I think you're right Jay.  I have one, but in TX, :(  Oh well.  I don't
know if the latest generation, 128kbps units would support
peer-to-peer.  I don't know why they wouldn't, but everything I read was
about making such connections with the very early, 19.2 or so versions (At
Berkley they made a mesh I believe).  Poletop units were distributed like
every .5 miles, so range is a similar issue there.  Just going "modem" to
"modem" in a peer-to-peer arrangement, without using the pole
infrastructure...would that violate any patents?  Just using equip. you
already purchased?

Also, as for Technomage's ideas about filtering out abusers via MAC
addr.  I don't know how all the "hops" will be configured to route
traffic, but MAC is only a valid thing within the LAN typically.  Each
router hop along the way will replace that part of the IP header with its
own MAC.  Don't "block" the MAC of your upstream router! ;)  If all
traffic was routed out one, or a select few, "approved" gateways, then you
might be able to address such issues there.  Make a giant 10. network
behind and NAT out the traffic that was "net" bound.

As for the ISP terms and such, would ASU, or a similarly minded type
organization provide such bandwidth?  At the Boys & Girls Clubs, we had a
rather small CIR on a fractional T1.  They ASU was the upstream provider,
and offered the "ISP" service at a very good rate, for nonprofits.  That
might still be available, but ASU can get into trouble, as it can't
"compete" with private businesses (as a state entity).  That's a bit of a
grey area, but might be worth investigating.  I'd think a University or
something similar would be in the best position to help with Internet
connectivity.

Or, maybe AZOTO, could oversee the freenet, and either take donations for
a T1 or something with fewer usage restrictions for a primary gateway to
the net, or allow free connections to the mesh (aka freenet ;) ) but
charge some sort of fee for internet access (that could be a nightmare to
manage), or something.

Also, as for Hans and the Strawberry idea, why not take some typical,
boring, land line connections, and build a freeswan IPSEC ESP tunnel
between two say DSL or cable lines or something?  Just for routing freenet
traffic over?  It'd be slower than the 802.11 links, but would get the job
done I'd think.  That would probably be the most sane way of hopping large
geographic areas (like jumping over El Paso and right into Richardson, TX
;) ).  Then you could technically connect your mesh with the Brisbane mesh
or whoever's.

Hopefully something I've said made any sense...I need to go get my AM
caffeine now ;)

C ya,

Wes

On Fri, 28 Dec 2001, Jay wrote:

> 
> 
> Hmmm...upon further thought, using some Ricochet stuff might not be a bad
> idea for the start of a FreeNet. I am an ex-Ricochet customer - now with a
> modem collecting dust. I recall reading that the modems can also "call"
> any other modem (as long as both are within the same metro service area)
> by "dialing" the serial number of the modem you want to call.
> 
> Assuming the transceivers are still active on the light poles, it wouldn't
> be too hard to use Ricochet modems as a point-to-point link between two
> *NIX boxes. Given the robustness of Linux's networking, any given Linux
> box could do long-haul via Ricochet to link "pockets" of 802.11b into one
> valley-wide massive FreeNet. Now that would be super-cool. :)  That way a
> user could access the FreeNet via 802.11b when he/she is in a "pocket,"
> but could also access the same FreeNet via Ricochet when 802.11b is not
> available.
> 
> I'm sure the Ricochet modems can be had cheap now. I think I remember
> seeing them for ~$20 somewhere...
> 
> Anyone out there with a dusty Ricochet modem (the "GS" model that did
> ~128kb/s) with a pole-top transceiver near by want to try and set up a
> point-to-point link with me?
> 
> ~Jay