This article provides a comprehensive overview of disk arrays and their role in high-performance, reliable secondary storage. Disk arrays, which use multiple disks to improve performance and reliability, have become widely used in modern computing. The article discusses the driving forces behind their popularity: performance and reliability. It introduces two key architectural techniques: striping across multiple disks to improve performance and redundancy to improve reliability. The article then describes seven RAID levels (0–6), comparing their performance, cost, and reliability. It also covers advanced topics such as improving RAID level 5 write performance, declustered parity, and data stripping in disk arrays. The article discusses six disk array prototypes and future research opportunities, including annotated bibliographies. The article begins by explaining the importance of disk arrays in the context of increasing processor speeds and the need for higher-performance secondary storage. It discusses the challenges of disk arrays, including the vulnerability to disk failures and the trade-offs between reliability, performance, and cost. The article then presents the basic concepts of data striping and redundancy, and describes the seven RAID organizations, their advantages and disadvantages, and their reliability, performance, and cost. It also discusses optimizations and variations to these organizations, as well as existing research in modeling disk arrays and future research directions. The article also addresses the reliability of disk arrays, considering factors such as system crashes, uncorrectable bit errors, and correlated disk failures. It explains how these factors can affect the reliability of disk arrays and how they can be mitigated. The article concludes by emphasizing the importance of understanding the trade-offs between reliability, performance, and cost when designing and implementing disk arrays. It also highlights the need for further research in this area to improve the reliability and performance of disk arrays.This article provides a comprehensive overview of disk arrays and their role in high-performance, reliable secondary storage. Disk arrays, which use multiple disks to improve performance and reliability, have become widely used in modern computing. The article discusses the driving forces behind their popularity: performance and reliability. It introduces two key architectural techniques: striping across multiple disks to improve performance and redundancy to improve reliability. The article then describes seven RAID levels (0–6), comparing their performance, cost, and reliability. It also covers advanced topics such as improving RAID level 5 write performance, declustered parity, and data stripping in disk arrays. The article discusses six disk array prototypes and future research opportunities, including annotated bibliographies. The article begins by explaining the importance of disk arrays in the context of increasing processor speeds and the need for higher-performance secondary storage. It discusses the challenges of disk arrays, including the vulnerability to disk failures and the trade-offs between reliability, performance, and cost. The article then presents the basic concepts of data striping and redundancy, and describes the seven RAID organizations, their advantages and disadvantages, and their reliability, performance, and cost. It also discusses optimizations and variations to these organizations, as well as existing research in modeling disk arrays and future research directions. The article also addresses the reliability of disk arrays, considering factors such as system crashes, uncorrectable bit errors, and correlated disk failures. It explains how these factors can affect the reliability of disk arrays and how they can be mitigated. The article concludes by emphasizing the importance of understanding the trade-offs between reliability, performance, and cost when designing and implementing disk arrays. It also highlights the need for further research in this area to improve the reliability and performance of disk arrays.