MQTT – Getting Started


MQTT Getting Started Guide provides a detailed explanation on MQTT for MQTT beginners.

What is MQTT?

MQTT stands for Message Queuing Telemetry Transport. Now, it is known as MQ-Telemetry Transport. MQTT is a networking protocol that is designed especially for M2M (Machine to Machine) communication for the transport of telemetry data. It is a messaging protocol and one of the main protocols for IoT Applications like smart homes, live chat applications, mobile messaging, etc. This protocol is a lightweight and bidirectional communication protocol designed to lessen the network bandwidth with limited/low bandwidth resources that are used to establish communication between multiple/constrained devices. There are millions of devices using and accessing the MQTT protocol every day, and it provides a stable and reliable communication foundation for all those millions of devices.

History of MQTT

Andy Stanford Clark (IBM) and Arlen Nipper think of designing the protocol MQTT, which had come true in 1999. The purpose of scheming this MQTT is for bridging to link sensors on an oil pipeline telemetry systems over satellite. It uses a publish-subscribe model that provides a stable and reliable network connection based on TCP/IP, a lightweight header for reducing transmission cost, and reliable message support. MQTT version 3.1.1 was released as an OASIS standard in October 2014. The upgrade of MQTT version 5 adds several new features and is released in March 2019. It meets the need of the industry and prepares for the changes in the development of industry in the future. MQTT is declared as an ISO and OASIS standard in 2010. MQTT 5.0 became the new OASIS standard in 2019.

MQTT Agenda

The protocol MQTT was invented in 1999, and its growth began many years later. Five years later, MQTT 3.1.1 was released with an OASIS and ISO standard. After 2010, the version changes from 3.1 to 3.1.1 then the OASIS standard sanctioned the new MQTT5 Specification. Then in March 2019, MQTT5 took place as the newly approved OASIS and ISO standard.


MQTT5 Agenda

MQTT 5 is the inheritor of MQTT 3.1.1. Its first public release was in January 2018, and its official release was in 2019, sanctioned with OASIS and ISO standards.

Why from MQTT3.1.1 to MQTT 5

The MQTT getting started guide describes the  inheritor of 3.1.1 is MQTT5. The answer is too simple, MQTT protocol defines a fixed header in the CONNECT packet that contains a single byte value for these protocol versions. MQTT 3.1.1 has the value “3” and 3.1.1 has the value “4”. Therefore the inheritor of 3.1.1 (MQTT5) has the value “5”. This number “5” is for both the protocol name and the value. MQTT5 introduces some changes and improvements compared to MQTT3.1.1.
Below, MQTT5 was explained to establish a clear overview of the functional objectives of MQTT messaging protocol.

MQTT 5 Goals

This simple MQTT getting started model provides the improvements for the purpose of meeting the needs of existing MQTT5 clients/users.

  • The MQTT5 goals are improving the ease of use to satisfy the needs of the long term users.

  • The MQTT5 goals are improving performance and supporting scalability without any overly complicated things.

There are five idle designs in MQTT 5 specification. They are, 

  • Enrichment of scalability for large-scale systems and support for small clients.

  • Increasing error reporting.

  • Formalize usual things, including capability.

  • Expandable techniques, including user properties.

  • Performance improvements for large and small scale clients.

MQTT Publish and subscribe/Unsubscribe



MQTT getting started guide gives a clear-cut view of PUBLISH packet and the publishing role. The process of sending messages is called Publishing. A client cannot publish a message directly to another client. But, the MQTT client can publish messages to the broker. Each message must contain a topic. An MQTT client can only publish messages to a single topic and cannot publish to a group of topics. The broker can filter the message under a topic and forward the message to the connected clients who are all subscribed to that topic. A PUBLISH packet is sent from a Client to a Server or vice versa to transport an Application Message. Each PUBLISH packet has a payload that contains the data (messages) transmitted in byte format. The sender (Publisher) decides whether it wants to send binary data, text data, or even the format needed by the client. 




