Why was EMQX Enterprise chosen for managing 10 million concurrent MQTT connections?

EMQX Enterprise was selected due to its Kubernetes-native architecture capable of handling massive scale with 10 million concurrent connections, sub-second latency, and full support for QoS 2 message delivery. It offers high availability, multi-region deployment, hardware-rooted authentication for security, and seamless integration with Kafka for real-time data streaming, addressing the limitations of public cloud IoT hubs and alternative brokers.

What challenges did the organization face with its previous IoT infrastructure?

The organization struggled with operational complexity from managing over 50 public cloud IoT hub instances, scalability ceilings with alternative brokers, and protocol limitations such as QoS 0-only delivery that increased latency and reduced reliability. Additionally, handling massive daily data throughput, large telemetry payloads, and ensuring secure, reliable device communication across a global fleet presented significant challenges.

How does EMQX ensure secure device communication in large-scale deployments?

EMQX employs hardware-rooted authentication using mTLS with hardware-backed certificates to authenticate devices at the silicon level, preventing cloning and credential theft. It also implements granular role-based access control (RBAC) and topic-based access control lists (ACLs) to restrict device interactions strictly to authorized management topics, ensuring robust security across the device fleet.

What benefits did the organization achieve after migrating to EMQX Enterprise?

Post-migration, the organization simplified its infrastructure by consolidating multiple IoT hubs into a single unified platform, achieved carrier-grade reliability with 99.99% uptime, enabled zero-loss updates via QoS 2 delivery, and realized sub-second latency in device communication. Additionally, real-time Kafka integration allowed scalable analytics on large data volumes, while Kubernetes automation reduced maintenance overhead and total cost of ownership.

How does EMQX support multi-region expansion for global device fleets?

EMQX supports multi-region expansion through its Multi-Region Ready architecture using Cluster Linking, which enables cross-region message routing and regional failover. This design ensures seamless global scalability, high availability, and disaster recovery by maintaining active clusters in multiple geographic regions such as APAC and EMEA to support worldwide connected device management.

Scaling Connected Device Management to 10 Million Concurrent Connections with EMQX

EMQX Team

Apr 23, 2026Use Cases

Scaling Connected Device Management to 10 Million Concurrent Connections with EMQX

Executive Summary
A leading global technology manufacturer operating a fleet of 50 million devices needed a mission-critical MQTT infrastructure to support remote device management and technical support operations. Facing scalability limits with cloud-native IoT solutions and complexity challenges with alternative brokers, the organization deployed EMQX Enterprise on Kubernetes to handle 10 million concurrent connections with sub-second latency and QoS 2 guarantees. The result: unified device management across North America with plans to scale globally.

The Evolution of Connected Device Management

For technology manufacturers operating at a global scale, connected device fleets are now core infrastructure. Managing millions of endpoints in the field requires real-time bidirectional communication between devices and cloud backends to enable:

Proactive device health monitoring through telemetry and diagnostic data
Secure remote updates for firmware, system software, and security patches
Technical support integration enabling rapid issue detection and resolution
Device compliance verification across distributed user bases

The Challenge

This organization's device support strategy generated unprecedented scale requirements:

Scale and Complexity

10 million concurrent connections across a global device fleet
Over 5 million messages daily with packet sizes up to 64KB
Daily data throughput reaching 128GB across distributed endpoints
Complex telemetry payloads (40KB+ per diagnostic agent scan) containing detailed hardware inventory and diagnostic information

Platform Limitations

The organization initially relied on 50+ public cloud IoT hub instances for cloud connectivity. While functional, this approach created operational complexity and limited flexibility. Previous evaluation of alternative brokers revealed scalability ceilings that couldn't support production requirements.

Protocol Requirements

Support operations demanded both QoS 0 and QoS 2 reliability guarantees, particularly critical for cloud-to-device commands requiring guaranteed delivery. The managed IoT hub's QoS 0-only limitation forced workarounds that added latency and complexity to update and remote-action workflows.

