I totally agree with Bob. Thanks to him for also providing the references. Hemant
-----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Bob Hinden Sent: Friday, March 28, 2008 2:34 PM To: Fred Baker (fred) Cc: Iljitsch van Beijnum; IETF IPv6 Mailing List; Brian E Carpenter; Pekka Savola Subject: Re: RFC3484 destination address selection rule 2 is buggy Fred, > > On Mar 19, 2008, at 4:56 PM, Brian E Carpenter wrote: > >>> - use a ULA source address if and only if the destination is a ULA >>> in the same prefix >> >> I think that is broken. There's a reason ULAs are defined as global >> addresses. > > but they are *not* global addresses. If they were, they would be UGAs, > not ULAs. The are defined as having global scope. They are unique and valid globally. They usually don't have global reachability. From RFC4193: >> 3.3. Scope Definition >> By default, the scope of these addresses is global. That is, they are >> not limited by ambiguity like the site-local addresses defined in >> [ADDARCH]. Rather, these prefixes are globally unique, and as such, >> their applicability is greater than site-local addresses. >> Their limitation is in the routability of the prefixes, which is >> limited to a site and any explicit routing agreements with other >> sites to propagate them (also see Section 4.1). Also, unlike site- >> locals, a site may have more than one of these prefixes and use them >> at the same time. The issue is about reachability, not scope defination. > What they are is local addresses with undefined scope. Routing may or > may not be stable between a system using a ULA and another system that > doesn't have an address in that prefix. The only way for the system > using the ULA to know is for it to have configuration that tells it, > or to receive a message from the remote system (one that it can know > with assurance doesn't have a spoofed source address) that is sent to > its ULA address. > > Hence,if it is selecting an address, it should go by the Robustness > Principle. It should accept a message from the peer sent to its ULA > address, but it should itself do something it has assurance will work > correctly. It should only send from a ULA if it knows that the peer > has an address in the same prefix, and it can only know that by > configuration or by observation that the peer is using the same > prefix. I think I am in general agreement of the consequences, but it isn't about scope, it is about reachability. Anytime there is a a choice of addresses to use, the same issue come to play. It's hard to know about reachability with out prior knowledge or trying it out and seeing what works. This applies to ULAs and other global scope unicast addresses, IPv4 and IPv6 addresses, etc., etc. Bob > > -------------------------------------------------------------------- > IETF IPv6 working group mailing list > ipv6@ietf.org > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 > -------------------------------------------------------------------- -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 -------------------------------------------------------------------- -------------------------------------------------------------------- IETF IPv6 working group mailing list ipv6@ietf.org Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6 --------------------------------------------------------------------
