[Standards-JIG] About Presence probe in 3921

Mridul mridul at sun.com
Wed Nov 15 20:21:42 UTC 2006


  Quoting from Initial Presence[1] in rfc 3921 :
"/Upon receiving initial presence from a client, the user's server MUST 
do the following if there is not already one or more available resources 
for the user (if there is already one or more available resources for 
the user, the server obviously does not need to send the presence 
probes, since it already possesses the requisite information):/"

I think we should remove the content within the brackets - servers MUST 
not attempt to do any form of caching of presence of remote entities : 
even if it is for different resource of the same bare jid.
The local server never possesses the requisite info - only the remote 
server does.

In the case of S2S, this would lead to incorrect presence, or worse 
presence leaks.
Within the same server - it does not really make sense, server knows 
statuses anyway : no need for explicit probe.
For more elaboration, please refer here[2].

The same is present in the bis drafts too.


[1] http://www.xmpp.org/rfcs/rfc3921.html#presence-resp-initial
[2] http://blogs.sun.com/mridul/entry/corner_cases_of_presence_probes

More information about the Standards mailing list