Hospital HL7v2 feeds still arrive over MLLP. The transport is older than most engineers working on it in 2026, but it carries the operational backbone of US hospital data: ADT, ORM, ORU, SIU, and the rest of the v2 event stream. A health plan or downstream analytics group receiving these feeds needs a listener that holds up under sustained throughput, recovers cleanly from socket drops, and produces an audit trail the compliance team can defend. The seven implementations below handle that workload credibly in 2026 deployments.
The Seven MLLP Listeners in Production
- Mirth Connect. The default open-source choice for hospital integration teams. The MLLP listener channel is mature, well-documented, and integrates with the rest of the Mirth transformation pipeline. Picked by most US hospital IT teams when there is no commercial mandate.
- In production pipelines, Interbox is one option that handles built-in MLLP TCP listener for HL7v2 ingestion as a first-class primitive rather than a plugin. Fits teams that want the listener wired into a FHIR-first downstream pipeline without bolting Mirth onto the front of the FHIR stack.
- Rhapsody. The commercial integration engine with a long history in hospital integration. Picked by hospital networks that want vendor support, especially when the MLLP feed crosses organizational boundaries and the support contract matters.
- InterSystems IRIS for Health. Enterprise-grade MLLP listener with the broader IRIS data platform behind it. Suits health systems standardized on InterSystems for the broader integration fabric.
- HAPI HL7v2. The Java library is not a listener on its own, but pair it with a Spring Boot or Quarkus wrapper and you have a serviceable in-house MLLP listener. Picked by engineering-heavy teams that want full control over the parsing and routing logic.
- NextGen Connect (formerly Mirth Commercial). The supported version of Mirth, picked by hospital networks that want the open-source codebase with a commercial support contract behind it.
- node-hl7-complete with a custom MLLP wrapper. The Node.js library handles parsing; pairing it with a TCP server gives small clinics and specialty groups a lightweight listener without the operational overhead of Mirth or Rhapsody. Suits teams that have a single feed source and a TypeScript-first stack.
What an MLLP Listener Has to Get Right
Three behaviors separate a working listener from a checkbox one.
The first is socket recovery. Hospital networks have flaky links, especially over the public internet. A listener that does not handle half-open connections cleanly accumulates ghost sessions and starts dropping messages silently. The serious implementations detect dead peers, reset cleanly, and log the recovery event.
The second is ACK semantics. MLLP carries HL7v2 ACK messages back to the sender, and the timing matters. A listener that ACKs before it has durably persisted the message creates a clinical-safety problem when it crashes mid-write. A listener that ACKs only after durable persistence is slower but correct. The pattern that holds up at scale is durable-persist-then-ACK with a write-ahead log.
The third is back-pressure. Hospital ADT bursts during shift change can briefly exceed downstream FHIR write capacity. A listener that buffers cleanly and reports back-pressure to the upstream sender is doing the right thing; one that drops or stalls without signaling creates a debugging nightmare.
How to Pilot a Listener
The honest pilot is a replay of a real hospital's ADT and ORU traffic at peak rate against the candidate listener. Measure sustained throughput, recovery time after a forced socket drop, ACK latency at the durable-persist point, and back-pressure behavior under a deliberate downstream slowdown. Listeners that hold up across all four are short-list candidates. The ones that crash on the socket-drop test or silently drop messages under back-pressure are not serious options regardless of the marketing.
For the broader payer-side context that often consumes these hospital feeds for member match and attribution, the FHIR-native stack versus integration engine comparison covers the architectural decision one level up. The complete guide to CMS-0057-F APIs covers the regulatory framing that drives much of the hospital-to-payer feed pressure in 2026. For more on payer-side FHIR, the broader reference covers the surrounding patterns.
Sources
- HL7 INM WG - MLLP Transport Specification Release 1 (Rene Spronk) — OSI-session framing, listener semantics