IoT (Internet of Things) generates high volume from many devices. Salesforce isn't naturally suited; integration architecture matters.
IoT shape:
- Many devices (thousands to millions).
- Each generating events frequently (seconds to minutes).
- Total event rate: extremely high.
- Salesforce ingestion limits don't fit raw IoT.
Architecture pattern: ingestion + filtering + Salesforce.
1. IoT platform.
- AWS IoT, Azure IoT Hub, Google IoT.
- Devices connect, send events.
- Handles device lifecycle, security.
2. Stream processing.
- AWS Kinesis, Azure Stream Analytics, Apache Kafka.
- Filter / aggregate / transform raw events.
- Sample down to interesting events only.
3. Aggregation.
- Hourly / daily summaries.
- Alarms / threshold-crossings.
- Reduce volume before Salesforce.
4. Salesforce sync.
- Aggregated data to Salesforce.
- Critical events as notifications / cases.
- Customer profile updates.
5. Field Service trigger.
- IoT alarm -> Field Service work order.
- Dispatcher allocates technician.
Use cases:
- Predictive maintenance — IoT detects equipment failure pattern; Salesforce dispatches.
- Smart products — usage data informs Salesforce customer profile.
- Field Service IoT — equipment talks to dispatch.
- Connected vehicles — vehicle health to fleet management.
Architectural decisions:
1. What goes to Salesforce?
- Aggregated, alarming, decision-driving events.
- NOT raw telemetry.
2. Latency requirements.
- Real-time alerts (high-stakes).
- Near-real-time aggregates (most use cases).
- Daily summaries (analytics).
3. Storage.
- Raw IoT in cloud warehouse.
- Aggregates in Salesforce.
- Long-term archival in cold storage.
4. Salesforce IoT product.
Salesforce had a dedicated IoT Cloud product (sunset). Now patterns use Salesforce + AWS IoT / equivalent.
Common pitfalls:
- Raw IoT to Salesforce — overwhelms governor limits.
- No filtering — every event reaches Salesforce.
- No aggregation — Salesforce drowns.
- Latency mismatch — real-time alarms delayed.
Senior architect insight: Salesforce as orchestration; cloud as ingestion. Don't try to make Salesforce ingest IoT; pre-process in cloud.
The senior framing: IoT + Salesforce is a hybrid architecture by necessity. Designed deliberately, it works. Forced into Salesforce alone, it fails.