Dual Matching Engine System for Low Latency Global Trading
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed computerized exchange systems face challenges with latency and bandwidth, particularly when offering 24-hour global trading with a central matching engine located in one physical location.
Innovation Solution
Implementing a dual matching engine system where a passive matching engine processes order entry messages and creates broadcast data, while an active matching engine handles order matching and sends confirmation data, thereby reducing bandwidth usage and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a single central matching engine is used for 24-hour global trading, then the system can serve traders worldwide, but latency increases due to the distance between the matching engine and connecting parties
Solution Approach 1:
The system divides the single central matching engine into multiple distributed matching engines located in different geographical regions. Each matching engine serves traders in its local region, reducing the distance data must travel and thereby minimizing latency while maintaining global trading capability.
Solution Approach 2:
The system introduces a message bus as an intermediary component that enables communication between distributed matching engines and trading terminals. The message bus facilitates efficient data exchange without requiring direct connections to a central engine, further reducing latency and enabling scalable global deployment.
2Device complexity
If a single central matching engine processes all order entry messages, then the system architecture is simple, but bandwidth consumption increases due to the massive amount of information that must be distributed to all participants
Solution Approach 1:
The system segments the order processing workload across multiple distributed matching engines. Each engine processes only the messages relevant to its local region, significantly reducing the total bandwidth required for message distribution compared to a single central engine that would need to transmit all messages to all participants.
Solution Approach 2:
Each distributed matching engine maintains local state and processes messages locally for its region, reducing the amount of information that needs to be transmitted across the network. Only necessary updates and confirmations are exchanged between engines and terminals, optimizing bandwidth utilization.
3Loss of time
If order entry messages are processed immediately at receiving terminals, then latency is minimized for local traders, but the matching engine cannot efficiently coordinate across global regions
Solution Approach 1:
The system segments the matching engine into region-specific instances that can immediately process messages for local traders, minimizing latency. Simultaneously, these segmented engines maintain coordinated communication through the message bus to ensure consistent matching across global regions, resolving the contradiction between immediate processing and coordinated matching.
Data Source
AI summary
The present technology relates to distributed computerized exchange systems for trading of financial instruments. In particular it relates to a passive matching engine and an active matching engine that cooperates in handling data messages such that less bandwidth is used and so that improvement in latency can be achieved.


