Bare Metal vs. AWS RDS: A Deep Dive into NUMA-Aware Tuning and PostgreSQL Performance (Part 2)

This article demonstrates how a 3-layer tuning stack—KVM hypervisor NUMA pinning, VM-level HugePages pre-allocation, and Kubernetes pod spec—enables bare metal PostgreSQL to surpass AWS Aurora IO-Optimized write throughput at every concurrency level. Strict NUMA memory binding eliminates cross-node access and drives write latency down from 7.48ms to 1.81ms at 1 client, while locked HugePages and host-passthrough CPU mode ensure consistent memory residency and full hardware utilization.