Where NAT disenfranchises the end-user ...

Thu Sep 6 18:53:04 UTC 2001

|> From: Jon Mansey [mailto:JMansey at interpacket.net]
|> Sent: Thursday, September 06, 2001 11:41 AM
|> 
|> Isnt part of the solution here for ppl to write NAT-aware 
|> applications?
|> 
|> I got this idea from a bugtraq post about gnutella that is able to 
|> detect and announce a different IP address than that of its actual 
|> private host IP based on what its internet-facing public IP is.
|> 
|> Im sure there are a host of reasons why this is not a good idea from 
|> a security POV, but its a start, no?
|> 
|> Im not disputing that a NATed connection should not be sold as "full 
|> Internet connectivity", I agree, but in terms of making it look and 
|> feel like one, I think we're close.

Two software development houses, playing nice with each other, is more rare
than two ISPs doing same. The bottom-line is that it comes right off the
bottom-line. You can't deny that it's extra work/cost and effects
time-to-market. Remember, in software development, it's the second 90% that
kills the project. Consider the straw that broke the camel's back. Then
consider that 80% of all software projects never make it to market. Then
consider that most developers are NOT network engineers. They expect the
network to *be there*, period.