Famous operational issues

Paul Ebersman list-nanog2 at dragon.net
Thu Feb 18 23:24:14 UTC 2021

warren> 2: A somewhat similar thing would happen with the Ascend TNT
warren> Max, which had side-to-side airflow. These were dial termination
warren> boxes, and so people would install racks and racks of them. The
warren> first one would draw in cool air on the left, heat it up and
warren> ship it out the right. The next one over would draw in warm air
warren> on the left, heat it up further, and ship it out the
warren> right... Somewhere there is a fairly famous photo of a rack of
warren> TNT Maxes, with the final one literally on fire, and still
warren> passing packets.

The Ascend MAX (TNT was the T3 version, max took 2 T1s) was originally
an ISDN device. We got the first v.34 rockwell modem version for
testing. An individual card had 4 daughter boards. They were burned in
for 24 hours at Ascend, then shipped to us. We were doing stress testing
in Fairfax VA. Turns out that the boards started to overheat at about 30
hours and caught fire a few hours after that... Completely melted the
daughterboards. They did fix that issue and upped the burnin test period
to 48 hours.

And yeah, they vented side to side. They were designed for enclosed
racks where are flow was forced up. We were colocating at telco POPs so
we had to use center mount open relay racks. The air flow was as you
describe. Good time. Had by all...

Both we (UUNET, for MSN and Earthlink) and AOL were using these for
dialup access. 80k ports before we switched to the TNTs, 3+ million
ports on TNTs by the time I stopped paying attention.

More information about the NANOG mailing list