Disk arrays were proposed in the 1980s as a way to use parallelism between multiple disks to improve aggregate I/O performance. Today they appear in the product lines of most major computer manufacturers. This paper gives a comprehensive overview of disk arrays and provides a framework in which to organize current and future work. The paper first introduces disk technology and reviews the driving forces that have popularized disk arrays: performance and reliability. It then discusses the two architectural techniques used in disk arrays: striping across multiple disks to improve performance and redundancy to improve reliability. Next, the paper describes seven disk array architectures, called RAID (Redundant Arrays of Inexpensive Disks) levels 0-6 and compares their performance, cost, and reliability. It goes on to discuss advanced research and implementation topics such as refining the basic RAID levels to improve performance and designing algorithms to maintain data consistency. Last, the paper describes six disk array prototypes or products and discusses future opportunities for research. The paper includes an annotated bibliography of disk array-related literature.

Parallel disk systems provide opportunities for exploiting I/O parallelism in two possible ways, namely via inter-request and intra-request parallelism. In this paper we discuss the main issues in performance tuning of such systems, namely striping and load balancing, and show their relationship to response time and throughput. We outline the main components of an intelligent file system that optimizes striping by taking into account the requirements of the applications, and performs load balancing by judicious file allocation and dynamic redistributions of the data when access patterns change. Our system uses simple but effective heuristics that incur only little overhead. We present performance experiments based on synthetic workloads and real-life traces.

Redundant disk arrays are an increasingly popular way to improve I/O system performance. Past research has studied how to stripe data in non-redundant (RAID Level 0) disk arrays, but none has yet been done on how to stripe data in redundant disk arrays such as RAID Level 5, or on how the choice of striping unit varies with the number of disks. Using synthetic workloads, we derive simple design rules for striping data in RAID Level 5 disk arrays given varying amounts of workload information. We then validate the synthetically derived design rules using real workload traces to show that the design rules apply well to real systems. We find no difference in the optimal striping units for RAID Level 0 and 5 for read-intensive workloads. For write-intensive workloads, in contrast, the overhead of maintaining parity causes full-stripe writes (writes that span the entire error-correction group) to be more efficient than read-modify writes or reconstruct writes. This additional factor causes the optimal striping unit for RAID Level 5 to be four times smaller for write-intensive workloads than for read-intensive workloads. We next investigate how the optimal striping unit varies with the number of disks in an array. We find that the optimal striping unit for reads in a RAID Level 5 varies inversely to the number of disks, but that the optimal striping unit for writes varies with the number of disks. Overall, we find that the optimal striping unit for workloads with an unspecified mix of reads and writes is independent of the number of disks. Together, these trends lead us to recommend (in the absence of specific workload information) that the striping unit over a wide range of RAID Level 5 disk array sizes be equal to 1/2 * average positioning time * disk transfer rate.

This paper reports on results and experiences from the COMFORT automatic tuning project. The objective of the project has been to investigate architectural principles of self-tuning database and transaction processing systems, and to develop self-tuning methods for specific performance tuning problems. A particular concern of the project has been to cope with workload dynamics and workload heterogeneity in multi-user systems. As a general guideline, an adaptive feedback control approach has been adopted, where observations of the current load characteristics are used to predict performance trends and to drive the dynamic adjustment of tuning parameters. As examples of these general principles, the paper discusses adaptive approaches to two specific tuning problems and the developed solutions. First, we present a self-tuning load control method that copes with overload caused by excessive lock conflicts that may occur during load surges. This conflict-driven load control method adapts t...

Large arrays of small disks are providing an attractive approach for high performance I/O systems. In order to make effective use of disk arrays and other multi-disk architectures, it is necessary to develop intelligent software tools that allow automatic tuning of the disk arrays to varying workloads. In this paper we describe an adaptive method for data allocation and load balancing in disk arrays. Our method deals with dynamically changing access frequencies of files by reallocating file extents, thus "cooling down" hot disks. In addition, the method takes into account the fact that some files may exhibit periodical access patterns, and considers explicitly the cost of performing the "cooling" operations. Preliminary performance studies based on real-life I/O traces demonstrate the effectivity of this approach.

Parallel disk systems provide opportunities for high performance I/O by supporting efficiently intra-request and inter-request parallelism. We review briefly the components of an intelligent file manager that performs striping on an individual file basis and achieves load balancing by judicious file allocation and dynamic redistribution of the data. The main part of the paper discusses our "disk cooling" procedure for dynamic redistribution of data which is based on reallocation of file fragments. We show that this heuristic method achieves excellent load balance in the presence of evolving access patterns. We report on two sets of experiments: a synthetic experiment which exhibits a self-similar skew in the data access patterns and a trace-based experiment where we study the impact of the file fragment size on the cooling procedure. 1 Introduction Parallel disk systems are of great importance to massively parallel computers since they are scalable and they can ensure that I/O is not ...

Abstract—We present a 3-competitive online algorithm for determining when to restripe a file for the case when the optimum stripe depth can take one of two possible values, depending on the application's behavior. We show that a good restriping strategy is to restripe a file when the cost incurred by delaying the restriping process equals the cost of restriping itself. We present an intuitive proof that the cost of this online algorithm is at most 3 times the cost of the optimum offline algorithm. Index Terms—disk striping, restriping, storage systems. I.