I built Aphelo to understand what happens to database latency when you move from simple std::map storage to high-concurrency production requirements.The biggest challenge wasn't raw speed, but "Stop-the-World" spikes during hash table resizes. I implemented Progressive Rehashing to amortize that $O(N)$ cost into $O(1)$ chunks.
Performance on consumer hardware:
SET: 172,716 RPS
GET: 137,859 RPS
I'm a 3rd-year student exploring systems internals, and I'd love feedback on the architecture—especially the dual-intrusive ZSet implementation.
I built Aphelo to understand what happens to database latency when you move from simple std::map storage to high-concurrency production requirements.The biggest challenge wasn't raw speed, but "Stop-the-World" spikes during hash table resizes. I implemented Progressive Rehashing to amortize that $O(N)$ cost into $O(1)$ chunks. Performance on consumer hardware: SET: 172,716 RPS GET: 137,859 RPS I'm a 3rd-year student exploring systems internals, and I'd love feedback on the architecture—especially the dual-intrusive ZSet implementation.