Nate wrote:
> That said, you'll need to do a site survey, to figure out what's
> actually going to block wireless signals. ...[snip]...  this'll tell
> you what type of s/n ratio you're getting to the AP everywhere you go. 

I believe you can just use "iwconfig <interface>" from the Linux
wireless tools packages.  i.e.

	#!/bin/sh -e
	# Run this script w/the first arg being the location (i.e.
	# "takewstat room221"
	iwconfig >> ${1:-"unspecified"}.stat

You should get signal and noise stats right there.  Sure, it's not all
pretty and graphed for you, but it's the data that's important.

> It's best to do this with the hardware they're planning on buying - if
> they want cheap-o AP's, then buy a cheap AP and use that. That way,
> you'll see where you need to place each AP to get optimal signal
> everywhere in the building.

Good advice.

> As far as authentication ...[snip]... 'course, there are a million
> other ways of doin this.. :)

I wonder if the University of Minnesota NTS department has published
their authentication software yet.  At the U, every student, faculty and
staff member receive X.500 accounts with which they can authenticate via
a web interface or system logins (UNIX).  The NTS department uses the
X.500 accounts to authenticate wireless users by tying in the DHCP
subnet with DNS forwarding and/or iptables/ipchains filtering. 

The user then goes to the website to authenticate.  When the
authentication is approved, the controlling daemon opens up the firewall
and routing rules for the DHCP IP address associated with that person.
It's pretty slick.

There was also an article in one of the recent System Administrator
magazine with something to that effect.

-- 
Chad Walstrom <chewie at wookimus.net>                 | a.k.a. ^chewie
http://www.wookimus.net/                            | s.k.a. gunnarr