Uploaded October 2021 | Updated September 2026, 1 week ago

---

In late 2017, Netflix did an AWS "This is My Architecture" video. It was one of the first. In the video they explained how they tackled the problem of failing over when disaster struck.

Specifically, they drilled down on a clever use of DNS records to create a very flexible system that allows services to continue in the event of some pretty significant disaster.

Few companies operate at Netflix's scale (even their scale back then) but this technique is broadly applicable for any team that needs to failover in the event of either disaster or even a simple failure.

Now, a few years later, I react to that video and see what's stood the test of time, what could be done simpler given today's technology, and generally critique the design against the AWS Well-Architected Framework in a mini AWS Well-Architected Review.

The original video: youtu.be/WDDkLOT8SCk

More on the AWS Well-Architected Framework is at aws.amazon.com/architecture/well-architected

Follow me on Twitter at twitter.com/marknca

Read my AWS Hero bio at aws.amazon.com/developer/community/heroes/mark-nunnikhoven

Learn more about the AWS Well-Architected Framework in my course on A Cloud Guru at markn.ca/courses/#mastering-the-well-architected-framework

Index:
- [00:00] Intro
- [00:24] Problem statement
- [02:10] Evacuating a region
- [03:10] Geo-routing
- [04:39] Intelligent failover
- [06:11] How a failover works
- [07:26] Record choice
- [08:13] Restoration
- [09:10] Tracking state
- [10:03] Can we improve?