How this runs on IBM LinuxONE
Reunite runs on IBM LinuxONE, a Linux server built on the IBM Z mainframe architecture (s390x). Reports are stored and matched on the server, not in anyone's browser.
What happens to a report
- A family sends a report from this website. It goes straight to the Flask application on LinuxONE.
- Authorities record people who have been found. They use a separate staff portal, and the server checks their access on every request.
- The matching engine runs on the same LinuxONE server. It ranks possible matches, each with a score and its reasons.
- A trained person checks every suggestion and confirms or rejects it. The software never confirms a match by itself.
- Families only ever see a status: Searching, Possible match under review, or Reunited. Details about found people stay on the server.
Loading benchmark results…
Why LinuxONE
Disaster data is personal and sensitive. Keeping storage, matching and the website on one secure server means fewer copies of the data and fewer places it can leak. The app uses plain Python and static web pages, so it runs on s390x without special build tools.