Re: [savi] A query and a kind of potential attack in draft-bagnulo-savi-fcfs-01

[marcelo bagnulo braun <marcelo at it.uc3m.es>]

Hi Zhang,

thanks for you comments.

please note that there is a new version of the draft 
http://www.ietf.org/internet-drafts/draft-ietf-savi-fcfs-01.txt
this has importnat modifications from the version you reviewed.
Especially, it only covers IPv6

ZhangDong escribió:
> Hi Marcelo,
>
> You mentioned special cases in section2.4 in fcfs-01. One of the special cases is o  Anycast i.e. multiple hosts using the same source address to send packets.
> I am puzzled.  As described in RFC2460:
>    o  An anycast address must not be used as the source address of an
>       IPv6 packet.
>
>    o  An anycast address must not be assigned to an IPv6 host, that is,
>       it may be assigned to an IPv6 router only.
> How does the case that multiple hosts using the same source address to send packets happen?
>   
AFAIU, irrespectivelly whatever the RFC says, anycast is widely used in 
IPv4 in particular in dns servers (especially root and tld servers)

> In fcfs-01, the word "anycast" means the case that multiple link-layer addresses (such as MAC)use one IP address?
>
>   
well, anycast means that multiple hosts use the same IP address

> The section 3.3.1 ARP-based Neighbor Unreachability Detection procedure in fcfs-01 designs a mechanism of NUD based on arp.
This section is no longer avialbale in the new version, please check if 
your concern is valid for the IPv6 version only

regards, marcelo


>  The savi device sends an ARP REQUEST packet as the NUD message. Suppose that an attacker has initiated arp cheating. The attacker responds the arp request whic is the NUD message.In this case, when the user really changes another IP address with the same MAC, because of  the NUD response of the attacker, the binding o
>   
> f IP and MAC in savi device wiil not be update. And then, the packets sent by the new IP address of the user will be discard by savi device.
>
> Will this problem happen?  Do i have misunderstanding?
>
> Thank you very much.
>
>
>
> Dong Zhang
>
>
>