Their Grid Had 100,000 Sensors and a Two-Megabit Pipe
NextEra Energy operates America's largest renewable generation fleet. Their SCADA systems were generating more data than their remote links could carry. Birth certificates hit 800 MB. Alert storms buried operations. Expanso sat in between and fixed it—without touching a single downstream system.
See How It Works- Client
- NextEra Energy
- Industry
- Energy & Utilities
- Use Case
- Edge Telemetry Processing for Distributed Renewable Generation Fleet
- Timeline
- Pilot site in weeks, fleet rollout ongoing
- ROI
- 70%+ bandwidth reduction on constrained links; zero protocol or application changes required
The Challenge
NextEra Energy operates more than 300 wind and solar generation facilities across more than 40 U.S. states. The fleet is geographically dispersed and connectivity is often brutal—some sites run on Starlink, others on cellular links topping out at 2 Mbps. Every site generates continuous sensor telemetry at 100-millisecond polling intervals, packaged as Sparkplug B MQTT through Ignition Edge and delivered to a HiveMQ broker cluster in the cloud. The protocol is industry-standard. The physics are unforgiving.
- 01 Birth certificate payloads exceeding 800 MB on some sites—saturating 2 Mbps uplinks for over an hour per event
- 02 Sparkplug B sequence gaps during saturation triggered retransmit storms, compounding congestion
- 03 Heterogeneous SCADA hardware emitting out-of-range values (door sensors: 1.5, temp transducers: 999°C) with no validation layer
- 04 No way to distinguish network hiccups from genuine equipment failures—phantom alarm storms burned operations bandwidth
- 05 Existing dead-banding at Ignition Edge was insufficient for worst-case connectivity scenarios
- 06 Any solution had to be invisible to downstream systems: historians, cloud SCADA, analytics pipelines
The Solution
Expanso deploys as a lightweight runtime between Ignition Edge and HiveMQ. It subscribes to the Sparkplug B stream, applies a configurable processing pipeline, and re-publishes downstream. The protocol is preserved end-to-end. Sequence numbers are maintained. Downstream systems see no change—because nothing about their interface changed.
Schema Enforcement & Validation
Every metric is validated against defined bounds before reaching the broker. Values that violate schema—a door sensor reporting 1.5, a temperature transducer returning 999°C—are intercepted and routed to a dead-letter queue. Clean data flows forward. Bad data is logged, auditable, and never silently dropped.
Intelligent Dead-Banding
Configurable change-detection thresholds per metric type: 1°C for temperature, 0.5 PSI for pressure. Messages that carry no new operational information are suppressed before they consume bandwidth. At 100ms polling intervals, this reduces message volume by an order of magnitude on slow links without losing any data that matters.
Payload Compression
Further compression at the pipeline layer produces meaningful savings on burst events. Birth certificates that previously saturated uplinks for over an hour now resolve in a fraction of the time—freeing bandwidth for real-time telemetry.
Fan-Out & Differentiated Routing
A single sensor stream feeds multiple destinations simultaneously: the real-time broker for operations, an archive for compliance, a filtered stream for anomaly detection. One pipeline definition deploys identically across hundreds of sites—or with per-site customization where warranted.
Full Pipeline Observability
Every message—forwarded, dropped, flagged—is accounted for. Operations teams see real-time metrics: messages per second in and out, bytes saved, validation failures by type, sequence integrity status. The difference between a network hiccup and real equipment failure becomes visible in seconds.
Fleet-Scale Deployment
Pipeline definitions are versioned and deployed centrally. A change to dead-banding thresholds or a new routing rule propagates to hundreds of sites without touching any of them individually. Every version of every pipeline ever deployed is auditable and reversible.
The Results
Expanso processes tens of millions of sensor readings per day across NextEra's generation fleet. On the most constrained sites, it changed what was operationally possible—not just cheaper, but reliable where reliability wasn't achievable before.
- Effective bandwidth utilization improved 70%+ on sites with 2 Mbps or less uplink
- Sparkplug B birth certificate events that previously saturated links for over an hour now complete in minutes
- Message volume reduced ~10× on typical sensor mixes through intelligent dead-banding—no operational data lost
- Schema validation catches and quarantines bad sensor values before they reach downstream historians
- Sequence gaps and retransmit storms dropped dramatically—phantom alert volume materially reduced
- Zero downstream application changes: historians, cloud SCADA, analytics pipelines all unchanged
- Single control plane manages pipeline definitions across the full fleet—versioned, auditable, rollback in one command
Running a fleet with constrained connectivity?
If your SCADA data is eating your bandwidth, your alerts are burying your team, or your remote sites go dark during network events—we've solved this before and we can solve it for you.