A big data set consists of a large collection of data that has a wide variety of data characteristics, has a high growth rate and has many complex characteristics. An infrastructure capable of managing large volumes of real-time data is needed to process complex data, which is unstructured. The computation of big data will therefore require the use of MapReduce, which simplifies the process. In Hadoop Distributed File System (HDFS), MapReduce is used to compute data sets. MapReduce algorithms are adaptable to various types of transformation. It's possible to review different architectures of the MapReduce method with Apache Flink. This final project utilizes unstructured text data for data management. Implementing MapReduce on Linux operating systems and designing HDFS applications. In Apache Flink, the MapReduce program is used to count words. The results of this study show that Flink MapReduce is faster at performing computation than Hadoop MapReduce by 38%.