Leveraging Storage Semantics to Enhance Data Security and Privacy

Data within a system travels through an I/O path from its generation in an application to its final storage on a device. Ensuring data security and privacy is a significant design concern, but heavily modulated storage stacks complicate understanding the data, thus presenting challenges to maintaining these properties. For example, the firmware in a storage device cannot interpret the semantics of an I/O request from the host, making it challenging to employ a semantic-aware malware defense in the storage device. Additionally, the evolution of storage media can weaken data privacy protection guarantees due to varying physical characteristics. Preserving the guarantee of data security and privacy requires understanding storage semantics, which provides insights into the data content and the architectural components within the storage system.