Ray,

I also agree that requirements and use cases seem the next logical step. I'd be 
willing to document the ones that we are contemplating. It would be useful to 
have someone more familiar with DSL/Fiber residential deployments do so there 
as well.

As has already been mentioned, this is a great conversation but it really 
belongs on the Homenet list at this point (copying that list in case people are 
not following both).

Jason

From: Ray Hunter <[email protected]<mailto:[email protected]>>
Date: Tue, 2 Aug 2011 06:12:12 -0400
To: Philip Homburg 
<[email protected]<mailto:[email protected]>>
Cc: "[email protected]<mailto:[email protected]>" 
<[email protected]<mailto:[email protected]>>
Subject: Re: [v6ops] default LAN routing protocol for IPv6 CE router

I think the earlier suggestion of Lee Howard to capture requirements for a home 
network first is a very good one.

In the current discussion, I don't think it's fully defined yet who operates 
your "generic home router" and under what conditions: whether the home router 
is one operated by the user onto a shared bridged IPv6 access network, or 
whether the home router spec will be used by an ISP for purchasing and 
operating the home router, and which then also has to take into account the 
user's requirements for more than one local LAN in the home, or whether there 
are 2 home router devices (one operated by the user and one operated by the 
ISP), or even more home router devices that also route between IPv6 prefixes.

Appropriate behaviour of an interface will depend very much on the management 
and deployment scenario.

In IPv4 + NAT that was pretty much irrelevant. The outbound path and address 
could be learned by acting as a DHCPv4 client and the reverse path was 
"learned" via NAT. And each cascaded device could in turn act as a DHCPv4 
server for its downstream neighbors. A simple "outbound only" firewall rarely 
did harm either. So simple cascading pretty much always worked without any 
configuration at all.

What if a user stacks multiple home routers in IPv6?
Do you assume / enforce a tree structure for a home network?
Do you assume / enforce a single-homed ISP model?

I know cascaded PD has been discussed many times in the past, but is there an 
RFC for how it should operate in practice?
RFC5877 says "Renumbering Still Needs Work"

So back to requirements, I think a list might conceivably consist of:
1) a home router should automatically discover its neighbors on a per port 
basis.
2) a home router should able to automatically detect whether each neighbor is 
in the same management domain or not
3) a home router should control peering in a secure manner
4) a home router should automatically learn prefixes and interface addresses 
(NB PD is only defined in rfc3769 to be to a single router attached to an ISP 
link, and sub-delegations within the site are NOT communicated back to the ISP)
5) Depending on its peering, a home router should determine it's position in 
the network relative to the CPE PD router (the root of the prefix delegation / 
home network assuming a tree structure).
6) Depending on it's position in the home network, a home router should 
automatically determine what functions it needs to enable per port (firewall, 
learning a delegated prefix, delegating further prefixes, management access 
allowed .... )

Most current IGP's would fulfill many of those requirements, but probably not 
all. The main problem is I don't think the above list is anywhere near 
definitive or widely accepted. So saying one protocol or mechanism is "better" 
than another is then a bit premature IMHO.

regards,
RayH

Philip Homburg wrote:

In your letter dated Tue, 02 Aug 2011 09:47:31 +0200 you wrote:


If the ISP wants to control a firewall or other equipment on a user's
premises then they probably need to provide and own the firewall or CPE.
Vice versa for the end user. The mix of a user owned firewall with an
ISP provided configuration (even if that's just enable/disable function)
just won't work practically in operations. And that's even before the
lawyers get involved. It's not an issue that the IETF should attempt to
solve/ automate using protocols IMVHO.



I don't think it is matter of an ISP wanting to control the user's firewall.

Suppose you have a generic home router, just a bunch of ethernet ports and
some proessing in a small box.

If the home router is attached to a WAN link, then it has to enable its
firewall to separate the WAN link from the internal links.

If the router is used internal to the user's network then it should not
enable its firewall, unless one of the ports is connected to the guest network
but then it should have a firewall tailered to guest networks.

How do you construct a router such that the router always knows what it
has to do, or at least is in some sense fail-safe?





________________________________
This E-mail and any of its attachments may contain Time Warner Cable 
proprietary information, which is privileged, confidential, or subject to 
copyright belonging to Time Warner Cable. This E-mail is intended solely for 
the use of the individual or entity to which it is addressed. If you are not 
the intended recipient of this E-mail, you are hereby notified that any 
dissemination, distribution, copying, or action taken in relation to the 
contents of and attachments to this E-mail is strictly prohibited and may be 
unlawful. If you have received this E-mail in error, please notify the sender 
immediately and permanently delete the original and any copy of this E-mail and 
any printout.
_______________________________________________
homenet mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/homenet

Reply via email to