Traffic Analysis in MapReduce

-MapReduce is a programming model, which can process the large set of data and produces the output. The MapReduce contains two functions to complete the work, those are Map function and Reduce function. The Map function will get assign fragmented data as input and then its emit intermediate data with key and send to this intermediate data with key to the Reducer, where Reducer will get the input from the Map function and then it compute the final output. In-between MapReduce the shuffle is there, which can be sort and shuffle the mapper inter mediate data and then it sends to the Reducer. In the Shuffle phase the large amount of traffic would occur and this can leads to even more traffic in the Reducer phase, because of this time and cost also increased, if one or two Mapper and Reducer is processing then it is ok, what if, the thousands of Map and Reducer are processing the job, no one has considered the traffic between the MapReduce, everyone is bothered only about load optimization and load balancing in Big Data, so in order to avoid the traffic in between the MapReduce in this we placed the Aggregator and Check function, in which checks the data come from the shuffle phase and then it removes the user irrelevant data and then reduces the data volume to process further and then sent to the Reducers, by placing the aggregator can reduces the traffic, cost and time and also it reduces the number of Reduce function to process the remaining task. This can be achieved by using the MapReduce distributed algorithm, and finally will show the performance analysis of the traffic reduction in the MapReduce.