Industry Voices—Banasik: 5G will clear the autobahn of software roadblocks

Autobahn
Catching software defects during development is an increasingly challenging task. (Pixabay)
Paul Banasik

Over the past few years, the wireless industry has been promoting 5G as an enabler for exciting new products and services such as autonomous vehicles, AR/VR, and massive IoT. Will cars need the low latency of 5G? Maybe, but the use cases and the impact at a business level are not clear yet. On the other hand, the automotive industry is starting to face a software dynamic—similar to the PC industry—that could threaten their profitability.

Proliferation of electronic controllers, increasing software complexity, shrinking development cycles, long product life cycles, and current laws are creating challenging conditions for the OEMs. The broadband capabilities of 5G could become very useful to them very soon, as they need the ability to upgrade software very quickly.

Electronic control units with embedded software first saw mainstream use in the 1980s. Early successes addressing economy, emissions, safety, convenience, and entertainment requirements economically led to increased use over the years. Today, an average automobile might have 40 to 80 embedded controllers, with some high-end models using up to 150 controllers per vehicle. The quantity of software in a vehicle is potentially up to 150 million lines of code! 

Sponsored by VoltDB

Webinar: The Hidden Inflection Point in 5G: When the Changing Definition of Real-Time Breaks Your Existing Tech Stack

Rethink your definition of real-time to match the changing reality brought to the forefront by 5G. Your users expect milliseconds, in-event decision making. Is your tech stack ready?

In comparison, Microsoft’s Windows 10 operating system has approximately 50 million lines of code, the F-35 jet fighter has 8 million to 9 million lines of code, and the space shuttle had roughly a half-million lines of code. With only a few historical exceptions, software defect rates have stayed relatively constant for a unit quantity of code. The rapid increase in the number of electronic controllers and the complexity of software in automobiles is leading to recall campaigns similar in number to those caused by purely mechanical/electrical issues.

Catching software defects during development is an increasingly challenging task. On one hand, the proliferation of controllers is adding to the software development load; on the other hand, the overall vehicle platform development cycle has shrunk from 8.4 years in the 1980s to 6.7 years currently. With more to do and less time to do it, all constrained by the payback of investments in tooling and manufacturing, the likelihood of adopting more rigorous and costly development methods looks quite low.

Including the production life of a vehicle platform, any provided warranties (3-10 years), and the statute of limitations for safety-related recalls (10 years per NHTSA), a manufacturer is potentially facing a 23-year timeline for managing a single vehicle platform’s software depending on what issues occur after the sale. Compounding this time issue is the expectation that even more electronic control units will be used in every vehicle over the next decade; the current fleet of 1 billion vehicles globally only represents a fraction of the potential problem. In other words, the situation looks like it will get much worse for the OEMs over time.

Laws are in effect preventing the sale of new cars if they haven’t had open recalls addressed first. Also, more recently, laws have been passed preventing rental car agencies from providing vehicles to customers if those fleet vehicles have an open recall. The financial motivation to update dealership inventory and rental fleets with new software as quickly as possible is quite high; nobody wants to lose revenue to a competing dealer or agency, and the logistics of cycling through cars manually or deploying a large array of Wi-Fi in each lot just for software updates looks unappealing. Making software updates part of a broader array of connected car services, including tie-in with dealer promotions, makes much more sense.

The wireless industry as a whole should continue to assess new use cases leveraging 5G’s inherent capabilities. At the same time the automotive industry is facing a fairly large problem now, and it looks to get even bigger over time. The number of controllers, the complexity of the code, the long development and product life cycles, the size of the global fleet, and the existing laws look like a great opportunity timed perfectly with the introduction of 5G.

Paul Banasik is a senior analyst at Mobile Experts LLC, a network of market and technology experts that analyze wireless markets. He currently focuses on specific IoT verticals and how they relate to the wireless industry. His background includes engineering and management experience in prime mover controls, power generation and measurement, fleet management, consumer electronics, automotive infotainment, and semiconductor applications. Paul previously worked at Qualcomm Incorporated, Eaton Corporation, Woodward Governor, and Cummins, and holds an electrical engineering degree from Kettering University.

Industry Voices are opinion columns written by outside contributors—often industry experts or analysts—who are invited to the conversation by FierceWireless staff. They do not represent the opinions of FierceWireless.

Suggested Articles

No surprise: Online business has driven postpaid and prepaid brands to close some retail stores permanently.

Including outdoor urban and rural environments, SCF expects 38.3 million small cells to be installed by 2026.

Dish Network met with the FCC to discuss its progress toward building a 5G network and its desire for using the 12 GHz band for 5G.