eBPF technology has roots in the Linux kernel. It is software designed to run programs in a privileged context within an operating system kernel in a sandboxed environment. This is done without touching kernel source code or loading kernel modules. A more complete description can be found on ebpf.io.
The use of the acronym eBPF is historic. BPF stood for Berkeley Packet Filter and eBPF was the extended version. eBPF now does lots more than packet filtering so the appropriateness of this meaning is defunct. The original BPF, mostly obsolete, is sometimes now called cBPF (classic BPF) to distinguish it from the eBPF concept.
A more technical guide to the technology can be found here on the Cilium website. Cilium is a tool built on eBPF used in Google Kubernetes Engine (GKE) and in EKS Anywhere on AWS (for on premise container management).
No comments:
Post a Comment