On Mon, Nov 24, 2008 at 15:46, Margaret Wasserman
<mrw at lilacglade.org> wrote:
On Nov 24, 2008, at 5:56 AM, Eric Klein wrote:
We need a team made up of both sides to sit down, spell out what are the functions of NAT (using v4 as a basis) and then to see if:
1. If they are still relevant (like number shortage from v4 is not the same issue under v6 for example)
2. Do they already exist in v6 without adding NAT
I know, I was one of the co-authors of that RFC. But there seems to be a differnce of opnion as to how well we covered all pervieved needs - otherwise this discussion would not be happening and everyone would say "oh yeah we don't need on NAT"
But as there are people that have a percieced need for topology hiding and port translation lets look at this compleatly and build what is needed the first time (and prevent the need for BEHAVE to rebuild it later)
Then we need to check:
1. Is there is a solution by using NAT
2. Is there is a better solution than using NAT
#2 was done in the gap analysis section of RFC 4864.
I'm not sure what you mean by #1, because if you start with a list of the functions of NAT, the fact that NAT can be used for those functions just follows, doesn't it?
#1 asks that if NAT will not fit the real need, why use it to fit the percieve need.
Only then can we make a proper and informed decission on what is needed and what is unneeded legacy.
I think we are ready to do this, based on the information in RFC 4864. Do you see anything missing from 4864 that needs to be analyzed further? If so, could you send specific points, and perhaps we can consider an update?
All of the points that have been made in this chain point to diffrent percieved requreiments, and there are multiple proposed solutions being tossed around.
