crazy project (was Re: Good ISP)

[Rusty Carruth]

> 
> It will remain this way until people get upset.  Most people just think
> that its this way and thats life.  So they cope.  Windows is an example of
> this.  I mean do you really think people like the blue screen of death?
> 
> ...
> I suppose some day they will make it so powerlines can do broadband then
> you get another great company like SRP or APS to choose from. ;)
> 

I'm seriously considering inventing an infrastructure that would bring back
the 'good old days' of the internet:

1 - redundant so that if one link breaks you can still get through (so
	it can survive a war - this was one of the original design goals
	of the 'net).

2 - cooperative

3 - un controllable by any one agency/agent/company/whatever.

4 - minimal setup costs, minimal monthly costs.

In fact, its actually more like the days of uucp than the days
of direct connection to the internet, except that you've got
lots more bandwidth.  (Some of us old pharts may remember
the days of uunet and uucp, ask one and they may or may not
talk your ears off ;-)


The idea is based upon using a grid, where each member of the
'gridnet' (just made that up ;-) talks (at something like either
10MegaBit or 100MegaBit) to their 4 neighbors, who talk to
THEIR 4 neighbors, etc.

The problems to be solved (that I know of right now, anyway ;-)
are:

1 - routing.  How do you get a packet from x,y to w,z, since
	there are a LOT of ways to get there?

2 - addressing.  How do you address all these nodes?

3 - latency. With so many nodes that I expect to be involved
	in one packet's travel latency is  a big issue.

4 - tying the grids together.  How do we bridge the gap
	between the different local grids? 

5 - connection to the existing internet.  Until and unless
	the current internet goes away, we'll need some way
	to talk to the 'rest of the net'.

6 - business plan.  How is this going to happen?  Start a
	business that tries to make money selling the interface
	boxes?  (What about startup costs???  Advertising costs???
	etc) Or start an 'open hardware' project that develops
	the protocol and the boxes and gives it all away? Or
	what???  (this includes financing issues!)

7 - 'news feeds'.  This ties in a lot with #5, and I expect that	
	for the most part we can assume that, as in the old days,
	people will share news with each other.  However, I'm
	a bit concerned about storage and bandwidth.  Perhaps
	news server companies will decide to charge less than
	$20/month for a news feed!

So, does anyone want to be involved in such a crazy project?
I *think* I've got some reasonable ideas for problems 1, 2,
3, 4, and 5, but that leaves 6 & 7 pretty much unsolved,
and until my ideas are verified as workable I'm not sure
I've totally solved them (and I *know* I'm not all the way
there with 1).

If there is interest, I'll set up a mailing list somewhere
and we'll start discussing the issues...  Just please reply
to ME not the list, as I usually save the list traffic
directly to a folder without reading it till MUCH later,
if I manage to find time...

rustyc <at> descomp <dot> com  (not that that helps much!)

(We DO teergrube incoming mail, but I've dropped my time 
down to 24 hours since I got lockups in 'antispam' with
the 1 week delay...)