UW-Madison Logo

The ADvanced Systems Laboratory (ADSL)

Semantically-Smart Disk Systems

One rich off-shoot of the gray-box work occurred when we had the following insight. The disk and RAID industry, for years, had been hampered for years by limited information: at the block level, all a device sees are reads and writes, with no knowledge as to the contents of each block, its importance, and other relevant characteristics. If a ``dumb'' disk could just know more about how such blocks are utilized, it could implement a surprising range of features to improve performance and reliability of the I/O system. As our work demonstrates, gray-box techniques enable such knowledge. By assuming certain details about the client of the disk (e.g., that it is a file system of a particular type), a semantically-smart disk can learn how it is being used and thus implement features previously unattainable at block level.
  • Semantically-Smart Disk Systems (FAST '03) The original paper, which introduces the idea, and shows its benefits through three case studies.
  • Improving Storage System Availability with D-GRAID (FAST '04). We show how knowledge of files and directories can lead to novel new approaches to data layout in RAIDs, which can greatly improve availability of the file system by exploiting such knowledge.
  • X-RAY: A Non-Invasive Exclusive Caching Mechanism for RAIDs (ISCA '04). We show how a RAID layer can determine what is likely cached in the OS above it, and thus avoid redundant caching, to improve performance.
  • Life or Death at Block Level (OSDI '04) Probably one of our deepest papers upon this topic, we show how to determine whether a block is live or dead, and exploiting such knowledge to build interesting new functionality such as a secure deletion facility, all at block level.
  • A Logic of File Systems (FAST '05) A formal framework to reason about semantic disks as well as file systems that arose from our empirical work, as we realized the need for a conceptual theory underpinning our previously heuristic approaches.
  • Database-Aware Semantically-Smart Storage (FAST '05) A final paper on the topic that highlighted that smarter disk systems could also be realized underneath database management systems (not just file systems), and that similar performance benefits could be achieved.