In low-latency finance, different threads (e.g., market data ingestion vs. order execution) must communicate without OS-level locks or blocking.