On Thu, 21 Feb 2002 [EMAIL PROTECTED] wrote:
>
>
> Title : IPv6 Router Advertisement Prefix Delegation Option
> Author(s) : N. Lutchansky
> Filename : draft-lutchann-ipv6-delegate-option-00.txt
> Pages :
> Date : 20-Feb-02
>
> This document defines the Prefix Delegation (PD) option used to
> delegate IPv6 address space to simple IPv6 sites. The PD option,
> which lists the global prefixes that a site may use to number its
> network, is attached to IPv6 Neighbor Discovery Router Advertisement
> messages that are sent across a point-to-point link from a provider's
> router to a site's border router. This document defines the
> mechanism by which a site router processes the PD option and
> configures each of its attached links allowing hosts within the site
> to obtain global addresses using address autoconfiguration.
A few comments.
2. Terminology
==> is necessary to define basic stuff like 'node' and router here. A
reference to RFC2460 or whatever should be sufficient?
3.3. Site router operation
Upon receiving a Router Solicitation message containing a Prefix
Delegation option, the router MUST process the message as described
in [DISCOVERY] and [ADDRCONF] before processing the PD option.
==> Router Advertisement, not Solication? PD should only be used in RA.
4. Prefix Delegation option format
Prefix Length 8-bit unsigned integer. The number of leading bits
in the Prefix that are valid. The value ranges
from 0 to 64.
==> The format supports anything from 0 to 128, when though some of those
make no sense.
5. Security considerations
Security issues regarding the Neighbor Discovery protocol are
discussed in [DISCOVERY].
==> doesn't PD bring up any new issues? Bring more weight to existing
ones? I bet it does :-). For example, if the point-to-point link is an
IPv6/IPv4 tunnel, it might be possible to inject RA packets with bogus PD
options..
General comments: this would affect how routers (that is, CPE) work wrt.
NDISC: basically the point-to-point link towards the ISP would have to
operate in "host" mode so it could sent RS's and be able to receive RA's.
I think changes to the current specification need to spelled out in a
separate chapter.
One might also consider whether CPE should immediately send out RA's with
a new prefix (and advertise the old one with lifetime of zero or whatever)
when the prefix delegated from upstream changes.
I'd move acknowledgements from introduction to a separate chapter, as is
usual.
--
Pekka Savola "Tell me of difficulties surmounted,
Netcore Oy not those you stumble over and fall"
Systems. Networks. Security. -- Robert Jordan: A Crown of Swords
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page: http://playground.sun.com/ipng
FTP archive: ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------
