I agree with Francis' comment on the /35 allocations.  If the /29
allocations were made in full from the outset there would be less
pressure at the other end...

In the case of a national educational network (my particular interest)
we currently have just 13 bits of network space to allocate to
institutions, if the institutions (Universities) get a /48 each.  At 
the same time it is bizarre that a University might get the same address 
space as a small end customer.  It is also likely that some Universities 
will command more than one /48, particularly if highly distributed.
If we have single (or multiple) subnets per room (at worst a cheap 
wireless access point per room), that /48 will look small very fast.

At present, the foreseeable allocations seem to offer little room for
the second-layer ISPs (e.g. the Universities).   It's almost as if we're
being painted into a corner by the initial 2001:: allocation and the
committment (rightly or wrongly) by the registries to ensure there are 
only 8K (13 bits) of top level routes in the DFZ, while at the same time
hedging their bets via the /35 strategy.

The last thing we want is end users running IPv6 NAT.

tim

On Fri, 7 Jul 2000, Francis Dupont wrote:

>  In your previous mail you wrote:
>    
> => many ISPs want to allocate /64 (or worse) to their customers... and
> shout a /48 per customer is far too large. I believe this is a consequence
> of the slow^N start, ie. the /35 rule (RIRs trim address space of ISPs,
> ISPs take back the burden to their customers).
>  The idea is to introduce a small site (/56) for "poor & little" customers
> and to make it the *default* allocation. IMHO this is an acceptable target.
> 
> Regards
> 
> [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