Margaret,

Not to mention the fact that if some-one were attempting to create 
applications/protocols which worked across a NAT boundary, having a consistent 
base of expected behavior for NAT implementations would be quite important, I 
would think.

Even where systems don't need to interoperate, standards are useful as a sort 
of "best practices" document. I've seen quite a number of RFC's of that type. 
For instance, something like RFC 4864. It's "informational", not a "standard" 
and I'll fully confess that the nuances of IETF's document classification 
system are lost on me. Without an enforcement mechanism, I don't see much 
practical difference between the two. Regardless, there is alot of value with 
RFC's of that nature, simply in terms of knowledge share for 
individuals/organizations seeking to implement their own solutions to common 
problems. That improves everyone's efficiency which I think benefits the 
community as a whole.

For instance, if someone with less experience then many of us here is being 
marketed a particular solution that is compliant with an RFC standard. They can 
go to that document and read and understand exactly how that solution is 
supposed to function. The known benefits and drawbacks of that solution. They 
have a good base for understanding whether that solution might be appropriate 
for them. I think that's a very important function RFC's can provide....beyond 
any interoperability concerns.



Christopher Engel
Network Infrastructure Manager
SponsorDirect
[email protected]
www.SponsorDirect.com
p(914) 729-7218
f (914) 729-7201

> -----Original Message-----
> From: Margaret Wasserman [mailto:[email protected]]
> Sent: Tuesday, October 26, 2010 10:55 AM
> To: Chris Engel
> Cc: 'Brian E Carpenter'; Gert Doering; Roger Marquis; [email protected]
> Subject: Re: [nat66] e2e [New Version Notification for
> draft-mrw-nat66-00]
>
>
>
> Hi Chris,
>
>
> On Oct 26, 2010, at 9:58 AM, Chris Engel wrote:
> > No one is arguing for a standard that NAT MUST be deployed across
> > the entire internet. People simply want a standard by which
> > organizations which CHOOSE to deploy it on their OWN network
> > boundaries can expect it to behave consistently.
>
> Some people have argued that we don't need a standard for
> NAT, because
> NATs don't have to interoperate.  That argument doesn't really apply
> to NAT66, though, because if all NAT66 devices implement the same
> algorithm, it would be possible to put multiple NAT66 boxes at the
> borders of a site, configure them with the same prefixes, and not
> worry about whether inbound traffic for a given flow runs
> through the
> same NAT66 box as outbound traffic.
>
> Margaret
>
>
_______________________________________________
nat66 mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/nat66

Reply via email to