Making the Internet faster. And sleeping well at night
What happens with network requests between your clients and servers? How to make them faster, not break things and keep quiet on-calls? This talk will show how to architect, engineer and operate complex fault-tolerant systems, using Netflix network optimizations work as an example.
A lot of people know about what Netflix is doing in AWS cloud, in microservices architecture and chaos engineering. They know about Netflix' approach to DevOps. Somebody is aware how Netflix delivers about 15% of global Internet traffic with its in-house CDN (Open Connect). Personalization and recommendations come from the cloud. Video and other static files — from CDN.
What happens if we will connect two systems? Use CDN to speed up dynamic requests to the cloud? Send requests from the client through the fastest way on the Internet?
This talk is about how the team of three engineers solved this problem. It’s about how to collect and use Internet performance metrics. How to redirect millions of requests through own infrastructure. How to do it in a tight schedule but without risk to clients. And how to make team receive less than one critical alert per week, although something changes and breaks on the Internet.
This talk is not for those who work in corporate giants. Presented principles and techniques should be known and practiced by every owner of Internet products. You will hear about metrics collection practices, fault-tolerant architecture and safe approach to development on the example of tasks that have been successfully solved in Netflix.