Re: [CGA-EXT] Comment on draft-jiang-csi-dhcpv6-cga-ps-03.txt

["沈烁" <shenshuo at cnnic.cn>]

Hi£¬Tony,
Thanks for the comments.
Your concern is important for a security solution. But this document is a problem
statement and it oversee possible solution space between dhcpv6 and cga. Security
considerations are closely related to actual solutions. For this particular
problem in section 4, we did propose a I-D in dhc wg: http://tools.ietf.org/id/draft-jiang-dhc-secure-dhcpv6-02.txt

Similarly as in SeND, CGA usage in dhcpv6 also need ADD like mechanism or
pre-deployment of public key. The benefit of using public key crypto protection
(compared with the symmetric crypto protection defined in dhcpv6) is discussed in
detail in section 3 and 4 of the I-D. You are more than welcomed to read and
discuss I-D.

Best,

Sean



ÔÚÄúµÄÀ´ÐÅÖÐÔø¾­Ìáµ½:
>From: Tony Cheneau <tony.cheneau at it-sudparis.eu>
>Reply-To: 
>To: shengjiang at huawei.com, Sean Shen <shenshuo at cnnic.cn>, tjc at ecs.soton.ac.uk
>Subject: Comment on draft-jiang-csi-dhcpv6-cga-ps-03.txt
>Date:Wed, 23 Sep 2009 11:28:20 +0200 (CEST)
>
>Hello,
> 
> I have a small comment on section 4 of your draft:
> " DHCPv6 message (from either a server, relay
>     agent or client) with a CGA as source address, can carry the CGA
>     Parameters data structure and a digital signature. The receiver can
>     verify both the CGA and signature, then process the payload of the
>     DHCPv6 message only if the validation is successful. In this way
>     DHCPv6 messages can be protected."
> Maybe I missed something, but what is the gain from a security point of 
> view ? An attacker can still generate its own CGA and craft misleading 
> DHCP messages. Maybe you implicitly implied there was a mechanism similar 
> to ADD of SEND or that the Public Key is learned in a previous RS/RA 
> message exchange and ADD has been performed on the CGA. Either way, the 
> text seems unclear on this point, can you clarify this text in the next 
> version ?
> 
> Best regards,
>  	Tony Cheneau
>