As Discussed by Anders Wallgren and SoftwareQualityEngineering.com
A few weeks ago, CloudBees CTO Anders Wallgren sat down with SQE.com’s Cameron Philipp-Edmonds to discuss the Internet of Things and the latest Honda recall, as reported by Bloomberg in July. Basically, Honda is recalling all 175,000 of the Fit and Vezel hybrid cars it has sold in Japan since last year because of a software flaw that causes unintended acceleration.
We live in a day and age where software has eaten the world. In fact, almost every company is becoming a software company. Furthermore, our own vehicles have become “smart” cars with somewhere around two to three hundred million lines of embedded software code, if not more, in the latest high-end cars. A lot of that is engine management and transmission systems, but the bulk of it these days is also in the entertainment systems. The point is, software is pervasive and today’s modern cars have so many different processors within them that they are essentially a network on wheels.
“When you had a recall in the past for a car, it used to be that they had to go in and replace some physical part,” explains Wallgren. “Now what they're doing is bringing the cars in and doing software upgrades to fix a bug, and that's very common.”
Having to bring a car into a service establishment and get the software upgraded in a manual fashion can be pretty expensive. The service establishments like it because they get a chance to sell oil changes and all kinds of other incremental services, but the consumer has to waste an hour or two waiting for that to happen. And for a company like Honda, this is a very expensive way to deliver software updates.
Wouldn't it be nice for Honda to have found this bug during the software development process, as a side effect of some testing that happened along the way? This drives right at the core of why Continuous Delivery is so important for the software delivery lifecycle: because it helps an organization take a more proactive approach to quality and testing. In this instance, the cost of fixing a bug is exponentially more expensive the later you get into the process.
To learn how this problem might have been averted, listen to the full conversation between Wallgren and Philipp-Edmonds from SQE.com: