On Wed, 1 Feb 2012, Brian E Carpenter wrote: > On 2012-02-01 08:14, Pete Resnick wrote: > > On 1/31/12 11:59 AM, George, Wes wrote: > >>> From: Noel Chiappa [mailto:j...@mercury.lcs.mit.edu] > >>> > >>> Is that wise? I thought (IIRC, and maybe I'm spacing) the > >>> whole reason for allocating this space was that 1918 space > >>> _couldn't_ easily be used for CGN because there were too many > >>> conflicting usages. > >>> > >> [WEG] yes, but the general sense I got from the ensuing discussion was > >> that no one expects anyone to actually adhere to that advice (ie MUST > >> NOT be used as substitute for 1918 space), and as soon as the space is > >> released, it'll be "cats and dogs living together, mass hysteria..." > >> because everyone and their cousin will start using it as 1918-bis > >> anyway, no matter whether the IETF wags their fingers at them or not. > > I have no doubt that this space will be (mis)used as additional > private ambiguous address space. But IMHO the text should make it > clear that this is the wrong way to use it and give the reasons > why - basically the same information as in the new text, but stated > exactly the other way round. For example > > Shared Address Space is IPv4 address space designated for Service > Provider use with the purpose of facilitating CGN deployment. > Shared Address Space is not intended to be used as additional [RFC1918] > space, because either or both of the following issues might arise: > > o Shared Address Space could also be used on the Service Provider side > of the CPE, with overlapping subnet or host addresses. > > o Some CPE routers behave incorrectly when using the same address block > on > both the internal and external interfaces. > > > Speaking as one of the bozos^h^h^h^h^h ADs whose comments (and suggested > > text) ended the document up here, let me suggest the slightly less > > pessimistic view from Wes's, and the reason that I think this > > *shouldn't* specifically update 1918: > > > > This *is* a special use address block that is akin to 1918. It is > > non-routable address space, just like 1918. But unlike 1918, this block > > is defined as "might be used by your ISP on your outside interface". So, > > people using it inside their networks (which, I agree with Wes, will > > happen, and like everything else on the net, will be done stupidly by > > some) have been told that this is *not* private use space and that they > > use it at their own risk and their CGN service might stop working if > > they use it in a way not described in this document. But I'd hate for us > > to allocate space to "CGNs only" when it's obvious that this can be used > > for a whole class of these sorts of things, and can be used by other > > people who build sane equipment that understands "shared" addresses can > > appear on two different interfaces. These aren't "private" addresses a > > la 1918, they're "shared", so it's not an addition to that space. Let's > > properly document what it is we're doing, giving people fair warnings. > > Exactly, hence my suggested text above.
+1 //cmh _______________________________________________ Ietf mailing list Ietf@ietf.org https://www.ietf.org/mailman/listinfo/ietf
