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

Aaron Falk falk at ISI.EDU
Mon Feb 11 17:12:54 PST 2002


--John Wroclawski wrote:
> Not that I disagree with you, but a couple of thoughts are a) better
> tools may provide more _different_ ways for ISP's to make more money, and
> some of them might have fewer unintended/undesirable side effects, and b)
> more tools might make it easier/cheaper for some other ISP to
> compete/exist/be cheaper, thus keeping prices down through competition.
> 
> This raises the question of what the tools are, and whether we can ever
> get to the point where the tools/economics/regulations support each
> aspect of ISP operations being as competitive as possible rather than
> locking everything into the least competitive part of the chain. But it
> does seem like tools might change the outcome somewhat, if not the desire.
> 
> John

So here's a naive notion that might qualify as a tool:

Suppose Linksys, D-Link, SMC, etc started making 6to4 boxes similar to the
NAT-ing 'DSL routers' they make today: idiot proof, self-configuring, and
-- most of all -- cheap.  

This could create a market for a presence-ISP service (as opposed to an
access-ISP service) which would give each user 128 global, public v6
addresses and terminate their 6to4 tunnel.  Who would use it?  Users who's
apps break under NATs (compiling a list might be helpful).  If
presence-ISPs can make money, it would demonstrate a market and the big
ISPs might see this as either a competative edge or as a value-added
service.  

Some might say it should be bundled with the basic ISP service but I can't
even *buy* consumer ISP access from my RBOC (Verizon) or cable MSO
(Adelphia) with fixed or multiple IP addresses and would pay for this
service.

What would it take to make this happen?  A compelling statement of apps
that break under NATs that would work with 6to4 might encourage the
investment necessary. 

--aaron



More information about the end2end-interest mailing list