The SUBSCRIBE packet is sent from the MQTT client to the MQTT server, which is used to create more than one Subscription. Let us have the clarification about the SUBSCRIBE packet and subscribe role.To receive messages on a particular topic, the Client must subscribe to that topic. The Server sends Publish packets to the customer to send published utility messages to topics that are compatible with these subscriptions. The SUBSCRIBE packet also specifies the maximum QoS with which the Server can send Application Messages to the Client. To send messages from the broker to the Client, the broker uses the same publishing method that is used by the publisher. Subscribe to multiple topics that are allowed using two wildcard characters(i.e., + and #). All subscriptions are accepted by the broker using a subscription acknowledgment message. An MQTT client can only subscribe to topics and cannot subscribe to a particular client. 




The contradiction of the SUBSCRIBE message is the UNSUBSCRIBE message. An UNSUBSCRIBE packet is sent by the Client to the Server, to unsubscribe from topics. It deletes the existing subscriptions on the broker. These Unsubscribe messages also have a list of topics from which the MQTT client wants to Unsubscribe. To confirm Unsubscribe, the broker sends an UNSUBSCRIBE to acknowledge the message to the MQTT client. They can assume that the subscriptions are deleted after the MQTT client receives the acknowledge message from the broker. Thus, the MQTT getting started guide exposes the UNSUBSCRIBE packet.

Client/Broker Connection

MQTT Client


Generally, any IoT device that speaks MQTT over a TCP/IP stack can be called an MQTT client. Both the publishers and subscribers are considered MQTT clients. The publisher and subscriber indicate whether the client is currently publishing messages or whether the client has subscribed to receive the messages. The one who publishes messages are called publishers and the one who subscribes to receive the information is called subscribers.

The publish-subscribe model is different from the client-server model. In the PUB/SUB model, more than one publisher publishes messages to one subscriber, or more than one subscriber receives messages from one publisher simultaneously. In this model, MQTT clients cannot contact each other directly. But, they communicate with the help of the MQTT Broker. Because, the client implementation of the MQTT Protocol is very straightforward and efficient. MQTT client is any device that connects to an MQTT Broker over a network connection.


MQTT Broker


The MQTT Broker acts as an intermediate between the MQTT Clients. It is the kernel of any publish/subscribe protocol. The MQTT Broker is proprietary software that runs on a computer (in the cloud) and can be self-generated or hosted by a third party. MQTT Broker helps the MQTT clients to share, manage data, and to control devices. The Broker has the responsibility for receiving all messages, filtering the messages under topics, and then distributing these messages to the subscribed clients. MQTT Broker cannot subscribe to an MQTT client. But the MQTT clients (subscribers) can subscribe to the other client (publishers). The Broker does not store any messages, while retained messages, persistent connections, and QoS levels can result in messages being stored temporarily on the broker.

Once the published message is sending to the MQTT Clients, then it is removed from the broker.  If no clients have subscribed to the particular topic or the subscribed clients are not currently connected, then the message is removed from the broker. But the broker can store data in the form of retained messages. This will enable new subscribers to get the last value directly. Also, when devices are called “Persistent sessions”, the broker monitors all the information of the sessions and eliminates vulnerable and insecure client connections.  The MQTT Broker is used to reduce network strain without compromising security. A backup broker can be set up to share a load of customers on multiple servers on the onsite, cloud, or in a combination of both.

MQTT Connection


The MQTT Connection is based on TCP/IP. Both the client and the broker need to have a TCP/IP stack. The MQTT Connection is always between one client and the broker. The MQTT Clients never connect directly. MQTT clients (publishers and subscribers) are mainly used to establish a connection between constrained devices. After a Network Connection is established between the MQTT clients, the first packet sent from the client to the server MUST be a CONNECT packet. To establish a connection first, the client sends a CONNECT message to the broker. When the broker receives the connect message, then it responds with a CONNACK message. If there is an enormous time gap between opening a network and sending the CONNECT control packet message, then the broker will close the connection. Once the connection is established, the broker keeps it open until the client sends the disconnect command or the network connection breaks of some network issues. The MQTT getting started guide gives the details of the MQTT connection over the client and the broker.