Infrastructure Constraints

The organization's Kubernetes-first strategy required:

Isolated namespace deployments for production workloads
Proven clustering for high availability and disaster recovery
Seamless integration with existing Kafka data pipelines
Multi-region expansion capability for APAC and EMEA operations

Why MQTT for Connected Device Management

MQTT emerged as the ideal foundation because it addresses all requirements that public cloud IoT hubs struggled to deliver at scale:

Lightweight protocol minimizes bandwidth for bandwidth-constrained connections
Bidirectional messaging enables both device-to-cloud telemetry and cloud-to-device commands in a single connection
Flexible QoS levels (0, 1, 2) allow optimization for different message types without protocol switching
Industry standardization ensures interoperability across hardware platforms and support systems
Persistent sessions enable offline buffering, which is critical when devices experience temporary connectivity loss

The EMQX Solution: Scalable Architecture for Global Device Management

To address these challenges, the organization deployed EMQX Enterprise in a Kubernetes-native architecture designed for massive, global device fleets.

Cloud-Native Infrastructure for Global Scale

Kubernetes-Native Management: Utilizing the EMQX Operator to automate cluster provisioning, scaling, and lifecycle management. This eliminates orchestration complexity while ensuring the system can handle 10 million concurrent connections with predictable performance.
High Availability & Multi-Region Strategy: The architecture features active clusters with dedicated HA/DR failover. It is designed as Multi-Region Ready, leveraging Cluster Linking to support planned expansion into APAC and EMEA through cross-region message routing and regional failover.

Hardened Enterprise Security

Hardware-Rooted Authentication: mTLS with hardware-backed certificates ensures every endpoint is authenticated at the silicon level, preventing device cloning and credential theft.
Granular Access Control: Comprehensive RBAC and Topic-based ACLs ensure that devices can only interact with their specific management topics, preventing cross-tenant data leaks.

Real-Time Data Integration & Pipeline

Real-Time Data Streaming: A pre-built Kafka Bridge streams device telemetry and support events directly into Kafka topics for analytics and customer portals.
System Synchronization: Supports bi-directional syncing with on-premises databases for internal reporting and historical analysis, alongside a Rule Engine for dynamic message filtering and enrichment based on device profiles.

Reliability & Efficiency for Critical Operations

Guaranteed Message Delivery: Full support for MQTT 5.0 and QoS 2 (Exactly Once) delivery, which is essential for the integrity of security patches and software updates.
Durable Sessions: Automatic message buffering prevents data loss during network transitions, ensuring telemetry is captured even if a device temporarily loses its connection.
Optimized Workload Distribution: Using Shared Subscriptions to load-balance command distribution across multiple backend support systems, preventing processing bottlenecks.

Results and Business Value

The migration to EMQX Enterprise transformed the organization’s device management from a fragmented infrastructure into a unified, high-performance backbone.

Operational Excellence

Infrastructure Simplified: Consolidated 50+ disparate IoT Hub instances into a single, unified management plane.
Global Readiness: Established a blueprint architecture for seamless expansion into APAC and EMEA markets without redesign.
Carrier-Grade Reliability: Guaranteed 99.99% uptime with automated failover, meeting the highest service level agreements.

Technical Performance

Zero-Loss Updates: Achieved exactly-once (QoS 2) delivery for critical patches, eliminating custom retry logic and edge-case failures.
Sub-second latency: Direct MQTT connections outperform REST-based IoT hub alternatives
Scalable data pipelines: Kafka integration enables real-time analytics on 128GB daily throughput without custom ETL

Cost Optimization

TCO Reduction: Eliminated surprise billing from usage-based cloud models with predictable, transparent scaling costs.
Lean Operations: Leveraged Kubernetes automation to slash manual monitoring and custom maintenance overhead.

Conclusion

This case demonstrates EMQX Enterprise's ability to serve as a global-scale mission-critical infrastructure for massive connected device fleets.

With QoS 2 reliability, horizontal scalability to 10 million concurrent connections, and seamless Kafka integration, EMQX helped this tech manufacturer consolidate infrastructure, boost reliability, and support worldwide growth while cutting operational complexity versus public cloud alternatives and proprietary brokers.

Talk to an Expert