[Standards] Proposed XMPP Extension: Stanza Repeaters

Richard Dobson richard at dobson-i.net
Wed Mar 19 10:42:37 UTC 2008


>> The repeaters concept doesn't solve the privacy list interactions 
>> either.
>
> Why not?  For each list of folks that a server wants to send presence 
> to, it creates a list.  If that list has different people on it for 
> different reasons, it creates different lists.  In the case of 
> presence, the server could, in the face of privacy lists being turned 
> on, create a presence-out list and a presence-probe list for the user 
> in question at each site where you have more than N users, where N 
> should be determined through some sort of analysis at where the 
> trade-off point is.  I'm betting it's at least 10.
>
> Yes, this could lead to **lots** of lists.  We better a) get the 
> protocol right, b) make sure it's going to help, and c) implement it 
> carefully.

Exactly the point I was trying to make.

> The following paragraph after that one qualified this a bit by 
> starting as "If multiple senders...".
>
> I should have written "The repeaters concept doesn't solve the privacy 
> list interactions either, if you are using multiple senders for the 
> same list".
>
> If you use multiple lists, one per user, the problem does not apply of 
> course.

Sure but the lists don't "have" to have multiple senders on them, its 
just nice in the protocol to allow that capability if it is needed, but 
notes probably need to be made in the spec on how it would work with 
privacy lists, as Joe says you would have however many lists per user 
the server needs defined in order to have the stanzas delivered 
according to the applied privacy lists, might well get a bit complicated 
and not actually gain you much if the privacy lists are being constantly 
changed but then that's up to the server logic to detect and only bother 
setting up repeaters where there is likely to be an actual real gain in 
bandwidth usage.

As you say I don't think that many people are using privacy lists or 
arn't likely to be defining complex sets of options and are likely to 
just be using them for blocking, so in real world usage it shouldn't get 
too complicated, but at least with repeaters you get the benefit of more 
optimised distribution of the stanzas but also allow privacy lists to 
still work.

Richard





More information about the Standards mailing list