Margaret,

> Margaret Wasserman wrote:
> - GUPI addresses may also be used to communicate over
>   private links with other GUPI-addressed networks.
>   In other words, several companies may use GUPI
>   addresses to communicate with each other over
>   a shared extranet.  These types of networks are
>   quite common in some industries for suppliers/
>   customers or data center/clients.  This wouldn't
>   and shouldn't require that multiple companies
>   share a GUPI prefix, just that they have routes
>   that point to each other.

Yes.

> - You may have different "levels" of GUPI addresses within
>   a single network...  Some devices may use addresses
>   that are filtered at the department level, some
>   may be filtered at the corporate level, and
>   some may be filtered at the extranet level, for
>   example.

Yes, but this is not specific to GUPI.

> - Some companies may pay their ISPs to globally route their
>   GUPI addresses.  I know that some people don't
>   want this to be possible, but I'm not sure why.

Explosion of the routing table. No-no.

>   I agree that we should only advise this if we can
>   come up with an aggregable method for allocating
>   GUPI addresses.

I disagree. We should not advise this for any reason.

The reason I proposed a method for aggregatable GUPIs is for the RIRs or
whoever would assign these addresses to get a jumpstart at doing it, as
it would be very similar to the final goal which is globally unique
*and* globally routable. But one problem at a time. We are not ready for
global PI yet, aggregatable or not.

Let me make myself clear that I will sabotage my own aggregatable
proposal if there are no guarantees about the non-routability.

The one thing that won't fly is to pervert the use of FEC0::/10 for
globally routable purposes. It is not why IANA allocated that prefix. It
would be simpler to ask for a new prefix, when time has come.

Besides, I will point out that if GUPIs become globally routable, people
that wanted private addresses will use 2002:0A00::/24 and we will be
back to square one except that we would have created a big PI mess in
the global routing table.

The fact that globally unique site-locals are aggregatable or not does
not change the need to enforce their un-routability.

What could reach consensus today is GUSL: Globally unique site-locals.
- Globally unique, free, no registration (Charlie Perkins)
- GUSL blackholed by default (Bob Hinden)
- GUSL BGP routes discarded by default (Michel Py)

Optional, if it reaches consensus and if someone implements them:
Globally unique, geographically aggregatable, registration needed. Lots
of ifs, should not delay process if consensus is reached on what is
above.

One problem at a time, please.


> There will continue to be application-level and mobility
> issues with these addresses, or any type of private or
> filtered addresses. The problems are reduced by the fact
> that the addresses are not ambiguous, but the problems
> are not all eliminated. However, it seems that people
> _will_ use filtering to create private networks. The
> best we can do is try to provide a solution that
> mitigates the damage.

I agree with the assessment, time to be realistic in what could reach
consensus.

Michel.


--------------------------------------------------------------------
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