Partitioning based multi-persistence model for multi-paradigm database

In the era of Internet of things and social media huge amount of data are generated day by day from numerous sources. The problem for each and every organization is to tackle with the volume, variety and velocity of data. This data is commonly referred to as “Big Data”. Most of the recent surveys provide the tools and techniques to deal with pace of data. In past decades traditional databases have shown tremendous growth in consistency, durability and isolation of data. But in current trends of data analytics, traditional database are fertile to deal with the volume, variety and velocity of database. In this paper, initially comparison is drawn to check the drawbacks of the existing system. To cope with the problems in existing system a partitioning-based technique is proposed and implemented using NoSQL Database. The final goal of this work is to identify the problems faced in traditional system and accordingly a partitioning based techniques are designed which will optimize the performance of query operation in terms of size and query time. To check the prediction accuracy around the regression line error estimation is used in the scenario. The measurement of speed difference between these proposed scenarios is depending on error estimation. Higher estimation is equal to higher performance. In this work the performance of MongoDB tuning technique is increased approximately by 65% as compare to the performance with default configuration of MongoDB. By using these techniques, the optimization of the system using MongoDB is increased compare to the Neo4j, Oracle 11g and traditional database. The final result shows that MongoDB performs best, followed by and Oracle 11g, Neo4j.