Cold-Chain and Placement Review

Cold-Chain and Placement Review

animation
edge
analytics-ml
TipData Dora, Data and analytics domain expertCheckpoint: Cold-Chain Edge Case

You now know:

  • The cloud-only fleet sends 16 kB/s, or 1.38 GB/day, before any filtering.
  • Level 3 summaries reduce the flow to 535 bytes/sec and 46.2 MB/day.
  • Gateway spending is justified when local alerts still work during cellular dead zones.
TipData Dora, Data and analytics domain expertCheckpoint: Placement Choice

You now know:

  • Edge placement is strongest for sub-second response, high-frequency data, intermittent links, privacy, and simple local rules.
  • Cloud placement fits minutes-to-hours analysis, stored summaries, broadband links, non-sensitive metrics, and cross-site training.
  • The illustrative rail example keeps selected diagnostic features while avoiding roughly 23 TB/day of raw upload.

Concept Check

NoteQuick Knowledge Check

Q: What is the key difference between data in motion (Levels 1-3) and data at rest (Level 4+)?

Data in motion is event-driven, streaming data that flows from sensors through edge processing. Data at rest is stored in databases for batch analytics and historical queries. Level 4 is the critical transition point where real-time streams become persistent storage.

Q: Why is Level 3 edge computing called the “transformation layer”?

Level 3 performs five critical functions: evaluation (check data quality), formatting (standardize schemas), expanding (decode proprietary protocols), distillation (reduce data volume through aggregation), and assessment (trigger local responses). It transforms raw device data into cloud-ready information.

Technical boundaries. The cold-chain flow comparison uses the cloud-only fleet before any filtering and Level 3 summaries. The roughly 23 TB/day of avoided raw upload belongs to the illustrative rail example.

Back to the chapter