Bruce K. Hillyer's research while affiliated with Bell and Howell and other places

Publications (24)

Conference Paper
Engineering diagnosis often involves analyzing complex records of system states printed to large, textual log files. Typically the logs are designed to accommodate the widest debugging needs without rigorous plans on formatting. As a result, critical quantities and flags are mixed with less important messages in a loose structure. Once the system i...
Article
The design and analysis of control strategies for high-capacity, reconfigurable optical transmission systems require an understanding of optical system dynamics involving the time-dependent interaction of many components. This paper describes system simulation software that couples continuous physical-layer models of optical transmission components...
Conference Paper
In this paper we present StarFish, a highly-available geographically-dispersed block storage system built from commodity servers running FreeBSD, which are connected by standard high-speed IP networking gear. StarFish achieves high availability by transparently replicating data over multiple storage sites. StarFish is accessed via a host-site appli...
Article
By "performance analysis of a storage system," we mean the application of a variety of approaches to predict, assess, evaluate, and explain the system's performance characteristics, along dimensions such as throughput, latency, and bandwidth. Several approaches are commonly used. One approach is analytical modeling, which is the writing of equation...
Conference Paper
Thermal instabilities of magnetic recordings pose a limit to the areal density that can be achieved with currently popular longitudinal magnetic recording. Perpendicular recording is expected to make possible higher areal densities than the current longitudinal ...
Article
It is widely believed that the IDE disks found in PCs are inexpensive but slow, whereas the SCSI disks used in servers and workstations are faster, more reliable, and more manageable. The belief that current IDE disks have performance and reliability disadvantages has been called into question by several recent reports. Thus we consider the possibi...
Article
The viability of storage outsourcing is critically dependent on the access performance of remote storage. We study this issue by measuring the behavior of a broad variety of I/O-intensive benchmarks as they access remote storage over an IP network. We measure the effect of network latencies that correspond to distances ranging from a local neighbor...
Article
The viability of storage outsourcing is critically dependent on the access performance of remote storage. We study this issue by measuring the behavior of a broad variety of I/O-intensive benchmarks as they access remote storage over an IP network. We measure the effect of network latencies that correspond to distances ranging from a local neighbor...
Article
Full-text available
We present a new technique for simulating fair coin flips using a biased, stationary source of randomness. Sequences of random numbers are of pervasive importance in cryptography and vital to many other computing applications. Many sources of randomness, such as radioactive or quantum mechanical sources, possess the property of stationarity. In oth...
Conference Paper
Full-text available
By “performance analysis of a storage system,” we mean the application of a variety of approaches to predict, assess, evaluate, and explain the system’s performance characteristics, along dimensions such as throughput, latency, and bandwidth. Several approaches are commonly used. One approach is analytical modeling, which is the writing of equation...
Article
An increasing number of database applications require on-line access to massive amounts of data. Since large-scale storage systems implemented entirely on magnetic disk can be impractical or too costly for many applications, tape jukeboxes can provide an attractive solution. Unfortunately, current implementations of tape jukeboxes deliver poor perf...
Article
New applications require online access to many terabytes of data, but a magnetic disk storage system this large requires thousands of drives. Magnetic tape is be a good alternative, except that the application demand for transparent data retrieval is not met by current tape systems because of their high access latency. This latency can be significa...
Article
Full-text available
We suggest a practical and economical way to generate random bits using a computer disk drive as a source of randomness. It requires no additional hardware (given a system with a disk), and no user involvement. As a concrete example of performance, on a Sun Ultra-1 with a Seagate Cheetah disk, it generates bits at a rate of either 5 bits per minute...
Article
Full-text available
In modern I/O architectures, multiple disk drives are attached to each I/O bus. Under I/O-intensive workloads, the disk latency for a request can be overlapped with the disk latency and data transfers of requests to other disks, potentially resulting in an aggregate I/O throughput at nearly bus bandwidth. This paper reports on a performance impairm...
Article
In modern I/O architectures, multiple disk drives are attached to each I/O controller. A study of the performance of such architectures under I/O-intensive workloads has revealed a performance impairment that results from a previously unknown form of convoy behavior in disk I/O. In this paper, we describe measurements of the read performance of mul...
Conference Paper
Full-text available
An increasing number of database applications require online access to massive amounts of data. Since large scale storage systems implemented entirely on magnetic disk can be impractical or too costly for many applications, tape jukeboxes can provide an attractive solution. The paper shows how the performance of tape jukeboxes can be improved acros...
Article
Full-text available
For a wide variety of computational tasks, disk I/O continues to be a serious obstacle to high performance. The focus of the present paper is on systems that use multiple disks per SCSI bus. We measured the performance of concurrent random I/Os, and observed bus-related phenomena that impair performance. We describe these phenomena, and present a n...
Article
Large data installations normally archive relatively inactive data to a near-line tape library. The tape library performs reasonably well for sequential-access retrieval workloads.
Conference Paper
New database applications that require the storage and retrieval of many terabytes of data are reaching the limits for disk-based storage systems, in terms of both cost and scalability. These limits provide a strong incentive for the development of databases that augment disk storage with technologies better suited to large volumes of data. In part...
Article
New applications require online access to many terabytes of data, but a magnetic disk storage system this large requires thousands of drives. Magnetic tape is be a good alternative, except that the application demand for transparent data retrieval is not met by current tape systems because of their high access latency. This latency can be significa...

