AI Infrastructure Security
Security for the platforms AI runs on: workload and agent identity, keys and PKI, the model supply chain, and the controls auditors sign off on.
- The AI Inference Networking Maturity Model
A vendor-neutral, five-level model for judging how ready your network is to serve AI inference at scale, scored across six dimensions.
- BGP for the AI Era: Multi-Region Routing for Inference Workloads
Design BGP and anycast routing for multi-region AI inference, tying route health to inference SLOs, regional GPU capacity, and predictable failover.
- eBPF in Production: Kernel-Level Observability and Security
Use eBPF for kernel-level observability and runtime security, with Cilium, Tetragon, bpftrace examples, and the trade-offs against sidecars and host agents.