DescriptionIt has come to a historical point where we should reconsider storage controller’s architecture to keep pace with the explosive growth of big data and fast emergence of new storage media technologies. We introduce a storage controller with built-in intelligence, referred to as DPU for Data-storage Processing Unit. DPU manages, controls, analyzes, and classifies data at the place where they are stored. The idea is to place sufficient intelligence closest to the storage devices that are experiencing revolutionary changes with the emergence of storage class memories. Machine learning logics are a major part of DPU that learns I/O behaviors inside the storage to optimize adaptively I/O performance, data reliability, and availability for a variety of applications. By learning production I/Os, page writes are clustered before they are programmed to flash blocks, resulting in 10x reduction of flash block erasures in flash SSD, dramatically extending the life time of flash memories. DPU efficiently implements in-situ data intensive processing tasks such as string/text search and data conversions, which alleviates I/O bus pressure. Our results have shown over 10x performance improvements for simple string search and data serialization/deserialization applications compared with traditional systems.