A Secure System for Accessing the Big Data Over the Scattered Cloud Information Center Using C-Hadoop

The growth of the data worldwide is extremely fast and the data growth statistics and predictions are really worth consideration when infrastructure, storage and retrieval are involved. Increasing amount of data is set to reach 175 zettabytes by 2025 and also increasing 51% of this data will exist in data centers while the remainder of 49% is expected to be stored on the public cloud. Sadly, there is a split in the predictive formats, and it states 80% of it will remain in unstructured format. Ultimately, storing and retrieving such a big size of data is not possible without the MapReduce concept. The MapReduce (MR) model which works on big data is best suited to process existing medical data and fine tune the prediction systems. This work focuses on securely accessing the medical data (in this case diabetes data) over the cloud and the data are structured followed by feature extraction and clustering using modified naïve bayes classifier (MNBC) for building a better prediction system. Since this method utilized the MR model for processing, the final classification is done and verified against using k-fold validation techniques.