[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: RS sending in draft-ietf-v6ops-ipv6-cpe-router-04
On Thu, 29 Apr 2010 09:11:51 -0500
"Hemant Singh (shemant)" <shemant@cisco.com> wrote:
>
> -----Original Message-----
> From: owner-v6ops@ops.ietf.org [mailto:owner-v6ops@ops.ietf.org] On
> Behalf Of Wojciech Dec (wdec)
> Sent: Thursday, April 29, 2010 3:37 AM
> To: Philip Homburg
> Cc: v6ops@ops.ietf.org
> Subject: Re: RS sending in draft-ietf-v6ops-ipv6-cpe-router-04
>
>
> >Indeed that may be a nice thing, however I'm sure you appreciate that
> the
> >practicalities of changing an rfc are very different from clarifying or
> >tightening some rfc behavior on a device meant to fulfill a specific
> role
> >(eg the CPE).
> >The behavior expected with respect to certain RA flags has been
> discussed
> >for years, with ambiguous conclusions. As such, it's very pragmatic
> that the
> >CPE draft lays down more clarity in some areas, esp. When they have the
> >potential to make the behavior of the system more deterministic. As
> such
> >these requirements are anything but random.
>
> I agree with Woj totally on this one. The IPv6 CE Rtr can tighten text
> rather than going back to changing any RFCs. For example, we could add
> text to the IPv6 CE Rtr saying the device sends up to three RS's and
> then given up IPv6 address acquisition if no RA is received. Cable
> modems in Docsis 3.0 have such a specification.
I never think DOCSIS should ever be used as an example of how to do
something properly. They made the fundamental mistake of making a layer
2 technology layer 3 aware. If only they'd kept to the layer they are
supposed to be operating at (i.e. layer 2), then they wouldn't have to
be rolling out DOCSIS v3 to support IPv6, at a very great expense to
themselves and, ultimately, their customers.
If that argument isn't clear, realise that everybody else in the
Ethernet/ADSL/frame relay/ATM/Token Ring/ArcNet/PPP/HDLC/SDLC world
isn't having to roll out a new revision of their layer 2 technology
just to support a new layer 3 protocol.
The IPv6 CE Rtr can also
> tighten text to say the RS must include a source address and if source
> address is included in the RS, note, one can also include the source
> link-layer address option in the RS as well. However, it is always good
> to first understand why a certain deployment cannot do without the RS
> being sent with a specified src-address rather than an unspecified
> address.
>
> Hemant
>