MQTT QoS Levels – (Quality of Service)


MQTT delivers simple messages according to the QoS levels. The MQTT broker can specify QoS levels for each Client-Server Connection. It is only used between the MQTT senders and receivers. The QoS levels are used to identify whether the message is received by the server from the Client or not. The QoS level is used when a client sends a subscription request to a particular topic. If a client subscribes to multiple topics, then they can use different QoS levels for different topics based on the need of the MQTT Clients.

The MQTT getting started guide makes a clear view of the QoS. There are three types of levels in QoS. They are,

  • QoS 0-At most Once

  • QoS 1-At least Once

  • QoS 2-Exactly Once

QoS 0 –> At most Once

The first level in QoS is 0. At this level, the message is sent to the receiver either once or never. There is no guarantee of a response from the receiver to the client. If the sender sends a message with QoS 0, then the recipient may or may not receive the message.  The client doesn’t know whether the message is sent to the receiver or not. This level is also called “Fire and Forget”. At this level, there is no guarantee of delivery and there is no acknowledgment by the broker.

QoS 1 –> At least Once

The second level in QoS is 1. At this level, the message must be sent to the receiver at least once. However, the message is delivered to the receiver multiple times until the acknowledgment is received from the broker. If the sender sends a message with QoS 1, then the subscribed recipient must receive the message under topics and then replies with a PUBACK packet. At this level, send is guaranteed, but the message may reach the broker more than once.

QoS 2 –> Exactly Once

The third level in QoS is 2, which is the highest level of service in MQTT. At this level, the message is received only once by their recipients. It is the safest QoS level. Between the sender and the receiver, there is a sequence of four messages. Here, the message has been sent to the receiver from the sender and the receiver has been sent an acknowledgment for that message. In this, the message was sent exactly once. So, most of the applications use this Quality of Service level.

Persistent Sessions / Message Queuing


A Persistent Session refers to an ongoing connection between a client and the MQTT message broker. Each MQTT client must store a persistent session or clean session. For example, when the MQTT Clients are connected to the broker, each client who is publishing and subscribing to the topic must contain the Persistent session or clean session. The persistent session is used while the internet is unreliable for the client. MQTT persistent session can be created by sending the Connect message and setting the clean session flag to 0. If the client does not have any sessions to send the CONNECT message, a new Persistent session will be created. If the client already has a session, the client will restart the session that already exists during the network connection. If the subscribed client suddenly disconnects from the broker, it will not receive any of the messages published on that topic from the publisher.

To overcome this problem, this persistent session or clean session is used. MQTT Broker supports clean session 0(false) and 1(true) for Persistent sessions. First, the MQTT client should connect to the broker, then it can request the broker for this persistent session. The broker stores all the client information. If the clients need to subscribe to the topics, they should use the QoS level of 1 or 2, and then the broker will publish the undelivered messages. After the disconnection, the client reconnects, then the broker will send all undelivered messages, which are all stored by the broker in a message queue manner to the client. When the message broker detects that a client is disconnected out of connection time, the persistent session expiration timer starts. When the client sends a CONNECT message, it sets the CleanSession flag to 1. If the persistent session expires before re-connection, then the stored messages in the broker are discarded.


Retained Messages


Retained messages is an important MQTT feature which enables to keep the latest key message on a topic for future client / subscribers. The publisher can ask the broker to keep the last message on the topic by setting the retained message flag. There can be only one retained message per topic at a time. The retained message is stored in the MQTT client (server). The retained messages in MQTT allow publishing MQTT clients to indicate the message stored for later delivery to subscribing clients. Retain messages can be retrieved by configuring the MQTT Retain bit. A client can publish or subscribe with any QoS level under a specific topic, but the retained message is only applicable whether the retained flag is set to true. The most common use of retained messages is to identify the last known messages to new subscribers.

