Ricochet modem site

[Mike O'Dell]

at the day  job, we invested in the Ricochet system
(can't remember the company name - not enough coffee yet).
Ricochet II was in deployment when it shut down. The 
v2 system was quite a bit better in that it had more
radios, some in a different band for the backhaul
fabric. That said, the v1 system was my only connection
to the net when i was down for the better part of 
2 months when i had my back surgery. in 1998, the net
was not as squanderous of bandwidth as it is now and
the v1 modem got a solid 19.2kbps almost all the time.

the big problem was the deployment economics. all the
modelling everyone does assumes average terrain looks
more like Silicon Valley or Oklahoma City than
Northern Virginia. They were very rudely surprised
that getting good coverage in NoVA required twice
as many poletops as the models had predicted.
When deployed at sufficient scale, the average poletop
density *would* look a lot more like the models
predicted, but that significantly changes the amount
of capital that must be invested before it becomes
clear that people will subscribe to the service.
and if an area ramps slower (even if just at first)
than modelled, the customer acquisition costs 
can skyrocket for that region since it is a region
by region rollout so customer acquisition marketing
has in one area doesn't have big impact on other
areas. 

when it became clear the rollout of the V2 network 
in the 35 NFL cities was going to take about $300mm 
before it started flattening to cash-flow break-even,
not even Microsoft had the stomach for that kind of
front-end-loading.

	-mo