Real-Time Telematics Stream Ingestion Specification

Role: Principal IoT Systems Architect specializing in connected commercial transport and real-time telematics platforms.

Context

• Fleet deployment footprint: {{fleet_size}} active commercial assets
• Ingestion frequency and throughput: {{telemetry_frequency_hz}} reporting cadence per vehicle
• Edge hardware runtime constraints: {{edge_gateway_hardware}}
• Target cloud ingestion infrastructure: {{cloud_ingestion_target}}
• Permissible stream delivery latency: {{max_network_latency_ms}} end-to-end
• Regulatory and telemetry standard: {{compliance_standard}}

Task

Produce an exhaustive technical architecture specification for an enterprise-grade vehicle telematics streaming subsystem, establishing protocol ingestion boundaries, edge-to-cloud failover strategies, stream processing topologies, and data partitioning models to ensure zero signal loss under degraded cellular conditions.

Method

1. Establish the edge gateway data lifecycle, detailing CAN bus message decoding, local buffer compaction, and opportunistic MQTT/QUIC transmission.
2. Define network transition state machines covering intermittent connectivity, edge dead-letter queues, and backoff burst transmission curves.
3. Architect the cloud ingest layer, specifying load balancing, protocol termination, and message broker partitioning keys based on vehicle telemetry schemas.
4. Design the stream processing pipelines for real-time geofencing, anomalous vibration detection, and odometer/fuel delta computations.
5. Formalize the dual storage architecture, delineating high-velocity time-series hot storage from compressed columnar cold analytical storage.
6. Specify telemetry payload data contracts using Protobuf/Avro, detailing field types, precision requirements, and forward/backward compatibility rules.
7. Detail identity management, vehicle mTLS certificate rotation, and perimeter security across all ingress interfaces.
8. Formulate observability metrics for ingestion lag, edge connection churn, and broker backpressure.

Constraints

• MUST guarantee exactly-once processing semantics for billing, driver logs, and safety event triggers.
• MUST NOT exceed {{max_network_latency_ms}} for critical safety alerts and geo-fence violations.
• Edge offline buffering MUST support up to 72 hours of persistent disconnected storage on {{edge_gateway_hardware}}.
• Architecture diagrams and structural definitions MUST use standardized ASCII or Mermaid flow topologies.
• All telemetry models MUST maintain compliance with {{compliance_standard}}.

Output format

1. Executive System Topology (Mermaid diagram + component registry)
2. Edge-to-Cloud Communication Spec (payload protocols, compaction, and reconnect strategies)
3. Stream Broker & Partitioning Schema (broker configurations, partition key distribution, retention)
4. Storage Tiering & Data Contracts (hot/cold schema specifications with Protobuf definitions)
5. Reliability, Security & Resiliency Matrix (mTLS, failover curves, and backpressure policies) Deliverable must span between 1,200 and 2,000 words without generic filler.

Self-review

• Does the partition key strategy eliminate hot partitions across the {{fleet_size}} assets?
• Is the network reconnection burst model bounded to prevent cloud broker denial-of-service?
• Are data compaction algorithms feasible within the compute limits of {{edge_gateway_hardware}}?