Philip Laine
Independant
About
Philip has been writing software in some way or another for the last 13 years. During that time he has worked on everything from iOS applications to developing application platforms for the automotive industry. Philip is a active open source contributor within the Kubernetes ecosystem. He used to work on Flux as a maintainer but has since moved on. Currently his main focus is working on Spegel.
Session
Cache me if you can: P2P Image Sharing in Kubernetes with Spegel
Session Recordings
Session Recording
The ability to pull container images quickly and reliably is critical for workload deployment and scaling. Pulling images from external registries can be slow, inefficient, and prone to failure due to network issues or downtime. In this session, we will explore different strategies for speeding up container image pulling in Kubernetes. We will compare the benefits and trade-offs of different image caching approaches, such as leveraging Harbor, preloading images on machine images, or adopting Spegel.
We will uncover how Spegel works seamlessly with Kubernetes to provide automatic local caching of images without explicit configuration, reducing the impact of external registry downtime and avoiding rate-limiting issues. Then, demonstrate how Harbor can be used as a pull through mirror. Attendees will learn to deploy these solutions to optimize their Kubernetes workloads, decrease pod startup times, and improve the performance of edge node deployments.