Re: draft-bourbaki-6man-classless-ipv6-00

[Tom Herbert <[e-mail]>]

On Sun, Jun 4, 2017 at 7:26 AM, Philip Homburg
<[e-mail]> wrote:
> In your letter dated Sun, 4 Jun 2017 07:13:21 -0700 you wrote:
>>That is not what we are proposing (in ILA at least). This is not
>>intended to work over the Internet. Mobility would be implemented
>>within a carrier network, external to that network it is transparent.
>>Logically, this provides the same functionality as encapsulation does
>>with the benefit of not incurring the overhead of encapsulation.
>
> The current IPv6 address architecture is basically a relatively densely
> allocated /64 prefix plus an extremely sparse 64-bit IID.
>
> With DHCPv6 you can today have dense IIDs that allow space for experiments.
>
> However, without a clear consensus that embedding identifiers in IPv6
> addresses is good idea, it strikes me as bad to use that as an argument
> to change how SLAAC works.
>
> At least for me, messing up the IPv6 address architecure just to save the
> overhead of encapsulation is not worth it.
>
Maybe it will be proven to be be worth it, maybe not. And this
probably true of other interesting things we could do with the large
IPv6 address space. But, I think imposing an artificial architectural
constraint that prevents us from even *trying* an alternative like
this is unfortunate. For all the talk that this draft forces IPv6 to
be like IPv4 I think it's the exact opposite at least in the this
case: /64 is preventing us from doing something different and
potentially innovative in IPv6 compared to IPv4.

Tom

>