Prefetcher for tuning mapreduce Framework in big data

Abstract

Big Data has been extended rapidly in many fields such as information systems. In recent years, the ability to process vast data has become a key aspect in driving business decisions. The big data query processing is the most significant challenging task to retrieve required data. Several researchers performed their research on user query processing in big data with less complexity. But, the query processing accuracy was failed to be improved. During the processing of query, the performance of data prefetching was not performed in an effective manner. This leads to increase the time consumption and error rate for performing the query processing with big data. In order to solve these problems, three proposed techniques are implemented with the aim of enhancing the accuracy of big data query processing with the aid of MapReduce framework for prefetching the data in a significant manner. newlineThe proposed Bootstrap Aggregative Prefetched Data Classification based Canonical Correlative MapReduce (BAPDC-CCMR) technique is introduced to enhance the performance of user query processing on big data with higher accuracy and lower time. In the proposed BAPDC-CCMR technique includes Bootstrap Aggregative Tuned Prefetching (BATP) process and Canonical Correlative MapReduce process to efficiently process the user queries. At first, BATP is employed to classify data from big dataset. During this process, C4.5 classifier is applied to categorize the input data. These results are combined to get robust classification outcome. The classified data is stored in the prefetch memory. This leads to minimize the time for big data query processing. In the proposed BAPDC-CCMR technique, the Canonical Correlative MapReduce process is designed to retrieve the user required queried data from the prefetch memory. Here, Map function is performed for mapping the requested query to the data in the memory. newline