I'm all for rate limits. I have no quibbles with this plan.
--
Nathanael Abbotts

Email: [email protected]
Wave: [email protected]
Twitter: @natabbotts (http://twitter.com/natabbotts)



On Wed, Nov 3, 2010 at 13:59, Joseph Gentle <[email protected]> wrote:

> Yep; that sounds good to me.
>
> Even leaving aside the robot case, it makes sense to implement that
> for spammy humans.
>
> I'd like this to be the default WIAB behaviour, but not required by
> the protocol.
>
> -J
>
>
> On Thu, Nov 4, 2010 at 12:45 AM, Tad Glines <[email protected]> wrote:
> > This conversation seems to have wandered a bit from the original topic.
> The
> > core problem raised is what to do about the possibility of two robots
> stuck
> > in a endless loop of programed replies (e.g. two echoy's). The simplest
> > solution is to rate limit deltas generated by robots and to automatically
> > drop them as a participant if they generate too many deltas in a fixed
> time
> > window. So for example, limit the burst rate to 2 deltas per second, if
> the
> > robot generates more than 110 deltas in 50 seconds, remove it as a
> > participant, and include a message to the wavelet author that the robot
> was
> > removed.
> >
> > As the creator/owner of a wave, I and no one else, should have control of
> > what happens in my wave. If someone doesn't like what they see, or isn't
> > interested in a particular category of changes, then can remove
> themselves
> > as a participant. I as a wave provider need to prevent a DOS attack on my
> > server, without being overly restrictive of what people can do with their
> > waves. If someone wants to see what happens when they put two separate
> chat
> > bots in a wave, that's their business. Providing a way for the robots to
> > know the server's rate limits allows the robots to self limit/self sensor
> > when needed. And the server can drop poorly behaving robots.
> >
> > The same rate limiting can also be applied to deltas generated by remote
> > wave servers. A sane limit might be 200 deltas per second (20 chars per
> > second * 10 participants). The federation protocol already has an error
> code
> > (RESOURCE_CONSTRAINT(5)) indicating that the sender needs to back-off. A
> > server can simply drop deltas on the floor and send a RESOURCE_CONSTRAINT
> > error to the sender if they arrive too fast.
> >
> > -Tad
> >
> > --
> > You received this message because you are subscribed to the Google Groups
> > "Wave Protocol" group.
> > To post to this group, send email to [email protected].
> > To unsubscribe from this group, send email to
> > [email protected]<wave-protocol%[email protected]>
> .
> > For more options, visit this group at
> > http://groups.google.com/group/wave-protocol?hl=en.
> >
>
> --
> You received this message because you are subscribed to the Google Groups
> "Wave Protocol" group.
> To post to this group, send email to [email protected].
> To unsubscribe from this group, send email to
> [email protected]<wave-protocol%[email protected]>
> .
> For more options, visit this group at
> http://groups.google.com/group/wave-protocol?hl=en.
>
>

-- 
You received this message because you are subscribed to the Google Groups "Wave 
Protocol" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to 
[email protected].
For more options, visit this group at 
http://groups.google.com/group/wave-protocol?hl=en.

Reply via email to