Scoping Scoped Addresses

[Bob Hinden]

Please excuse the pun in the title, but I wanted to get your attention :-)

[Working group chair hat on]

I have been trying to make some sense of this discussion. The only obvious conclusion is that there is not a consensus in the working group on how site-local addresses should be used.

Some people think that site-local is an important feature with many uses, others think they are bad and should not be used. Some think they provide security, some do not. Some thing they help with renumbering, some do not. Some thing they help avoid IPv6 NAT's, some think they encourage IPv6 NAT's. Etc., etc. The only thing that is clear is there are a significant number of people who have different views on this topic. It's doubtful that one group will convince the other group. One positive result of the discussion was that the issues and benefits are better understood. The real question for the working group is what to do next.

Now that the IPv6 Address Architecture was approved as a Draft Standard and the Default Address Selection document was previously approved, we have site-local addresses in IPv6 and a set of default rules for how an implementation selects them. What we don't have is how they should be used or not-used.

Even though there is no consensus on how site-local addresses should be used, I think there is enough people who want to use site-local that it is reasonable that the w.g. should continue trying to flesh out site-local usage as well as pitfalls of usage.

Here is a proposal for how to proceed from here that tries to take into account both sides of the discussion.

1) Node Requirements should not require any multi-site implementations. The only site-local requirement should be limited to what is currently in the address selection rules and for routers to configure site-locals just like any other unicast prefix. Vendors are free to go beyond this in their products, but the IETF won't require it.

2) People who think the usage of site-local is harmful should write a draft titled something like "Use of Site-local Addresses Considered Harmful". The people in the other camp can comment to make sure the arguments are accurate.

3) People who want to use site-local addresses should work on completing the "IPv6 Scoped Address Architecture" document (and other docs if needed). I think a good focus for this would be to focus on the simplest cases. Topics to cover need to include site border routers, adding site-local addresses in the DNS, routing protocols, the use of firewalls to enforce site boundaries, and guidelines on how applications might want to select between global and site-local addresses. The people in the other camp can review this work and make sure the technical content is accurate.

I believe this approach should help provide the larger community (e.g., vendors, ISP's, enterprise operators, etc.) the information they need to make an informed decision on the usage of site-locals.

Bob

p.s. I will also send out a few personal comments on site-locals in a separate email.

--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------