InterSystems EMPI Benchmark Report

InterSystems EMPI™ is a next-generation enterprise master person index – an automated, easily integrated solution for identity resolution. It is available as a standalone solution or can seamlessly pair with InterSystems HealthShare Unified Care Record®, merging multi-source data into a single, accurate longitudinal health record.

We have recently completed a performance and scalability benchmark of our latest version of InterSystems EMPI. The test profiles consisted of two distinct workloads – Bulk Ingestion and Person Search. Testing focused on several key areas including bulk data ingestion, large data volume, and a high rate of sustained transactions per second (TPS). This document describes the observed characteristics and also provides general configuration and sizing guidelines for InterSystems EMPI.

The benchmark simulates workloads that have been designed to match typical production-like deployments starting with massive bulk loading demographics of a 308-million persons data set generated using Synthea (the most popular synthetic data generator for FHIR). Once ingestion is completed, various person search workloads were conducted mixing concurrency levels to demonstrate scale and performance along with differences between deterministic and probabilistic search methods. Details of the workload simulations are described in the Workload Description and Methodology section of this report.

The 2025.2 version of InterSystems EMPI has demonstrated a sustained average bulk ingestion throughput of over 5400 transactions per second. This high pace of ingestion allowed for over 308 million persons to be loaded in ~13.5 hours. Person search round trip average latencies are as low as 100ms for deterministic and 400ms for probabilistic searches for an aggregate average 263ms person search response time demonstrating massive national-level scale of population volumes with consistent sub-second response time searches.