Hello Chuck,
The customers I have come across with buffering problems on 3750 were
either not effectively configuring QoS - sometimes the best config is no
QoS actually. The buffers get divided up when you enable "mls qos" and
if you don't use all the queues as they are set up, they will not be
optimal, especially for the default class.
The big scenarios you need buffering are speed difference (regardless of
average rate!) or many-to-one host at the same time, otherwise you don't
need significant buffers at all.
Since you have the same speed interfaces, how many transactions do you
think will be happening at once? That's where I would focus.
Regards,
John Gill
cisco
On 7/26/11 11:16 AM, Chuck Church wrote:
Anyone,
Been following the 3750 vs. other switches for datacenter use thread
here. We're putting together a small system using a couple IBM DS3512
storage units, with 6 physical hosts containing maybe 25 VMs total. My
original plan was etherchannel (LACP) from the hosts to a 3750X stack of 2
switches, and LACP to the SANs as well. SAN vs. user traffic probably on
separate channels. Everything is 1 gig copper. I'm assuming all hosts and
SAN do channeling, need to confirm with our server guy. Anyway, I think in
the grand scheme of things, this is a pretty small setup. This port buffer
issue has me a bit concerned however. I'd like to keep the etherchannel and
2 physical chassis for load-balancing plus redundancy reasons. Seems like
my options are:
Pair of Nexus 5K switches - price probably too high, no immediate need for
10 gig.
Pair of 4948 - does away with port buffer issue, but no multichassis
etherchannel
4507, dual sups, dual line cards - This seems like it would work, but pricey
as well.
3750X stack - backplane seems fine, port buffers seem to be the worry.
I guess what I'm asking, for those with 3750/3750E/3750X and seeing
performance issues, is it because you're a much larger setup, or am I likely
to regret the 3750s as well? I don't have much baseline on this now, as
we're adding more hosts and VMs, so current numbers will grow.
Thanks,
Chuck
_______________________________________________
cisco-nsp mailing list [email protected]
https://puck.nether.net/mailman/listinfo/cisco-nsp
archive at http://puck.nether.net/pipermail/cisco-nsp/
