Re: [v6ops] Scope of Unique Local IPv6 Unicast Addresses (Fwd: New Version Notification for draft-gont-6man-ipv6-ula-scope-00.txt)

[David Farmer <[e-mail]>]

On Sun, Feb 14, 2021 at 1:57 PM Ted Lemon <[e-mail]> wrote:

> On Feb 13, 2021, at 7:45 PM, David Farmer <farmer=[e-mail]>
> wrote:
>
> I think I'm hearing,
>
> 1. Rewrite RFC4007 to only discuss zone indexes for the Link-Local address
> on an interface, maybe rename then too. Junk the rest, it only confuses
> things.
>
> 2. Rewrite RFC4291 to only refer to Link Scope (for Link-Local Addresses
> and the Loop-Back Address) and Global Scope for all other Unicast addresses
> except the unspecified address, eliminate all inference to other unicast
> scopes and the size of scopes. All uniqueness for Link-Local address comes
> from IID uniqueness provided by DAD, and there should be no discussion of
> uniqueness regarding Global Scope in RFC4291, it is only discussed in
> RFC4007. Also, eliminate the discussion of IID scope. Leaving Multicast
> scope alone.
>
> However, unless we can agree on how to deal with the 64-bit boundary, or
> not to touch it, touching RFC4291 will only end in flames.
>
> Is it possible to do anything with RFC4007 without ending in flames?
>
>
> From my perspective at least, this conversation has been really useful,
> but I don’t see a cause to rewrite any documents. Being forced to confront
> my assumptions as a result of Fernando’s questions resulted in me
> discovering that a lot of things I hadn’t realized were clearly specified
> are actually clearly specified. The dissonance between RFC 4193 and RFC
> 4007 is interesting, but RFC 6724 addresses this dissonance quite
> satisfactorily. It’s possible that there’s an opportunity to write a small
> ops document here to make clear how to populate the policy table
> automatically based on prefixes seen on an interface, but I don’t think we
> have a Really Big Problem to solve here.
>
> If people are confused, it’s probably because they haven’t read these
> documents.
>

Actually, I think you are correct, complete rewrites are unnecessary.

I still would like to suggest we tighten up the language regarding scope in
RFC4291 once it eventually gets rewritten. As for RFC4007 a minor update
clearing up one point I think would really help. But, let me summarize from
the beginning first;

Prior to RFC3979, there with three (3) IPv6 Address Scopes, Link, Site, and
Global (AKA Internet). The Link and Site Scopes require Zone Indexes or
Zone Identifiers to distinguish an address between Zones of the same Scope
because the same address can exist in each of Zone of a Scopes. For the
Global Scope, Zone Indexes or Zone Identifiers are unnecessary as there is
only one instance or Zone of the Global Scope by definition.

RFC3979 deprecated Site-Local addresses and effectively deprecating the
Site Scope too.

RFC4193 created ULA and specified them as part of the Global Scope, meaning
that Zone Indexes or Zone Identifiers are not associated with ULA addresses
like other addresses within the Global Scope, and nothing more than that.

However, as Fernando points out, ULA cannot be unique across the
entirety of the Internet, and they are only unique within a local domain of
coordination, coordinated by or among local administrator(s). Furthermore,
the reachability of ULA, and other special-use IPv6 addresses intended for
local-use, SHOULD be restricted to this local domain of coordination only.

Whereas, GUA is unique across the entirety of the Internet and is
coordinated by the Internet Registry System, as described in RFC7020.
The reachability of GUA is unrestricted, however, it is subject to the
routing policies of ISPs and other network operators.

As the Internet does not guarantee universal reachability for GUA and only
local reachability is intended for ULA, and other special-use IPv6
addresses intended for local-use, therefore, there is no universal view or
understanding of what is included or contained within the Global Scope.

So I recommend a small update to RFC4007, further quantify the necessary
uniqueness;

Global scope, for uniquely identifying interfaces anywhere in the Internet
*or within the span of reachability for local-use addresses.*

Thanks
-- 
===============================================
David Farmer               Email:[e-mail]
Networking & Telecommunication Services
Office of Information Technology
University of Minnesota
2218 University Ave SE        Phone: [phone]
Minneapolis, MN 55414-3029   Cell: 612-812-9952
===============================================