Citations

... Contrarily to this paper, most studies consider a more complex tape geometry, usually a serpentine. Hillyer and Silberschatz [15] focus on low-level hardware information (key points) to evaluate several heuristics. Sandsta and Midtstraum [24] propose a low-cost function to approximate the seeking time between two points of the tape. ...
... For example, MEMS devices can interact with existing system [4]. While disk arrays are widely used in the high-end storage systems, a lot of research work focuses on the modeling and simulation of disk arrays567. Among the work, the Pantheon storage system simulator [7, 8] was built to support the rapid exploration of design choices of storage systems in HP AutoRAID advanced disk array technology [9] and TicherTAIP parallel RAID architecture [10]. ...
... In order to reduce disk I/O overheads, several techniques and optimizations have been proposed. This literature includes work on optimizing the scheduling of disk requests [33] [3] [17] [15], disk arrays [10] [4] [34], and optimizing disk writes using logs [23] [13] [32]. Despite these techniques and optimizations, low disk throughput is still a serious problem for data-intensive servers, such as Web proxies, email and news servers, multimedia servers, and database servers. ...
... Previous studies have recognized the importance of constructing I/O system performance models. Various analytical and simulation models have been constructed for disk drives [5] [16] [25] [28] [36], disk arrays [8] [33], OS prefetching [6] [29] [31], and memory caching [15]. However, performance models for individual system components do not capture the inter-dependence of different components and consequently they may not accurately predict the overall application performance. ...
... Outside the realm of disk-based storage, simulation models of hypothetical MEMS-based storage devices were developed by Griffin et al. [73] and Madhyastha and Yang [112]. Various analytic models and simulation models for tape systems are presented by Drapeau and Katz [49], Golubchik, Muntz, and Watson [71], and Hillyer and Silberschatz [82]. ...
... iSCSI [13], iFCP [9], and NBD [8] have been used as a protocol for building IP-SANs. IP-SAN protocols realize connectivity to remote storage devices over a conventional TCP/IP network, but they still have several issues to be solved-in particular, performance issues [7,10]. IP-SAN protocols generally utilize TCP (Transmission Control Protocol) for data delivery, which results in low performance in a longfat network. ...
... Moreover, as part of a huge infrastructure, often the more storage that is used, the cheaper per unit it becomes. Storage may be better managed and cost-effective when handled as an outsourced service, which has contractual properties, instead of capital-intensive locally managed storage [34]. This is significant in three ways: (1) handling large storage requires expertise. ...
... This algorithm ensures fairness and avoids starvation. The system's performance can be further improved by appropriately scheduling batches of requests in cartridges to minimize seek times [12], [13]. This issue, however, is beyond the scope of this article. ...
... Une solution pour résoudre certains problèmes posés par cette approche est d'enrichir le modèle de la machine pour prendre plus d'éléments en compte. De nombreux travaux dans la littérature présentent des modèles analytiques précis pour le système de cache [WSMW00], la hiérarchie mémoire [JCSM96], les performances des entrées/sorties [SHS00] ou même un ordinateur complet [AE98,SS96]. ...
... As the primary site and backup site usually disperse in geographically distant locations, how to transmit the data correctly and fast to the backup site is the key issue of remote mirroring. This paper mainly focuses on the data consistency of remote mirroring [5], [6], which ensure the integrity and accuracy of data. ... Seneca [2] uses a log barrier technology to ensure data consistency, however it needs a complex protocol. ...