DBSeer, a software system developed by MIT researchers, can improve cloud computing performance as well as reduce hardware requirements by 95%. It makes use of techniques in order to create more accurate models of resource demands and performance in database-driven applications. This is basically the content of the research presented at the Biennial Conference on Innovative Data Systems Research.

Lead author Barza Mozafari said that when there are so many user requests, the system tends to slow down. Furthermore, the high speed memory of database server will be used to store data due to the increased demand. DBSeer will be able to monitor any chances in system performance due to increased user requests. Machine-learning techniques will be able predict the outcome of such changes. Using the “gray box” model, the algorithm is also able to take into consideration the oddities of database systems in order to meet the increased demands.

IDC, on the other hand, released a report last year which predicted 41% cloud services growth for the next 4 years. By 2016, IT cloud services spending will be $100 billion and that from 2012 to 2016 the compound growth rate annually will be 26%.