1.
Basics of Big Data
Analytics &
Hadoop
Ambuj Kumar
Ambuj_kumar@aol.com
http://ambuj4bigdata.blogspot.in
http://ambujworld.wordpress.com

2.
Agenda
 Big Data –
 Concepts overview
 Analytics –
 Concepts overview
 Hadoop –
 Concepts overview
 HDFS
 Concepts overview
 Data Flow - Read & Write Operation
 MapReduce
 Concepts overview
 WordCount Program
 Use Cases
 Landscape
 Hadoop Features & Summary

3.
What is Big Data?
Big data is data which is too large, complex and dynamic for any conventional data tools to capture,
store, manage and analyze.

4.
Challenges of Big Data
• Storage (~ Petabytes)1
• Processing (Timely manner)2
• Variety of Data (Structured,Semi
Structured,Un-structured)3
• Cost4

5.
Big Data Analytics
Big data analytics is the process of examining large amounts of
data of a variety of types (big data) to uncover hidden patterns,
unknown correlations and other useful information.
Big Data Analytics Solutions
There are many different Big Data Analytics Solutions out in the
market.
 Tableau – visualization tools
 SAS – Statistical computing
 IBM and Oracle –They have a range of tools for Big DataAnalysis
 Revolution – Statistical computing
 R – Open source tool for Statistical computing

6.
What is Hadoop?
 Open-source data storage and processing API
 Massively scalable, automatically parallelizable
 Based on work from Google
 GFS + MapReduce + BigTable
 Current Distributions based on Open Source and Vendor Work
 Apache Hadoop
 Cloudera – CDH4
 Hortonworks
 MapR
 AWS
 Windows Azure HDInsight

7.
Why Use Hadoop?
Cheaper
Scales to Petabytes
or more
Faster
Parallel data
processing
Better
Suited for particular
types of BigData
problems

8.
Hadoop History
In 2008, Hadoop became Apache Top Level Project

9.
Comparing: RDBMS vs. Hadoop
Traditional RDBMS Hadoop / MapReduce
Data Size Gigabytes (Terabytes) Petabytes (Hexabytes)
Access Interactive and Batch Batch – NOT Interactive
Updates Read / Write many times Write once, Read many times
Structure Static Schema Dynamic Schema
Integrity High (ACID) Low
Scaling Nonlinear Linear
Query
Response Time
Can be near immediate Has latency (due to batch
processing)

10.
Where is Hadoop used?
Industry
Technology
Use Cases
Search
People you may know
Movie recommendations
Banks
Fraud Detection
Regulatory
Risk management
Media
Retail
Marketing analytics
Customer service
Product recommendations
Manufacturing Preventive maintenance

11.
Companies Using Hadoop
 Search
Yahoo,Amazon, Zvents
 Log Processing
Facebook,Yahoo,
ContextWeb.Joost,Last.fm
 Recommendation Systems
Facebook,Linkedin
 DataWarehouse
Facebook,AOL
 Video & Image Analysis
NewYorkTimes,Eyealike
------- Almost in every domain!

12.
Hadoop is a set of Apache
Frameworks and more…
 Data storage (HDFS)
 Runs on commodity hardware (usually Linux)
 Horizontally scalable
 Processing (MapReduce)
 Parallelized (scalable) processing
 Fault Tolerant
 Other Tools / Frameworks
 Data Access
 HBase, Hive, Pig, Mahout
 Tools
 Hue, Sqoop
 Monitoring
 Greenplum, Cloudera
Hadoop Core - HDFS
MapReduce API
Data Access
Tools & Libraries
Monitoring & Alerting

13.
Core parts of Hadoop distribution
HDFS Storage
Redundant (3 copies)
For large files – large
blocks
64 or 128 MB / block
Can scale to 1000s of
nodes
MapReduce API
Batch (Job) processing
Distributed and Localized
to clusters (Map)
Auto-Parallelizable for
huge amounts of data
Fault-tolerant (auto
retries)
Adds high availability and
more
Other Libraries
Pig
Hive
HBase
Others

14.
Hadoop Cluster HDFS (Physical)
Storage
Name Node
Data Node 1 Data Node 2 Data Node 3
Secondary
Name Node
• Contains web site to view
cluster information
• V2 Hadoop uses multiple
Name Nodes for HA
One Name Node
• 3 copies of each node by
default
Many Data Nodes
• Using common Linux shell
commands
• Block size is 64 or 128 MB
Work with data in HDFS

15.
MapReduce Job – LogicalView

16.
Hadoop Ecosystem

17.
Common Hadoop Distributions
Open Source
Apache
Commercial
Cloudera
Hortonworks
MapR
AWS MapReduce
Microsoft HDInsight

18.
HDFS :Architecture
Master
NameNode
Slave
Bunch of DataNodes
HDFS Layers
NameNode
Storage
…………
NS
Block Management
NameNode
DataNode
DataNode DataNode DataNode DataNode DataNode
DataNode
Name
Space
Block
Storage

19.
HDFS : Basic Features
Highly fault-tolerant
High throughput
Suitable for applications with large data sets
Streaming access to file system data
Can be built out of commodity hardware

20.
HDFS Write (1/2)
Client Name Node
1
2
Data Node
A
Data Node
B
Data Node
C
Data Node
D
A2 A3 A4A1
3
Client contacts NameNode to write data
NameNode says write it to these nodes
Client sequentially writes
blocks to DataNode

21.
HDFS Write (2/2)
Client Name Node
Data Node
A
Data Node
B
Data Node
C
Data Node
D
A1
DataNodes replicate data
blocks, orchestrated
by the NameNode
A2
A4
A2 A1
A3
A3 A2
A4
A4 A1
A3

