[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

ASN-based prefixes for leaf ASes

I'm starting here with the posit that the global routing table will contain
at least one prefix (/48 or shorter in the case of IPv6) from each ASN.
Obviously if this doesn't hold, the following discussion isn't valid...

Leaf ASes in the IPv6 world are expected to get a /48 prefix allocation
(from an RIR) as they do in the IPv4 world.  Given the size of a /48, it's
reasonable to expect the vast majority of these ASes will only need a single
prefix for IPv6, whereas in IPv4 they might have hundreds.  Since these ASes
already have an ASN and only need a single prefix, it seems logical to
assign a prefix to each AS based on their ASN.

This would reduce pressure on the "normal" prefix allocation process used by
non-leaf ASes.  In fact, a "normal" prefix need only be allocated to a leaf
AS when they demonstrate that their ASN-based prefix has been fully utilized
or bona fide technical reasons why the AS cannot use a single prefix.  A
non-leaf AS may (?) use ASN-based addresses for their internal use, but must
not assign any such addresses to customers.  To reduce advertisements,
non-leaf ASes may choose not to use their ASN-based prefix and instead give
themselves an internal allocation from one of their "normal" prefixes.

Given the number of 16-bit ASNs currently assigned, it seems reasonable to
assume an extension to 32-bit ASNs will occur within the lifetime of IPv6.
Thus, a single /16 prefix for all ASN-based allocations would be required.

The largest problem I see with this is that ASNs are permanently assigned,
or at least so far no recovery process is defined and no rent is charged.
This means that ASN-based allocations would be similarly permanent.  This
might be good or bad depending on your view of the RIRs.



Stephen Sprunk        "Stupid people surround themselves with smart
CCIE #3723           people.  Smart people surround themselves with
K5SSS         smart people who disagree with them."  --Aaron Sorkin

IETF IPv6 working group mailing list
Administrative Requests: https://www1.ietf.org/mailman/listinfo/ipv6