[IP] Re: Harvard, BBN Use Streetlamps to Light Up Wireless Network

[David Farber <dave@farber.net>]

Begin forwarded message:

From: Brett Glass <brett@lariat.net>
Date: May 13, 2007 4:18:47 PM EDT
To: dave@farber.net, Ip ip <ip@v2.listbox.com>
Subject: Re: [IP] Re: Harvard, BBN Use Streetlamps to Light Up  
Wireless Network

Bob Frankston writes:

> First a technical detail. This assumes that the street lamps have
> power even when the lights are off. Many lights are run in series.

Actually, the lights are not run in series (the loss due to resistance
would be too high). And standard industry practice is to put a separate
photocell on each pole. This provides redundancy; if one light goes dark
due to a failed photocell, the rest don't. Metricom, which lost a  
billion
dollars trying to build similar networks, relied on knowledge of how
these systems were wired and constructed.

Unfortunately, mesh networks constructed in this way do not scale  
well due
to problems with spectrum usage. This is even more so with systems based
on 802.11b, which has too few non-overlapping channels to prevent  
interference
between mesh nodes. The most likely result of this project will be a
semi-reliable network (with problems due to interference with consumer
devices) that covers a small area and disrupts private wireless  
broadband
services in the area. (Those services should really be allocated  
their own
spectrum, but at the moment virtually all of them have been forced, by
shortsighted policy, to use the same unlicensed spectrum as cordless
phones and indoor Wi-Fi.)

--Brett Glass



-------------------------------------------