Microservices, APIs, and the Autonomous Web

What if you could write and deploy services onto the WWW and be confident that they wouldn’t crash or corrupt data — even when no humans are monitoring them? What if you was able to define and maintain APIs for the Web and get instant feedback whenever your changes threaten to break some client that you didn’t even know existed? And what if you could build services that could automatically resolve dependencies at runtime — even negotiate connection and cost profiles — without any human getting involved?

Sound fantastic? Unrealistic? Impossible?

It turns out, we have all the technology to accomplish this on the open web today, but few are working to put it all together in a way that makes it all safe, cheap, and easy to do.

In this talk, we’ll cover the patterns and practices needed to build autonomous and safe services. We’ll also review an API design aesthetic that ensures continued evolvability without needing to break clients already able to adapt to reasonable changes. Finally, we’ll outline an "open discovery" model that allows services to self-register, find each other, and complete handshaking and connection details — all at runtime live — without the need to direct human intervention.

This session is for anyone responsible for designing, architecting, programming, and maintaining APIs on the Web. The Autonomous Web is fantastic. It is also real and possible. Let’s see what it looks like and start building it today.