Re-, >From where I stand, I don't see NEW issues that are specific to the priority >option.
I updated the text to record this: NEW: Misbehaving intermediate nodes may alter the content of the S46 Priority Option. This may lead to setting a different IPv4 service continuity mechanism than the one initially preferred by the network side. Also, a misbehaving node may alter the content of the S46 Priority Option and other DHCPv6 options (e.g., DHCPv6 Option #64 or #90) so that the traffic is intercepted by an illegitimate node. Those attacks are not unique to the S46 Priority Option but are applicable to any DHCPv6 option that can be altered by a misbehaving intermediate node. Cheers, Med De : Ben Campbell [mailto:[email protected]] Envoyé : mardi 27 septembre 2016 16:37 À : BOUCADAIR Mohamed IMT/OLN Cc : The IESG; [email protected]; Yong Cui; [email protected]; [email protected] Objet : Re: Ben Campbell's No Objection on draft-ietf-softwire-unified-cpe-06: (with COMMENT) Hi, thanks for the quick response! One comment below. Ben. On 27 Sep 2016, at 4:44, [email protected]<mailto:[email protected]> wrote: - section 3: "This may lead to setting a different IPv4 service continuity mechanism than the one initially preferred by the network side" Are there consequences of that that should be discussed? E.g. bid-down attacks, ability to direct packets via a compromised path, etc? (I'm not saying there are; I'm just asking.) [Med] We didn't include examples of such consequences because those attacks depend on the modification of other DHCPv6 options that are not defined in this document. For example, the ability to direct packets via a compromised path will require the modification of the content of DHCPv6 Option #64 or #90 to redirect packets to an illegitimate AFTR/BR. What about the following change: OLD: Misbehaving intermediate nodes may alter the content of the S46 Priority Option. This may lead to setting a different IPv4 service continuity mechanism than the one initially preferred by the network side. NEW: Misbehaving intermediate nodes may alter the content of the S46 Priority Option. This may lead to setting a different IPv4 service continuity mechanism than the one initially preferred by the network side. For example, a misbehaving node may alter the context of the S46 Priority Option and other DHCPv6 options (e.g., DHCPv6 Option #64 or #90) so that the traffic is intercepted by an illegitimate node. That helps. I was thinking more in terms of the S46 priority option specifically, but I guess it doesn't hurt to mention others. If one can modify the DHCPv6 options in general, one can do plenty of mischief without S46 Priority. Does adding S46 priority add any new issues to that? I suspect the answer is "no", but if so it would be worth saying that.
_______________________________________________ Softwires mailing list [email protected] https://www.ietf.org/mailman/listinfo/softwires
