Building for Millions of Devices, Not Dozens
As IoT adoption matures, scaling is no longer an aspiration; it’s a survival requirement.
Connecting thousands of devices was once impressive. Today, manufacturers and service providers are targeting millions of concurrent connections, each transmitting data in real time to drive analytics, compliance, and operational efficiency.
At this scale, the MQTT protocol has become the backbone of IoT communication. But achieving true scalability takes more than just deploying a broker. It requires an architecture designed specifically for industrial IoT realities; high fan-in traffic, unpredictable connectivity, and enterprise-grade security.
That’s where Cumulocity’s Managed MQTT Service comes in.
Why MQTT is the Engine of Scalable IoT
MQTT’s lightweight footprint makes it perfect for resource-constrained devices and unstable networks. It minimizes message overhead and tolerates frequent disconnects; ideal for sensors, meters, and edge devices that don’t live in perfect conditions.
But while MQTT simplifies device communication, scaling MQTT for tens of millions of devices introduces new challenges.
Traditional message brokers were never designed for IoT’s “asymmetrical” workloads; where millions of devices publish data to just a handful of consuming applications. Generic brokers treat all connections equally; IoT systems need something different:
- Massive fan-in performance — handling millions of device-to-cloud messages per second.
- Device isolation and tenant separation — so one compromised device can’t affect others.
- Flexible payload handling — to support everything from JSON to binary telemetry.
These requirements demand more than an MQTT broker. They need a full IoT-first messaging service.
Cumulocity’s Third-Generation MQTT Service
Cumulocity (trusted by more than 1,000 enterprises and managing over 25 million connected devices worldwide) has built a third-generation Managed MQTT Service from the ground up for IoT scale.
Proven Scale and Reliability
The platform has been validated for 100 million connected devices and one million messages per second; end to end, from device to cloud.
That validation isn’t theoretical. It’s based on live production-grade testing in realistic IoT fan-in scenarios, where millions of devices push unique data streams to fewer back-end consumers.
Cumulocity’s architecture achieves this through:
- Partitioned subscriptions, dividing device data across logical streams to balance load.
- Tiered architecture: The MQTT protocol engine is decoupled from the back-end message persistence and routing layer, allowing independent scaling for optimal resource usage.
- Limit and quota enforcement: Ensures the deployed resources are not overloaded, and that “noisy neighbours” (tenants or devices) are not a problem.
- Protocol independence, so MQTT can work alongside LwM2M, REST, or other standards within the same platform.
And all of this runs under a 99.5% SLA uptime, proven in global deployments across manufacturing, utilities, and connected equipment.
Scaling Without Breaking the Bank
Performance is only half the story. Linear scaling often comes with linear cost. But Cumulocity’s architecture is designed for horizontal elasticity.
Because persistence and routing are handled by a shared enterprise-grade messaging backbone, the system scales without bloating infrastructure. The same core supports device messaging, analytics, and application data processing; avoiding duplicate components and wasted compute.
This is scalability that grows with your business, not your cost base.
Simplicity at Scale
Scaling IoT isn’t only about throughput; it’s about maintainability.
Cumulocity’s MQTT Service reduces complexity for developers and operators through:
- Low-code configuration and out-of-the-box tools like Device Management, Cockpit, and Analytics Builder.
- Flexible topic and payload models, supporting both SmartREST and custom formats.
- AI-powered Data Preparation (coming soon) that automatically maps raw device data into Cumulocity’s domain model for instant analytics.
On the edge, thin-edge.io support extends MQTT functionality for constrained environments, while Cumulocity Edge delivers full MQTT capability in disconnected or on-premise setup, with identical security and performance characteristics as the cloud.
Built for IoT, Not Retro-Fitted for It
Cumulocity’s Managed MQTT Service isn’t an enterprise message bus adapted for IoT. It’s purpose-built for device communication at massive scale, designed by the same engineers who helped bring MQTT to market. That means:
- Validated global scale: 100M devices, 1M messages/sec.
- 99.5% availability: enterprise reliability at IoT speed.
- Multi-protocol support: MQTT, LwM2M, REST, and beyond.
- Security by default: tenant isolation, PKI authentication, TLS encryption.
- Ease of onboarding: device registration and mapping in minutes.
The Bottom Line
Scaling IoT is hard. Scaling it securely, efficiently, and affordably is harder.
Cumulocity’s Managed MQTT Service takes the guesswork out of scale, delivering a production-proven architecture that’s been tested for 100 million connections; not a lab demo, but a live-ready system built for industrial realities.
If your IoT strategy is moving beyond the pilot stage, this is the foundation you need to scale with confidence.
Learn more about Cumulocity’s Managed MQTT Service and see how you can scale your IoT network from thousands to millions; securely, simply, and at proven global scale.
Talk to an Expert
Have questions or want to dive deeper? Our experts are ready to help you find the best solution for your needs. Speak with an expert now.