[e2e] using p2p overlays to overcome recursive NATs/realms

[Christian Huitema]

> >On the other hand, there are different ways to do
> >this, with different engineering results. Trying to restrict the
number
> >of addresses has been proven to backfire: you get the deployment of
home
> >NATs, the ISP does not do more business, and the network becomes
> >brittle. Why don't we look at this as an engineering issue, and
provide
> >the community with better tools?
> 
> Maybe I'm not saying it well, but this is the point I'm trying to
make.
> 
> However, there *is* constant pushback against better tools from
> engineering
> folks who seem to think that their operator customers and employers
won't
> buy into better, more flexible tools at any price, and prefer to
follow
> the
> "evil" strategies like selling IP addresses at high prices, or
refusing to
> sell them at all, blocking ports, restricting "servers", etc.

Dave, you should really cool down a little. I have yet to see network
engineers refusing to use better tools when the tools are available (and
are actually better, indeed). I have seen a number of marketing
departments engaging into misguided strategies, but the solution is
seldom to brand them evil -- dialogue tends to work better. It is much
easier to get the right decisions once the technical options are clearly
explained.

-- Christian Huitema