This release takes MQTT a step further by adding native support for the A2A protocol over MQTT, enabling AI agents to register, discover, and collaborate directly through the broker without additional infrastructure.
This article demonstrates how EMQX Neuron bridges these three gaps through **SQL feature engineering + flexible algorithm integration**, providing an end-to-end solution from data acquisition to edge inference.
Discover how EMQX Neuron leverages edge computing, data aggregation, and contextual intelligence to eliminate the Cloud Ingress Tax, dramatically reducing IoT data costs, cloud compute expenses, and network overhead for Industrial AI deployments.
Explore why MQTT is emerging as the missing infrastructure layer for Agentic AI, enabling scalable multi-agent coordination, real-time messaging, and distributed AI systems.
Learn how to export MQTT data from EMQX to Amazon S3 in Parquet format and analyze it with DuckDB for cost-efficient, SQL-driven IoT analytics.
Learn how to modernize your IIoT data pipeline by shifting from legacy gateways to an intelligent AI edge with EMQX Neuron for real-time SQL filtering, 99% bandwidth savings, and AI-ready data normalization.
This final part assumes your infrastructure and Erlang VM are already hardened, as described in Part 2. We now move to the layers most SREs and security teams interact with day‑to‑day: TLS termination, MQTT‑level authentication and authorization, administrative access, and disaster recovery.
This second article focuses on the foundation beneath EMQX: the Linux kernel, network stack, and Erlang VM. If Part 1 explained why security and reliability converge for a stateful MQTT broker, this part shows where that convergence actually bites in production: file descriptors, TCP behavior under load, firewall rules, and Erlang distribution security.
Learn how to implement MQTT 5.0 in iOS apps using CocoaMQTT for real-time, battery-efficient, and secure IoT messaging with Swift.
A deep dive into decoupled compute and storage architecture: how Parquet columnar storage, multi-level caching, and compression reshape IoT data storage.
We are continuously evolving EMQX Cloud to provide a more robust, observable, and flexible MQTT platform. Our latest update introduces expanded availability for EMQX Tables, simplified observability with Datadog, and a preview of our new SSE Gateway.
This version's core highlights include three new drivers: SNMP, GE Historian, and OPC AE, enabling EMQX Neuron to cover full-scenario requirements from network device monitoring, historical data access, to alarm and event management. Additionally, features such as Northbound Application Import/Export, Tag Management, AI-Powered Rule Writing, and Built-in Modbus TCP Simulator significantly enhance configuration efficiency and intelligence.
This article is the first in a three-part series on hardening EMQX for large‑scale IoT production use. Here we frame security as a core reliability concern, explain why MQTT’s stateful nature changes the threat model, and end with a practical checklist you can apply directly to your own clusters.