whoa

[James Rogers]

On 9/8/02 3:16 PM, "Gary Lawrence Murphy" <garym@canada.com> wrote:
>>>>>> "J" == James Rogers <jamesr@best.com> writes:
> 
>   J> An example: Being able to model RF propagation in three
>   J> dimensions for a metro area when deploying wireless networks.
>   J> By having every single tree and building detail and similar,
>   J> you can "see" even tiny dead spots due to physical blockage and
>   J> signal attenuation.
> 
> Hmmm, just as I thought.  In other words, it has no practical uses
> whatsoever ;) ... do the biz guys in your office /really/ think WISPs
> are really going to shell out /their/ money to find a house or two
> they can't reach?  Experience suggests (a) they won't care and (b)
> they will even sign up that errant house and then give them a
> run-around blaming the dead-spot on "unsupported vendor equipment".


Errrr....the biz guys in my office don't care what the "WISPs" want to do
with their little WiFi networks.  And the bandwidth shadows in most cities
are surprisingly large and common.  They aren't selling the software, which
is pretty pricy as it happens.  They are using it to optimize next
generation wireless canopies over metro areas and fiber networks on a large
scale.  There are an essentially infinite number of metro wireless
configurations, some of which generate far more dead or marginal spots and
others which are very expensive to operate (due to backhaul transit
considerations) or both.  This software can be used as a tool to optimize
the canopy coverage and minimize the actual transit costs since the wireless
is tied into fiber at multiple points.

The canopies we are talking about aren't short-range wifi technologies, but
a mixture of long-range high-performance wireless networking, with bandwidth
measured in tens to hundreds of mbits and ranges measured in miles (up to
well over a hundred miles on the extreme end).  At those ranges and
bandwidth levels, the cost of providing the network can easily vary by an
order of magnitude or more depending on how you manage RF shadows and
proximity to fiber access points.  The idea ultimately is to optimize the
cost and performance such that no existing network infrastructure providers
can remotely compete and maintain profitability.  This is a surprisingly low
bar, and it is about time networks were designed with this level of
large-scale optimization (cost per mbit, maximizing coverage, and effective
bandwidth available per unit area) in any case.  And for this company's
long-term plans, this type of capability will be absolutely necessary to
keep things sane.

Or at least investors find this capability very sexy and compelling,
especially since we have this lovely visualization engine tied into the
system (CLI batches never have the same effect, even if it is more
efficient).

-James Rogers
 jamesr@best.com