Real-time Reference Data Joins
Live market ticks and trade events arrive normalized. They carry various identifiers, a price, a quantity, and not much else. The context, including counterparty and instrument details, lives in slowly-changing dimension tables maintained somewhere else. 3forge joins streaming events to those reference tables in memory as they arrive, so downstream consumers always receive a fully enriched record without round-tripping to a lookup service.
The Challenge
Thin Events, Heavy Lookups, Stale Joins
Real-time consumers need each event enriched with reference data, counterparty, instrument, facility, the moment it arrives. That reference data lives in separate systems on their own update cycles, so keeping enrichment both fast and current is a standing tension rather than a settled problem.
Lookups add latency
Enriching each event against an external reference system adds a hop per message, and under burst load those lookups can become the throughput ceiling.
Reference data changes intraday
Credit ratings, corporate actions, and instrument masters move during the day, so a reference snapshot refreshed on a schedule can trail the live stream.
Keeping consumers consistent
When risk, surveillance, and reporting each enrich on their own, keeping their joins aligned so they agree on the same record takes deliberate effort.
Late-arriving reference data
An event can arrive for an instrument or counterparty created moments earlier, before it appears in the reference set, and downstream controls still need a correct answer.
The 3forge Approach
Streams and Tables, Joined in Memory, Always Fresh
3forge models live feeds and reference tables in the same runtime. The Relay streams events into real-time tables, while reference is cached in adjacent real-time tables that are refreshed based on timers. Joins are expressed in AMIScript against these in-memory structures and re-evaluated incrementally on every tick, with late-arriving rows reconciled automatically. Every downstream consumer reads the same enriched record from the same place.
Key Capabilities
What You Can Build with 3forge
In-Memory Stream-to-Table Joins
Join every incoming event against resident reference tables in microseconds, with no per-event hop to an external lookup service.
Live Reference Subscriptions
Subscribe directly to upstream reference systems and apply inserts, updates, and deletes to the resident dimension as they happen, so intraday changes are visible immediately.
Single Enriched Output
Publish one enriched stream that risk, BI, and operations consume identically, eliminating per-consumer join drift.
Late-Arrival Reconciliation
Bulk updates events with missing references automatically when the missing data appears.
Multi-Table Enrichment
Chain joins across counterparty, instrument, book, and legal-entity dimensions in a single pass without writing intermediate stages to disk.
Point-in-Time Replay
Replay any historical event through the join engine using the reference state as it was at the original timestamp, for accurate reconciliation and audit.
Get started
Ready to see it live?
Book a 30-minute demo with a 3forge solutions engineer and see real-time reference data enrichment running on your shape of feed.