optical gear cooling requirements

Martin Hannigan hannigan at gmail.com
Wed Mar 4 01:52:44 UTC 2015


Remember the Ascend MAX TNT and the sideways left-right airflow? The
preferred method of deployment was three tall in a two post rack, mid
mount. At the end of about the 10th row you could literally cook a steak
and subsequently burn out the gear beyond that point. We fashioned our own
dividers to keep airflow from forming a jet down the line via ACE hardware
angle brackets that we cored to EIA screw spec and sheets of 1/4"
plexiglass mounted between. Cheap as heck and worked like a champ. Airflow
redirection can work.

Building a plenum for this could be relatively cheap and easy with a small
amount of sheet metal and a pair of tin sheers. Don't forget the warranty;
I'd check for anything explicit around airflow.

Let us know what you do, love the innovation.



On Tue, Mar 3, 2015 at 9:27 AM, Alex Rubenstein <alex at corp.nac.net> wrote:

> The rock has turned over for a moment and I have crawled out. It is good
> to see the sunlight from time to time.
> Those who know me know my life has gotten away from networking and that
> sort of thing, and I am fully immersed in datacenter design and
> construction for IT type loads (blades, compute, disk, etc.). However, I am
> presented with the challenge of having to deal with some optical gear
> (Ciena stuff, mainly). My question: have the optical folks woken up and
> made things cool front to back, or are they still in to the bottom to top
> world?
> Comments and lambasting: go
> Thanks!

More information about the NANOG mailing list