โ† Back to the chapter

Technical detail ยท Edge-Fog Use Cases: Bandwidth Optimization

Review meaning, timing, custody, and priority

Bandwidth problems appear when reduction changes meaning, timing, order, custody, or priority. A summary may erase an exception; a queue may replay stale records as fresh; compression may drop diagnostic detail; a fog buffer may protect telemetry while starving commands.

At the meaning boundary, state what filtering and aggregation lose, which exceptions bypass reduction, and which decisions remain valid. At the timing boundary, batching, retry, backoff, and replay must expose freshness, expiry, and ordering. At the custody boundary, local retention preserves source, timestamp, sequence, quality flags, owner, loss markers, and cloud acceptance state. At the priority boundary, alerts, commands, acknowledgements, updates, diagnostics, and routine telemetry need an explicit order on a constrained uplink.

Consider the vibration gateway on an intermittent cellular link. Normal windows can be compressed and uploaded later. High-severity anomalies bypass the routine queue, carry sequence and freshness metadata, and leave a local record if upload fails. If storage fills, apply the declared policy: for example, keep the newest routine summaries while retaining incident evidence and marking the loss boundary. Otherwise, cloud users may mistake delayed replay for current plant state or miss discarded lower-priority records.

Make every reduction reviewable even when the original data cannot be recovered. If raw detail is gone, the record must still show what was reduced, why that was acceptable, which exception path exists, and what field change forces a retest. A queue-drain log alone cannot prove original event times, earlier drops, duplicate handling, or correct delayed-data display.

โ† Back to the chapter