Re: Routing loop attacks using IPv6 tunnels

[Brian E Carpenter <brian.e.carpenter@gmail.com>]

On 2009-09-15 04:25, Templin, Fred L wrote:
> Brian,
> 
>> -----Original Message-----
>> From: Brian E Carpenter [mailto:brian.e.carpenter@gmail.com]
>> Sent: Friday, September 11, 2009 6:27 PM
>> To: Templin, Fred L
>> Cc: v6ops; Christian Huitema; ipv6@ietf.org; secdir@ietf.org
>> Subject: Re: Routing loop attacks using IPv6 tunnels
>>
>> On 2009-09-12 11:12, Templin, Fred L wrote:
>>> Brian,
>>>
>>>> -----Original Message-----
>>>> From: Brian E Carpenter [mailto:brian.e.carpenter@gmail.com]
>>>> Sent: Friday, September 11, 2009 4:06 PM
>>>> To: Templin, Fred L
>>>> Cc: Christian Huitema; v6ops; ipv6@ietf.org; secdir@ietf.org
>>>> Subject: Re: Routing loop attacks using IPv6 tunnels
>>>>
>>>> On 2009-09-12 09:13, Templin, Fred L wrote:
>>>>
>>>> (much text deleted)
>>>>
>>>>> Otherwise, the best solution IMHO
>>>>> would be to allow only routers (and not hosts) on the
>>>>> virtual links.
>>>> This was of course the original intention for 6to4, so
>>>> that any misconfiguration issues could be limited to presumably
>>>> trusted staff and boxes. Unfortunately, reality has turned out
>>>> to be different, with host-based automatic tunnels becoming
>>>> popular.
>>> Thanks. I was rethinking this a bit after sending, and
>>> I may have been too premature in saying routers only
>>> and not hosts.
>>>
>>> What I would rather have said was that mechanisms such as
>>> SEcure Neighbor Discovery (SEND) may be helpful in private
>>> addressing domains where spoofing is possible. Let me know
>>> if this makes sense.
>> Except for the practical problems involved in deploying SEND.
> 
> Can it be said that there is any appreciable operational
> experience with SEND yet? Are there implementations?

I'd like to know that too.

> 
>> We still have an issue in unmanaged networks.
> 
> By "unmanaged", how unmanaged do you mean? 

I was thinking of home networks, the kind where Teredo or
6to4 starts up spontaneously. Probably not a concern for
ISATAP sites.

    Brian

> ISATAP is
> intended for networks where there is at least some modicum
> of cooperative management. We want that it can also be used
> in "loosly" managed networks where there is an overall mutual
> spirit of cooperation but where site-internal link-layer
> address spoofing may still be possible. Can SEND be used
> for that, or do we need something else in addition (e.g.,
> a nonce with every message)?
> 
> Thanks - Fred
> fred.l.templin@boeing.com
> 
>>     Brian
>> --------------------------------------------------------------------
>> IETF IPv6 working group mailing list
>> ipv6@ietf.org
>> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
>> --------------------------------------------------------------------
>