When Tuesday came, and a lot of people suddenly stared into a time without Facebook (News - Alert), it was undoubtedly a difficult time. The service went down for about an hour, at last report, and not surprisingly a lot of thoughts turned to hackers. This time, though, the cry of “hacker!” was premature, and as it turned out, quite wrong, as Facebook revealed the cause of its hour's outage: internal injuries.
Facebook released a statement not long after the outage had hit, revealing that the cause of the shutdown was not “...the result of a third-party attack but instead occurred after we introduced a change that affected our configuration systems.” Unusual for Facebook, certainly, but not out of line; these kinds of things can happen, particularly when dealing with computers.
Still, however, the impact of that change was widespread and often heavily-felt. The hour without a Facebook page represented what may have been the largest outage in nearly five years, dating back to September 24, 2010, when the site reportedly went out for two and a half hours. Reports from Web tracker DynaTrace noted that over 7,500 websites featured services affected by the Facebook outage, and the outage extended to not only Facebook but also to Facebook-owned Instagram, a development that would lead to some losses.
With Facebook down, many took to Facebook's chief competitor, Twitter (News - Alert), to complain about the outage, with the hashtag #facebookdown generating plenty of action. Given that 1.35 billion people, at last report, have some connection with the service, it was a noticeable, if ultimately survivable, outage that may have hit at the worst possible time: when large portions of New England had shut down anyway during a snowstorm, and were likely to start using the service in earnest.
Analyst Debra Aho Williamson, who serves with research firm eMarketer (News - Alert), offered up some comment on the outage, noting that it was “...over quickly...” and “...easily fixed and life came back to normal fairly quickly,” comparing it to, once again, the snowstorm targeting New York City. The projections for damage, meanwhile, ranged from slight losses of traffic to slighter losses of ad revenue. Worse, the outage came a day ahead of Facebook's plans to release quarterly earnings figures.
It's a point that comes up every so often: nothing has 100 percent uptime, ever. Even by most standards, the five-nines concept—99.999 percent—is considered optimal when it comes to uptime on systems. The loss of Facebook for an hour every couple of years or so is much better than that, and we should hope for similar uptime in the future. Of course, it's also a good idea to have a backup plan ready to address such outages before same happen. Whether it's a power loss or the loss of a cloud-based system or anything like that, having a little redundancy in the system is just good practice. We hear a lot about eliminating redundancy, classifying it as waste, but when there's just a little of it on hand, it becomes a perfect way to bridge the gaps in emergencies. Like a spare tire or renter's insurance, its value isn't immediately felt, but when it becomes necessary, it's worth its weight in gold and beyond.
So the big Facebook outage of 2015 has come and gone, and for many, it's hopefully the last for another five years or so. But as is so often the case, we must be prepared for our technology to betray us, and respond with appropriate backup measures.