This paper presents the TickerTAIP architecture and an evaluation of its behavior. We demonstrate the feasibility by an existence proof; describe a family of distributed algorithms for RAID parity calculation; discuss techniques for establishing request atomicity, sequencing and recovery; and provide a performance evaluation of the TickerTAIP design space in both absolute terms and by comparison to a centralized RAID implementation. We conclude that the TickerTAIP architectural approach is feasible, useful, and effective. *Princeton University, Princeton, NJ, **University of Illinois, Urbana-Champaign, IL, ***University of Wisconsin, Madison, WI Also published as Operating Systems Research Department report HPL-OSR-92-6 1

This paper introduces the FAWN—Fast Array of Wimpy Nodes—cluster architecture for providing fast, scalable, and power-efficient key-value storage. A FAWN links together a large number of tiny nodes built using embedded processors and small amounts (2–16GB) of flash memory into an ensemble capable of handling 700 queries per second per node, while consuming fewer than 6 watts of power per node. We have designed and implemented a clustered key-value storage system, FAWN-DHT, that runs atop these node. Nodes in FAWN-DHT use a specialized log-like back-end hash-based database to ensure that the system can absorb the large write workload imposed by frequent node arrivals and departures. FAWN uses a two-level cache hierarchy to ensure that imbalanced workloads cannot create hot-spots on one or a few wimpy nodes that impair the system’s ability to service queries at its guaranteed rate. Our evaluation of a small-scale FAWN cluster and several candidate FAWN node systems suggest that FAWN can be a practical approach to building large-scale storage for seek-intensive workloads. Our further analysis indicates that a FAWN cluster is cost-competitive with other approaches (e.g., DRAM, multitudes of magnetic disks, solid-state disk) to providing high query rates, while consuming 3-10x less power. Acknowledgements: We thank the members and companies of the CyLab Corporate Partners and the PDL

Multimedia information systems have emerged as an essential component of many application domains ranging from library information systems to entertainment technology. This is because these systems utilize a variety of human senses to provide an effective means of communicating information. However, most implementations of these systems (based on a workstation) cannot support a continuous display of high resolution audio and video data and suffer from fkequent disruptions and delays termed hiccups. This is due to the low U0 bandwidth of the current disk technology, the high bandwidth requirement of multimedia objects, and the large size of these objects which mquks them to be almost always disk resident. In this paper, we describe a parallel multimedia information system and the key technical ideas that enable it to support a real-time display of multimedii objects. These techniques are as follows. First, we decluster a multimedia object across several disk drives, enabling the system to utilize the aggregate bandwidth of multiple disks to retrieve an object in real-time. Second, the workload of an application is distributed evenly across the disk drives in order to maximize the processing capability of the system. To support simultaneous display of several multimedia objects for different users, we describe two altemative approaches. The first approach multitasks a disk drive among several requests while the second replicates the data and dedicates resources to each individual request. We investigate the trade-offs associated with each approach using a simulation model. Our results demonstrate the superiority of the replication approach.

The reliability measures in today’s disk drive-based storage systems focus predominantly on protecting against complete disk failures. Previous disk reliability studies have analyzed empirical data in an attempt to better understand and predict disk failure rates. Yet, very little is known about the incidence of latent sector errors i.e., errors that go undetected until the corresponding disk sectors are accessed. Our study analyzes data collected from production storage systems over 32 months across 1.53 million disks (both nearline and enterprise class). We analyze factors that impact latent sector errors, observe trends, and explore their implications on the design of reliability mechanisms in storage systems. To the best of our knowledge, this is the first study of such large scale – our sample size is at least an order of magnitude larger than previously published studies – and the first one to focus specifically on latent sector errors and their implications on the design and reliability of storage systems.

This paper presents cooperative prefetching and caching --- the use of network-wide global resources (memories, CPUs, and disks) to support prefetching and caching in the presence of hints of future demands. Cooperative prefetching and caching effectively unites disk-latency reduction techniques from three lines of research: prefetching algorithms, cluster-wide memory management, and parallel I/O. When used together, these techniques greatly increase the power of prefetching relative to a conventional (nonglobal -memory) system. We have designed and implemented PGMS, a cooperative prefetching and caching system, under the Digital Unix operating system running on a 1.28 Gb/sec Myrinetconnected cluster of DEC Alpha workstations. Our measurements and analysis show that by using available global resources, cooperative prefetching can obtain significant speedups for I/O-bound programs. For example, for a graphics rendering application, our system achieves a speedup of 4.9 over a non-prefetc...

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.

Redundancy based on a parity encoding has been proposed for insuring that disk arrays provide highly reliable data. Parity-based redundancy will tolerate many independent and dependent disk failures (shared support hardware) without on-line spare disks and many more such failures with on-line spare disks. This paper explores the design of reliable, redundant disk arrays. In the context of a 70 disk strawman array, it presents and applies analytic and simulation models for the time until data is lost. It shows how to balance requirements for high data reliability against the overhead cost of redundant data, on-line spares, and on-site repair personnel in terms of an array’s architecture, its component reliabilities, and its repair policies.

High performance applications involving large data sets require the efficient and flexible use of multiple disks. In an external memory machine with D parallel, independent disks, only one block can be accessed on each disk in one I/O step. This restriction leads to a load balancing problem that is perhaps the main inhibitor for adapting single-disk external memory algorithms to multiple disks. This paper shows that this problem can be solved efficiently using a combination of randomized placement, redundancy and an optimal scheduling algorithm. A buffer of O(D) blocks suffices to support efficient writing of arbitrary blocks if blocks are distributed uniformly at random to the disks (e.g., by hashing). If two randomly allocated copies of each block exist, N arbitrary blocks can be read within dN=De + 1 I/O steps with high probability. In addition, the redundancy can be reduced from 2 to 1 + 1=r for any integer r. These results can be used to emulate the simple and powerful "single-disk multi-head" model of external computing [1] on the physically more realistic independent disk model [33] with small constant overhead. This is faster than a lower bound for deterministic emulation [3].

This white paper describes a new project exploring the design and implementation of “self- * storage systems:” self-organizing, self-configuring, self-tuning, self-healing, self-managing systems of storage bricks. Borrowing organizational ideas from corporate structure and automation technologies from AI and control systems, we hope to dramatically reduce the administrative burden currently faced by data center administrators. Further, compositions of lower cost components can be utilized, with available resources collectively used to achieve high levels of reliability, availability, and performance. 1

We are developing a compiler and runtime support system called PASSION: Parallel And Scalable Software for Input-Output. PASSION provides software support for I/O intensive out-of-core loosely synchronous problems. This paper gives an overview of the PASSION Runtime Library and describes two of the optimizations incorporated in it, namely Data Prefetching and Data Sieving. Performance improvements provided by these optimizations on the Intel Touchstone Delta are discussed, together with an outof -core Median Filtering application. 1 Introduction There are a number of applications which deal with very large quantities of data. These applications exist in diverse areas such as large scale scientific computations, database applications, hypertext and multimedia systems, information retrieval and many other applications of the Information Age. The number of such applications and their data requirements keep increasing day by day. Consequently, it has become apparent that I/O performance ...