Interesting People mailing list archives

The 'Wi-Fi At Conferences' Problem


From: Dave Farber <dave () farber net>
Date: Fri, 9 Oct 2009 18:39:07 -0400





Begin forwarded message:

From: Jim Thompson <jim () netgate com>
Date: October 9, 2009 18:18:58 EDT
To: dave () farber net
Cc: Dewayne Hendricks <dewayne () warpspeed com>
Subject: Re: [IP] The 'Wi-Fi At Conferences' Problem



The majority of the problem with Wi-Fi for 2000 people is that the parties who 'install' the Wi-Fi Access Points typically don't know what they're doing.

For instance, one of the *classic* mistakes is that the installers will hear the often-stated, "three non-overlapping channels" (in the 2.4GHz band) and assume that this means that they can install access points on channels 1, 6, and 11.

While its true that these channels don't "overlap" in terms of transmit mask, the radios used in the APs *and the clients* don't have enough selectivity to "ignore" (if you will) a radio operated on the adjacent "non- overlapping" channel. The result is many, if not most packets (or frames) end up being smashed in the receiver. Remember that this happens for the clients (iPhones, notebooks, etc) too. The result: extremely low 'throughput'. I've witnessed Apple make this mistake (using their own staff and their own gear) at events as large as MacWorld.(*)

Another mistake is under-provisioning bandwidth. A T-1 line can be saturated by the average conference attendee, especially if the conference content
is boring.   The IETF meetings normally get a DS-3 or better.

Having a DHCP server "run out" of addresses is just rank amateurism. While the 802.11 (wifi) specification (from IEEE) "allows" 2037 stations (clients) to associate with a single AP, I can't think of any commercially- available AP that will support this number of simultaneous "clients".

Jim
(*) its also the proximate cause of failure on Vivato's first- generation product as well as other "Wi-Fi switches" that used to be marketed by various firms.


On Oct 9, 2009, at 11:18 AM, David Farber wrote:



Begin forwarded message:

From: dewayne () warpspeed com (Dewayne Hendricks)
Date: October 9, 2009 10:57:17 AM EDT
To: Dewayne-Net Technology List <xyzzy () warpspeed com>
Subject: [Dewayne-Net] The 'Wi-Fi At Conferences' Problem

[Note:  This item comes from friend David Isenberg.  DLH]

The “WiFi At Conferences” Problem
by Joel Spolsky
Thursday, October 08, 2009
<http://www.joelonsoftware.com/items/2009/10/08.html>

Why does WiFi work so poorly at tech conferences?

I assume that WiFi wasn’t really designed to handle a big ballroom with 2000 people, all trying to connect with their laptops and ce ll phones at the same time. Sometimes I feel like I’m lucky if it works in my apartment. So I never thought it was even possible to get it to work at a large, technically-savvy conference. At Stack Overflow DevDays, yesterday in Boston, the bandwidth seemed OK but the DHCP server ran out of addresses. This didn’t seem to be some thing that anyone could fix. The conference organizers (er, me and Greg) were incredibly busy trying to, you know, organize the conf erence, so spending time tracking down the mysterious ISP and maki ng them fix their router was impossible.

It’s almost getting boring to read the conference reports complain ing about this. Almost every conference, even the ones put on by f ancy tech companies, has trouble. I never assume WiFi is going to work whenever I’m in a room with that many techies.

[snip]RSS Feed: <http://www.warpspeed.com/wordpress>




-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com





-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/
Powered by Listbox: http://www.listbox.com

Current thread: