RPC Services for Sharing Trading Events Across Processes

Summary

The document explains how a trading application can use a remote procedure call (RPC) service to share events and handle requests across separate processes. It frames RPC as a way to work around Python’s global interpreter lock limiting CPU-bound work in a single process, while allowing each process to retain its own isolated memory. A server process connects to a trading interface and pushes events to client processes on the same machine or across a local network.

The guide outlines how to load and start the server and client components in graphical or script-based setups. The server owns the trading connection and client processes connect through an RPC gateway, without entering account details again. It also describes sample server and client launch patterns, including a server that can run with or without a graphical interface. The material is an integration walkthrough rather than a performance study: it gives no measurements of throughput, latency, reliability, or scaling, and it does not discuss network security or failure recovery.

Key ideas

• RPC lets separate processes exchange trading events and requests despite isolated process memory.
• A server process connects to the trading interface and distributes events to client processes.
• Clients connect through an RPC gateway and use the server’s existing trading connection.
• The setup can run on one machine or across a local network, with graphical and script-based launch options.
• The guide provides no benchmarks or detailed guidance on security and fault handling.

Tags

From a private course collection; the original is not published.