Watch out! The nanoservices are comIng.
Welcome to the world of nanoservices: smaller than a microservice, bigger than a function, they are the perfect unit of software. Nanoservices are flexible, manageable, and scalable and a great way to do serverless computing. Matthew Clark explains how to get nanoservices right, drawing on his experience at the BBC, which now has over a thousand in production.
Talk Title | Watch out! The nanoservices are comIng. |
Speakers | Matthew Clark (BBC) |
Conference | O’Reilly Velocity Conference |
Conf Tag | Build Resilient Distributed Systems |
Location | London, United Kingdom |
Date | October 18-20, 2017 |
URL | Talk Page |
Slides | Talk Slides |
Video | |
Welcome to the world of nanoservices: smaller than a microservice, bigger than a function, they are the perfect unit of software. Nanoservices are flexible, manageable, and scalable. They are the Lego bricks that let you build huge, complex systems. And they’re a great way to do serverless computing. . .unless you get them wrong, in which case they are expensive, slow, and out of control. Matthew Clark explains how to get nanoservices right, drawing on his experience at the BBC, which now has over a thousand in production, which power many of its biggest web pages. The BBC’s coverage of the 2017 UK general election and 2016 Olympic Games, for example, combined hundreds of nanoservices to create fast and scalable websites and apps. Matthew explores nanoservices, discussing what they are, their pros and cons, and the platform required to support them and outlining key differences from microservices, including how they are deployed and supported. Nanoservices are not deployed in isolation. They are distributed and closely connected (but still loosely coupled) to reduce overhead and maximize cloud efficiency, scaling, and resilience. Nanoservices can be updated independently and quickly, and in that regard, they are ideal for implementing true continuous delivery. In short, nanoservices are a novel approach to large-scale web and data stacks that challenges how we architect, manage, and operate at scale.