This document provides an overview of Apache Hadoop, including its architecture, components, and ecosystem. Hadoop is an open-source framework for distributed storage and processing of large datasets across clusters of commodity hardware. It consists of HDFS for storage, MapReduce for processing, and YARN for resource management. Related projects in the Hadoop ecosystem include HBase, Hive, Pig, Flume, Sqoop, Oozie, Zookeeper, and Mahout.
3. Hadoop Tutorial
1
https://data-flair.training/big-data-hadoop/
1. Hadoop Tutorial
Apache Hadoop is an open source, Scalable, and Fault tolerant framework written in Java. It
efficiently processes large volumes of data on a cluster of commodity hardware. Hadoop is not only a
storage system but is a platform for large data storage as well as processing.
We will learn in this Hadoop tutorial about Hadoop architecture, Hadoop daemons, different flavors
of Hadoop. At last, we will cover the Hadoop components like HDFS, MapReduce, Yarn, etc.
2. What is Hadoop?
Hadoop is an open-source tool from the ASF – Apache Software Foundation. Open source project
means it is freely available and we can even change its source code as per the requirements. If certain
functionality does not fulfill your need then you can change it. Most of Hadoop code is written by
Yahoo, IBM, Facebook, Cloudera.
It provides an efficient framework for running jobs on multiple nodes of clusters. Cluster means a
group of systems connected via LAN. Apache Hadoop provides distributed processing of data as it
works on multiple machines simultaneously.
By getting inspiration from Google, which has written a paper about the technologies it is using
technologies like Map-Reduce programming model as well as its file system (GFS). Hadoop was
originally written for the Nutch search engine project. When Doug Cutting and his team were working
on it, very soon Hadoop became a top-level project due to its huge popularity.
Apache Hadoop is an open source framework written in Java. The basic Hadoop programming
language is Java, but this does not mean you can code only in Java. You can code in C, C++, Perl,
Python, ruby etc. but it will be better to code in java as you will have lower level control of the code.
Hadoop efficiently processes large volumes of data on a cluster of commodity hardware. Hadoop is
developed for processing huge volume of data. Commodity hardware is the low-end hardware; they
are cheap devices which are very economical. Hence, Hadoop is very economic.
Hadoop can be setup on a single machine (pseudo-distributed mode), but it shows its real power with
a cluster of machines. We can scale it to thousand nodes on the fly ie, without any downtime.
Therefore, we need not to make the system down to add more nodes in the cluster. Follow this guide
to learn Hadoop installation on a multi-node cluster.
Hadoop consists of three key parts –
Hadoop Distributed File System (HDFS) – It is the storage layer of Hadoop. HDFS is the
most reliable storage system on the planet.
Map-Reduce – It is the data processing layer of Hadoop. MapReduce is the distributed
processing framework, which processes the data at lightning fast speed.
YARN – Yet Another Resource Negotiator, It is the resource management layer of Hadoop.
Yarn manages resources on the cluster
4. Hadoop Tutorial
2
https://data-flair.training/big-data-hadoop/
3. Why Hadoop?
Let us now understand why Hadoop is very popular, why Hadoop capture more than 90% of big
data market.
Apache Hadoop is not only a storage system but is a platform for data storage as well as processing.
It is scalable (as we can add more nodes on the fly), Fault tolerant (Even if nodes go down, data is
processed by another node).
Following characteristics of Hadoop make it a unique platform:
Flexibility to store and mine any type of data whether it is structured, semi-structured or
unstructured. It is not bounded by a single schema.
Excels at processing data of complex nature. Its scale-out architecture divides workloads
across many nodes. Another added advantage is that its flexible file-system eliminates ETL
bottlenecks.
Scales economically, as discussed it can deploy on commodity hardware. Apart from this its
open-source nature guards against vendor lock.
4. Hadoop Architecture?
After understanding what is Apache Hadoop, let us now understand the Hadoop Architecture in detail.
Hadoop works in master-slave fashion. There are master nodes (very few) and n numbers of slave
nodes where n can be 1000s. Master manages, maintains and monitors the slaves while slaves are the
actual worker nodes. In Hadoop architecture the Master should be deployed on a good hardware, not
just commodity hardware. As it is the centerpiece of Hadoop cluster.
Master stores the metadata (data about data) while slaves are the nodes which store the actual data
distributedly in the cluster. The client connects with master node to perform any task. Now in this
Hadoop tutorial, we will discuss different components of Hadoop in detail.
5. Hadoop Tutorial
3
https://data-flair.training/big-data-hadoop/
5. Hadoop Components
There are three most important Apache Hadoop Components. In this Hadoop tutorial, you will learn
what is HDFS, what is MapReduce and what is Yarn. Let us discuss them one by one:
5.1. HDFS – Strong Layer
Hadoop HDFS or Hadoop Distributed File System is a distributed file system which provides storage
in Hadoop in a distributed fashion.
In Hadoop Architecture on the master node, a daemon called namenode run for HDFS. On all the
slaves a daemon called datanode run for HDFS. Hence slaves are also called as datanode. Namenode
stores meta-data and manages the datanodes. On the other hand, Datanodes stores the data and do
the actual task.
HDFS is a highly fault tolerant, distributed, reliable and scalable file system for data storage.
First Follow this guide to learn more about features of HDFS and then proceed further with the Hadoop
tutorial.
