A REVIEW ON HADOOP BASED DATA CLUSTERING

Abstract

Clustering problems are becoming more challenging as the quantity of data and unstructured data handling complexities is increasing. Such big amount and complex data that cannot handle by todays DBMS tools is known as Big Data. Processing time for data is directly proportional to increase in data. To access (Read/write) big data efficiently data clustering is a good solution. Because of it differentiates between dissimilar data. Routine data clustering algorithms are considered as NP-Hard problem. To deal with such situation huge research is going on with respect to parallelization of resources and algorithms.

However, parallelization gives the worst results when data is dependent on each other and dedicated setup is required. Alternative to Parallel architecture is a distributed environment and that can utilize remote system processing capacity on the fly. Hadoop as a distributed environment and framework runs the same code on partitioned data, and finally gathers the result at one place. Hence, to solve the data clustering problems distributed environment is very helpful. Data classification or data distribution in distinct classes has an obligation to reduce the dependency. The availability of similar data at a single point from a big data is known as Data Clustering. To solve the data-clustering problem of big data distributing environment has proposed. This dissertation report gives application survey and research scope to inspire researcher to solve Big data-clustering problems using distributed computing architecture.