–
Understand Facebook’s Flaw
–
Facebook and its other services, such as Instagram and WhatsApp, do not they were just down
on Monday (4), but they also ceased to exist for the internet.
After a chaotic day, the company published an explanation
on the downfall of its services, written by Facebook’s vice president of infrastructure, Santosh Janardhan. after almost 7 hours
off the air, the executive said that “configuration changes to the backbone routers that coordinate network traffic between our data centers caused problems that disrupted that communication.”
facebook disappeared from the internet
But, after all, what does this mean? Alberto Azevedo, CEO of information security company CYB3R, says Facebook has simply ceased to exist.
“Facebook wasn’t just down, it wasn’t a situation where I was trying to access the site and the server was down. Facebook literally went off the internet yesterday. It went out and back,” he says. “Facebook’s failure was very complex indeed, which is why it was very serious.”
Alberto explains that the internet is made in layers, and that yesterday’s problem happened in a network layer, which is deep. Each internet distribution node communicates with a few other nodes, which communicate with others, and so on. “It’s like this: I know my neighbors, but my neighbors know the neighbors there, and it goes on until it arrives in an entire city”, explains the specialist.
And for communication to take place between these nodes, a server must announce that a certain service exists. “For me to exist on the internet, I need to advertise: I am here, my addresses are these and I am in these places”, says Alberto.
After an error in a configuration, the Facebook server simply stopped announcing that it existed. This information stopped circulating between the nodes and, for the end user, the result was a drop in service.
“That server that made the ads saying ‘I’m here’, said ‘I’m nowhere else.’ It revoked all the ads from where Facebook was. Once you revoke the ads, the internet forgets to you, it’s as if you no longer exist there,” explains Alberto.
could have been worse
Facebook did not disclose the exact reason that triggered the series of errors, but Alberto says it could have been something simple, like a mistake in an everyday task. If someone executed a wrong command, for example, it’s enough for the failure to happen.
The problem was only so serious because it happened in a deep layer of the internet. “The internet has seven layers. The protocols that we know, which is HTTP, HTTPS, where the internet actually works, is all in layer seven. This BGP protocol [onde ocorreu o erro] it is in layer four, a network layer. In other words, it works down there”.
“Usually 90% of problems can be solved remotely. Now, when you have a problem to solve at a network layer, you have to be on the server,” says Alberto. That’s why the Facebook has moved teams to its servers
– and the employees still had to deal with building access issues
because even the badges didn’t work.
Given the size of the failure, it was possible that Facebook services returned to the air with an error or took too long to recover. “Facebook has the best of the best at work. If you think about the magnitude of the problem, the fact that they were back on the air in seven hours is impressive. out of the air”, evaluates the specialist.
– .
