Cold starts are definitely a hot topic when it comes to serverless, on any FaaS platform. Suggested remediations such as tricks to make them warmer are all around. But why are cold starts happening at all? Is it a broken design in the infrastructure or just a very hard thing to implement?
Let me walk you through a few concepts that make serverless platforms react the way they do, from a container and sandboxing in Linux, up to packaging and orchestration. Join me in this serverless internals talk to dig deep in the infrastructure, understand more, followed by live examples and interesting experiments.
Cold starts are troubling many developers – after this talk, the audience will understand why they happen. Another objective is to give an in-depth look into FaaS platform internals including a live demo of what’s actually going under the hood when hitting cold starts.
Basic understanding of FaaS and familiarity with serverless latency