Chronicle Matching Engine Framework

A Chronicle Matching Engine instance encapsulates a single processing pipeline, which may manage one or more order books, and governs the interfaces between the pipeline and the broader exchange environment. At its core, each instance features a single event loop that controls and coordinates all activity within the system.

The architecture of the Matching Engine is designed to meet key non-functional requirements, including:

• Low, predictable latency with high throughput

• High availability and disaster recovery (HA/DR) with minimal downtime

• Live upgrades

• Offline querying and analytics

• Deterministic behavior to enable reproducibility offline

By leveraging a queue to decouple the instance from other feeds across the exchange environment, the system offers flexibility in distributing components to optimise resource utilisation. This approach also enables the segregation of services based on latency profiles, for example, drop copies, which are typically less latency-sensitive than market data, can be handled outside of the critical path.