any
case, we
probably need to clarify the spec.
Regards,
David Fischer
Drummond Group.
----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>
----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>
----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>
----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>
----------------------------------------------------------------
To subscribe or unsubscribe from this elist use the subscription
manager: <http://lists.oasis-open.org/ob/adm.pl>