Kafka is a stream processing software developed by LinkedIn previously and now functioning under Apache foundation. Kafka is written in Scala, it is a publish-subscribe based messaging system. 

In this tutorial, you will gain knowledge on concepts like Kafka Introduction, messaging system, terminologies, workflow, cluster setup, use-cases and real-time applications.

Enroll in our Apache Kafka Certification course today and develop a strong foundation in Apache Kafka.

The following topics will be covered in the article

What is Apache Kafka

Apache Kafka is a messaging system which allows interaction between producers and consumers through message-based topics. Kafka is becoming popular because of the features like easy access, immediate recovery from node failures, fault-tolerant, etc. These features make Apache Kafka suitable for communication, integrating the components of big data systems. It is integrated with Apache spark and storm for analyzing the streamed data. This tutorial teaches you the Kafa basics, Advantages, Disadvantages, workflow, Installation and basic operations. In the end, we will conclude with real-time applications of Kafka.

Kafka Messaging System

The main job of the messaging system is to transfer the data from one application to another. As a result, applications can concentrate on data, not on how to share it. Messages are distributed using a message queuing system. The consumer consumes messages present in the queue by following specific messaging patterns, they are:


In this pattern. Messages remain in the queue. Anyone can consume the messages present in the queue, but only one consumer can consume a specific message. When the consumption is over, that particular message goes away from the queue.


This pattern is popularly known as pub-sub, and it is the most widely used messaging pattern. In this pattern, Messages remain in a topic. In this system, publishers are those who produce the messages and subscribers are those who consume the messages. To explore this system, we can take Dishtv as an example. Dishtv publishes various channels, and anybody may subscribe to their group of channels and have them at any time.

Kafka Advantages

Reliable: It is reliable because of the features like distributed, replicated and fault-tolerant.

Scaleable: The messaging system of Kafka works without any time outs.

Durable: Kafka is durable because messages exist on the disk right away with the help of “Distributed commit log”.

Throughput: Throughput of Kafka is high, and it remains high even if a huge amount(TB) of messages are saved.

Kafka Terminology

Kafka terminologies

Topics: A flow of messages associated with a specific category is known as a topic. Data or Messages are stored according to topics.

Partition: Topics are divided into partitions so that it can manage any random data.

Partition offset: The messages which are partitioned will have an individual sequence id, which is known as Partition offset.

Replicas of Pattern: Replicas are also known as “backups” of a partition. They are not at all used for reading or writing data. Preventing data loss is the job of replicas. 

Brokers: Brokers are defined as a system established for managing the published data. Each broker may be allocated to zero or more partitions of a topic.

Cluster: The Kafka's who has more than one broker is known as Kafka cluster. Clusters are utilized to handle the existence and replicance of message data.

Producers: Producers are known as publishers of the messages to more than one Kafka topics. Producers may send messages to brokers or partition.

Consumers: consumers are known as readers of the messages. consumers consume the messages by extracting the data from brokers,

Leader: The node responsible for all the reads and writes of a given partition is known as a leader. Each partition will have one server operating as a leader.

Follower: Follower is known as the node which follows the leader's instructions. 

Kafka Cluster Architecture

Kafka Cluster

The above diagram shows the Kafka cluster architecture. The elements of the Kafka cluster architecture can be explained in the following way:

Broker: Usually Kafka cluster contains several brokers to preserve load balance. As Kafka clusters do not have states, they take zookeeper’s help to sustain cluster state. In a second, the number of reads and writes are handled by the single kafka broker instance. Zookeeper does Kafka broker leader election.

ZooKeeper: The primary responsibility of Zookeeper is to manage and synchronize the Kafka broker. Zookeeper will notify the producer and consumer regarding the existence of the new broker or breakdown of the broker. According to this notification, the producer and consumer will take the decision and starts synchronizing their activities with another broker.

Subscribe to our youtube channel to get new updates..!

