[Standards] throttling

Justin Karneges justin-keyword-jabber.093179 at affinix.com
Sun Feb 13 00:09:26 UTC 2011


On Saturday 12 February 2011 12:14:34 Matthew Wild wrote:
> On 12 February 2011 19:56, Rene Treffer <treffer+xmpp at measite.de> wrote:
> > On 02/11/2011 11:40 PM, Justin Karneges wrote:
> >> On Friday 11 February 2011 14:04:59 Matthew Wild wrote:
> >>> My opinion on this is that we don't need application-layer throttling
> >>> mechanisms. If a server wants to punish a peer, it can simply stop
> >>> reading from the connection for a while. The peer doesn't have to know
> >>> about this (such a notification MAY be useful for UI purposes, but I
> >>> personally doubt it).
> >> 
> >> The trouble is that throttling and keepalive pings don't play well
> >> together. It is easy to imagine a client today that uses XEP-0199 pings
> >> to the server every minute, but then gets throttled by the server for
> >> over a minute.  The result is that sending too fast means you get
> >> disconnected.  This is pretty terrible if there's no way to know what
> >> counts as "too fast".
> > 
> > This looks suspicious. If a server drops / delays pings I'd say it's a
> > server bug, not a client not implementing a XEP. You should seriously
> > think about handling pings differently if you feel like dropping them.
> 
> A server can choose to spend CPU time and RAM processing stanzas from
> the client, or it can choose to ignore that client for a while. If the
> client is flooding the server with stanzas, it only seems right it
> should be throttled, no? If it's being throttled, we have no way of
> knowing if a ping is in the socket buffer unless we spend CPU time
> parsing, etc.
> 
> I'm not convinced that we need a solution for this (I'm not sure if in
> the real world, a server would actually stop reading from a client for
> 60s or more). However if consensus is that this is something we need
> to fix, I think Justin's on the right track and I wouldn't oppose
> standardising "whacks" (at last!).

jabberd1 penalizes for over a minute, which is the inspiration for this 
discussion.  Newer servers are not nearly as aggressive.  But, from where I 
sit, those design decisions seem arbitrary and there's nothing to say that 
other servers couldn't adopt similar policies.

-Justin



More information about the Standards mailing list