Monitor Devices Over an Unreliable Network

Design a central system that communicates with devices in a power-grid setting over the **unreliable public Internet**. Devices must report their status to the system, and those reports may be delayed or arrive out of order. Explain how the system should exchange messages, maintain a useful view of device status, and recover after connectivity problems. Clarify whether the required communication is limited to monitoring and status requests or also includes state-changing commands. The latter should be treated as a conditional extension; no power-control algorithm is specified. Fleet size, required freshness, device capabilities, message format, and protocol are open requirements to clarify. Label the assumptions behind your proposed design. ### What a Strong Answer Covers - Device identity, message identity, retries, and the distinction between acknowledging receipt and completing a requested action. - A rule for handling duplicate and out-of-order reports, including device restarts. - The difference between the newest received observation and a fresh observation of the device's present state. - How the interface represents stale or unknown status during a network partition. - Durable ingestion, reconnection, and checks that exercise delayed reports and lost acknowledgments. ### Follow-up Questions - What happens when an old report arrives after a newer one, or after the device has restarted? - If the central system receives no response, what can it conclude about the device itself?