Regards Brian Carpenter
On 05/08/2012 16:17, Michael Thomas wrote: > Brian E Carpenter wrote: >> On 05/08/2012 07:58, Ray Hunter wrote: >>> I disagree. The context of my message is that there should be some >>> identifier that can disambiguate the namespace per Homenet. >> >> That's what I meant too. The only point is to avoid ambiguity in the >> namespace. The only reason for using a ULA prefix to create a unique >> identifier string is to avoid the bother of inventing a new method of >> creating a unique identifier string. The name has no relation to >> addressing or routing whatever. If you prefer some other way of creating >> a string with a very high probability of uniqueness, that's fine. > > So I think I understand the point of a ULA like suffix. However, that doesn't > translate into a name that anybody would use. Locally, you could just have it > as part of the DNS search list, I suppose, but that's limited to resources > that are accessed locally which .local more or less works now. So it seems a > small gain with some ugliness to boot. I don't think that elminating ambguity in scope is a small gain. The best defence we have against ambiguous addresses is an unambiguous namespace. > What I really want though is to have resources that I can access regardless of > where I am though. I don't see how this is of any help at all for that > problem. At least it will *tell* you that the FQDN is meaningless where you are, if you are out of the homenet. > The only thing I see is that either I make the name globally accessible -- in > which > case I'd use a real DNS name, Absolutely - a local namespace is always second best to a genuine FQDN. > or to make myself topologically part of the .local[site] > namespace. Right - and giving it a unique name can only help in that process. Brian Isn't that what this all boils down to? TANSTAAFL? > > Mike > > PS: the 40 bits of a ULA was constrained by v6 prefix length. if we were > to do this > for sitelocal, we'd probably want to add some bits to better cope > with the birthday paradox. > iirc, 64 bits is a lot safer. > _______________________________________________ homenet mailing list [email protected] https://www.ietf.org/mailman/listinfo/homenet
