On Sat, 27 Jul 2002, Matthew S. Hallacy wrote:

> <bold statement>
> I don't think it matters what anyone wants.
> </bold statement> <clarification>
>
To some extent this is true.
What _can_ be done is independent of what people want.

> Let's examine the 2 topics being discussed:
>
> 1) Overlay network
> 2) Hotspots
>
> What does #1 provide?
>
> a) A twin cities-wide network where content can be shared (such as the ftp
> mirrors at real-time).
> b) internet access via various access points that choose to have a default
> gateway and a route to the 'net
> c) hotspots that are not in-range of the rest of the wireless
> network, but are connected to it via tunnels over the internet (ie: GRE)
>
I am becoming even more convinced that this is a prerequisite
to #2. What good is a hotspot with no connectivity?

> What does #2 provide?
>
> a) internet access via various access points that choose to have a default
> gateway, and a route to the 'net
> b) connectivity to other hotspots via tunnels over the internet (ie: GRE)
>
> Now, what have people expressed a wish for (owner viewpoint):
> 1) reimbursement for cost of equipment (non profit)
This makes sense for shared equipment, and there are ways to
work with this.
> 2) profit (make money, isp-style)
That would involve another group. Any "profits" from the
operation of a TCWUG wireless network should go back
into the WUG and the network. It would be nice for the WUG
to have a small budget for community outreach, presence
at community gatherings (fairs, parades, etc.).

I don't see profits coming out of this though, we will
be best served if we simply try to share the _costs_
of the connectivity among ourselves.
And it is not going to be a cheap project, though there
are opportunities for cost savings in quite a few
areas.

> 3) connectivity to other nodes

This is the core of the whole matter.

>
> What do users want:
> 1) connectivity
> 2) alternative to slow downloads over the 'net that are from sites that
> are local (ie: real-time's ftp mirror)
I do not think that a TCWUG network should have "users" as such
until we decide that we are ready for them. Setting up a network
to be used by a relatively small group of geeks is much different
from setting up a network for general public use.

I mean, how many regular users could setup IPv6 on MSWindows?
>
>
> Conflicts:
>
> The /only/ conflict I see is where it comes to people profiting from their
> access points, I can see where people would have an issue in the following
> scenario:
>
> User connects to access point #1
>
> Access point #1 connects to <TCWUG wireless network>
>
> User gets on the net via gateway Z, which is charging User for bandwidth.
>
> access point #1 feels used because he's not making any money off it.
>
>
> Personally, I do not feel that there's any profit to be made from a network
> like this, I've worked for an ISP that tried, they are now bankrupt.

I agree with this, which is why I keep suggesting the Co-op model.
The best we can reasonably hope for is fair cost sharing. If we
happen to do better than that, cool, but nobody should be getting
hung up on "profitability" or on anyone other than the upstream
connectivity providers getting more out of the system than they
put into it.

>
> I do believe it's completely possible to build the network, keep it stable,
> and very usable. I see the network as an overlay network regardless of what
> anyone calls it (car, truck, motorcycle.. they're all still an automobile)
> different access points will decide, on their own, weither or not they're
> going to allow internet access from their access point. There's no need
> to discuss it, because it's not a group decision. The overlay vs. hotspot
> question is moot, access points that wish to connect to each other will,
> access points that either cannot, or will not, connect to the rest, won't.
>
> Any questions, or something I missed?
>
> What I Think We Need To Figure Out:
>
> Mobile IP
> Global Routing Protocol (BGP?)
> Global Network Authentication (ie, nocatauth or something custom)
>


-- 
Daniel Taylor
dante at plethora.net