IoT and ERP
IoT (Internet of Things) integration with ERP describes the architectural and operational patterns connecting sensor-equipped devices, machines, vehicles, buildings and products with the enterprise business systems. Modern ERP rarely processes raw IoT streams directly; dedicated IoT platforms (Microsoft Azure IoT, AWS IoT, Siemens Industrial Edge, Bosch IoT Suite) handle ingestion, processing and aggregation, with ERP consuming structured results for business decisions. The architecture has matured substantially since the IoT hype peak around 2018-2019.
Reference architecture
Four-layer architecture is standard. (1) Device layer: sensors, machines, vehicles, wearables with embedded compute. Communication via BLE, LoRaWAN, NB-IoT, 5G, Ethernet, OPC UA depending on use case. (2) Edge layer: local gateways aggregate device data, perform local processing, manage device fleets. Edge computing (AWS IoT Greengrass, Azure IoT Edge, Siemens Industrial Edge) reduces bandwidth and latency. (3) Platform layer: cloud-based IoT platforms handle device management, time-series storage, stream processing, ML model deployment. Major platforms: Microsoft Azure IoT Hub plus Azure IoT Central, AWS IoT Core plus AWS IoT Analytics, Google Cloud IoT, Siemens Insights Hub, PTC ThingWorx. (4) Application layer: aggregated metrics flow to ERP, BI tools, EAM systems for business decisions. ERP-side integration via REST APIs or event-streaming.
ERP-relevant IoT use cases
- Production monitoring — machine-state, OEE (Overall Equipment Effectiveness), throughput metrics feeding production planning
- Predictive maintenance — ML on sensor data triggers ERP maintenance work orders
- Asset utilisation — industrial equipment, fleet vehicles, leased equipment with usage-based billing
- Connected products — sold products in customer hands generating service-and-maintenance opportunities back to ERP
- Cold-chain monitoring — temperature-sensitive logistics for food, pharma
- Energy monitoring — factory-level energy consumption per machine for CSRD reporting
- Smart-building integration — facility management with occupancy, HVAC, security data feeding ERP-side facility operations
- Tank-level monitoring — automated reorder triggering for bulk-supplied materials
ERP-vendor IoT offerings
SAP IoT: SAP Asset Intelligence Network, SAP Predictive Asset Insights, SAP Edge Services. Direct integration with S/4HANA Plant Maintenance. Microsoft: Azure IoT Hub plus Dynamics 365 Field Service plus Dynamics 365 Connected Operations. Tight integration through common data platform. Oracle: Oracle IoT Cloud Service plus Oracle Asset Monitoring alongside Oracle Cloud ERP. Siemens: Siemens Insights Hub (formerly MindSphere) plus Industrial Edge, integrating with SAP and other ERPs. Specialist IoT platforms: PTC ThingWorx, GE Digital APM, AVEVA Insight, Litmus Automation, Particle, Losant. Mid-market typically starts with hyperscaler IoT (Azure IoT, AWS IoT) plus specialist platforms for specific use cases; full enterprise-IoT platforms (PTC ThingWorx) suit large-scale industrial transformations.
Practical guidance
Three patterns for successful IoT-ERP integration. (1) Home with specific business outcomes: predictive maintenance for high-value assets, OEE improvement on bottleneck lines, cold-chain compliance for shipped products. Generic 'digital transformation' IoT investments without specific outcomes consistently underdeliver. (2) Aggregate before integrating: do not feed raw IoT streams into ERP. The IoT platform aggregates and contextualises; ERP receives structured business events (machine failure, threshold breach, completed production run). (3) Invest in operational discipline: IoT estates need ongoing operations (firmware updates, certificate management, device monitoring) that scale with device count. Underestimating this operational burden is the most consistent IoT-deployment failure mode.
Data architecture and scalability
IoT-ERP architectures must handle data volumes that exceed classical ERP scale by orders of magnitude. A single production line with 100 sensors at one-second sampling generates millions of data points per day; a fleet of 1,000 connected vehicles produces hundreds of millions per day. Time-series databases (InfluxDB, TimescaleDB, Amazon Timestream, Azure Data Explorer) handle the high-ingestion-rate write workloads classical relational databases cannot match. Stream processing (Apache Kafka, Apache Flink, AWS Kinesis Data Streams, Azure Stream Analytics) processes data on the fly, computing aggregations and detecting events without persisting every raw measurement. Data tiering: hot data (recent, frequently queried) in fast storage, warm data (weeks to months old) in cheaper storage, cold data (long-term archive) in object storage. Downsampling: raw measurements aggregated to minute, hour, day rollups for long-term storage while keeping query performance.
ERP consumes the structured outputs of this pipeline — alerts, aggregated KPIs, predicted maintenance windows — not raw measurements. The data-architecture discipline of separating IoT-scale workloads from ERP-scale workloads is essential. Mid-market organisations attempting to push raw IoT data into ERP databases consistently encounter performance problems and data-archive complexity. The right pattern: IoT platform handles ingestion and analytics; ERP consumes structured business events with low frequency.
Related Topics
Frequently Asked Questions
Should we use ERP-vendor IoT or specialist platforms?
For tight integration with the specific ERP-side processes (SAP S/4HANA Plant Maintenance, Dynamics 365 Field Service), the ERP-vendor IoT typically wins on integration depth and TCO. For cross-system or cross-ERP IoT scenarios, specialist platforms (PTC ThingWorx, Siemens Insights Hub) offer breadth that ERP-vendor IoT cannot match.
How does IoT change OT/IT integration?
IoT blurs the historical OT (Operational Technology) versus IT (Information Technology) boundary. OT systems (PLCs, SCADA, MES) historically operated isolated from corporate IT. IoT pulls OT data into IT analytics and decision-making layers, requiring new security models, data-governance frameworks and cross-team collaboration. The cultural change is often more challenging than the technology.
What about IoT cybersecurity?
Critical and underinvested. IoT devices have small attack surfaces individually but vast attack surfaces collectively. NIS-2 obligations explicitly cover IoT supply-chain risks. Best practice: zero-trust network access, certificate-based device authentication, secure-boot and signed firmware, network segmentation, SIEM integration. Mid-market organisations should not deploy production IoT without these controls in place.