[afnog] New IPv6 Address Block Allocated to the RIPE NCC

[Jeroen Massar]

[Filiz, I guess an explanation why certain organizations are getting big
blocks is in place as there seem to be quite some negative comments on it]

Brian Candler wrote:
> [cc list trimmed]
> 
> On Sat, Dec 31, 2005 at 10:51:51AM +0100, Jeroen Massar wrote:
>> Cheered to early, it's just a big block again:
>>
>> inet6num:       2a01:c000::/19
>> netname:        FR-TELECOM-20051230
>> descr:          France Telecom
> 
> Wonder how FT managed to justify a /19? That's the same as 8,192 ISPs each
> with a standard /32. Or assuming an 80% utilisation rate, it means they are
> planning to connect 429 million customers with a /48 each over the next two
> years. Right.

Something along those lines, but I am pretty sure they met all of the
constraints that RIPE has put on them. Also note that IPv6 allocations
are based on an HD ratio and not utilization.

Don't forget the aggregation part and that France has about 60 million
inhabitants(*). Also calculate into that growth for the future.

Then again France Telecom is in the same league as:
2003::/19 	[de] DE-TELEKOM-20050113
2001:2000::/20 	[eu] EU-TELIANET-20040510
2001:8000::/20 	[au] TELSTRAINTERNET41-AU-20041202
2400::/20 	[kr] KORNET-KRNIC-KR-20050601
2400:2000::/20 	[jp] SBB-IPv6-20050712
2001:5000::/21 	[xx] EU-EN-20040910
2001:a000::/21 	[jp] NTTWEST-IPv6-JPNIC-JP-20041201

And a number of other /22's. FT is not the first to get a /19, the
Germans (83m population) already beat them with that ;)

> That's going to throw a spanner in the works - now every two-bit wannabe ISP
> will want a /19 to compete in the routing tables.

*EHM* I don't think you should call France Telecom a "two-bit wannabe
ISP"... Anybody who can justify address space and pays the bills gets
the address space. Simple(tm)

> I guess the advantage is
> that it will make IPv6 address depletion happen just as fast as IPv4, and so
> we'll have a new chance to design a protocol which actually solves the
> problems of IPv4...

I guess you don't see how vastly large the IPv6 address space is.
All current allocations are happening from 2000::/3(**), the "Global
Unicast" space. A /19 is only (19-3 = 16 -> 2^16 ->) 1/65536th of that
part of the address space.

Now lets say that 65k companies get such a prefix then the 2000::/3
prefix is full and everybody knows that this attempt failed. We can then
move forward to the other 7 /3 prefixes and make a much better policy
for those. Also note that there is already work being done on changing
the HD ratio which will result in smaller allocations being given out to
endsites.

Greets,
 Jeroen

* http://www.cia.gov/cia/publications/factbook/geos/fr.html
  http://www.cia.gov/cia/publications/factbook/geos/gm.html
  http://www.sixxs.net/tools/grh/
** http://www.iana.org/assignments/ipv6-address-space

-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 238 bytes
Desc: OpenPGP digital signature
Url : http://listserv2.cfi.co.ug/pipermail/afnog/attachments/20060103/081396b0/signature.bin