[ previous ] [ next ] [ threads ]
 From:  "Mitch \(WebCob\)" <mitch at webcob dot com>
 To:  "Danne" <danne at djesigns dot com>, m0n0wall at lists dot m0n0 dot ch
 Subject:  RE: [m0n0wall] Traffic Shaping issue
 Date:  Mon, 26 Jul 2004 16:21:42 -0700
> trying that AIM direct connection, my latency to their 768k connection
> was under 40ms. With traffic shaping turned off, I could hit my max
> speed. With it on, I couldn't.

One more test to try... use larger ping packets. Try packets that are around
your MTU. Say 1000 bytes? See what happens then?

> I've never seen the traffic shaper introduce enough latency to reduce my
> uploads while ping testing. Always stayed the same regardless of being
> on or off.

Guess it depends on how they work and what "fudge factors" they have built
in. My understanding of traffic shaping is that it basically works by
throttling the acks to cause the other side to slow it's sending... If you
lied to it to say you had a bigger channel than you do, you might find the
delay goes away... I've heard that if there isn't a margin of error all
(which I think is how some of them work) that there CAN be momentary
interruptions in service due to saturation of the link - but the way this
one is designed to work, it should NEVER saturate completely, but that does
have some overhead - I think.

> BTW, thanks for following up with me on this! It's very appreciated!

No worries - it helps me learn too.  ;-)