Since the invention of the movable head disk, people have improved I/O performance by intelligent scheduling of disk accesses. We have applied these techniques to systems with large memories and potentially long disk queues. By viewing the entire buffer cache as a write buffer, we can improve disk bandwidth utilization by applying some traditional disk scheduling techniques. We have analyzed these techniques, which attempt to optimize head movement and guarantee fairness in response time, in the presence of long disk queues. We then propose two algorithms which take rotational latency into account, achieving disk bandwidth utilizations of nearly four times a simple first come first serve algorithm. One of these two algorithms, a weighted shortest total time first, is particularly applicable to a file server environment because it guarantees that all requests get to disk within a specified time window. 1.

In this paper, we present the Cello disk scheduling framework for meeting the diverse service requirements of applications. Cello employs a two-level disk scheduling architecture, consisting of a classindependent scheduler and a set of class-specific schedulers. The two levels of the framework allocate disk bandwidth at two timescales: the class-independent scheduler governs the coarse-grain allocation of bandwidth to application classes, while the class-specific schedulers control the fine-grain interleaving of requests. The two levels of the architecture separate application-independent mechanisms from application-specific scheduling policies, and thereby facilitate the co-existence of multiple class-specific schedulers. We demonstrate that Cello is suitable for next generation operating systems since: (i) it aligns the service provided with the application requirements, (ii) it protects application classes from one another, (iii) it is work-conserving and can adapt to changes in wor...

Modern disk drives read-ahead data and reorder incoming requests in a workload-dependent fashion. This improves their performance, but makes simple analytical models of them inadequate for performance prediction, capacity planning, workload balancing, and so on. To address this problem we have developed a new analytic model for disk drives that do readahead and request reordering. We did so by developing performance models of the disk drive components (queues, caches, and the disk mechanism) and a workload transformation technique for composing them. Our model includes the effects of workload-specific parameters such as request size and spatial locality. The result is capable of predicting the behavior of a variety of real-world devices to within 17% across a variety of workloads and disk drives.

Managing large amounts of storage is difficult and becoming more so as both the complexity and number of storage devices are increasing. One approach to this problem is a self-managing storage system. Since a self-managing storage system is a real-time system, it requires a model that quickly approximates the behavior of the storage device in a workload-dependent fashion. We develop such a model. Our approach to modeling storage devices is to model the individual physical components of the device, such as queues, caches, and disk mechanisms, and then compose the component models. Each component model determines its behavior from the specification of the entering workload and the lower-level device behavior. To support the lower level component model in determining its behavior, each component model creates a modified workload specification to support the manner that the physical component would modify the entering workload. Modifying the workload specification allows us, for example, to capture the altered spatial locality that occurs when queues reorder their requests. Our model predicts the device behavior in terms of response time within a relative error ranging from 2% to 30% for interesting subsets of the domain of devices and workloads. To demonstrate this, the model has been validated with synthetic traces of parallel scientific file system workloads and video-on-demand applications and traces of transaction processing applications. Our contributions to the area of performance modeling for storage devices include the following: - An infrastructure for developing a composite model. The infrastructure supports the development of more complicated devices and workloads than we have validated. - Methods to approximate the mean seek time and rotational latency of a disk mechanism using measures of workload spatial locality. - Methods to approximate the miss probability and the full- and partial- hit probabilities in an I/O system's data caches using measures of workload spatial locality. - Methods to approximate the queue delay for non-FCFS scheduling algorithms using a description of the workload arrival process. These methods can be composed to provide analytic estimation procedures for the behavior of a subset of current storage devices.

Integration—supporting multiple application classes with heterogeneous requirements—is an emerging trend in networks, file systems, and operating systems. In this paper, we evaluate two architectural alternatives—partitioned and integrated—for designing next generation file systems. Whereas a partitioned server employs a separate component file system for each application class, an integrated file server shares its resources across all application classes. We evaluate the performance of these two architectures with respect to sharing of disk bandwidth. We show that although the problem of sharing disk bandwidth in integrated file systems is conceptually similar to that of sharing network link bandwidth in integrated services networks, the arguments that demonstrate the superiority of integrated services networks are not applicable to file systems. Furthermore, we show that: (i) an integrated server outperforms the partitioned server in a large operating region and has slightly worse performance in the remaining region, (ii) the capacity of an integrated server is larger than that of the partitioned server, and (iii) an integrated server outperforms the partitioned server by up to a factor of 6 in the presence of bursty workloads.

. In this paper we study three popular on-line disk scheduling algorithms, FCFS, SSTF, and LOOK, using competitive analysis. Our results show that, in a competitive sense, the performance of LOOK is better than those of SSTF and FCFS. As a by-product, our analysis also reveals quantitatively the role played by the size of the window, which in our model is a waiting buffer that holds a fixed number of requests waiting to be serviced next. The window, in some sense, offers the lookahead ability which is mentioned in several on-line problems. 1. Introduction Disk scheduling is a problem that is of practical importance and theoretical interest in the study of computer systems, in particular, in the areas of databases and operating systems. The goal of disk scheduling is to devise a policy for servicing disk requests in an on-line fashion so as to minimize the total disk access time. Traditionally, measuring the performance of a disk scheduling algorithm often resorts to either probabili...

this paper, we present a priority inversion disk scheduling algorithm where both priority and seek optimization are considered. The algorithm is evaluated through detailed simulation and its performance is compared with traditional scheduling algorithms. We then identified the conditions when our algorithm performs better. The simulation results show that under these conditions, we achieve close to 50% reduction in total system response time. Based on this, an adaptive disk scheduling algorithm was constructed which consistenly outperforms the traditional approach. 1 Introduction

Abstract not found

This paper focuses on evaluation of the effectiveness of optimization at various layers of the IO path, such as the file system, the device driver scheduler, and the disk drive itself. IO performance is enhanced via effective block allocation at the file system, request merging and reordering at the device driver, and additional complex request reordering at the disk drive. Our measurements show that effective combination of these optimization forms yields superior performance under specific workloads. In particular, the impact on IO performance of technological advances in modern disk drives (i.e., reduction on head positioning times and deployment of complex request scheduling) is shown. For example, if the outstanding requests in the IO subsystem can all be accommodated by the disk queue buffer then disk level request scheduling is as effective as to close any gaps in the performance between IO request schedulers at the device driver level. Even more, for disk drives with write through caches, large queue depths improve overall IO throughput and when combined with the best performing disk scheduling algorithm at the device driver level, perform comparably with an IO subsystem where disks have write-back caches. 1

His interests include operating systems, distributed systems, user interfaces, and computer-aided design. He and his students have developed several wideI).-used programs for computer-aided design, including Magic, Caesar, and Crystal. Ousterhout is now leading the development of Sprite, a network operating system for highperfon'r, an_e workstations. Ousterhout is a recipient of the ACM Grace Murray Hopper Award,