How we verify that each integration layer holds up under real mainframe traffic, edge cases, and long-running sessions.
How we define the boundary of a legacy integration, what we test, and where our responsibility ends.
We treat any system that predates your current corporate network standard as legacy. This usually means mainframes running CICS, IMS, or custom transaction managers, but it also covers older Unix or AS/400 environments. The key is not the age of the hardware, but whether the system uses proprietary protocols, fixed-length records, or batch-oriented interfaces that modern services cannot call directly.
No. Heirapi builds an API bridge that sits in front of the existing mainframe and translates modern REST or event-driven calls into the native protocol. The mainframe stays as the system of record. We do not migrate your COBOL or PL/I applications, and we do not rewrite the core transaction logic. Our work is the secure technical layer that exposes the data and functions you choose, without disturbing the host environment.
You own the API specification, the code, and the deployment artifacts. We hand over the full repository, including the protocol mapping definitions, the authentication configuration, and the operational runbooks. We can stay on for a support retainer, but the intellectual property and the right to modify the bridge are yours from day one. There is no lock-in to a proprietary runtime.
Testing is split into three phases. First, we run a connectivity test against your staging mainframe to confirm the protocol translation works. Second, we execute a data mapping audit, comparing sample records from the mainframe with the JSON or XML output of the API. Third, we run a load test that simulates the expected peak call volume. You receive the test report and the sample payloads for your own verification before we go live.
If the mainframe uses a custom protocol with no public specification, we start with a packet capture and a session trace. We document every field and handshake we observe, then share that analysis with your team for confirmation. If the protocol is too obscure to map safely, we will tell you before we write any code. We do not guess when it comes to transaction integrity.
Short notes from engineering and operations teams about how the API layer held up during rollout and after steady state. No slogans, just what changed on their side.
We pointed the new CRM at the mainframe through the Heirapi bridge and retired the nightly export job. The same data now arrives in under a second, and the ops team stopped patching the old sync script.
The security review was the longest part, and that was mostly our own checklist. Once the mutual TLS and credential vaulting were in place, the actual cutover took an afternoon. We kept the old batch process running for two weeks as a fallback, but never needed it.
What convinced me was the error handling. When the mainframe had a slow transaction, the bridge returned a clear timeout instead of hanging the whole request. That alone saved us from a few late-night calls.
We had a custom protocol on the legacy side that no one fully documented. Heirapi mapped it field by field and gave us a test harness before the first deployment. The integration went live without a single data mismatch.
The monitoring dashboard made the difference for our audit team. They could see every request that crossed the bridge, including the mainframe session IDs. That level of traceability was something we did not have with the old point-to-point links.