Getting Started with MQTT
Written by The HiveMQ Team
Published: April 24, 2020
Introduction to MQTT
MQTT is a publish/subscribe protocol that is lightweight and requires a minimal footprint and bandwidth to connect an IoT device. Unlike HTTP’s request/response paradigm, MQTT is event driven and enables messages to be pushed to clients. This type of architecture decouples the clients from each other to enable a highly scalable solution without dependencies between data producers and data consumers.
The key benefits of MQTT are:
- Lightweight and efficient to minimize resources required for the client and network bandwidth.
- Enables bidirectional communication between devices and servers. Also, enabling broadcasting messages to groups of things.
- Scales to millions of things.
- Specifies Quality of Service (QoS) levels to support message reliability.
- Supports persistent sessions between device and server that reduces reconnection time required over unreliable networks.
- Messages can be encrypted with TLS and support client authentication protocols.
Use Cases of MQTT
MQTT is used in many industries and applications. HiveMQ has published a number of case studies from industries such as automotive (BMW), telecommunications (Liberty Global), energy (Fortum), public safety (Hytera), connected products (Awair, Matternet). Read all the stories of our customers here.
Basic Concepts of MQTT
At the core of MQTT is the MQTT broker and the MQTT clients. The broker is responsible for dispatching messages between the sender and the rightful receivers. An MQTT client publishes a message to a broker and other clients can subscribe to the broker to receive messages. Each MQTT message includes a topic. A client publishes a message to a specific topic and MQTT clients subscribe to the topics they want to receive. The MQTT broker uses the topics and the subscriber list to dispatch messages to appropriate clients.
An MQTT broker is able to buffer messages that can’t be dispatched to MQTT clients that are not connected. This becomes very useful for situations where network connections are unreliable. To support reliable message delivery, the protocol supports 3 different types of quality of services messages: 0 - at most once, 1 - at least once, and 2 - exactly once.
There are two versions of the specification: MQTT 3.1.1 and MQTT 5. Most commercial MQTT brokers now support MQTT 5 but many of the IoT managed cloud services only support MQTT 3.1.1. We highly recommend new IoT deployments to use version 5 due to the new features that focus on more robust systems and cloud native scalability.
For a more in depth description of the protocol, we recommend that you read the MQTT Essentials series of articles or review the companion video series. MQTT 5 Essentials series also provides an in depth introduction to the MQTT 5 specific features.
There are many open source clients available in a wide range of programming languages. HiveMQ offers the HiveMQ MQTT Client that is developed in Java. Eclipse Paho also offers C/C++, Python and a variety of other implementations. A detailed list of clients can be found at mqtt.org.
MQTT brokers are offered in open source, commercial implementations and managed cloud services. HiveMQ offers two commercial editions: HiveMQ Professional and HiveMQ Enterprise, a managed cloud MQTT service: HiveMQ Cloud, and an open source version: HiveMQ Community Edition. A detailed list of brokers can be found at mqtt.org.
Example Implementation of MQTT
To better illustrate how MQTT works, we present a simple example implementation that uses HiveMQ Cloud. To test the implementation on an actual cluster, you first need to sign up for HiveMQ Cloud Free plan, which allows you to connect 100 IoT devices for free. In addition, no credit card is required during sign up. As soon as you sign up and log in to HiveMQ Cloud, your free HiveMQ Cloud Free cluster is running and ready. If you are a first time HiveMQ Cloud user, HiveMQ Cloud automatically redirects you to the getting started section in the management view of your cluster. The following example was tested on Java 11.
You can also find this example, and a publicly hosted one in our ready-to-go Github repository along with an example for our public broker.
In this example we will have a temperature and brightness sensor connected to a Raspberry Pi that will send the sensor data to an MQTT broker. Another device will run a control center that receives the MQTT data.
Step 1 - Implement sensor client
The first step is to create the MQTT client that publishes the sensor data. In this example, we will use a thermometer and brightness sensor. We will also use HiveMQ Cloud as MQTT broker. After you signed up for HiveMQ Cloud, navigate to the Details section on the Overview tab of your cluster. There you will find your hostname. Copy this hostname to replace it in the example code snippet below.
To connect an MQTT client securely to your cluster, you must create MQTT credentials for the cluster.
Switch to the Access Management tab of you HiveMQ Cloud Basic cluster, define your MQTT credentials and select +Add.
These are also the credentials, which you have to replace in the code for
The above code snippet does the following:
- It creates the MQTT client. It is important to use a unique identifier. We use automatic reconnect because the sensor internet connection may be unstable.
- It connects the client to
"<your_host>". As a will is specified, the broker automatically publishes the “sensor gone” message if the sensor loses its connection.
- Then it periodically publishes (simulated) brightness and temperature data.
Step 2 - Implement subscribing client
The next step is implementing the subscribing client, which consumes the values on the topics
The above code snippet does the following:
- It creates the MQTT client.
The initialization of the MQTT client instance is almost the same as for the sensor, except we use
controlcenter-as prefix for the client id.
- It connects the client to your specified host in
<your_host>. We use a session expiry interval of 1 hour to buffer messages when then control center is offline.
- It subscribes the client to all topics starting with
homeso we use the multi-level wildcard
#in the topic filter. All arriving messages are printed out with topic and payload. When the sensor loses its connection, it prints the topic
home/willand the payload “sensor gone”.
Now that you have a good introduction to MQTT we suggest the following:
- Read the MQTT Essentials and MQTT 5 Essentials series of articles to learn more technical details about the protocol. Do check out MQTT Glossary to know the key MQTT terminologies at a glance.
- Upgrade from your 100 free connections to a different option, depending on the functionality you want to use.
- If you prefer a locally-hosted version, you can try out the evaluation license of our HiveMQ commercial edition. We also have the HiveMQ MQTT Client, the MQTT CLI for easy testing of MQTT systems, and a Getting Started tutorial.
- Contact HiveMQ to discuss your requirements for MQTT and IoT messaging. Our experts have lots of experience helping companies build reliable and scalable IoT applications.