Trading at light speed: designing low latency systems in C++ - David Gross - Meeting C++ 2022

In this monologue podcast, David Gross discusses building low-latency C++ trading systems, emphasizing the increasing retail flow in financial markets and the need for speed in trading. He outlines four key areas: data models, concurrent access, system tuning, and performance measurement. Gross advocates for designing with performance in mind, balancing strategy and tactics, and highlights the importance of data locality and size for cache efficiency. He shares strategies for sharing data between applications, including the use of SecLock for state sharing and a custom ring buffer for event distribution, and stresses the importance of continuous performance measurement in production to maintain system speed and robustness.