Hi all, TL;DR: Pub/sub and multicast hallway discussion happening at IETF104 (possibly during the hackathon?). Slides here: https://github.com/EricssonResearch/coap-pubsub-profile/blob/master/Pubsub-multicast.pdf Contact me if interested. As mentioned during the CoRE interim, I have started to think on how to progress the security for pub/sub work. For the people not following, there is currently one draft in Ace that describes a profile of Ace for authorization and key distribution + communication protection for CoAP pubsub [1]: https://tools.ietf.org/html/draft-palombini-ace-coap-pubsub-profile-03.
While looking at how to move forward that draft, some things came up: first of all, it would be nice to use multicast to broadcast notifications from broker to subscribers, for performance reasons. Secondly, the ace-coap-pubsub document miss a way to protect unaware nodes to get unwillingly subscribed by attackers spoofing their IP address. In fact, ace-coap-pubsub does protect the publication, but does not set up the “authorization for subscribers” mechanism, or any other DoS protection mechanism. These two points might seem parallel and independent, but one influence the others: depending on how multicast notifications are set up, we might reuse existing mechanisms that might protect against unauthorized nodes being sent notifications from the broker. I put up some ideas in slides and was hoping to get some discussion started during the hackathon (if possible): https://github.com/EricssonResearch/coap-pubsub-profile/blob/master/Pubsub-multicast.pdf and/or in the mailing list. As you can see, I try to explain the problem and come up with possible solutions based on the existing drafts. These are of course just very high level draft solutions, and require more discussion. Any feedback welcome! Francesca [1] https://tools.ietf.org/html/draft-ietf-core-coap-pubsub-08
_______________________________________________ Ace mailing list Ace@ietf.org https://www.ietf.org/mailman/listinfo/ace