Re: [v6ops] draft-ietf-6man-rfc4291bis prohibiting non-/64 subnets

> On 24 Feb 2017, at 11:33, Gert Doering <gert(_at_)space(_dot_)net> wrote:
>
> Hi,
>
> On Fri, Feb 24, 2017 at 06:04:30PM +0900, Erik Kline wrote:
> > IMHO having /64 as the logical unit of allocation to network leaves
> > is a very good thing.
> Why, exactly, except "because it was decided to be that way, many years
> ago"?
>
> (Not that I have any plans to fight that particular windmill, but /64
> never made sense to me, after all the more interesting aspects of 8+8
> never happened, and thus, effectively, IPv6 today is "IPv4 with longer
> addresses" as far as "hosts attaching to networks" and "routing" is
> concerned...)
>
> Wasting half the address space and then having to start arguments on
> the amount of subnets available to home users ("can we give them a
> /48, or will we run out?", "can we give ISPs enough space so they can
> give all their users a /48, or do we need to make this a /56?", "how
> much conservation is required by ISPs?") is major silliness - something
> like a /96 would have served the aspect "more machines than you can
> imagine per subnet" perfectly well.
And you can see RFC 5739 on how a /64 subnet is required for each road warrior 
connected to a VPN.  You’d need a /56 or /48 to do what a /24 does in IPv4.

Yoav

signature.asc
Description: Message signed with OpenPGP