In Germany, the speed and determinism of FlexRay are helping it make solid inroads. Audi is among the users, and BMW has been increasing its usage since it put the first FlexRay transceiver on its X5 in 2006.

“BMW just presented its rollout plan for FlexRay. In 2010, they have 35 ECUs with FlexRay,” said Jeroen Keunen, General Manager Integrated In-Vehicle Networking for NXP Semiconductors. “But 2015, that will grow to 44.”

NXP, which claims to be the leader in FlexRay chip shipments, sold more than a million devices last year, bringing its total shipments to 2 million. That one-year volume highlights increasing momentum: it took NXP three years to sell the first million FlexRay devices.

The standard also has widespread support from suppliers. All the major automotive microcontroller and transceiver providers have FlexRay offerings, and there is also broad support across the tools supply chain. Proponents say the market is beginning to extend beyond Germany.

“It isn’t just German Tier 1 suppliers who are investing in FlexRay,” Keunen said. “Japanese Tier 1s are designing with it, though many suppliers are using it primarily for German customers. The important thing is that once they get to their second or third ECU, they’re well along in the learning curve.”

Though more engineers are moving ahead on their learning curve, the majority of the industry is sticking with CAN. It is well known and inexpensive, so many engineers meet demands for more bandwidth by adding another CAN bus. That approach has been used extensively for safety applications, which have very tight timing requirements.

“The majority of the market is OK with using CAN,” said Raad Konja, Chief Engineer, Airbag ECUs & Satellite Sensors at TRW. He noted that timing can be managed by adding some software. “If you need to use an extra software layer and send additional bits for something like a checksum, it’s not a big deal,” Konja said.

Though FlexRay was designed by an automotive consortium, it may face competition from ethernet, which leverages the huge volumes of consumer electronics to drive speeds upward and pricing downward. BMW plans to deploy ethernet for subsystems such as cameras by 2013. Many contend that once ethernet enters the vehicle, its usage will expand quickly.

“There’s a lot of interest in jumping over FlexRay and going to ethernet,” said Paul Fox, Marketing Director at Renesas. “FlexRay is very expensive, while commercial usage has made ethernet very inexpensive. Ethernet has a lot of bandwidth. Even in powertrains, there’s interest in using it.”

Though ethernet doesn’t have FlexRay’s deterministic capabilities, it provides far higher bandwidth that could meet many near real-time requirements. Commercial-grade ethernet now runs at 100 Gbit/s, compared to FlexRay’s peak 20 Mbit/s rate.

While ethernet may eventually have an impact, CAN is the incumbent FlexRay must push aside. Even though TRW’s Konja feels multiple CAN buses are a viable alternative, he isn’t writing FlexRay out. Germany’s use is building the volumes that drive semiconductor pricing, making the bus more cost-effective in applications that need the utmost in performance.

“Chassis safety will be the first area that will need the benefits of FlexRay. It’s a bit more expensive, but that’s mostly a matter of volume,” Konja said.

That cost issue is currently a big stumbling block. Proponents contend that FlexRay may actually reduce overall costs even before chip costs are driven down by rising volumes.

“You can get rid of two or three CAN buses with one FlexRay network,” said Hans Roelofs, FlexRay Product Manager for NXP. “That means that, on a system level, we’re already at a break-even point.”

He also contends that reducing the number of buses will improve reliability. A single network with fewer gateways and nodes will reduce component counts, boosting reliability.