[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Using state and routing inbound traffic

On 08/05/2005 05:07:53 PM, Daniel Hartmeier wrote:
On Fri, Aug 05, 2005 at 03:47:57PM -0600, Chris 'Xenon' Hanson wrote:

If the theory is correct, the graphs will nicely show so, and you can
make a nice little web page which we can refer to the next time
argues about rate-limiting incoming traffic. If the graph for server
interface deviates noticably from the one for the client interface
the server does not converge to a steady stream), that would lay the
theory to rest.

You've got to figure there's some wiggle, which is why you don't set your queue to be the full bandwidth of the WAN link. You leave some "slop".

The amount of wiggle tells you how much bandwidth you need to reserve
for "slop", the bandwidth you're trading off for improved convergence
time.  The amount of time the wiggle goes over the threshold of the
amount of "slop" you set tells you the porporition of the time
the whole business is "broken".

However, your test gives only best case I'd think.  A real-world
case would have more wiggle as there'd be more connections than
just one using up the bandwidth and they'd be starting and stopping.

Karl <[email protected]>
Free Software:  "You don't pay back, you pay forward."
                 -- Robert A. Heinlein