<snip>
> Without that, the impression is that everyone is just silently agreeing
> (other than those who say that the draft can't change because it already
> exists) with the points being made, but I am pretty sure that's not the
> case.  But I have no idea at all why.   Maybe someone can convince me that
> there's actually some benefit that I just can't see.
<snip>

I, for one, am silently agreeing with you.  As best I can tell, here are the
arguments involved (I will present each argument followed by my opinion on the
argument):

1) The u bit only makes sense if we keep the 64-bit boundary, so we must keep
   the 64-bit boundary just in case we one day figure out how to use the 
   u bit.

   This argument should be removed because even Michael Py states that it is 
   a red herring.  I.e., even Michael Py states that the strength of his 
   argument for keeping the 64-bit boundary does not rest on the possible uses
   of the u bit.

3) EUI-64 is good.  It shouldn't be removed!

   No one is advocating this.  In fact, what we are advocating does not have
   any influence on how anyone else chooses to operate.  I am simply asking
   for the freedom to use the last 64 bits in whatever way I choose.

3) The draft should not be updated.  Instead, stability is crucial.

   The draft is already being updated.  Furthermore, removing the 64-bit
   boundary does not appear to have an affect on any existing implementations.

3) Current implementations do not enforce the 64-bit boundary, nor do they 
   enforce the uniqueness implied by the u bit.

   I am slightly unclear about these two points.  I have seen that 
   implementations do not inforce the uniqueness implied by the u bit; 
   however, it has already been decided to ignore the u bit in this argument.
   I am not sure if implementations forbit prefixes longer than 64-bits.
   Based on what I have read on this list, they don't.  If this is true, than
   it is fair to say implementation experience has proven that the 64-bit
   boundary is not necessary.  Implementation experience is a very valid 
   reason to update the draft.

4) Current practice does make use of prefixes longer than 64 bits.

   This increases the strength of the argument in number 3.  Current practice 
   should indeed have an influence on whether or not this draft should be 
   updated.  In other words, because using a /126 has been proven to be safe
   for point-to-point links, there is no reason to forbid it.

4) Allocation efficiency for point-to-point links.

   Some people feel that the the overhead of using a /64 for a point-to-point
   link is small, and quite justifiable if it simplifies things.  Others feel
   that space should not be wasted, and that any simplification achieved is 
   not justifiable, in this case.  I feel that the amount of time you spend
   arguing a point should be based on the relative importance of the point.
   Having said that, I'd like to say...nothing (i.e. the importance of this
   point does not justify further argument) :)

5) I want to have maximum freedom with my /64.

   This is the most important argument, in my mind.  I don't think any benefit
   gained from establishing a 64-bit boundary can possibly justify loosing 
   the freedom of not being confined to that boundary.  I think that this 
   point will become increasingly important in the long run.  Furthermore, I 
   think any arguments in favor of the 64-bit boundary will likely become less
   important in the long run.

6) Not establishing a 64-bit boundary will make the lives of ASIC designers
   difficult.

   I know nothing about designing ASIC's.  However, I presume that the header
   chaining used in IPv6 is an even more difficult challenge for ASIC
   designers.  Hence, I doubt this argument has real merit.

7) Using a 64-bit boundary may permit certain memory optimizations since 
   every prefix will be at most 64 bits long.

   This argument has been disproved in this thread twice.

In summary, I feel that experience has shown that the 64-bit boundary is 
unnecessarily confining and should be removed.

Best Regards,
-jj

-- 
o*s*o*pho*bi*a n. A common fear among embedded systems programmers.
--------------------------------------------------------------------
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