[c-nsp] Latency on ATM IMA circuits

[John Neiberger]

All of these locations are in the same geographic area, not more than 20
miles apart.

Round trip times to the first hop on the IMA are averaging 35 ms, which
seems awfully high. I would expect less then 10 ms to the first hop, but
perhaps my expectations are too high.

Hmm...interesting. I just took a look at the RTT to the first hop at my
T1 sites and it's pretty high there, as well. It's averaging close to 30
ms just to the first hop! Man, that's slow. It's interesting that a ping
to another T1 site is only a few milliseconds longer than to my first
hop. I would expect the first hop RTT to be considerably lower than that
of other sites.

I guess I need to do some research to figure out how high my
expectations should actually be.

Thanks,
John
--

>>> Brent Van Dussen <vandusb at attens.com> 2/21/05 3:32:56 PM >>>
Geographic differences wouldn't account for that added latency would 
it?  What do pings look like to the first hop of your IMA link?

-Brent


At 02:26 PM 2/21/2005, you wrote:
>I'm beginning to migrate our sites to an MPLS-based VPN product. All
but
>one of our sites has a full T-1, but one site has an ATM IMA circuit
>composed of four T1s. When I ping from a T1 site to another T1 site I
>see an average round-trip time of about 32 ms. However, when I ping
from
>or to the site with the IMA, the round-trip time ranges between 65
and
>80 ms. That's a pretty big jump! And one I wouldn't have expected.
>
>Is this type of added latency to be expected with IMA circuits? This
is
>the first one I've ever used before so I have no frame of reference.
I'm
>wondering if the multiplexing functions add some latency that isn't
>present on our other circuits.
>
>Any thoughts?
>
>Thanks!
>John
>--
>_______________________________________________
>cisco-nsp mailing list  cisco-nsp at puck.nether.net 
>https://puck.nether.net/mailman/listinfo/cisco-nsp 
>archive at http://puck.nether.net/pipermail/cisco-nsp/