HDFS is developed to handle huge volumes of data. The file size expected is in the range of GBs to TBs.
A file is split up into blocks (default 128 MB) and stored distributedly across multiple machines. These
blocks replicate as per the replication factor. HDFS handles the failure of a node in the cluster.
5.2. MapReduce – Processing Layer
Now it’s time to understand one of the most important pillar if Hadoop, i.e. MapReduce. MapReduce
is a programming model. As it is designed for large volumes of data in parallel by dividing the work
into a set of independent tasks. MapReduce is the heart of Hadoop, it moves computation close to
the data. As a movement of a huge volume of data will be very costly. It allows massive scalability
across hundreds or thousands of servers in a Hadoop cluster.
6. Hadoop Tutorial
4
https://data-flair.training/big-data-hadoop/
Hence, MapReduce is a framework for distributed processing of huge volumes of data set over a
cluster of nodes. As data is stored in a distributed manner in HDFS. It provides the way to Map–
Reduce to perform distributed processing.
5.3. YARN – Resource Management Layer
YARN – Yet Another Resource Negotiator is the resource management layer of Hadoop. In the multi-
node cluster, it becomes very complex to manage/allocate/release the resources (CPU, memory,
disk). Hadoop Yarn manages the resources quite efficiently. It allocates the same on request from any
application.
On the master node, the ResourceManager daemon runs for the YARN then for all the slave
nodes NodeManager daemon runs.
Learn the differences between two resource manager Yarn vs. Apache Mesos. Next topic in the
Hadoop tutorial is a very important part i.e. Hadoop Daemons
6. Hadoop Daemons
Daemons are the processes that run in the background. There are mainly 4 daemons which run
for Hadoop.
Hadoop Daemons:
Namenode – It runs on master node for HDFS.
Datanode – It runs on slave nodes for HDFS.
ResourceManager – It runs on master node for Yarn.
NodeManager – It runs on slave node for Yarn.
These 4 demons run for Hadoop to be functional. Apart from this, there can be secondary NameNode,
standby NameNode, Job HistoryServer, etc.
7. Hadoop Tutorial
5
https://data-flair.training/big-data-hadoop/
7. How Hadoop works?
Till now we have studied Hadoop Introduction and Hadoop architecture in great details. Now let us
summarize Apache Hadoop working step by step:
i) Input data is broken into blocks of size 128 MB (by default) and then moves to different nodes.
ii) Once all the blocks of the file stored on datanodes then user can process the data.
iii) Now, master schedules the program (submitted by the user) on individual nodes.
iv) Once all the nodes process the data then the output is written back to HDFS.
8. Hadoop Flavors
This section of Hadoop Tutorial talks about the various flavors of Hadoop.
Apache – Vanilla flavor, as the actual code is residing in Apache repositories.
Hortonworks – Popular distribution in the industry.
Cloudera – It is the most popular in the industry.
MapR – It has rewritten HDFS and its HDFS is faster as compared to others.
IBM – Proprietary distribution is known as Big Insights.
All flavors are almost same and if you know one, you can easily work on other flavors as well.
8. Hadoop Tutorial
6
https://data-flair.training/big-data-hadoop/
9. Hadoop Ecosystem Components
In this section, we will cover Hadoop ecosystem components. Let us see what all the components form
the Hadoop Eco-System:
Hadoop HDFS – Distributed storage layer for Hadoop.
Yarn Hadoop – Resource management layer introduced in Hadoop 2.x.
Hadoop Map-Reduce – Distributed processing layer for Hadoop.
HBase – It is a column-oriented database that runs on top of HDFS. It is a NoSQL database which does
not understand the structured query. For sparse data set, it suits well.
Hive – Apache Hive is a data warehousing infrastructure based on Hadoop and it enables easy data
summarization, using SQL queries.
Pig – It is a top-level scripting language. Pig enables writing complex data processing without Java
programming.
Flume – It is a reliable system for efficiently collecting large amounts of data from many different
sources in real-time.
Sqoop – It is a tool design to transport huge volumes of data between Hadoop and RDBMS.
Oozie – It is a Java Web application uses to schedule Apache Hadoop jobs. It combines multiple jobs
sequentially into one logical unit of work.
Zookeeper – A centralized service for maintaining configuration information, naming, providing
distributed synchronization, and providing group services.
9. Hadoop Tutorial
7
https://data-flair.training/big-data-hadoop/
Mahout – A library of scalable machine-learning algorithms, implemented on top of Apache Hadoop
and using the MapReduce paradigm.
Refer this Hadoop Ecosystem Components tutorial for the detailed study of All the Ecosystem
components of Hadoop.
10. Conclusion
In conclusion to this Hadoop tutorial, we can say that Apache Hadoop is the most popular and
powerful big data tool. Hadoop stores & processes huge amount of data in the distributed manner on
a cluster of nodes. It provides world’s most reliable storage layer- HDFS. Batch processing engine
MapReduce and Resource management layer- YARN.
This conclusion is not the end but a foundation to learn Apache Hadoop. Here are the next steps after
you are through with this Apache Hadoop Tutorial:
1. Internal Working of Hadoop
2. Hadoop Distributed File System
3. MapReduce Tutorial
4. YARN - Tutorial
5. Install Hadoop 2 on Ubuntu