SlideShare a Scribd company logo

Hadoop introduction

•
•
Chirag Ahuja
Chirag Ahuja

This document provides an introduction to Hadoop, including its ecosystem, architecture, key components like HDFS and MapReduce, characteristics, and popular flavors. Hadoop is an open source framework that efficiently processes large volumes of data across clusters of commodity hardware. It consists of HDFS for storage and MapReduce as a programming model for distributed processing. A Hadoop cluster typically has a single namenode and multiple datanodes. Many large companies use Hadoop to analyze massive datasets.

Related slideshows

Hadoop Technology
Hadoop TechnologyHadoop Technology
Hadoop Technology
 
Hadoop distributions - ecosystem
Hadoop distributions - ecosystemHadoop distributions - ecosystem
Hadoop distributions - ecosystem
 
Apache hadoop technology : Beginners
Apache hadoop technology : BeginnersApache hadoop technology : Beginners
Apache hadoop technology : Beginners
 
1 of 19
A SOFT INTRODUCTION
Contents Hadoop Introduction Ecosystem Architecture HDFS Map-Reduce Characteristics Flavors
A large ecosystem
Zookeeper (Coordination) Oozie (Workflow) Sqoop (Data Exchange) Hbase (Columnar Store) HIVE (SQL-Query) PIG (Scripting) Mahout (Data Mining) MapReduce (Distributed Processing Framework) HDFS (Hadoop Distributed Framework) Flume (Data Exchange) Avro (Data Serialization System) STORM KAFKA TAJO SCALA IMPALA (Real Time Processing)
Who uses Hadoop ?
Hadoop An open source project from the Apache Software Foundation It provides a software framework for distributing and running applications on clusters of servers It is inspired by Google's Map-Reduce programming model as well as its file system (GFS) Hadoop was originally written for the Nutch search engine project
Hadoop Hadoop is open source framework written in Java It efficiently process large volumes of data on a cluster of commodity hardware Hadoop can be setup on single machine, but real power of Hadoop comes with a cluster of machines, it can be scaled from single machine to thousand nodes on the fly Hadoop consists of two key parts – Hadoop Distributed File System (HDFS) and Map-Reduce
Architecture Master Secondary Master User . . . . . . . . . Slaves Hadoop Cluster
HDFS HDFS is a highly fault tolerant, distributed, reliable, scalable file system for data storage HDFS stores multiple copies of data on different nodes; a file is split up into blocks (default 64 MB) and stored across multiple machines Hadoop cluster typically has a single namenode and number of datanodes to form the HDFS cluster
HDFS
Map-Reduce Map-Reduce is a programming model designed for processing large volumes of data in parallel by dividing the work into a set of independent tasks Map-Reduce is a paradigm for distributed processing of large data set over a cluster of nodes
Hadoop Daemons Namenode (Master) (HDFS) SecondaryNameNode (3rd system / Slave) JobTracker (Master) (MR) DataNode (Slave) (HDFS) TaskTracker (Slave) (MR)
Hadoop Daemons HADOOP CLUSTER Storage Layer Computation Layer Phase HADOOP DAEMONS ARCHITECTURE Map-Reduce Job Tracker Task Tracker Task tracker Task tracker MapReduce jobs are submitted on jobtracker HDFS NameNode DataNode DateNode DataNode NameNode Stores Meta-data
task tracker task tracker task tracker task tracker data node data node data node job tracker name node data node task tracker data node Master Slaves Hadoop Cluster Map-Reduce Layer Storage Layer
Characteristics Open-source ◦ Code can be modified according to business requirements Distributed Processing ◦ Data is processed parallely on cluster of nodes in distributed manner Fault Tolerance ◦ Failure of nodes or tasks are recovered automatically by the framework
Characteristics Reliablity ◦ Data is reliably stored on the cluster of machine despite machine failures High Availability ◦ Data is highly available and accessible despite hardware failure Scalablility ◦ New hardware can be added to the nodes ◦ Horizontal Scalablility – new nodes can be added on the fly
Characteristics Economic ◦ Runs on cluster of comodity hardware Easy to use ◦ No need of client to deal with distributed computing , framework takes care of all the things Data Locality ◦ Move Computation to data instead of data to computation
Flavors Apache Cloudera MapR IBM Pivotal Connectors ◦ Almost all the databases have provided their connector with Hadoop for fast data transfer
QUESTIONS ??

