The Internet and computers have made life in the twenty-first century easier and more convenient. In some cases, computers and software have saved lives. But sometimes things go wrong and software fails. In some cases, software failure is catastrophic, even causing the loss of human lives. The examples of software failure listed below demonstrate that however “smart” computers or software are designed to be, human oversight is still necessary to prevent disaster from occurring.
In June 1996, the European Space Agency launched the Ariane 5 rocket into space. However, a control software glitch cut its flight short during a test launch. The rocket self-destructed, exploding into a ball of fire half a minute after its initial takeoff. Fortunately no one was injured or killed in the incident,
Computers run amok caused Knight Capital to lose half a billion dollars in half an hour in 2013, almost running the company into bankruptcy. A bug in the software allowed computers to buy and trade shares with absolutely no human oversight. Although the flaw was eventually fixed, the company lost seventy-five percent of its market value in just two days as result of that programming glitch.
In a tragic case during the 1980s, five patients died from radiation poisoning after a software glitch caused them to receive massive overdoses of X-Ray therapy from Therac-25 radiation therapy machines. The bug was believed to have been created by a glitch that commanded the machines to perform two operations at once. This horrible incident illustrates the essential nature of redundancy and oversight in mission critical operations, and the disaster that can occur when this safeguard is neglected.
The Amazon cloud crash of 2012 was triggered by an act o God – severe electrical storms. However, the problem was made much worse by the discovery of previously undetected software glitches. The Amazon crash raised questions that still persist about the security of data stored in remote servers. These questions affect virtually everyone in the developed world, because cloud storage involves email messages, photos and social media. While no lives were lost in the Amazon crash, it is not difficult to imagine that losing medical records stored in the cloud would create potentially life threatening risks.
Northeastern US Power Outage
In 2003, much of the northeastern United States went dark for days during what was at that time the second most widespread blackout in all of human history. The disaster began as a localized blackout that went undetected, and then triggered a series of cascading blackouts. The culprit was ultimately traced to a software bug in General Electric’s monitoring program – a similar error that caused tragic X-Ray overdoses in the 1980s.
Earlier in 2013, a glitch in the Sabre reservation software system grounded the entire American Airlines fleet. Before the incident, several airlines had been merged, and as a result two existing systems were patched together. Apparently something went wrong during the patching operation, which involved programs written on different platforms, in two different programming languages.
Zoe Fenner is a software developer. She likes to share her best findings with other developers on tech blogs. Learn more about HighSpeedInternetProviders by clicking the link.