Blog

MapReduce – Hadoop Implementation

MapReduce is a framework that is used for writing applications to process huge volumes of data on large clusters of commodity hardware in a reliable manner. This chapter takes you through the operation of MapReduce in the Hadoop framework using Java.

MapReduce Algorithm

Generally, the MapReduce paradigm is based on sending map-reduce programs to computers where the actual data resides.

  • During a MapReduce job, Hadoop sends Map and Reduce tasks to appropriate servers in the cluster.
  • The framework manages all the details of data-passing like issuing tasks, verifying task completion, and copying data around the cluster between the nodes.
  • Most of the computing takes place on the nodes with data on local disks that reduces the network traffic.
  • After completing a given task, the cluster collects and reduces the data to form an appropriate result, and sends it back to the Hadoop server.

Inputs and Outputs (Java Perspective)

The MapReduce framework operates on key-value pairs, that is, the framework views the input to the job as a set of key-value pairs and produces a set of key-value pairs as the output of the job, conceivably of different types.

The key and value classes have to be serializable by the framework and hence, it is required to implement the Writable interface. Additionally, the key classes have to implement the WritableComparable interface to facilitate sorting by the framework.

Both the input and output format of a MapReduce job is in the form of key-value pairs −

(Input) <k1, v1> -> map -> <k2, v2>-> reduce -> <k3, v3> (Output).

InputOutput
Map<k1, v1>list (<k2, v2>)
Reduce<k2, list(v2)>list (<k3, v3>)

MapReduce Implementation

The following table shows the data regarding the electrical consumption of an organization. The table includes the monthly electrical consumption and the annual average for five consecutive years.

JanFebMarAprMayJunJulAugSepOctNovDecAvg
19792323243242526262626252625
198026272828283031313130303029
198131323232333435363634343434
198439383939394142434039383840
198538393939394141410040393945

We need to write applications to process the input data in the given table to find the year of maximum usage, the year of minimum usage, and so on. This task is easy for programmers with a finite amount of records, as they will simply write the logic to produce the required output, and pass the data to the written application.

Let us now raise the scale of the input data. Assume we have to analyze the electrical consumption of all the large-scale industries of a particular state. When we write applications to process such bulk data,

  • They will take a lot of time to execute.
  • There will be heavy network traffic when we move data from the source to the network server.

To solve these problems, we have the MapReduce framework.

Input Data

The above data is saved as sample.txt and given as input. The input file looks as shown below.

19792323243242526262626252625
198026272828283031313130303029
198131323232333435363634343434
198439383939394142434039383840
198538393939394141410040393945

Example Program

The following program for the sample data uses the MapReduce framework.

package hadoop;
 
import java.util.*;
import java.io.IOException;
import java.io.IOException;
 
import org.apache.hadoop.fs.Path;
import org.apache.hadoop.conf.*;
import org.apache.hadoop.io.*;
import org.apache.hadoop.mapred.*;
import org.apache.hadoop.util.*;
 
public class ProcessUnits
{
   //Mapper class
   public static class E_EMapper extends MapReduceBase implements
   Mapper<LongWritable,  /*Input key Type */
   Text,                   /*Input value Type*/
   Text,                   /*Output key Type*/
   IntWritable>            /*Output value Type*/
   {
      //Map function
      public void map(LongWritable key, Text value, OutputCollector<Text, IntWritable> output, Reporter reporter) throws IOException
      {
         String line = value.toString();
         String lasttoken = null;
         StringTokenizer s = new StringTokenizer(line,"\t");
         String year = s.nextToken();
         
         while(s.hasMoreTokens()){
            lasttoken=s.nextToken();
         }
         
         int avgprice = Integer.parseInt(lasttoken);
         output.collect(new Text(year), new IntWritable(avgprice));
      }
   }
   
   //Reducer class
       
   public static class E_EReduce extends MapReduceBase implements
   Reducer< Text, IntWritable, Text, IntWritable >
   {
      //Reduce function
      public void reduce(Text key, Iterator <IntWritable> values, OutputCollector>Text, IntWritable> output, Reporter reporter) throws IOException
      {
         int maxavg=30;
         int val=Integer.MIN_VALUE;
         while (values.hasNext())
         {
            if((val=values.next().get())>maxavg)
            {
               output.collect(key, new IntWritable(val));
            }
         }
      }
   }
       
   //Main function
       
   public static void main(String args[])throws Exception
   {
      JobConf conf = new JobConf(Eleunits.class);
              
      conf.setJobName("max_eletricityunits");
              
      conf.setOutputKeyClass(Text.class);
      conf.setOutputValueClass(IntWritable.class);
              
      conf.setMapperClass(E_EMapper.class);
      conf.setCombinerClass(E_EReduce.class);
      conf.setReducerClass(E_EReduce.class);
              
      conf.setInputFormat(TextInputFormat.class);
      conf.setOutputFormat(TextOutputFormat.class);
              
      FileInputFormat.setInputPaths(conf, new Path(args[0]));
      FileOutputFormat.setOutputPath(conf, new Path(args[1]));
              
      JobClient.runJob(conf);
   }
}

Save the above program into ProcessUnits.java. The compilation and execution of the program are given below.

