Many automotive companies have a long and noble history. They also tend to have a fairly rigid internal structure. You have your powertrain engineers over here, your safety engineers over there, and your infotainment engineers somewhere else altogether.
As vehicle functions become more and more sophisticated, their functionality increasingly relies on information from across numerous vehicle domains. It makes sense to many head-unit vendors to try and integrate some driver assist functions, such as surround view. Head units have a large, powerful processor, and access to a large screen. What they don’t typically have is access to the camera data over a high-speed digital bus.
One view of the future is that Ethernet will become the dominant solution for moving high-speed data around the vehicle. This is across all vehicle domains, and the vision crucially crosses the infotainment/rest of vehicle divide.
Rick Kreifeldt, Vice President, Global Automotive Research and Innovation at Harman Automotive is an enthusiastic evangelist for this vision. Perhaps, given that he is also the AVnu Alliance chairman and president, this is understandable.
In a recent discussion with Strategy Analytics, Rick outlined his view of an Ethernet-enabled future for automotive. He admitted that AVnu has focused on high-end professional audio/visual applications to date – but claimed that this was part of the intended roadmap. 2011 will see a vehicle OEM along with one or two Tier One suppliers and key automotive semiconductor vendors join the Alliance, he asserts.
He sees the potential for Ethernet in the vehicle as huge, with a future vehicle architecture potentially containing only Ethernet and low-speed CAN buses. No LIN. No FlexRay. No MOST.
It’s still far from certain how soon – or if at all – this vision will be realized. As explained already, and highlighted by Rick, the internal structure of many automotive companies is an active block on pursuing this approach.
Strategy Analytics believes that the first major casualty may well be MOST. Although volumes are likely to grow further for a few years (a volume OEM is expected to launch soon, and Audi and Mercedes Benz are committed to MOST 150 rollout) – by 2015 or so MOST could well be in decline.
The bigger – and as yet far from answered – question, however, is the relevance of this to mainstream vehicles. MOST and FlexRay have been solutions that have found application almost exclusively on high-end vehicles and prestige brands. Will the conventional CAN/LIN vehicle architecture be a bottleneck for the Ford Fiesta class vehicles of 2020? If Ford can implement Sync using nothing faster or more exotic than USB 2.0 and CAN, is Ethernet an engineer’s solution to a problem that largely doesn’t exist?
When it comes down to it, Rick sees Ethernet in the car as coming down to the silicon ecosystem. The range of products and suppliers – together with his prediction of Ethernet’s future ubiquity in vehicle diagnostic systems – could tip the balance.
“There’s a convincing economic argument for Ethernet-enabling each vehicle ECU based upon flash/re-flash times alone” he reckons. “Hey, if it’s in there and already paid for, you’d be crazy not to use it.”
We’ll see. There are still big blocks and vested interests that are far from sold on Ethernet. Some of those heavily involved in Ethernet are also forging their own paths and own standards. It would still seem unlikely to us that a 2020 Fiesta class vehicle will have an Ethernet-enabled architecture.
But if you’re a vendor with significant LIN, MOST or FlexRay revenues, it’s certainly something worth thinking about. And if you’re an OEM or Tier One with high and wide internal walls, it could be the push that gets those barriers a-tumbling down.