Re: [sig-policy] Need to understand logic behind assigning /64 IPv6 addr
On Sep 17, 2011, at 10:16 AM, Mark Tinka wrote:
> On Sunday, September 18, 2011 01:05:08 AM Owen DeLong wrote:
>
>> In some older stacks, there was an issue where a /127
>> could have problems due to a deprecated use of a
>> "unknown" address which could allow the host on one side
>> of the link to claim both addresses, but that shouldn't
>> be the case with any up to date implementation.
>
> I'm not saying it's bad, I'm glad there are folk happy to
> use /31 in v4 and /127 in v6. We just prefer /30 and /126,
> respectively :-).
>
> Whatever cuts your mustard...
>
I use /30 in v4 and /64 in v6 as designed.
>> You're welcome to do as you wish, but, I see no point in
>> doing so. It just complicates your administration
>> without providing any benefit.
>
> Ummh, how?
>
Multiple prefix sizes, address fragmentation, etc. Admittedly, it's
a small complication, but, it is a complication.
Further, it violates the principle of least surprise as your organization
scales and brings in new engineers.
>>> I'm just not sure how many devices you can address on a
>>> point-to-point link that has only two devices on either
>>> end of it.
>>
>> Any value of n where 2≤n≤18e+18 if you use a /64.
>> Any value of n where 2≤n≤65,536 if you use a /112
>> Any value of n where 2≤n≤4 if you use a /126
>> and n=2 if you use a /127
>
> I meant if you're assigning anything shorter than a /126 or
> /127, as many folks do, particularly with /64's.
>
So did I. I was being a little tongue in cheek/snarky with just
presenting the math on the number of addresses, but the
reality is that there may be some cases where having
multiple addresses for one end of a point to point or the
other (or both) may prove useful. These are admittedly
rare.
Nonetheless, it is actually very convenient not to have to
count hosts and simply use /64 everywhere.
Owen