32C3 – Vehicle2Vehicle Communication with IEEE 802.11p

One feature some proponents are pushing for future 5G networks are ultra short reaction times for ultra critical communication, for example between cars. What I failed to understand so far in this discussion was why for car to car communication a fixed network infrastructure and a backhaul network was necessary!? After all, car to car communication mainly makes sense for cars that are in close proximity to exchange information about potential dangers such as emergency braking, breakdowns and their current status such speed, direction etc., etc. It seems that my skepticism was not unfounded because unknown to me and perhaps also to the 1 ms 5G proponents, decentralized solutions not requiring a network infrastructure already exist.

While Europe and the US seem to be on different paths (once again) on higher layers of the protocol stack, both approaches are based on the IEEE 802.11p extension of the Wi-Fi standard. In this "Wi-Fi" flavor, there are no central access points, no fixed equipment and no backhaul of any kind. On top of this physical layer, event and context information is exchanged. An interesting challenge is how to ensure that messages are sent from "real" vehicles and not from rouge devices that want to disrupt traffic, e.g. by sending messages about emergency breaking etc. while at the same time ensuring privacy, i.e. to send messages anonymously to prevent tracking.

The concept that car companies have come up with is a public key infrastructure and cars equipped with a master certificate by car manufacturer. Based on the master certificate, temporary certificates are signed by a certificate authority which are then included in 802.11p messages sent by cars. Vehicles receiving messages can then validate the message by checking the temporary certificate which does not contain the car's identity and which are changed frequently. Rouge devices that do not have a master certificate can't get temporary certificates, at least in theory, and therefore can't include proper temporary certificates in their messages. That makes me wonder of course how hard it might be in the future to get a valid certificate by extracting it from an on-board computer of a vehicle. SIM cards of mobile devices have provided pretty good security over the past decades so there is at least some hope that the master certificates can be stored safely.

For more details, here's the talk on this topic from 32C3.

One thought on “32C3 – Vehicle2Vehicle Communication with IEEE 802.11p”

  1. “why for car to car communication a fixed network infrastructure and a backhaul network was necessary”

    Well, if all you have is a hammer (business model of selling somewhat arbitrarily limited access to said infrastructure), every problem starts to look like a nail.

Comments are closed.