An anatomy of a Google downtime: how much do we rely on online services?
The unimaginable has happened. Google went offline shortly this November 6th. Actually, Google does experience limited outages one time or another, although given Google’s reputation as a big technology company, this kind of thing is usually considered unthinkable by users. Can’t access your Gmail for just five minutes, and you’re probably pissed. What more if Google searches or other services don’t work?
Especially now that our mobile devices are always connected and reliant on Google services, any outage or downtime will be easily noticed. Now that businesses are moving their operations to the cloud — such as with Google Apps — any downtime will lead to decreased productivity and even lost revenues.
A couple of days ago, Google had a limited outage of about 27 minutes in some regions. According to network engineer Tom Paseka, who wrote an explanation at the Cloudflare blog, this was due to “route leakage.” Tom writes that the Internet’s “glue” is basically built on trust. The Border Gateway Protocol, which announces which IP addresses belong to each autonomous network, supposedly establishes routes from one network to another. In the case of the Google outage, an upstream provider based in Indonesia had a glitch, which then spread across the Internet at large, due to the wrong routing information being sent to other providers.
Said upstream providers were quickly notified of the situation, and were able to address the issue. In all but three minutes of the fix, the Internet went back to normal for the affected parties. Tom estimates that about 5% of Internet users were affected, mostly from Hong Kong and nearby countries.
This was not a malicious hacking attempt, says Moratel, which was the provider in question. Rather, it was a hardware glitch.
All this underscores the vulnerability of our connected systems in the light of Internet outages, whether these are enterprise systems or consumer devices that rely on information from the cloud. With everything going to the cloud — our emails, documents, messaging services, and whatnot — any outage can lead to severe repercussions. Or at the very least, we could be subject to inconveniences.
This leads me to ask. If Google were to go offline for a day, would you be able to cope? Will you miss anything on your smartphone or tablet if you can’t get online, or can you get by with doing stuff on the device itself, independent of an Internet connection?