Hi Cameron,

You may need spend time to review it later, it's different from 6RD and
DS-Lite,

Secondly, I am wondering whether it could fit for mobile CPE case.

Best regards,

-Hui

2011/7/28 Cameron Byrne <[email protected]>

>
> On Jul 28, 2011 5:08 AM, "Paco Cortes" <[email protected]>
> wrote:
> >
> > Hi Wojciech,
> >
> > two comments on the 3gpp sections of draft-dec-stateless-4v6-02.
> >
> > First, the definition of "Rx" is missing in section 4.2.1.
> >
> > Second, the table in section 4.2.2 claims "No discernible impact" for
> > "AF Application Function" for the 4V6 translation mode. This is not
> > fully correct.
> > The PCC architecture support a plurality of AFs communication over Rx
> > with the PCRF, and most of them "interact" in some way with the end
> > user payload.
> > An AF might terminate, proxy or transparently inspect a subset of the
> > UE traffic (for instance control signalling, like SIP).
> > An AF might also terminate, proxy or transparently inspect all UE
> > traffic (application server acting as AF, deep packet inspection
> > device, ...).
> > And the most important point: An AF might located in different
> > locations inside the operators networks, or even outside the operators
> > network.
> > Referring to your own "Figure 2" in the draft, this means that for
> > IPv4 traffic an AF might "intercept" traffic between the "PDN-Gw" and
> > the "S46 Gateway", i.e. with an IPv6 header, BUT it might also
> > intercept traffic behind the "S46 Gateway", i.e., with an IPv4 header.
> >
> > I propose to change the impact to "No impact for IPv6. Feature to map
> > IPv4-IPv6 addresses may be needed only in case of  IPv4-only
> > applications".
> >
> > I'm aware that it would be theoretically possible to "hide" this
> > impact from an AF by using a more complex PCRF implementation that
> > "translates" IPv4 addresses in Rx message (including TFTs) into the
> > mapped IPv6, but this would imply limitations, and would just be one
> > implementation possibility. I believe that this draft should not
> > assume a specific implementation, unless clearly documented.
> >
>
> +1, I believe this section causes confusion.
>
> I have not had time to do a full review.
>
> Tunneling traffic breaks pcc and a new bearer type is a tall order. We
> already visited these issues when concluding 6rd and ds-lite were not good
> fits for 3gpp.
>
> Cb
>
> > br,
> >  Paco
> > _______________________________________________
> > Softwires mailing list
> > [email protected]
> > https://www.ietf.org/mailman/listinfo/softwires
>
> _______________________________________________
> Softwires mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/softwires
>
>
_______________________________________________
Softwires mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/softwires

Reply via email to