Netflix: Life on AWS
A lot of developers and IT folks still harbor distrust for Amazon's cloud (and maybe just public cloud infrastructure in general). You hear about the outages once in a while and some companies decide that they can't afford any outages. The guarantees AWS and other cloud vendors make don't seem to mean anything.
What we forget is that some of the giants of the internet are all-in on AWS. Netflix is probably the biggest success story for cloud deployment. Almost all of their customer-facing services run on AWS's public cloud. So why is Netflix such a big deal? This:
"Netflix user accounts for 33 percent of all downstream traffic in North America during the peak hours between 9 p.m. and 12 a.m." – Don Reisinger
Netflix has learned a lot of tricks over the years for keeping their services running smoothly despite AWS's blemishes. Here are three of their major strategies:
- Micro-services: These smaller bits of functionality keep any failures contained and minimize the negative effects of cloud outages.
- Using Lots of AZs: Netflix builds in a ton of redundancy by deploying each service onto a minimum of 3 availability zones.
- The Simian Army: A monitoring platform with open source tools that are designed for managing tests and other operations on AWS at massive scale.
If you want to hear more about Netflix and all the internal tools they create, listen to the Ship Show podcast from this week and tune in for part two next week.