Panel Discussion: eBPF: A New Era in Cloud Infrastructure Tools
summary
eBPF runs sandboxed programs inside the kernel to instrument and change behavior for observability, networking and security; real examples: Cilium/Tetragon benchmarks show ~6 million file reads observed with under 2% CPU, and Meta routes every packet through eBPF since 2016. Tooling includes BPF Trace for ad hoc probes, higher-level projects like Cilium and tetragon to avoid kernel coding, and vendor stacks (Datadog, Polar Signals) that hide complexity. Operational notes: verifier compatibility back to 5.3, high privileges required, and careless full-system tracing causes unacceptable overhead.
AI-written summary · about
chapters (18)
- 0:00eBPF panel introduction
- 1:26What is eBPF and why it matters
- 2:49eBPF as cloud infrastructure platform
- 4:36Why eBPF is high performance
- 6:24Cloud performance challenges
- 7:49eBPF production challenges
- 9:59Should developers learn eBPF
- 10:41Learning eBPF tools and resources
- 13:54Contributing to eBPF
- 15:32Making eBPF easier
- 17:19eBPF for network tuning
- 19:29eBPF vendor adoption
- 21:16eBPF criticisms addressed
- 24:06eBPF futures and innovations
- 25:16eBPF for Windows potential
- 27:21Closing remarks
- 28:02eBPF platform development
- 31:10Q&A
related talks
40mPanel Discussion: eBPF: A New Era in Cloud Infrastructure Tools
32mExploring the Power of eBPF for System Observability
29mBuzzing Across the Cloud Native Landscape with eBPF
32meBPF: The Next Power Tool of SREs
30mObservability in Kubernetes: Unlocking the Power of eBPF
36mPanel: Exploring eBPF Use Cases in Cloud-Native Security
35mAn Introduction to eBPF - Superpowers for Networking, Observability & Security
25mWelcome + Opening: Why Everyone is Excited about eBPF in Cloud Native in 2021