[ietf78-tech] thoughts on the terminal room...

Geert Jan de Groot geertj at psg.com
Wed Aug 11 05:02:40 PDT 2010


On Mon, 09 Aug 2010 17:27:10 -0700 Joel Jaeggli wrote:
> I've been trying to push us to think differently about the the terminal
> room for a while.

I hoped for Rob and Sjoerd to chime in, after all, they've spent a lot
of time at the helpdesk and had a clear view of the people there..

The number of people connecting to "wire" was surprizing. I spoke to
a few, and the motivation was wifi problems (which they admitted were
due to their own brokenness and for which they did not want to trouble
the helpdesk, thankfully..), or "corporate laptop policies" resulting
in same. At the same time, I agree that one switch with wires (instead
of wriring a whole room) would probably do - perhaps one of the
8-port or a little bigger. That could be in the hallway, especially
if it was to be a cheap unmanaged switch.
Too bad we no longer have the "link event" logs on the switches
from Maastricht so we can count actual usage. 
Perhaps something for Beijing (or even, add a cacti graph to count
terminal room wired network connections..)

Having the printer "watched" probably made it survive longer,
an anonymous place in the hallway may be too risky. Not sure if
the helpdesk minds having it in it's vicinity, given the amount
of questions on this, this time around. It too, would make wiring simpler.
Note that the main pain is still power wiring in that room 
and we're not reducing that by much.

AfNOG does not have a terminal room, but usually a few places to sit down 
and chat, and this works well. We always suffer a sudden evaporation 
of power bars (and so does the hotel, a power strip used to 
light a desklamp invariably finds that lamp unpowered after a geek session),
but we don't get many questions on IETF-terminal-room-style setups
and the questions we do get are not easily covered by 
a standard terminal room seat (usually power plug adapter / 
broken power supply / dead hard disk).

I'd suggest giving it a try, and it's easy enough to undo this
by moving tables around if neccessary.

On a similar but different topic, I was able to configure the network
printer on my XP toy while it was connected to a different subnet,
which is what caused a few people to contact the helpdesk.
The processs is a bit counterintuitive ("local printer port" ->
"TCP/IP network printer port").
I think this should be verified (on XP and W7), explicitely documented
(for XP and W7), and a few setup CD's copied for the helpdesk
once we know what printers we'll have.
Alternatively, perhaps a samba box can be set up to "beacon"
the printer service on all relevant subnets if that's what ExPee needs?
While I apologise to Rob, I'm sure he does not mind a further slowdown
in helpdesk volume if the printer call load would go down because of this.

I did miss the boarding pass printer rush - did it happen?

Geert Jan


More information about the ietf78-tech mailing list