[132061] in North American Network Operators' Group
Re: Low end, cool CPE.
daemon@ATHENA.MIT.EDU (=?utf-8?Q?Bj=C3=B8rn_Mork?=)
Fri Nov 12 07:55:46 2010
From: =?utf-8?Q?Bj=C3=B8rn_Mork?= <bjorn@mork.no>
To: nanog@nanog.org
Date: Fri, 12 Nov 2010 13:55:27 +0100
In-Reply-To: <20101112014100.GA97649@ussenterprise.ufp.org> (Leo Bicknell's
message of "Thu, 11 Nov 2010 17:41:00 -0800")
Errors-To: nanog-bounces+nanog.discuss=bloom-picayune.mit.edu@nanog.org
Leo Bicknell <bicknell@ufp.org> writes:
> - IPv6 support, native or tunnel to tunnelbroker.net type thing.
This is far too diffuse. You'll get a "yes, we've got IPv6".
You should at least add
- IPv6 packet filtering and policy management (at least simple access
lists)=20
- DHCPv6-PD client running over PPP or ethernet (possibly bridged DSL)
WAN interface(s)
- Ability to split the delegated prefix into a /64 for every LAN and
loopback interface, preferably fully configurable
- Configurable RA on LAN interfaces, using the dynamically allocated
prefixes
- (wishlist) configurable ifid's on the LAN and loopback interfaces as
an alternative to using EUI-64
- WAN link addressing using whatever is available of SLAAC, DHCPv6
IA_NA or link local. Specifically: Using SLAAC for the WAN link
should be possible without sacrificing any router functionality on
the CPE.
=20
and probably a lot more. DNS resolver handling needs a chapter on it's
own....=20=20
The point is: We've been asking for "IPv6" for too long. That's just
one bit in a packet header. We need to start asking for the features we
expect, which is a lot more than that bit.
Bj=C3=B8rn
