Optimizing Data Migration and Classification Accuracy in Parallel Hybrid Storage Systems Using Efficient Data Mining Techniques

A hybrid parallel storage system is a system which has a hierarchy of storage; each is a primary or secondary storage device, which has significantly different rates of reading the data. The more one goes up the storage hierarchy, greater will be the speed of data access. Moving critical application data to higher levels can greatly lessen application I/O wait time. In general, it assists in cutting down the total time consumption of the app to finish a process. This study presents a prototype of a two-level parallel hybrid storage system with SSD and HDD. The proposed system uses data mining methods to systematically evaluate and classify the application data. While the application runs and requests data, the solution at the same time works to reactively predict what data will be useful next. Data blocks that are classified as high priority are automatically elevated to SSD within the same hybrid cloud storage disk group. With the help of simulation, the data migration strategy backed up by data mining will improve user productivity. The application execution time decreases significantly if any of the application’s data access trace is particular and over 57.21% is used. New hybrid storage architectures paired with predictive data management help applications run faster than ever before.