I built a Rust in-memory DB that hits 1.6M RPS. What's your biggest scaling bottleneck?

Hey everyone,

I’m Yasir, a systems engineer and founder. For the past few months, I’ve been obsessed with how much multi-core server power is wasted by legacy, single-threaded databases (like Redis), which ultimately leads to crazy AWS bills for startups.

To fix this, I built VelocityDB — a shared-nothing, thread-per-core database engineered from scratch in Rust utilizing Linux io_uring.

The Benchmark:

I recently stress-tested it on a basic 2-core AWS t3.medium free-tier instance. It processed 25 Million requests in 70 seconds (~357,000 RPS) with under 3MB of memory footprint. On a standard 16-core bare-metal server, it scales linearly to over 1.6M+ RPS.

I am launching VelocityDB here on Product Hunt tomorrow! 🚀

But before the launch, I want to ask the technical founders, CTOs, and systems engineers here:

What is the biggest database or infrastructure bottleneck you are currently facing? Is it latency, cross-core locking, memory leaks, or just massive cloud bills?

Would love to hear your real-world experiences and discuss architecture!