Eli5: How does netflix work perfectly almost all the time using chaos monkey ?

450 views

Or what is chaos monkey?

In: Technology

Anonymous 0 Comments

Chaos monkey is a testing tool. Imagine if the network was made of separate computers, the “Chaos Monkey” would wander around and turn them off randomly and then turn them back on randomly. That’s the concept behind the tool. If your software can detect the computers being turned off, and work around them without the customer noticing any missed frames of their movie, you “pass the Chaos Monkey test”.

in reality, there is no monkey and there are not switches. All the computers are virtual machines running their processes in a big data center. The actual CM test tool just kills processes at random and lets the operating system restart them. It’s the same idea implemented more reasonably.

A number of special fault-tolerant designs are used, and Amazon Web Services has published a number of papers describing how they work.