This column is not about Facebook (I promise)
I had no intention of writing about Facebook again so soon after last week. The back-and-forth between the company and politicians were on full display. Then the whistleblower story blew - and online opinion-makers did their thing. I’m not even getting into any of that here. I said what I was going to say.
What was the other thing that happened to Facebook this week? The outage - and what happened next. That’s what I want to talk about.
Facebook is just the unlucky example that demonstrates how fragile the Internet still is.
On Oct. 4, some poor soul pressed a button and made Facebook disappear for 6 hours. And then it happened again.
What went wrong? Due to a configuration error during regular maintenance, Facebook's routers withdrew information from the rest of the Internet on how to find its own servers.
That meant that your browser knew where their server was, but the network didn't know how to get to it. Think about it like this: you want to go to the store to buy something. You know where the store is. However, you have no idea which road will take you there.
People got outraged that this free service, which harnesses their data and alters their lives, was unavailable.
This type of outage can be caused by error or by a malicious act. Sometimes, it’s a combination of both.
In 2008, in Pakistan, the government asked an ISP to censor YouTube within the country. The ISP complied, altered the router configuration and voilà, YouTube was inaccessible around the world.
Similarly in 2014, a Turkish ISP tried to reroute all domestic traffic destined to YouTube and Twitter to its own servers to fulfil their government's request. Instead, it broadcast to the rest of the Internet that it was the best place to route all traffic. The Internet obliged.
“This is how the Internet breaks,” wrote Internet Society Chief Internet Technology Officer Leslie Daigle.
Was this outage a first for Facebook?
Of course not. Another outage happened in 2019 when due to a server configuration Facebook experienced a day-long outages or service disruptions.
Were there others who had outages due to bad configuration?
Many! Here are a few examples, from some of the biggest tech companies around - which you might assume just never made these kinds of errors.
November 2020: Amazon's cloud AWS experienced outage for hours. These were some of the casualties of that outage: 1Password, Acorns, Adobe Spark, Anchor, Autodesk, Capital Gazette, Coinbase, DataCamp, Getaround, Glassdoor, Flickr, iRobot, The Philadelphia Inquirer, Pocket, RadioLab, Roku, RSS Podcasting, Tampa Bay Times, Vonage, The Washington Post, and WNYC.
March 2021: Microsoft's cloud Azure was down due to the Active Directory misconfiguration during migration.
We’ve all experienced technology malfunctions many times.
All these examples should remind us that all the amazing technology around us was created by us and it is a reflection of human ingenuity and flaws at the same time. Sometimes, those flaws cause popular online services to go down. Sometimes the flaws are much deeper.
It is a common fallacy, even among founders of tech companies, who ought to know better, that technology will solve all problems. Certainly, technology gives us all kinds of solutions. But it is still up to companies and their leaders to define challenges and develop strategies, before throwing technology at problems. Anytime an executive recommends that an algorithm do the thinking for them, they've lost the plot.
It is still us who have to do the heavy lifting. And that is the recurrent pattern for now, forever.