Industrial Edge Telemetry Ingestion Architecture Specification

Role: Principal Industrial IoT Systems Architect with 15+ years designing resilient backend data pipelines.

Context

• Target deployment scale: {{plant_count}} distributed industrial facilities
• Primary industrial protocols: {{telemetry_protocols}}
• Sustained message velocity: {{peak_events_per_second}} events per second
• Factory network offline tolerance: {{max_network_downtime_hours}} buffer duration
• Cloud analytics target: {{cloud_target_infra}}
• Tiering and cold storage lifecycle: {{retention_tiering_policy}}

Task

Author a comprehensive backend architecture specification for an industrial telemetry ingestion pipeline that guarantees zero data loss during network partitions, converts disparate factory protocols into unified schemas, and streams clean time-series data to {{cloud_target_infra}}.

Method

1. Define the edge ingestion gateway topology, specifying protocol conversion brokers for {{telemetry_protocols}}.
2. Formulate the local store-and-forward caching mechanisms capable of handling {{max_network_downtime_hours}} of line disconnection.
3. Specify network bandwidth optimization strategies including batching, compression, and deadband filtering.
4. Design the cloud message broker partitioning strategy to sustain {{peak_events_per_second}} with zero consumer lag.
5. Establish the canonical data schema using Avro or Protobuf with strict payload validation rules.
6. Detail the stream-processing enrichment layer for metadata injection (asset ID, site location across {{plant_count}} sites, sensor thresholds).
7. Define the persistent storage topology and data lifecycle matching {{retention_tiering_policy}}.
8. Model end-to-end security protocols including mTLS, edge device credential rotation, and air-gapped gateway isolation.

Constraints

• MUST specify deterministic backpressure algorithms between edge brokers and cloud consumers.
• MUST NOT permit uncompressed or unvalidated JSON payloads over the WAN link.
• Edge forwarder memory footprint MUST NOT exceed specified hardware boundaries during network drops.
• All timestamp handling MUST use nanosecond-precision UTC with monotonic clock reconciliation.

Output format

1. Executive Architecture Summary (max 150 words)
2. Protocol Normalization & Edge Topology (ASCII diagram + structural breakdown)
3. Ingestion Pipeline & Storage Specifications (detailed component table with throughput and latency targets)
4. Resilience & Network Disconnection Matrix (failure modes, recovery actions, and data sync ordering)
5. Security & Validation Envelope Spec (schema definition and auth flow)

Self-review

• Did I account for backpressure and queue overflow during a {{max_network_downtime_hours}} network outage?
• Are protocol transforms for {{telemetry_protocols}} fully detailed with schema validation boundaries?
• Does the storage tiering plan directly enforce {{retention_tiering_policy}}?