Sandbox to confidential containers: A brief on the evolution of container isolation
summary
Containers run as host processes using namespaces and cgroups, so sandboxing adds a kernel boundary: kata-containers puts each pod in a lightweight VM and gVisor implements a user-space kernel that intercepts syscalls. Confidential computing moves isolation into hardware with TEEs/enclaves, cryptographic attestation and a key-broker flow; the CNCF confidential containers project provides a vendor‑neutral runtimeClass and attestation stack. Vendors report performance hits under ~10%; kata needs nested virtualization so cloud use often requires alternate deployment modes.
AI-written summary · about
chapters (14)
- 0:00Introduction
- 0:43Why container isolation matters
- 2:05Security risks of shared kernel
- 3:10Container runtime sandboxing
- 5:42Limitations of sandboxing
- 6:24Confidential Computing basics
- 8:31Hardware-based trust model
- 9:15Confidential Computing requirements
- 9:57Confidential Containers project
- 11:01Confidential containers in action
- 13:33Isolation layers explained
- 14:38Balancing security and cost
- 15:42Why hardware-level isolation
- 16:44Closing and references
related talks
38mConfidential Containers Explained
01h 14mHands-on Workshop: Confidential Containers: Bringing Confid...
14mYour Containers Aren’t Alone: Demystifying Container Isolation
31mYour Containers Aren’t Alone: Demystifying Container Isolation
49mCloud Native Confidential: Isolating Kubernetes Pods with Confidential Computing
31mConfidential Containers – Sensitive Data and Privacy in Cloud Native Environments
38mCloud native Confidential: Isolating Kubernetes Pods with Confidential Computing
20mKata * TEE = A Lego-like Two-way Sandbox for Seamless Security and Privacy