Re: WG Last Call: draft-ietf-v6ops-isp-scenarios-analysis-01.txt

[Ronald van der Pol <Ronald.vanderPol@rvdp.org>]

On Fri, Feb 06, 2004 at 07:58:44 +0200, Pekka Savola wrote:

> This is a WG Last Call for comments on sending
> draft-ietf-v6ops-isp-scenarios-analysis-01.txt, ...

Overall, I think the document is "good enough". I have three minor
points.

Section 4.3.1 has this sentence in the fifth paragraph:
"Therefore, the use of same process is recommended especially for      
large ISPs intending to deploy, in the short-term, a fully dual-     
stack backbone infrastructure."
I needed to read it several times. I think it says: "when deploying
dual-stack in short-term, same process IS-IS is recommended." Maybe
the sentence could be rephrased slightly to make it easier to read.
Also, isn't same process IS-IS recommended in general because of
operational benefits? It is easier to manage one protocol and topology
than several. The need to run separate processes should be the exception.

In section 5.2 it might be useful to explicitly describe the situation
where a service is limited by ACLs to e.g. customers. When upgrading
this service to dual-stack, the corresponding IPv6 ACLs should be
added to avoid exposing the service worldwide. So, I am thinking
about prefix filtering and "allow all" defaults.

It would be useful to have a "recommendation" section which summarizes
the missing features, i.e. the items the IETF should work on. That's
the main purpose of this document, isn't it? I could only identify:
- STEP/TSP
- tunneling when dynamic IPv4 addresses are in use
- authenticating when the customer connection network is shared
- multi-homing

	rvdp