Ex-Jane Street Exec Explains How They Cut Trading System Latency by 100x
ivan_bezdomny · x · 2026-08-05
Brett Harrison recounted leading the rebuild of Jane Street's core trading system in 2015, successfully cutting system latency by two orders of magnitude.
He shared key underlying optimization techniques:
- Zero Allocation: Programs pay a steep latency penalty when allocating memory for objects on the heap. The simplest solution is to avoid dynamic memory allocation entirely.
- OCaml Constraints: Jane Street famously uses OCaml, a strongly typed language that relies on a dynamic garbage collector by default. While most competitors use manual memory management languages, Jane Street's tech culture strictly mandated that all risk-sensitive code be written in OCaml.
- Core Library Overhaul: To achieve zero allocation in a garbage-collected language, multiple tech groups collaborated to build custom, zero-allocation core libraries.
More from Infra
- Groq Unveils LPX Architecture: Combines Its LPU with NVIDIA GPUs for Inference — GroqInc · 2026-08-05
- Garage Solar-Powered 384GB Xeon Rig for Remote AI Coding — angadsg · 2026-08-05
- AMD CEO Lisa Su to Analyst: Your Data Center AI Number is Probably Too Low — firstadopter · 2026-08-05
- Open-Sourced Recipe: Running a 27B Local Agent 24/7 on a Single RTX 5090 — max_paperclips · 2026-08-05
- Europe Pledges €30B for AI Gigafactories, Only €1B Actually Committed — sanjaykalra · 2026-08-05
- Modal Optimizes Serverless Architecture to Reduce Network Latency — AAAzzam · 2026-08-05