Hi,
The disadvantages of NAT is pretty well known but I was trying to see if I
can extract any feature of NAT that is useful. I am not saying that use NAT
as it is, but see if we can learn any thing positive from it, beside saying
it has problems. These are the two things which I think we can learn from
NAT:

i) I see discussion here of the application trying to decide the address. I
don't think it is a good idea to couple application with addresses, (Nat
being one example, though some may say invalid example, it is a real life
example.).  For example in IPv4 world if we had decided that the application
had to use mac address if the peer is in same subnet else use the IP
address. I think it is better to avoid such concepts unless we see some
other big advantage out of it.(Perhaps you also don't like the idea of
application dealing with addresses).

ii)If you can model NAT in a different way, you can see that it gives a
different way of route aggregation. An enterprise might be using lots of
private address but so many private address are aggregated into few public
address. This is a different way of address aggregation at least very
different from what I am aware of. This address aggregation as existing
currently in NAT is certainly a problem as the private addresses overlap but
this architecture can be modified

I think if we can modify the applications and IP stacks (as we have to do
for IPv6), the problems of NAT can be solved too.  I also think not many
people will like to modify there system for NAT. (An example for solving NAT
problem: To make the address unique, the application passes both the private
address and the public address to the peer. The peer uses both the address
in the packet and on the receiving side the gateway changes the public IP
address to the private. If you consider this two address in pair it will be
unique. I am aware there are other issues, but they can be solved too.)

Vijay

----- Original Message -----
From: "Keith Moore" <[EMAIL PROTECTED]>
To: "Vijay Amrit Agrawal" <[EMAIL PROTECTED]>
Cc: <[EMAIL PROTECTED]>
Sent: Friday, June 28, 2002 7:32 PM
Subject: Re: Maybe: IPv6 Scoped Addresses and Routing Protocols


> > Why not try to make IPv6 such that, it can work well even when
> > technology like NAT is deployed i.e. the end address is different from
the
> > transport address.
>
> the problem with NAT isn't just that the address changes in transit;
> it's also that NATs split up the network into multiple addressing
> realms and make addresses ambiguous.  there's really no way to solve
> that problem.
>
> Keith
> --------------------------------------------------------------------
> IETF IPng Working Group Mailing List
> IPng Home Page:                      http://playground.sun.com/ipng
> FTP archive:                      ftp://playground.sun.com/pub/ipng
> Direct all administrative requests to [EMAIL PROTECTED]
> --------------------------------------------------------------------

--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to