One Day. One Track. One Community.
worked with different flavours of Cloud tech mostly at IBM
*Operating serverless at scale: perks for developers and lessons learned being a serverless platform vendor. * Serverless is amazing, but surprisingly it still involves operations, to a lesser extend for developers and the bigger for the vendors. How to further reduce that burden for adopters and to run a multi-region public serverless platform at the same time? In this talk we are going to touch base about a new way of orchestrating serverless applications - Composer for Apache Openwhisk and take you behind the bike shed showing you how we operate Cloud Functions at IBM. We will introduce a new open-source serverless tool, Composer, developed by IBM Research that aims to resolve some of issues developer face in the serverless community. It will identify common architectures that struggle with using serverless platforms and think about how to re-implement them with these new approaches. We will also share the experience we have collected over the last two years running real customer workloads. We're going to expose all details users never have to think about (but might be still curious). It'll discuss all the challenges we faced and is filled with stories of debugging distributed systems, hardware disasters and the relentless pursuit of bitcoin miners.