Couple more thoughts ...
>From: Paul Vixie [mailto:[EMAIL PROTECTED] >Sent: Tuesday, June 12, 2007 17:54 >To: 'IETF IPv6 Mailing List' >Subject: Re: Revising Centrally Assigned ULA draft > >> ... It is meant to be a private address space, to be routed / >> routable as private address space should be - specifically NOT on the >public internet. > >what we mean it to be routed to is less important than what people who >use it will actually route it to. let's focus for now on examples >involving many networks run by folks with diverse goals. any time you >propose a rule like "not meant to be XYZ" you have to be able to say how >that rule will get enforced, and what the internet will look like if >enforcement doesn't happen. Fair enough, and certainly the world is full of lots of different people / environments, all with diverse goals. Some good, some bad - many subjective :). Is simply recommending that providers not accept reachability announcements for ULAs, the same way it is recommended that they not accept RFC1918 announcements, not good enough (and for the same reasons)? > >> ... and, golly, can't we be smarter this time around? > >so far, not. > I wouldn't be that negative; the fact that we are attempting to solve a problem is a good sign, no? (providing a method to accomplish a nearly-collision-free private address space ... which (to me) seems to be a noticeable improvement.) >> >Should routers not forward ULAs under any circumstance? >> >> Routers, meaning any routers? Of course they should - the whole point >> is to have private IPs that are routable (unlike link-local >addresses). > >that's another voice heard from. > I think I am merely passing along the sentiment of the authors of the RFC, albeit poorly paraphrased. To quote: Abstract This document defines an IPv6 unicast address format that is globally unique and is intended for local communications, usually inside of a site. These addresses are not expected to be routable on the global Internet. That says, to me anyway, that they are to be routable - but not globally routable. >> Routers, meaning out in the DFZ - of course not, this is private >> address space. (Except maybe to black-hole them ,that is) > >is the dfz the only place these routes shouldn't go? how will this be >enforced, if cooperating connectees to the dfz all want to do it anyway? > >if we can't agree on "what's a site" then can we ask "what's >``private''" ? I am not fully sure we need to define "site". Even in the abstract above (wherein it mentions "site") the meat of the statement seems to focus on the "private" aspect, which I believe we have recommendations and policies in place to manage (again, a la RFC1918). Does everyone follow them, no. Are they perfect, no. Do we have enough policy, and experience with that policy, to move forward - I would think so. ... ? Again - just MHO. /TJ -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------