Since these addresses by definition have no geographical or topological
significance, and the idea is that they should be assigned as a public
good (i.e. approximately at cost price with any profit being used for
the public good), I'm still at a loss as to why we would need to
distribute the registry function. In any case, that is an IANA matter.
Brian
Aidan Williams wrote:
>
> Hi Guys,
>
> Geoff points out in his document that distribution of the "centrally
> assigned" Unique Local IPv6 Unicast Addresses space FC00::/8 is tricky
> if the registries all have to coordinate with each other when selection
> occurs.
>
> Also, assigning address blocks seems undesirable because the desire to
> cripple the aggregability of this address space and hence cripple its
> global routability.
>
> A mechanism for partitioning the available space amongst a set of
> registries that isn't "block-like" could therefore be a useful thing.
>
> A suggestion: IANA can allocate vectors of numbers to each registry
> that span the entire address space.
> For example: In the allocatable range 0-2^41, each set of numbers
> described by
> S = { n < 0-2^41: n = V + k*G where k=0,1,2, ...}
> where:
> G is a largish number that partitions the available space, and
> V (< G) identifies the vector that is allocated to a registry
> are non-overlapping, non-aggregable allocations.
>
> IANA would allocate integer Vs to registries and keep a database of the
> Vs already assigned. The Vs can be sequentially allocated. Registries
> allocate from the S-es (vectors of numbers) generated by the Vs they are
> given without the need to coordinate with each other. The value G
> controls the size of the vector (I'm not sure off-hand what a sensible
> number might be).
>
> Technically, a router could work out V by examining the address (ie
> the_41_bit_part mod G will be the same for all allocations in the same
> vector) and thus perform aggregation by routing on the value of V,
> however if G is not a power of 2 (perhaps a prime) this would be an
> expensive operation.
>
> - aidan
>
> --------------------------------------------------------------------
> 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]
> --------------------------------------------------------------------
--
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Brian E Carpenter
Distinguished Engineer, Internet Standards & Technology, IBM
NEW ADDRESS <[EMAIL PROTECTED]> PLEASE UPDATE ADDRESS BOOK
--------------------------------------------------------------------
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]
--------------------------------------------------------------------