22.
HDFS Read
Client Name Node
1
2
Data Node
A
Data Node
B
Data Node
C
Data Node
D
A1
3
Client contacts NameNode to read data
NameNode says you can find it here
Client sequentially
reads blocks from
DataNode
A2
A4
A2 A1
A3
A3 A2
A4
A4 A1
A3

23.
HA (High Availability) for
NameNode
NameNode (StandBy)
DataNode
NameNode (Active)
Active NameNode
Do normal namenode’s operation
Standby NameNode
Maintain NameNode’s data
Ready to be active NameNode
DataNode DataNode DataNode DataNode

24.
MapReduce
 MapReduce job consist of two tasks
 Map Task
 Reduce Task
 Blocks of data distributed across several machines are
processed by map tasks parallel
 Results are aggregated in the reducer
 Works only on KEY/VALUE pair

25.
MapReduce:Word Count
Deer 1
Bear 1
River 1
Car 1
Car 1
River 1
Deer 1
Car 1
Bear 1
Bear 2
Car 3
Deer 2
River 2
Can we do word count in parallel?
Deer Bear River
Car Car River
Deer Car Bear

26.
MapReduce:Word Count Program

27.
Data Flow in a MapReduce
Program in Hadoop

28.
Mapper Class
Package ambuj.com.wc;
import java.io.IOException;
import java.util.StringTokenizer;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Mapper;
public class WordCountMapper extends
Mapper<LongWritable, Text, Text, LongWritable> {
private final static LongWritable one = new LongWritable(1);
private Text word = new Text();
@Override
public void map(LongWritable inputKey, Text inputVal, Context context)
throws IOException, InterruptedException {
String line = inputVal.toString();
StringTokenizer tokenizer = new StringTokenizer(line);
while (tokenizer.hasMoreTokens()) {
word.set(tokenizer.nextToken());
context.write(word, one);
}
}
}

29.
Reducer Class
package ambuj.com.wc;
import java.io.IOException;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Reducer;
public class WordCountReducer extends
Reducer<Text, LongWritable, Text, LongWritable> {
@Override
public void reduce(Text key, Iterable<LongWritable> listOfValues,
Context context) throws IOException, InterruptedException {
long sum = 0;
for (LongWritable val : listOfValues) {
sum = sum + val.get();
}
context.write(key, new LongWritable(sum));
}
}

30.
Driver Class
package ambuj.com.wc;
import org.apache.hadoop.conf.Configuration;
import org.apache.hadoop.conf.Configured;
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.io.LongWritable;
import org.apache.hadoop.io.Text;
import org.apache.hadoop.mapreduce.Job;
import org.apache.hadoop.mapreduce.lib.input.FileInputFormat;
import org.apache.hadoop.mapreduce.lib.input.TextInputFormat;
import org.apache.hadoop.mapreduce.lib.output.FileOutputFormat;
import org.apache.hadoop.util.Tool;
import org.apache.hadoop.util.ToolRunner;
public class WordCountDriver extends Configured implements Tool {
@Override
public int run(String[] args) throws Exception {
Configuration conf = new Configuration();
Job job = new Job(conf, "WordCount");
job.setJarByClass(WordCountDriver.class);
job.setMapperClass(WordCountMapper.class);
job.setReducerClass(WordCountReducer.class);
job.setInputFormatClass(TextInputFormat.class);
job.setMapOutputKeyClass(Text.class);
job.setMapOutputValueClass(LongWritable.class);
job.setOutputKeyClass(Text.class);
job.setOutputValueClass(LongWritable.class);
FileInputFormat.addInputPath(job, new Path(args[0]));
FileOutputFormat.setOutputPath(job, new Path(args[1]));
job.waitForCompletion(true);
return 0;
}
public static void main(String[] args) throws Exception {
ToolRunner.run(new WordCountDriver(), args);
}
}

31.
A view of Hadoop
Client Job
Data Node
Task
Tracker
Task
Task
Task
Job Tracker Name Node
Data Node
Task
Tracker
Task
Task
Task
Data Node
Task
Tracker
Task
Task
Task
MasterSlave
Blocks HDFS
MapReduce

32.
Use Cases
 Utilities want to predict power consumption
 Banks and insurance companies want to understand risk
 Fraud detection
 Marketing departments want to understand customers
 Recommendations
 Location-Based Ad Targeting
 Threat Analysis

33.
Big Data Landscape

34.
Hadoop Features & Summary
Distributed frame work for processing and storing data
generally on commodity hardware.Completely open
source and written in Java.
 Store anything
 Unstructuredor semi structured data,
 Storage capacity
 Scale linearly, cost in not exponential.
 Data locality and process in your way.
 Code moves to data
 In MR you specify the actual steps in processing the data and drive the out put.
 Stream access: Process data in any language.
 Failure and fault tolerance:
 Detect Failure and Heals itself.
 Reliable,data replicated, failed task are rerun , no need maintain backup of data
 Cost effective: Hadoop is designed to be a scale-out architecture operating on a cluster of commodity
PC machines.
The Hadoop framework transparently for customization to provides applications both reliability, adaption
and data motion.
Primarily used for batch processing, not real-time/ transactional user applications.

35.
References - Hadoop
 Hadoop:The Definitive Guide,Third Edition by Tom
White.
 http://hadoop.apache.org
 http://www.cloudera.com
 http://ambuj4bigdata.blogspot.com
 http://ambujworld.wordpress.com

36.
ThankYou