How to build reliable systems (with unreliable components)

In this episode of VM End-to-End, Developer Advocates Carter Morgan and Brian Dorsey are joined by Reliability Advocate Steve McGhee to discuss why we have many smaller VMs instead of one large machine, as well as the benefits of scaling horizontally. Watch as they talk about the layers of abstraction, the rule of 9s, live migration, and how planning for failure can create more reliability!

0:00 – Intro
1:34 – How can reliable systems be built from unreliable components?
5:00 – What are the layers of compute?
8:59 – Wait…reliability is like the bulkheads of a ship?
10:48 – The rule of 9s and how to reason about reliability
17:05 – Summary
19:20 – What’s next?

VM End To End Playlist →
Subscribe to Google Cloud Tech →


product: Cloud – General; fullname: Carter Morgan, Brian Dorsey;

Duration: 00:20:03
Publisher: Google Cloud
You can watch this video also at the source.