/48 vs. /56 for customers, was XS4ALL Introduces native IPv6 for DSL customers
Jeroen Massar
jeroen at unfix.org
Fri Aug 27 10:14:45 CEST 2010
On 2010-08-27 10:00, Shane Kerr wrote:
> Daniel,
>
> On Fri, 2010-08-27 at 09:49 +0200, Daniel Roesen wrote:
>> On Fri, Aug 27, 2010 at 09:16:09AM +0200, Marco Hogewoning wrote:
>>> Als a lot of folks out there once heard about the 48 and are still
>>> sticking to it, even when the policy differs. Saves ourself a lot
>>> of discussion and waving policy.
>>
>> Unless you run out of /48s and have to go back to RIPE NCC to get more
>> address space. And then you're judged on /56s, not /48s for efficiency.
>
> Hm... that's a good point.
>
> With at "normal" /32 allocation you can only give 65k customers a /48,
> which is non-trivial but also almost certain to happen sooner or later.
> So any ISP of size will run out of address space if they use /48, and
> then will have to shift to /56 eventually.
>
> Best to use /56 initially? Perhaps this should be a strong
> recommendation somewhere?
No, they just have to do a proper address plan when they request their
allocation and get something bigger.
I mean, it is not THAT hard, you have at the current moment X customers.
You can maximally grow your market to Y customers if all the competition
is gone, figure out the chance that that happens and you know what your
largest set is ever going to be. Round the X up to how much you expect
to be in say 5 years, and presto, convert that into number of /48's
round that nicely to a factor of 2 and you have the number of bits that
you need. Write a proper address plan and submit to the NCC.
How else do you think two ISP's came up with /19's?
Heck, if you are a government you can get a /13! ;)
>From a conservation point of view though, if one would state "I give
/56's to home-customers, /48's to companies" and that would be taken as
a universal thing, I think it would be fine. Then again, does it really
matter on the grand scale, 2000::/3 won't fill up that easily...
(fast forward a 25 years, and that will bite me probably ;)
Greets,
Jeroen
More information about the ipv6-ops
mailing list