Glitchy airbags point toward a growing problem for all companies

There’s another big problem for automotive airbags, and this time it’s an issue that points to mounting trouble for car makers: systems that have been mostly mechanical for more than a century are increasingly run by problem-prone computer software.

Last year, of course, we learned that up to 100 million airbags made by Japan’s Takata Corp. were subject to unexpected deployment because of unstable chemistry that can make them ignite with explosive force.

The latest problem, reported this month by General Motors, affects far fewer vehicles but is, potentially, no less deadly. GM recalled 4.3 million vehicles after discovering a software defect.

“The sensing and diagnostic module that controls air bag deployment has a software defect that may prevent the deployment of frontal air bags in certain ‘rare circumstances,'” the company said.

Roger Lanctot, associate director in the global automotive practice at Strategy Analytics, told that, “of course GM, like every other auto company and its suppliers, conducts extensive tests … of all their systems prior to deployment.”

However — and this is the key — Lanctot noted that the glitchy airbag software is “a symptom of a much larger problem — managing the growing mountain of software in cars.”

This is a phenomenon we wrote about here back in June: Software is driving automakers crazy.

As Lanctot told eetimes: “The software governing the operation of an airbag is pretty sophisticated, but in the future, safety systems will require even more software as we seek to avoid crashes altogether.”

If I may drag out the soap box for a moment: This is a Code Red situation for car companies. Surely, as Lanctot says, GM and other automakers are testing these systems, but the reported string of problems in cars begs the question of whether development testing is rigorous enough.

Do GM engineers have the capability of testing software from the outset of development? Can dev teams work in parallel, testing all along the way against simulated versions of all dependencies? Do they have the tools to simulate virtually unlimited use cases?

One hopes they’re savvy enough to be using Service Virtualization techniques to ensure the cars they sell are as safe as possible.

Another complication of selling cars driven by software is the need to roll out updates for on-board systems on a regular basis. After all, cars today are built to last 15 or 20 years. How will automakers be certain that the software they push out in 10 years won’t crash older systems? Again, Service Virtualization can help — and not just in building better cars.

Tools that allow limitless app testing — and along the way allow for faster, less-expensive development — should be a central part of every company’s IT infrastructure these days.

If delivering better product, faster and cheaper is a concept that appeals to you, you should check out a free webcast being held this Wednesday.

The session, set for 2:30 Eastern Time in the United States, is titled “Driving Continuous Delivery with Service Virtualization.” It will explore how companies are leveraging this technology to:

  • Replicate production-like dependencies in testing
  • Do more “negative testing” by allowing users to control response times and other inputs
  • Virtualize third-party services to eliminate the costs associated with repeated system queries
  • Enable parallel development
  • Reproduce production defects so your team can fix problems faster

If you can’t join the webcast at the appointed hour, check the same link for a replay. You’ll be glad you did.

In today’s unforgiving business environment, can you really afford not to explore a tool that’s proven to help dev/test teams do better work?