Retained Messages are sent to new subscribers on a specific topic, then the broker processes the request and publishes the messages to the configured clients. If the server needs to clear the retain flag before forwarding, then the client will indicate it by itself. To delete a retained message, the client sends a retained message with a payload empty to a specific topic then the server deletes the retained message under that topic. Then the message expiration interval is set in the Retained Message. After the expiration interval of time, the retained message will be deleted. The retained message feature is useful for keeping the client’s last message, and especially useful when the state is constantly unchanged. Service quality does not affect retained messages. The Retained message feature is paired with LWT (last will and testament) feature in MQTT.


Last Will and Testament (LWT) and Will Delay

Last Will and Testament (LWT)

Last Will is a message stored on a specific MQTT Broker to specific clients. The Last Will message is a normal MQTT message that has the following fields will topic, will message, QoS, and retain flag. If there is an unexpected shutdown, then the Last Will and Testament are used to notify the users or MQTT clients. The MQTT client is used to publish the message. The fundamental techniques of the LWT are if the connection breaks, then using the Last Will message will help to notify the interested subscribers that they have lost the connection. For example, when the broker detects the disconnection, it sends the Last Will message to all subscribers of that specific topic. Making the right use of LWT in MQTT is, it avoids a lot of unwanted monitoring messages.

Will Delay

Will Delay provides the ability to delay sending a specific will message after the connection is interrupted. After the disconnection, if either the Will Delay interval has vanished, or the session ends, then the Will message must be published.  The network connection will be open for the MQTT client before the Will delay interval has faded. The Will Message should not be published, when the Will Delay interval is longer than the session expiry interval. The Will publish works as expected when the Will Delay interval is shorter than the session expiry interval. The server delays publishing the client’s Will Message until the Will Delay interval has expired or elapsed, which one of the above happens first. Thus the MQTT getting started guide delineates the Last Will and Will Delay.



MQTT getting started model explains the keep-alive in a control packet. Keep-alive is the interval of time between the start and end of the transmission process of the MQTT Client. Keep-alive is a two-byte integer that is measured in seconds. The MQTT client must have the responsibility to check whether the keep-alive value does not exceed the time interval between the sending MQTT control packets. If the keep-alive value is 0, then the keep-alive mechanism will be turning off and also, the client does not send MQTT control packets on any particular schedule. The actual value of the keep-alive is application-specific. The maximum value is 18 hours 12 minutes, and 15 seconds. The server does not receive an MQTT control packet till the Keep-alive time interval reaches one and a half times. Since the keep alive-value is non-zero, the client must close the network connection. If Keep-alive is in the control packet, then make sure that the connection between the broker and the client is still open and also confirm that the broker and the client are connected.


MQTT over Web Sockets

WebSocket is a computer communications protocol. It provides bidirectional simultaneous transmission over a single TCP  connection. With MQTT over WebSockets, every browser can be an MQTT device. The MQTT is transporting over a web socket connection. The WebSocket URL is used to connect the client and the server. A single WebSocket data frame contains multiple MQTT control packets, which does not create any impact on the MQTT protocol.

The following conditions should apply to use WebSockets in MQTT,

  • If the MQTT control packets are ready to transport over a WebSocket Connection, then it must be sent in WebSocket binary frames. Apart from this, any other frames that are received by the recipient must close the network connection.

  • The MQTT receiver control packets can not be aligned within the WebSocket frame boundaries.

  • The client should add “MQTT” to the list of WebSocket sub-protocols it provides. The server selected and returned the WebSocket sub-protocol name in “MQTT”.


User Property    

Properties (MQTTv5) is probably one of the most important combinations of the MQTT protocol explained by this MQTT getting started guide, and they are available in most MQTT message types. In the MQTT control packets like CONNECT, CONNACK, PUBLISH, PUBACK, PUBREC, PUBREL, PUBCOMP, SUBSCRIBE, SUBACK, UNSUBSCRIBE, UNSUBACK, DISCONNECT, and AUTH packet has a set of User Properties in the last field of the variable header.  The CONNECT packet also contains the options set of Properties in the Will Properties field with a payload. The User Property field is part of the properties field. A property defines its usage and data followed by a default value, and it can be used to provide additional information to the MQTT clients.

