Interoperability
A dropped line doesn’t lose a message.
Hospital internet is not always reliable. The bridge is built to absorb the gaps.
The problem
The link between a hospital and the cloud is not guaranteed. Connectivity drops: a router reboots, an ISP flaps, a line is cut during construction. In that window, an analyser is still producing results and an HIS is still emitting messages. Anything that depended on a live connection at that instant would be lost.
Queue on the outage, flush on reconnect
- The mechanismorophin keeps a **durable local outbox**, a persistent on-disk queue that survives a process restart. Messages from your machines are written to the outbox and held there; when the connection to the Pensieve backend returns, the outbox delivers them. A network outage becomes a delay.
The honest limit
The on-prem bridge is designed to absorb machine-to-cloud outages. The staff web app is a different matter: it is a cloud application, so a member of staff at a workstation still needs connectivity to use it. We are direct about that boundary rather than implying the whole platform runs offline. The durable outbox protects the interop traffic; it is not a claim that the hospital keeps working with the internet down.