On 11 February 2011 21:41, Peter Saint-Andre <[email protected]> wrote:
> In Brussels we discussed the throttling feature of XEP-0198. It seems
> that no one has implemented it, and it's not clear to me if we need it.
> The proposal was to move it from XEP-0198 to a new XEP, but if we don't
> think it's necessary then we can simply remove it.
>
> What do people think?
>
> 1. Throttling is beautiful, but let's put it in a separate spec.
>
> 2. Who needs throttling? Let it die!
>

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).

Also the mechanism currently in 0198 is very simplistic, and assumes
the throttling is done based on the number of unhandled stanzas. A
more sensible throttling mechanism would instead take into account the
size of the stanzas instead.

In practice throttling/karma algorithms are typically proprietary to
each server/service, and I don't see a reason they ought to be
standardized or indeed have any interaction with the protocol flow at
all.

Regards,
Matthew

Reply via email to