Producer: Producer main job is to move data to the brokers. If producers find the existence of new broker, automatically they will forward the data to that broker. Producers do not wait for the recognition of the broker, they will send messages as soon as a broker can manage.

Consumer: As brokers are stateless, consumers have to record how many messages are spent by utilizing Partition offset. If the consumer recognises a specific message offset, it indicates that all the previous messages are consumed by that consumer. For consuming the next message, the consumer sends an asynchronous request to the broker to make a set of bytes ready. The consumer can rewind or skip anywhere in the partition by using message offset.

Kafka WorkFlow

Kafka allows both pub-sub and queue-based messaging system. In both systems, producers only job is to send the messages and consumers job is to choose any messaging system depending upon the requirement. Let us see the steps required for a consumer to select a messaging system.

Pub-sub messaging Workflow

  • Producers send messages to the topic frequently.
  • After receiving the messages from the producer, broker stores all the messages in the partitions which are arranged for that particular topic. Broker’s other responsibility is to distribute the messages uniformly. For example, if a broker gets four messages, and there are four partitions, then the broker will store one message in each partition.
  • Consumer subscribes to a particular topic.
  • When the consumer subscribes to a topic, Kafka will give the current offset of the topic to the consumer and stores the offset in the ZooKeeper’s ensemble.
  • Consumer frequently requests Kafka for new messages.
  • After receiving the messages from the producer, Kafka forwards them to consumers.
  • After receiving the messages, the consumer will process them.
  • After processing the messages, consumers send recognition to the Kafka broker.
  • After receiving the acknowledgement, it updates the offset value and updated value is stored in ZooKeeper. 
  • The above workflow is repeated until the consumer is requesting the resource.
  • At any moment, the consumer can rewind/skip to the required offset of a topic and read all ensuing messages.

Queue Messaging Workflow.

In queue messaging system rather than a single consumer, group of consumers who are having the same ‘group ID’ will subscribe to a topic. Consumers who are subscribed to a topic with same “group ID” are considered as one group, and the messages of that topic are shared among them. The real workflow of this system is as follows:

  • At regular intervals, producers send messages to a topic.
  • Like the earlier strategy, messages are stored by Kafka in the partitions designed for that particular topic.
  • One consumer subscribes to a particular topic say “topic1” with “Group ID” say “Gr1”.
  • The communication between Kafka and consumer is carried out in the same manner as the pub-sub messaging system until the new consumer subscribes that topic.
  • After the arrival of the new consumer, Kafka changes its function to share mode, to share the messages between the old consumer and new consumer. This process of sharing is continued till the number of consumers equals the number of partitions designed for that topic.
  • When the number of consumers surpasses the number of partitions, then the new consumer will not receive any subsequent message until any current consumer unsubscribes. This situation emerges because, in Kafka, every consumer is allocated to at most one partition. When all the partitions are allocated to current consumers, then new consumers have to wait for the partition.
  • This property is known as a consumer group. Likewise, Kafka will offer the finest of both systems effectively.

Kafka Cluster Setup

Before installing Apache Kafka, we should check whether Java is installed in our system. To check the java installation, we should follow the below step :

Step-1: Checking Java

$ java -version

 If java is installed in your system, java installed version is displayed. If java is not installed, you have to follow the below steps for installing java

Step 1.a:  Downloading JDK

JDK can be downloaded by clicking on the below URL:


By clicking the above URL, you can download JDK based on your system configuration.

Step 1.b:  Extracting the JDK files.

After downloading JDK in your system, open downloads folder and check whether JDK is stored or not. After checking, extract the files from the JDK archive file(tar or rar) by using the below commands:

$cd /my/dir/downloads/path
$tar -zxf jdk-8u241-win-x32.gz.

Step 1.c:  Move to my directory.

Java files are made accessible to every user by extracting java files to my directory.

