While some software problems merely anger airline passengers — and we had some of those recently — others can kill. Such was the case in Spain last month, when an Airbus A400M Atlas military transport went down with four people aboard. An investigation determined three of four engines had shut down, and the crew was unable to restart them.
According to the industry blog The Manufacturer, Airbus Chief Strategist Marwan Lahoud said the problem was a “quality issue in the final assembly” of the plane’s software. The blog interpreted that to mean that while original code had no errors, either new problems emerged in installation or the code “was not sufficiently tested in its new hardware environment.”
That’s precisely the type of problem that can be prevented through the use of service virtualization practices, which allow the testing of software in any hardware environment – even if you have no access to that environment.
Dangerous defect in Acura brakes pinned on software problem
Acura has recalled nearly 50,000 MDX SUVs and RLX sedans from 2014 and 2015 because of problems that can cause the brakes to be automatically and unexpectedly applied.
According to an Associated Press report, the vehicles have a safety system that uses radar to automatically apply the brakes if a radar device detects the car is about to hit an object. Because of a software issue, systems in the recalled vehicles can become confused and brake if another car in front of you speeds up. The glitch has caused rear-end collisions when vehicles brake unnecessarily and for no reason.
Michelle Krebs, a senior analyst for Autotrader, told AP that “unfortunately, some of these [problems] will only be discovered in real-world settings with real people behind the wheel.”
Au contraire, Ms. Krebs. With service virtualization-aided testing, engineers can simulate any real-world condition in the safety of their labs. We’re not sure if such techniques were employed by Honda, but it and other automakers should take heed. As systems get smarter and smarter, testing will need to get better and more rigorous at every level.
Southwest website meltdown stresses importance of live-like load tests
Dirt-cheap Southwest Airlines tickets should have been selling like hotcakes. Instead, traffic on the carrier’s website was flat as a fallen soufflé.
The site cratered under the weight of extraordinary demand. Many travelers had no luck getting through. Southwest was forced to extend its sale. However, it had lost the moment because competitors had the opportunity to match prices.
“It was very slow. It was pending it said it did not know the URL,” Mike Weingart, president of the Southwest Chapter of the American Society of Travel Agents, told KTRK.
“I found some dates, but they were not subject to the sale, and when it was all said and done, I was ticketless.”
This was not Southwest’s first trip on the website crash carousel. They, and other enterprises, should be doing a few more tests to be sure their systems can handle the peak traffic. Service virtualization allows live-like simulation of traffic many times heavier than you get ever anticipate.
200 bikes go missing when software code causes company to lose track of them
A company that rents bicycles across London lost track of more than 200 two-wheelers because of a software glitch.
According to The Standard, the general manager of Santander Cycles said the missing bikes were issued without a code that allows them to be tracked.
“We recently experienced a software issue that meant some cycles were released from a small number of docking stations without a valid hire code,” said James Mead.
Teenagers being teenagers, it didn’t take long for them to figure out the problem, and a cheap opportunity for a free ride. The problem was discovered after teens from estates in Westminster were seen using and dumping the red bicycles.