[lug] OT: Wiring recommendations?

[Elyse Grasso]

On Tuesday 19 March 2002 11:46 am, you wrote:
> On Tue, Mar 19, 2002 at 10:51:16AM -0700, Elyse Grasso wrote:
> > We use wireless networks at work, and I must say I'm not terribly 
impressed 
> > by the throughput or the reliability we've seen: our training lab network 
> > died twice during a class we were teaching last week (possibly 
interference 
> > from equipment in the doctor's office next door?) which was very 
embarassing 
> > and annoying. Anyway, I'm already equipped and connected for a wired lan 
at 
> > home (just with exposed cables) and I'm spoiled by the connection speeds.
> 
> Does the Dr. have a microwave oven, and was it lunchtime?  :-)
> 
> Ahh... the joys of unlicensed (i.e. unprotected) RF spectrum.  (GRIN)
> 
> However, most 802.11b setups can be configured to do CTS/RTS which is
> sometimes called the "anti-microwave oven mode" or similar... heh heh.
> It slows throughput a bit.
> 
> Dumb question... what application software were you training on, what
> speed are your cards synching up at, and has anyone done any throughput
> tests on the wireless side of the LAN?
> 
> (I guess what I'm saying here is, who installed it, and why didn't they
> give you a baseline test?  Oh ... the dreaded "T" word that no one seems
> to do anymore.  Did they do any planning to determine if your training
> application was a good fit for wireless or was it just convenient not to
> run wires?)
> 
> Also, remember even though wireless is spread-spectrum and all that,
> don't let the mumbo jumbo hide the fact that it's still half-duplex and
> worse, two stations that can't hear each other can collide creating the
> "hidden node" problem experienced by wireless data networks since they
> were created.  Cisco cards and some others allow for tweaking of the
> output power of both the access points and the cards themselves,
> allowing for customizing and fine-tuning the RF patterns for a
> particular building or application.  As much as I love the Lucent cards
> (okay, Aguiere or WaveLAN, or Orinoco or whatever they are now) they
> don't appear to have this feature easily accessible.
> 
> All the old RF rules apply, the 802.11b cards just put it in
> easy-to-install packaging and few people have played with wireless
> networks prior to using 802.11b due to its popularity.
> 
> -- 
> Nate Duehr <nate at natetech.com>

I think the failures were mid-morning and mid-afternoon, and our own 
microwave doesn't seem to affect anything. 

The training lab uses wireless because it's portable: if a customer wants 
on-site training, the laptops and wireless cards and wap pack up into 2 big 
cases, then get unpacked in a conference room, and poof! we have a training 
lab all installed and configured and ready to go. At the moment we're using 
bottom of the line Linksys wireless gear (this was all put together on a 
shoestring) and I'm sure the configuration is sub-optimal, to put it mildly. 
We have lots of problems with collisions (made worse by the behavior of 
Win2K, which most of our clients want us to train on). 

I'll check the Linksys docs for CTS/RTS mode. Thanks. 

The training lab student machines dual-boot Win2k and KRUD, but we can't yet 
do Linux training on them.  Dell 2500s have a broken apm implementation (apm 
crashes in the BIOS at Linux boot time), which stops the pcmcia ports from 
working. However, 7.2 recognizes the builtin ethernet ports (7.1 did not), so 
if we get a training job that wants Linux client machines we may need to 
invest in a hub and some cables. 

Our current training courses are for Rational ClearCase and associated 
products, on Windows and Unix