Hi Nick,

On 24/08/11 1:06 AM, "Nick Hilliard" <[email protected]> wrote:

> On 23/08/2011 12:56, Arturo Servin wrote:
>> On 23 Aug 2011, at 03:20, Terry Manderson wrote:
>>> Just as 169.254.0.0/16 is intended as Not Routable. Do you disagree with
>>> that?
>> 
>>       No, but you cannot tag this one in IANA IPv4 Address Space Registry.
> 
> I'm not convinced by this ID.  The "Routable" column isn't flexible enough
> for the IANA v4 allocations list because IANA deals with multiples of /8

Actually IANA deals with any prefix length it is told to by the IETF. The
smallest one being a /24 (documentation prefix) the largest one being a /4
... in IPv4 terms.

So far I could not find any RFC which specified in any way what the prefix
length of the entry in the IANA registry should be. Certainly the RIRs are
allocated in /8 chunks. No denial there. And also given that the IETF
reservations vary in prefix length. My guess is that the prefix length
selection was arbitrary for parsimony. But if anyone has a specific pointer
It would be most appreciated.

> (for ipv4 allocations), while individual non-routable prefixes can be as
> small as /24.  This alone calls into question whether the basic premise of
> the draft is worth pursuing.

I think it justifies it more. If everything was considered routable, and no
differences of prefix length existed then there would be no point. But that
just isn't the case.

> 
> Would it not be better to approach this problem from a completely different
> direction?  E.g. to formally recommend that end users consider everything
> routable except address space defined in RFC 3330 + its successors + the
> equivalent for ipv6?

In a way this draft does exactly that. Just explicitly - and in truth that
is always what I thought was the great thing about the IETF. If you are
going to write something, or change something - make it explicit.

> And then perhaps to put in some scary wording about
> the terrible consequences of using address space on your network that has
> been assigned to other organisations?
> 

That is a uniqueness point that I'm not addressing.

Cheers
Terry

_______________________________________________
GROW mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/grow

Reply via email to