Software Fail Watch 2016, Quarter One header background image

Software failures

Software Fail Watch 2016, Quarter One

Compared to the slew of software bugs that hit the news in late 2015, the first quarter of 2016 has been relatively quiet. This seems to match the data patterns we have observed in the past year of collecting software bugs, something I was able to confirm while giving Microsoft’s recently released Sanddance app a try.

First off: the app. The introductory tour with data sets from the Titanic was fascinating, though of course the data set chosen is one with just the right type of information to really showcase Sanddance’s capabilities. Once I had imported in my own data, I found that some of the viewing and organization options became significantly less exciting. Also, during my quick testing session I wasn’t able to sort out why the program would occasionally create multiple “bins” for one category, or how I could arrange the Month column chronologically instead of alphabetically.  Aside from these small hang ups, I had a blast exploring the program and playing around with the data I had collected.

Ok, back to the software fails.

As you can see from these two screenshots, the year-on-year trends for Q1 in 2015 and 2016 seem quite consistent. In both years February had smaller numbers than January (residual from the holidays, I am assuming), whereas March shows a significant uptick in numbers as consumer trends start to pick up again.

Once again the category with the highest reported bugs is Government. This year however, a disturbing amount of the stories have come from the United States elections, with many reports of software improperly counting or discrediting votes and voters.

As to the notable stories of the quarter, there are many. Here are 5 interesting stories from the 107 bugs we have collected so far. Want to see the rest? You can download the full list here.

Whistle-blower sheds light on Blue Cross health insurance system failures

Blue Cross Blue Shield of North Carolina, USA experienced a large “system failure” in early January 2016, resulting in 25,000 customers being enrolled in the wrong health insurance. While it had been apparent from the flood of complaints that the system was not working, it was not until a whistle-blower stepped forward that the story got interesting. According to the source, internal documents prove that leadership at the health insurance titan were aware of widespread software flaws, but decided to move forward with implementation regardless. It’s one thing to be surprised by a bug discovered in production – it’s another thing entirely to pretend they aren’t there. Source.

Software bug assists in bank heist

This story comes in two parts: one software bug related, one not. The first part to hit the news in mid-March detailed how a group of hacker-thieves hijacked the Bangladesh Bank system to steal funds. The group successfully transferred $81 million in four transactions, before making a spelling error that tipped off the bank, causing another $870 million in transfers to be canceled.

The software bug comes in with the $81 million the thieves did successfully steal. According to Bangladesh Bank authorities, a printer is set up to automatically print read-outs of transactions made. The glitch in the system (whether coincidental or created by the thieves), interrupted the automatic printing process, so that is was only several days later that the transfer receipts were even discovered – giving the thieves plenty of time to cover their tracks. Source 1 and 2.

Satellite failure causes global software repercussions

A 25-year-old satellite failed this past January, sparking a software bug that lasted for 13 microseconds (0.000013 of a second). As momentary as the error was, it caused surprisingly large repercussions to global positioning systems. As the cited article states, “The rule of thumb is that for every nanosecond of error, you could be out by as much as a foot…an error of 13 microseconds or 13,000 nanoseconds works out as just under four kilometers.” GPS systems all over the world, from weather to defense, found their systems thrown off for several hours before operations were restored to normal. Other systems said to have been affected by the error included select radio stations, which took several days to be fully restored. Source 1 and 2.

Pentagon’s F-35 Jet Fighter glitches…again

The abject failure of the F-35 fighter jet has become a popular news subject. It is no secret that the plane’s tests were met with disaster over 2015, but it seems that the software bugs just continue to compound. The most recent problem discovered was a failure in the F-35’s radar, essentially rendering a pilot blind until the radar is restarted. Given the fact that the F-35 consists of over 8.3 million lines of code, it is not surprising that problems exist. What is surprising is how the added time and monumental budget (over $400 billion) don’t seem to have been a help. Source 1 and 2.

Hive thermostat app sends users’ homes to 32C (90F)

As “smart” as all the devices are, the Internet of Things is bound to have problems. British Gas’s, “Hive” smart thermostats made the latest media round in late February, when an error continually set user’s homes to a boiling 32C (90F). Aside from the danger the heat posed to vulnerable persons, many angry customers complained regarding the state of their future energy bill due to the glitch. This is second recent blow to the Hive system, riding on the heels of a report in late 2015 stating how hacking the system would provide a “burglar’s dream” of information. Source 1.

More on Software Failures

Software Fail Watch: 2015 in Review

Software Failures of 2015: Quarter Three

Software Failures of 2015: Quarter Two

Software Failures of 2015: Quarter One

Software Failures of 2014: Finance Edition


Check out our Tosca Testsuite fact sheet today to see how you can make sure you never experience a newsworthy software failure.