Behind the auto industry's quest to replace the aging CAN standard with a new in-vehicle networking standard for transmitting data from autonomous car sensors
all of which has to be shared among components in the car and used to make split-second driving decisions. http://www.nytimes.com/... Terri Delfino / @terri_d : Self-driving cars will generate 4 tb of data/hr and need approx 300 teraflops of computing power - that's a trillion operations per second. The catch? “No one's solved the problem of how do I stream all that data to the central processor?” http://www.nytimes.com/... @nytimesbusiness : There's a potential windfall for whoever wins the race to bring high-speed networks to cars, which will create a market for millions of new connecting chips, gateways, and controllers to support the new standard. Here are the companies working on this: http://www.nytimes.com/... NYTimes Tech / @nytimestech : Autonomous cars, with all their data needs, will be rolling supercomputers. That'll require a faster internal data network. http://www.nytimes.com/... NYTimes Tech / @nytimestech : New advanced driver assistance systems like automatic emergency braking, electronic stability control and lane-keeping assistance demand instant communications. So automakers have been supplementing their networks. Here's the latest: http://www.nytimes.com/... NYTimes Tech / @nytimestech : Automakers are aware of the coming data tsunami. “They think we need 300 teraflops of computing power,” said a senior director at chip supplier Xilinx. A teraflop is a trillion operations per second, meaning every vehicle would be a rolling supercomputer. http://www.nytimes.com/... NYTimes Tech / @nytimestech : Self-driving cars need systems that process live information about road conditions, weather, objects around them, traffic and street signs — all of which has to be shared among components in the car and used to make split-second driving decisions. http://www.nytimes.com/... NYTimes Tech / @nytimestech : Cars need to get faster — not on the road, but on the inside. Speed has always been part of the mystique of the automotive business. But cars have been notoriously slow when it comes to handling information. Our reporter explains the details: http://www.nytimes.com/...
Context & Ripple Effects
The CAN bus that has carried control messages between a car's components for decades was never built for what autonomous driving demands: as Terri Delfino notes, self-driving cars can generate around 4TB of sensor data per hour and need on the order of 300 teraflops of compute, and no one has yet solved how to stream all of it to a central processor. That gap is why the industry is racing to define a successor in-vehicle networking standard — and why the Times frames it as a potential windfall for whoever wins, given a market of millions of new connecting chips per year.
The timing matters because the car is already drowning in silicon: carmakers now integrate 100-150+ chip-based electronic control units per vehicle, an architecture that multiplies wiring and makes centralized, high-bandwidth data sharing harder. Whoever sets the new network standard effectively decides how data moves inside every future car — which is why the fight over in-car data flows, from the $300B-$800B connected vehicle data marketplace to manufacturers' control over performance and maintenance data (repair-access disputes), keeps surfacing in adjacent coverage.
First-order effects
- Chip and networking suppliers — Intel, Qualcomm, Xilinx and Broadcom among them — are competing to own the new in-vehicle interconnect, turning a commodity wiring question into a high-stakes design win worth millions of sockets annually.
- Automakers face an architecture decision now: keep federating dozens of ECUs over legacy buses or commit to a central-compute topology whose network standard is not yet settled.
Second-order effects
- Whoever supplies the winning network also sits on the pipe for all sensor and vehicle data, strengthening the manufacturer-side grip on data flows that already limits independent repair and feeds the largely unregulated connected-vehicle data market.
- A centralized network standard pressures suppliers of individual ECUs and legacy bus components, shifting value from distributed controllers toward the central processor and its interconnect vendor.
Third-order effects
- If the pattern holds, the car converges on the same structure as the data center — a few powerful compute nodes on a coherent high-speed fabric rather than 100-plus scattered microcontrollers — making the interconnect standard as strategically important to autos as it is to servers.
- Because the network determines who can read and route in-car data, the standard-setting process becomes a de facto governance decision over the vehicle data economy, likely drawing regulatory attention as that market scales.
The trend: In-vehicle networking is being rebuilt around centralized, high-bandwidth compute for autonomy, and the vendor that sets the replacement standard for CAN will capture the chokepoint of the software-defined car.