The CONNECT message will have different property fields than the PUBLISH message because property fields depend on the message type. The User property allows the same name to appear more than once and also allows it to appear multiple times to represent multiple names, value pairs when forwarding the messages to the client and server. It is mainly used for sending metadata between clients. User properties can be effective to allow the user to extend the MQTT protocol and to all messages and responses. They only have the meaning of the user’s activation because the user’s properties are defined by the user.

Session/Message Expiry   

The session expiry interval defines the period that the broker stores the session information of that particular MQTT client. Session Expiry is a four-byte integer that is represented in seconds. If the value is 0, then the session expiry interval is absent, and also the session ends when the network connection is closed. The session expiry interval in the CONNECT packet is used by the MQTT clients when the session expiry interval is absent. If the session expiry interval more than once included in the MQTT control packets and also in the CONNECT packet, the session expiry interval is zero, then it is to be considered as a protocol error. It doesn’t treat it as a valid DISCONNECT packet when the non-zero session expiry interval is received by the server. In the event of server disconnection, the session expiration interval should not be sent to the server. The message expiry interval is similar to the Session expiry interval. It is a new feature introduced by MQTT 5.0 protocol. Both the session expiry and message expiry interval can be used to improve capacity on the MQTT broker.

Reason codes/Acknowledgement


A Reason code is a new feature introduced by MQTT 5.0 protocol and here we have described this in MQTT getting started guide. It is a one-byte unsigned value that is used for the result of an operation indication. The successful completion of an operation in reason codes indicates less than 0*80. The reason code for success is 0. The reason code values greater than 0*80 that indicate failure. Many MQTT packets now include reason codes. A reason indicated that a predefined protocol error occurred in the code. These reason codes are usually carried in acknowledgment packets and allow the client and broker error conditions. The reason codes are sometimes called negative acknowledgment.  The below control packets allow the subscriber to determine if the requested function is successful. The control packets have a single reason code as part of the variable header are CONNACK, PUBACK, PUBREC, PUBREL, PUBCOMP, DISCONNECT, and AUTH. The control packets have a list of one or more reason codes in their payload are SUBACK and UNSUBACK.



 Reason Code











Normal disconnection




Granted QoS 0




Granted QoS 1




Granted QoS 2




Disconnect with Will Message




No matching subscribers




No subscription existed




Continue authentication








Unspecified error




Malformed Packet




Protocol Error




Implementation specific error




Unsupported Protocol Version




Client ID (Identifier) not valid




Bad User Name or Password




Not authorized




Server unavailable




Server busy








Server shutting down




Bad authentication method




Keep Alive timeout




Session taken over




Topic Filter invalid




Topic Name invalid




Packet Identifier in use




Packet Identifier not found




Receive Maximum exceeded




Topic Alias invalid




Packet too large




Message rate too high




Quota exceeded




Administrative action




Payload format invalid




Retain not supported




QoS not supported




Use another server




Server moved




Shared Subscriptions not supported




Connection rate exceeded




Maximum connect time




Subscription Identifiers not supported




Wildcard Subscriptions not supported


Server Disconnection

Server disconnection happens when the client sends the DISCONNECT packet with a reason code. The disconnection is the final step in the MQTT process. The DISCONNECT is the last MQTT control packet sent from the client or the server. Server disconnection indicates the reason why the Network Connection is being closed.  Before closing the Network Connection, the Client or Server MAY send a DISCONNECT packet. If the Client without sending a DISCONNECT packet first with Reason Code 0x00 (Normal disconnection) but the Network Connection is closed and the Connection has a Will Message, this Will Message is published.

