Acceleration of HadoopMapReduce using in-memory Computing

摘要

In now a day, the data in real time is increasing exponentially. This data is generating from every corner of the earth viz., social networks, sensors mainly from IoT (trending technology), e-commerce site, GPS signals etc. This data may be in form of structured, semi -structured or unstructured. Currently, tech companies, for example, Facebook, Amazon, Twitter, You Tube and Google handle big data sets around terabytes or petabytes of data per day. Therefore, this data is to be analyzed or processed. It is not easy to process the whole data. The first solution to such big data problems is Hadoop. Hadoop uses Hadoop Distributed File System (HDFS) to store the data. Though Hadoop gives solution to big data problems, it takes more time to produce the results. The most important constraint in this 21st century is time. In this paper, acceleration of Hadoop using Apache Ignite Filesystem that acts in in-memory instead of HDFS.