Re: a different perspective on NAT

Eric S. Johansson wrote:

> folks seem to be living a bit of a fantasy in that if we somehow magically
> migrated to IPv6, all of a sudden applications like speak freely or 
> BitTorrent
> would somehow work without interference.  I will argue that no matter what
> version of IP we use, security perimeter measures be it a traditional 
> subnet
> firewall or firewall plus address translation, will still be impediments to
> peer-to-peer protocols.

I could not agree more strongly with this.  While the original
motivation of NAT was conservation of address space, its side
effect--constraining peer to peer traffic among individual
users--will be and, in fact, already is seen as a benefit
by many in a position to influence the future of the Internet.
Forcing individuals to interact through a server, rather than
peer to peer, creates a central point of control over their
activity, as Napster so vividly demonstrated.  Call me a pessimist,
but I find it difficult to believe that once such a degree of
control has been exerted over the activity of individual Internet
users, it will be readily relinquished.  A political analogy is all
of the various "emergency" measures adopted by governments at the
expense of their citizens' freedom which remain in force long after
everybody has forgotten the emergency which motivated their enactment.

>  Instead, I think we should be focused on the
> development of some mechanism allowing peer-to-peer services to transit
> firewalls in a "safe" manner.  Ideally this is something that should happen
> automatically without requiring any application change but obviously 
> that might
> not be possible.  We also need to consider that it may be impossible to 
> do any
> form of tunneling or pinholing safely and maybe we need to look at a 
> different
> protocol infrastructure that can be distributed.  For example, proxy by 
> CGI on a
> Web service "owned" by a end-user.

That's the real challenge.  Voice and video, of course, are substantially
more difficult that other kinds of peer to peer traffic since they
require low latency and consistent packet delivery times.  Many proxy
schemes which might work for multi-player games (something I know
*absolutely nothing* about, so I may be speaking nonsense here), would
not be able to provide a suitable quality of service for audio or
video, and the bandwidth requirement (which I worked out in the EOL
announcement) is a real challenge for potential server sites, which
makes some means of distributing it very attractive.


                      * * *

To unsubscribe from this mailing list, send E-mail containing
the word "unsubscribe" in the message body (*not* as the
Subject) to [email protected]