Re: [lisp] Nonce in PubSub

Hi Erik,

I see your point. The draft assumes a security association between the
ITR and the MS in order to authenticate the Map-Notifies. I think this
addresses your valid concern on spoofed Map-Notifies. How this
security association is established is a different discussion :)

Note also that, ideally, the Map-Notifies sent as publications should
each have a different nonce so the MS can easily correlate them with
the Map-Notify-Acks received as responses.

Best,
Alberto

On Thu, Nov 16, 2017 at 11:17 PM, Erik Nordmark <nordmark@acm.org> wrote:
> On 11/17/2017 10:49 AM, Alberto Rodriguez-Natal wrote:
>>
>> Just to clarify what was discussed in the meeting. The nonce used in
>> the Map-Request requesting the subscription will be used in the
>> Map-Notify that confirms the subscription. This is at top of page 6 in
>> the draft.
>>
>> Similarly, a Map-Notify sent as publication will be ack'ed by a
>> Map-Notify-Ack using its nonce.
>
>
> Albero,
>
> My understanding from Dino's comment at the make was that in his
> implementation the map-notify has the nonce from the original map-request.
>
> The reason I asked about this is that there are some additional security
> benefits if the map-notify has a nonce which corresponds to what the xTR had
> sent in the map-request. Otherwise you need some other mechanism to guard
> against receiving spoofed map-notifies.
>
>    Erik
>
>>
>> https://tools.ietf.org/html/draft-rodrigueznatal-lisp-pubsub-01
>>
>> Thanks,
>> Alberto
>>
>> _______________________________________________
>> lisp mailing list
>> lisp@ietf.org
>> https://www.ietf.org/mailman/listinfo/lisp
>>
>