         Password: (type the password of the root user)
$mkdir /my/jdk
$mv jdk-13.0.2_windows-x64_bin.tar.gz /my/jdk/

Step 1.d: Setting environment Variables and Path

For setting environment variables and path, you should follow the below commands.

export JAVA_HOME = /user/jdk/jdk-13.0.2
export PATH=$PATH: $JAVA_HOME/bin 

Step 2: Installing the Zookeeper Framework

Step 2.a: Downloading Zookeeper

By visiting the below link, you can download the zookeeper framework.


Zookeeper’s latest version is 3.5.6

Step 2.b: Tar file Extraction

We can extract files from tar files by following the below commands.

$cd my/
$tar -zxf zookeeper-3.5.6.tar.gz
$cd zoo3.5.6
$mkdir mydir1

Step 2.c: Creation of configuration file

By executing command vi “conf/zoo.cfg” we can open the configuration file which is named as “conf/zoo.cfg”

$ vi conf/zoo.cfg
mydir1 Dir = /path/for/zookeeper/mydir1

After this configuration file is saved, zookeeper can be started.

Step 2.d:  Starting Zookeeper server

$ bin/zkserver.sh start

When you execute the above command, the System displays the following response:

$ JMX is enabled default
$ Using config: /user1/../zookeeper-3.5.6/bin/ ../conf/zoo.cfg
$ Starting zookeeper … STARTED

Step 2.e: Connecting to the Zookeeper server.

For connecting to the zookeeper server, you should execute the following command.

$ zkCli.sh

After executing the command, your Zookeeper server is started.

Step 2.f: Stopping the server of Zookeeper

After completing your work,   zookeeper server  is stopped by executing the below command:

$ bin/zkserver.sh stop

We have successfully finished the installation of java and zookeeper. Now, let us see the installation of Apache Kafka.

Step 3: Installing Apache Kafka

To install apache Kafka into our system, the following steps are executed

Step 3.a: Download Apache Kafka 

 For installing the Apache Kafka into your system, you have to download it from the following URL


Step 3.b: Extracting Tar file

By executing the below commands, tar file is extracted.

 $ cd my/
$ tar -zxf kafka_2. tar.gz
$cd kafka_2.

You have successfully downloaded the Apache Kafka into your system.

Step 3.c:  Starting Apache Kafka server

The Apache Kafka server is started by executing the below command.

$ bin/kafka-server-start.sh config/server.properties

When we execute the above command, your Apache Kafka will be started. You will see the parameters of the server like timeout, protocol version, roll hours, etc.

Step 3.d: Stopping the Apache Kafka server.

The Apache Kafka server can be stopped by executing the following command.

$ bin/kafka-server-stop.sh config/server.properties

Kafka use-cases

1. Messaging

2. Tracking Website Activity

3. Metrics

4. Log Aggregation

5. Stream Processing

6. Event Sourcing

7. Commit Log

Real-time Application of Kafka


Apache Kafka is widely used in Twitter Platform. Due to Apache Kafka users can send and receive tweets. In twitter, logged users can view and post tweets, but users who are not logged can only see the tweets. For its stream processing, twitter uses Storm-Kafka.


For data streaming and operational measures, Apache Kafka is used in LinkedIn. Many products of LinkedIn like LinkedIn Newsfeed, LinkedIn Today, take the help of Apache Kafka. The durability factor of Apache Kafka helps LinkedIn to use it.


Netflix is another online platform which uses Apache Kafka for carrying out its services. It uses Apache Kafka only for event processing and video streaming.

Apache Kafka is also used by Mozilla, Oracle and many other enterprises.


Apache Kafka, introduced by Apache, plays a vital role in managing real-world data feeds. It offers fault tolerance if any machine fails. Kafka is speedy, it does two million writes/sec. It provides messaging in two different ways. Its simple terminology makes us understand the procedure for message passing. This tutorial is enough to acquire basic knowledge about Apache Kafka. For more information about Apache Kafka, please attend Apache Kafka Training.