Payload Format / Content-Type

Payload Format

The Payload Format is part of any MQTT packet that can contain a payload. Allow the Payload format to be specified when a message is published. This can be forwarded to the receiver of the message. Some MQTT control packets contain a payload as the final part. The control packets that must require the payload format fields are CONNECT, SUBSCRIBE, SUBACK, UNSUBSCRIBE, UNSUBACK. In the PUBLISH control packet, the Payload format field is an optional one. Because it is considered as an Application Message. The payload field exists only in the will property of PUBLISH packet and CONNECT packet. A server must send the payload format indicator unchanged, to all subscribers who receive the application message. The receiver can confirm that the payload is in the format indicated, and if it doesn’t send a PUBACK, PUBREC, or causal code then the payload format is incorrect.



Content-Type exists only in the WILL property of PUBLISH packet and CONNECT packet. Allow the MIME (Multipurpose Internet Mail Extensions – the content type of the data sent by a server-side to a client-side) style to be specified when a message is published. This can be forwarded to the receiver of the message. A server must send the content message to all subscribers, without changing the content type. The Content-type is more than once included in the packet for that, it is to be considered a protocol error.  The Content-type value is defined by sending and receiving the application messages. Cannot truncate the content type during message forwarding. Thus the MQTT getting started guide clarifies the payload format and content type.

Request/Response pattern

Formulate request/response system within MQTT provides response titles and contact data attributes to allow response messages to be returned to the publisher of a request. Add the client’s capability to get configuration information from the server about how to configure response headers. Some applications may want to run request/response contacts via MQTT. This version of MQTT has three properties that can be used for some purpose: response topic, correlation data, and request/response information.

An MQTT client publishes a requested message to a topic that is forwarded to all subscribed clients. There are multiple responders subscribed to the same topic name. The responder takes action based on the request message and then publishes response topics to the clients. The request/response message can be of any QOS, the responder use Session Expiry Interval and shared subscription. The requester has the responsibility to ensure the necessary authority to publish and subscribe to the topic name it sets in the response topic property. It is the responsibility of the respondent to ensure that responder has the authority to subscribe to the requested topic and publish the respondent topic. 

Request-Response interaction process: 

  • An MQTT client (the requester) publishes a Request Message with a response topic.

  • The other clients (responders) subscribed to a topic that matches the topic name sent by the requester, then other MQTT clients received the request message.

  • The responder publishes a response message to the topic name which was existed in the request message also takes suitable action based on the request message.

  • In the regular application, the requester subscribes to the response header so that the response message is received. However, some other clients may subscribe to the response header, where the response message is received and processed by that client. Like a request message, multiple clients or anyone can subscribe to the response message.

Shared Subscription

The MQTT 5.0 protocol introduces a new feature is called shared subscriptions. It equates to the load balancing function of multiple subscribers. Adding shared subscription support will allow the subscriber to load the balance. It can be identified using a special style of topic filter. The shared name must not contain “/”, “+”, or “#” special characters. $ share prefix indicates that a shared subscription. Shared subscriptions can be associated with multiple subscribing MQTT sessions. The publication of messages that match its topic filter is sent only to one of its subscribing sessions. This feature can be useful where several clients share the processing of the publication in parallel. Generally, applications use the shared name to refer to a group of subscription sessions.

Shared subscriptions are defined in terms of the MQTT server instead of a session. Once there is a shared subscription, other sessions can be subscribed to the same shared subscription topic filter. It only sends one of the sessions to which can subscribe to a shared subscription. Each shared subscription is independent of one another. It is possible to have two shared subscriptions with overlapping filters. In such cases, a message that applies to both shared subscriptions will be processed separately. When sending an Application Message to a Client, the Server MUST give the maximum QoS for the Client’s subscription because different subscribing clients ask for different requested QOS levels.


Subscription Identifier & Subscription Options

Subscription Identifier

During the subscription, the client can specify a subscription identifier. The subscription identifier is followed by the integer value byte that represents the subscriber’s identifier. The subscription identifier has a value of 1 to 268,435,455. If the subscription identifier value is zero, then it is a protocol error. The Subscription identifier allows the client to determine which subscription or subscriptions caused the message to be delivered. When the client can specify a subscription identifier during a subscription, the broker establishes and stores the mapping relationship between this subscription and the subscription identifier when it successfully creates or changes the subscription. When you need to send PUBLISH packets to this client that matches this subscription, the broker will return the subscription identifier with this PUBLISH packet. Multiple Subscription identifiers will be added if the publication is the result of identical items for more than one subscription. In this case, the order is not significant.


