When a site like Facebook goes down, it immediately prompts speculation. With three billion users across the globe, there were plenty of people who, unable to indulge in their favourite social network, could take to alternatives to speculate on why the site was no longer available.
Explanations ranged from cybercrime, organised hacktivism, to a single thirteen-year-old Chinese child. The truth, though, proved a little more mundane: when it comes to a choice between conspiracy and cock-up, the cock-up is always a good bet.
In short, a maintenance command set off a chain of events that resulted in the outage. And, ironically, the outage also appears to have hampered Facebook’s attempts to rectify the problem.
During maintenance, a command intended to assess capacity accidentally shut down the Border Gateway Protocol (BGP), which manages traffic between the servers. According to Facebook, an internal system that should have audited and prevented this command failed because of a “bug”.
Facebook run their own Domain Name Servers (DNS), the computers that tell internet traffic where their servers are, and, by design, they will stop directing traffic to data centres that are not responding. But because none of the data centres could be reached, it meant Facebook’s DNS removed all references of itself. For a few hours, as far as the internet’s infrastructure knew, there was no such thing as Facebook or any of its associated companies.
What were the consequences?
The immediate consequences were that Facebook products stopped working. Much of the online concern was that Facebook, WhatsApp, and Instagram were down. But other Facebook properties also failed. Oculus owners were unable to use their headsets, and anyone who used Facebook as a sign-in method on third-party sites were unable to access those accounts. It also had financial effects on the company. Facebook lost about 5% of its market value, around $50 billion (£36 billion) and Mark Zuckerberg saw his personal wealth drop by $6 billion (£4 billion or around 15 million of his custom-made t-shirts).
But there were some practical effects at Facebook too. The disconnection of their servers also disabled their internal communications. Facebook’s employees were unable to communicate with each other, and the outside world, using Facebook channels. This loss of communication appears to have hampered their ability to respond to the outage.
While Facebook has not confirmed exactly how the issue was addressed, there are some suggestions that those with the authority to resolve the problems were off-site, and therefore unable to access servers that were now cut off from the world. And there were reports that when engineers arrived, they struggled to get in because the security systems were linked to the disconnected servers. Facebook has not confirmed this, but did opaquely point out, “these facilities are designed with high levels of physical and system security.”
Is it a cover-up?
Facebook, of course, would say it was a simple mistake, even if it had major consequences. But those who like to believe in conspiracies have pointed to the timing, with the coverage of the outage detracting attention from the testimony of a Facebook whistle-blower (as well as preventing many from sharing it).
Frances Haugen has been testifying before a Senate committee following her leak of tens of thousands of internal Facebook documents. She claims the company prioritised profit over people. Her allegations included that the company ignored evidence its platforms were harming children’s mental health, was promoting hate speech, and knowingly allowed misinformation to be promoted before the 6 January Washington riots.
Facebook has rejected these claims. Mark Zuckerberg issued a statement saying they were “just not true” and “don’t make any sense.” However, they are part of a long line of allegations that Facebook has failed to clean up the platform because division attracts attention and visits, and which in turns generates advertising revenue.
For most people affected, the six-hour outage will have been an inconvenience. And they will be back on the platforms they were using before. Indeed, given Facebook’s global reach, there will be some people who didn’t even notice the issue because they were at work or asleep while it was all happening.
However, it will not be back to normal for everyone. Many Facebook users will have explored alternatives during the outage. Sites like Twitter, and messaging services like Telegram and Signal, all saw surges in use. And while there will only be some who never return to Facebook, there will be many who will reduce their reliance on a single platform.
The biggest danger for Facebook are the additional questions this raises. Facebook and all its services collapsed because of a single point of failure. While they will have addressed this to prevent a recurrence, it highlights that, despite being created for its resilience, the internet can still be remarkably fragile.
The loss of a social network may be inconvenient, but it caused very real difficulties for those relying on services like WhatsApp, which is frequently used for critical communication. Those who are concerned about the size of the platform or thier reliance on it now have a little bit more to be concerned about.