More Related Content

Hadoop introduction

  • 2. Contents Hadoop Introduction Ecosystem Architecture HDFS Map-Reduce Characteristics Flavors
  • 4. Zookeeper (Coordination) Oozie (Workflow) Sqoop (Data Exchange) Hbase (Columnar Store) HIVE (SQL-Query) PIG (Scripting) Mahout (Data Mining) MapReduce (Distributed Processing Framework) HDFS (Hadoop Distributed Framework) Flume (Data Exchange) Avro (Data Serialization System) STORM KAFKA TAJO SCALA IMPALA (Real Time Processing)
  • 6. Hadoop An open source project from the Apache Software Foundation It provides a software framework for distributing and running applications on clusters of servers It is inspired by Google's Map-Reduce programming model as well as its file system (GFS) Hadoop was originally written for the Nutch search engine project
  • 7. Hadoop Hadoop is open source framework written in Java It efficiently process large volumes of data on a cluster of commodity hardware Hadoop can be setup on single machine, but real power of Hadoop comes with a cluster of machines, it can be scaled from single machine to thousand nodes on the fly Hadoop consists of two key parts – Hadoop Distributed File System (HDFS) and Map-Reduce
  • 8. Architecture Master Secondary Master User . . . . . . . . . Slaves Hadoop Cluster
  • 9. HDFS HDFS is a highly fault tolerant, distributed, reliable, scalable file system for data storage HDFS stores multiple copies of data on different nodes; a file is split up into blocks (default 64 MB) and stored across multiple machines Hadoop cluster typically has a single namenode and number of datanodes to form the HDFS cluster
  • 10. HDFS
  • 11. Map-Reduce Map-Reduce is a programming model designed for processing large volumes of data in parallel by dividing the work into a set of independent tasks Map-Reduce is a paradigm for distributed processing of large data set over a cluster of nodes
  • 12. Hadoop Daemons Namenode (Master) (HDFS) SecondaryNameNode (3rd system / Slave) JobTracker (Master) (MR) DataNode (Slave) (HDFS) TaskTracker (Slave) (MR)
  • 13. Hadoop Daemons HADOOP CLUSTER Storage Layer Computation Layer Phase HADOOP DAEMONS ARCHITECTURE Map-Reduce Job Tracker Task Tracker Task tracker Task tracker MapReduce jobs are submitted on jobtracker HDFS NameNode DataNode DateNode DataNode NameNode Stores Meta-data
  • 14. task tracker task tracker task tracker task tracker data node data node data node job tracker name node data node task tracker data node Master Slaves Hadoop Cluster Map-Reduce Layer Storage Layer
  • 15. Characteristics Open-source ◦ Code can be modified according to business requirements Distributed Processing ◦ Data is processed parallely on cluster of nodes in distributed manner Fault Tolerance ◦ Failure of nodes or tasks are recovered automatically by the framework
  • 16. Characteristics Reliablity ◦ Data is reliably stored on the cluster of machine despite machine failures High Availability ◦ Data is highly available and accessible despite hardware failure Scalablility ◦ New hardware can be added to the nodes ◦ Horizontal Scalablility – new nodes can be added on the fly
  • 17. Characteristics Economic ◦ Runs on cluster of comodity hardware Easy to use ◦ No need of client to deal with distributed computing , framework takes care of all the things Data Locality ◦ Move Computation to data instead of data to computation
  • 18. Flavors Apache Cloudera MapR IBM Pivotal Connectors ◦ Almost all the databases have provided their connector with Hadoop for fast data transfer