Exploiting type-awareness in a self-recovering disk

Abstract

Data recoverability in the face of partial disk errors is an important prerequisite in modern storage. We have designed and implemented a prototype disk system that automatically ensures the integrity of stored data, and transparently recovers vital data in the event of integrity violations. We show that by using pointer knowledge, effective integrity assurance can be performed inside a block-based disk with negligible performance overheads. We also show how semantics-aware replication of blocks can help improve the recoverability of data in the event of partial disk errors with small space overheads. Our evaluation results show that for normal user workloads, our disk system has a performance overhead of only 1-5% compared to traditional disks.

A survey on data security in data warehousing: Issues, challenges and opportunities

“…Another approach for correcting corrupted data consists on applying error correction codes such as Hamming codes. Data storage systems have been proposed, able to recover from data block corruption, using error correcting codes, replication and remapping of bad blocks, such as [32,38]. Other systems use these features and add encryption techniques for distributing storage [25].…”