Container

Optimizing Performance with Unikernels: Exploring Container Runtimes for Serverless Workloads with Knative Benchmarking

In our previous posts, we walked through the process of configuring various low-level container runtimes in Knative using the RuntimeClass feature of K8s. We detailed the setup for isolation mechanisms like gVisor, with a special focus on Kata and its associated hypervisors, including AWS Firecracker and QEMU. Additionally, we delved into the capabilities of unikernels, showcasing the power of urunc in the serverless realm.