Compilation and Execution of ProcessUnits Program

Let us assume we are in the home directory of Hadoop users (e.g. /home/hadoop).

Follow the steps given below to compile and execute the above program.

Step 1 − Use the following command to create a directory to store the compiled java classes.

$ mkdir units

Step 2 − Download Hadoop-core-1.2.1.jar, which is used to compile and execute the MapReduce program. Download the jar from mvnrepository.com. Let us assume the download folder is /home/hadoop/.

Step 3 − The following commands are used to compile the ProcessUnits.java program and to create a jar for the program.

$ javac -classpath hadoop-core-1.2.1.jar -d units ProcessUnits.java
$ jar -cvf units.jar -C units/ .

Step 4 − The following command is used to create an input directory in HDFS.

$HADOOP_HOME/bin/hadoop fs -mkdir input_dir

Step 5 − The following command is used to copy the input file named sample.txt in the input directory of HDFS.

$HADOOP_HOME/bin/hadoop fs -put /home/hadoop/sample.txt input_dir

Step 6 − The following command is used to verify the files in the input directory

$HADOOP_HOME/bin/hadoop fs -ls input_dir/

Step 7 − The following command is used to run the Eleunit_max application by taking input files from the input directory.

$HADOOP_HOME/bin/hadoop jar units.jar hadoop.ProcessUnits input_dir output_dir

Wait for a while till the file gets executed. After execution, the output contains a number of input splits, Map tasks, Reducer tasks, etc.

INFO mapreduce.Job: Job job_1414748220717_0002
completed successfully
14/10/31 06:02:52
INFO mapreduce.Job: Counters: 49
 
File System Counters
   
   FILE: Number of bytes read=61
   FILE: Number of bytes written=279400
   FILE: Number of read operations=0
   FILE: Number of large read operations=0
   FILE: Number of write operations=0
 
   HDFS: Number of bytes read=546
   HDFS: Number of bytes written=40
   HDFS: Number of read operations=9
   HDFS: Number of large read operations=0
   HDFS: Number of write operations=2 Job Counters
   
   Launched map tasks=2
   Launched reduce tasks=1
   Data-local map tasks=2
       
   Total time spent by all maps in occupied slots (ms)=146137
   Total time spent by all reduces in occupied slots (ms)=441
   Total time spent by all map tasks (ms)=14613
   Total time spent by all reduce tasks (ms)=44120
       
   Total vcore-seconds taken by all map tasks=146137
   Total vcore-seconds taken by all reduce tasks=44120
       
   Total megabyte-seconds taken by all map tasks=149644288
   Total megabyte-seconds taken by all reduce tasks=45178880
 
Map-Reduce Framework
   
   Map input records=5
       
   Map output records=5
   Map output bytes=45
   Map output materialized bytes=67
       
   Input split bytes=208
   Combine input records=5
   Combine output records=5
       
   Reduce input groups=5
   Reduce shuffle bytes=6
   Reduce input records=5
   Reduce output records=5
       
   Spilled Records=10
   Shuffled Maps =2
   Failed Shuffles=0
   Merged Map outputs=2
       
   GC time elapsed (ms)=948
   CPU time spent (ms)=5160
       
   Physical memory (bytes) snapshot=47749120
   Virtual memory (bytes) snapshot=2899349504
       
   Total committed heap usage (bytes)=277684224
 
File Output Format Counters
 
   Bytes Written=40

Step 8 − The following command is used to verify the resultant files in the output folder.

$HADOOP_HOME/bin/hadoop fs -ls output_dir/

Step 9 − The following command is used to see the output in the Part-00000 file. This file is generated by HDFS.

$HADOOP_HOME/bin/hadoop fs -cat output_dir/part-00000

Following is the output generated by the MapReduce program −

198134
198440
198545

Step 10 − The following command is used to copy the output folder from HDFS to the local file system.

$HADOOP_HOME/bin/hadoop fs -cat output_dir/part-00000/bin/hadoop dfs -get output_dir /

Leave a Reply