Subscription Options


In the MQTT v5,  the server behavior is changed by using subscription options. Provide primarily limited subscription options to allow applications. This includes the options not to send messages appearing on this client (No Local) and options for handling messages retained in the subscription. In MQTT v5, subscription options support QoS, No Local, Retain As Publish, Retain Handling. It may be helpful not to send retained messages to an existing subscription when reconnected, and the client will not know if the subscription was completed on the previous connection for the session. It is useful not to send saved messages due to a new subscription where a customer wants to receive change notifications and does not need to know the initial states. Subscription Options give the maximum QoS level at which the Server can send Application Messages to the client. The subscription identifier and subscription options are clearly explained by this MQTT getting started guide.


Topic Alias

Topic Alias is a new feature introduced by the MQTT 5.0 protocol, which is a mechanism for reducing the size of published packets by reducing the size of the topic field. It will reduce the MQTT packet overhead size by allowing the topic name to be abbreviated to a small integer. The Client and Server independently specify, how many topic aliases they allow. A Topic Alias is an integer value that is used to identify a topic rather than a topic name. This is useful when the topic names are long and the same topic names are repeatedly used within a network connection. The mappings of the topic alias exist only within a network connection and will only last for the lifetime of that network connection. If these mappings are used by the Client and Server, then they are independent of each other. In most cases, they are not essential but if many customers are frequently released they can significantly reduce network traffic.

Flow Control

Flow control is the process of managing the rate of data transmission between MQTT clients. It is useful for controlling the number of unacknowledged PUBLISH packets. Allow the client and the server to specify the number of pending trusted messages (QoS > 0) independently. Then the sender will stop sending such messages to study below this quota. The client or server should not receive more than its initial dispatch quotation value of zero. At other MQTT control, packets must continue to process and to respond even if the client and server quotation value is zero. This is used to limit the rate of credible messages and to control how many are on the path at once. It is mainly used for controlling the number of pending reliable messages.

Maximum Packet Length

Here we have defined the maximum packet length. This Packet Length specifies the size of the whole packet including the header and the data sent on that packet. Allow the client and server independently to determine the maximum packet size they support. Sending a large Packet to the session partner is an error.

Optional Server Function Availability

Optional Server Function Availability defines the set of features that the server does not allow, and provides the client with a mechanism to specify this function from the server. The main features of Server function availability are Maximum QoS, Retain Available, Wildcard Subscription Available, Subscription Identifier Available, and Shared Subscription Available. If the server declares these features are not available, then the clients using these features are considered an error. The fact that a server cannot enable a feature in earlier versions of MQTT is made possible by notifying the client that the functionality is not authorized. This feature allows us to report such preferred behavior and adds specific reason codes when the client uses one of these features.

Enhanced authentication


Enhanced authentication is used to implement authentication in a challenge-response manner. It provides a mechanism to enable authentication with username and password including mutual recognition. Enhanced authentication relies on the authentication method and authentication data to complete the entire authentication process. The authentication method is the agreement between the client and the server on the object of the data sent in the authentication data. The data transfers and processes are required for the client and server to complete the authentication. This allows SASL (Simple Authentication and Security Layer) style authentication to be used if supported by both the client and the server and can re-authenticate a client within a connection. MQTT offers the implementation of enhanced authentication. This implementation of enhanced authentication is optional for both clients and servers.

Server Keeps Connected


This allows the server to specify the Keep-alive values the client has to use also allows the server to set the maximum allowable keep-alive.


Hope this article provides a detailed explanation on MQTT enabling early adopters to kick-start with MQTT.

Have More Questions?

We are with all ears waiting to hear from you. Post us with your questions and feedback.