RFID Cloud Platform Integration

RFID Cloud Platform Integration: AWS IoT, Azure IoT Hub, and GCP

Modern RFID deployments rarely terminate at on-premises middleware. Cloud IoT platforms — AWS IoT Core, Microsoft Azure IoT Hub, and Google Cloud IoT Core (now Pub/Sub-based) — provide the scalability, managed infrastructure, and analytics integration that enterprise RFID deployments require. This guide covers the integration architecture, protocol mapping, and platform-specific trade-offs.

Why Cloud for RFID

Traditionally, RFID middleware ran on-premises, feeding an on-premises ERP or WMS. Cloud integration becomes compelling when:

• Multiple sites need consolidated visibility (retail chains, 3PLs)
• Analytics workloads (read-rate trending, anomaly detection) need elastic compute
• Trading partners require access to EPCIS events without VPN
• IoT sensor data (temperature, humidity) must be stored long-term alongside RFID events
• Machine learning models (inventory forecasting, shrinkage detection) need historical read data

Architecture Overview

Reader (LLRP) → Edge Gateway → Cloud IoT Platform → Event Processing → Data Store
                                     |
                              Device Management
                              (OTA updates, health)

The edge gateway is the critical component: it normalises LLRP reader events, performs local filtering (deduplication, zone aggregation), and forwards structured events to the cloud. It also provides resilience — buffering events during connectivity loss.

Edge processing before cloud upload reduces cloud ingress costs significantly: a busy portal reader generates 100,000+ raw read events per day, but after deduplication and aggregation, the relevant business events may be 1,000–10,000 EPCIS events.

AWS IoT Core Integration

AWS IoT Core is the dominant choice for US-headquartered enterprises and those already on AWS.

Integration path:

MQTT topic structure for RFID:

rfid/{site_id}/{reader_id}/events      → raw tag reads
rfid/{site_id}/{reader_id}/status      → reader health
rfid/{site_id}/aggregate               → zone-level aggregated events
rfid/epcis/{site_id}                   → formatted EPCIS events

MQTT QoS 1 (at-least-once delivery) is appropriate for RFID events — a duplicate tag read is preferable to a missing read. QoS 2 (exactly-once) adds latency overhead that is rarely justified.

AWS IoT Greengrass for edge processing:

AWS Greengrass runs on the edge gateway, executing Lambda functions locally before forwarding to IoT Core. Greengrass components relevant to RFID:

• Local MQTT broker (for air-gapped operation)
• Stream Manager (local buffering, upload retry)
• Local inference (ML models for anomaly detection)

For large retail deployments (500+ stores), Greengrass fleet deployment enables OTA configuration updates — changing antenna dwell times or reader power without on-site visits.

Azure IoT Hub Integration

Azure IoT Hub is preferred for Microsoft-stack enterprises (Dynamics 365, Azure Data Factory, Power BI).

Azure IoT Hub device provisioning:

Each RFID edge gateway registers as an IoT Hub device using Device Provisioning Service (DPS). DPS supports X.509 certificate-based authentication — appropriate for enterprise deployments where device identity must be cryptographically bound.

Integration with Dynamics 365:

Azure Logic Apps provides pre-built connectors for Dynamics 365. EPCIS receiving events from dock-door portals can automatically trigger purchase order receipt in Dynamics without custom middleware development.

Google Cloud IoT / Pub/Sub

Google deprecated Cloud IoT Core in August 2023. The replacement architecture uses:

• Cloud Pub/Sub for device-to-cloud message ingestion (via MQTT bridge or direct publisher)
• Dataflow for streaming RFID event processing
• BigQuery for RFID event storage and analytics
• Cloud Run for EPCIS REST API

GCP is favoured when the analytics workload dominates (BigQuery ML for demand forecasting) or when the organisation is Google Workspace-centric.

Protocol Translation Table

Security Considerations for Cloud RFID

• TLS 1.2+ for all MQTT connections (IoT Core and IoT Hub enforce this)
• Mutual TLS or JWT/SAS token authentication for edge gateways
• Topic-level authorization — individual reader gateways should only publish to their own topic namespaces
• PII in EPCs — if EPCs encode individual-identifiable information, ensure cloud storage is in-region and encrypted at rest

For RFID-specific security architecture, see RFID Security Threats.

Platform Selection Guide

See also: RFID ERP Integration, EPCIS Implementation, Edge Computing in RFID.