共查询到20条相似文献,搜索用时 46 毫秒
1.
We describe a hybrid mirrored disk organization patented by LSI Logic Corp. and compare its performance, reliability, and performability with traditional mirrored RAID1 disk organizations and RAID(4+?), ?≥1. LSI RAID has the same level of redundancy as mirrored disks, but also utilizes parity coding. Unlike RAID1, which cannot tolerate all two disk failures, LSI RAID similarly to RAID6 is 2?Disk Failure Tolerant (2DFT), but in addition it can tolerate almost all three disk failures, while RAID1 organizations are generally 1DFT. We list analytic expressions for the reliability of various RAID1 organizations and use enumeration when the reliability expression cannot be obtained analytically. An asymptotic expansion method based on disk unreliabilities is used for an easy comparison of RAID reliabilities. LSI RAID performance is evaluated with the Read-Modify-Write (RMW) and ReConstruct Write (RCW) methods to update parities. The combination of the two methods is used to balance data and parity disk loads, which results in maximizing the I/O throughput. The analysis shows that LSI RAID has an inferior performance with respect to basic mirroring in processing an OLTP workload, but it outperforms RAID6. LSI RAID in spite of its higher Mean Time to Data Loss (MTTDL) is outperformed by other RAID1 organizations as far as its performability is concerned, i.e., the number of I/Os carried out by the disk array operating at maximum I/Os Per Second (IOPS) until data loss occurs. A survey of RAID1 organizations and distributed replicated systems is also included. 相似文献
2.
While aggregating the throughput of existing disks on cluster nodes is a cost-effective approach to alleviate the I/O bottleneck
in cluster computing, this approach suffers from potential performance degradations due to contentions for shared resources
on the same node between storage data processing and user task computation. This paper proposes to judiciously utilize the
storage redundancy in the form of mirroring existed in a RAID-10 style file system to alleviate this performance degradation.
More specifically, a heuristic scheduling algorithm is developed, motivated from the observations of a simple cluster configuration,
to spatially schedule write operations on the nodes with less load among each mirroring pair. The duplication of modified
data to the mirroring nodes is performed asynchronously in the background. The read performance is improved by two techniques:
doubling the degree of parallelism and hot-spot skipping. A synthetic benchmark is used to evaluate these algorithms in a
real cluster environment and the proposed algorithms are shown to be very effective in performance enhancement.
Yifeng Zhu received his B.Sc. degree in Electrical Engineering in 1998 from Huazhong University of Science and Technology, Wuhan, China;
the M.S. and Ph.D. degree in Computer Science from University of Nebraska – Lincoln in 2002 and 2005 respectively. He is an
assistant professor in the Electrical and Computer Engineering department at University of Maine. His main research interests
are cluster computing, grid computing, computer architecture and systems, and parallel I/O storage systems. Dr. Zhu is a Member
of ACM, IEEE, the IEEE Computer Society, and the Francis Crowe Society.
Hong Jiang received the B.Sc. degree in Computer Engineering in 1982 from Huazhong University of Science and Technology, Wuhan, China;
the M.A.Sc. degree in Computer Engineering in 1987 from the University of Toronto, Toronto, Canada; and the PhD degree in
Computer Science in 1991 from the Texas A&M University, College Station, Texas, USA. Since August 1991 he has been at the
University of Nebraska-Lincoln, Lincoln, Nebraska, USA, where he is Professor and Vice Chair in the Department of Computer
Science and Engineering. His present research interests are computer architecture, parallel/distributed computing, cluster
and Grid computing, computer storage systems and parallel I/O, performance evaluation, real-time systems, middleware, and
distributed systems for distance education. He has over 100 publications in major journals and international Conferences in
these areas and his research has been supported by NSF, DOD and the State of Nebraska. Dr. Jiang is a Member of ACM, the IEEE
Computer Society, and the ACM SIGARCH.
Xiao Qin received the BS and MS degrees in computer science from Huazhong University of Science and Technology in 1992 and 1999, respectively.
He received the PhD degree in computer science from the University of Nebraska-Lincoln in 2004. Currently, he is an assistant
professor in the department of computer science at the New Mexico Institute of Mining and Technology. He had served as a subject
area editor of IEEE Distributed System Online (2000–2001). His research interests are in parallel and distributed systems, storage systems, real-time computing, performance
evaluation, and fault-tolerance. He is a member of the IEEE.
Dan Feng received the Ph.D degree from Huazhong University of Science and Technology, Wuhan, China, in 1997. She is currently a professor
of School of Computer, Huazhong University of Science and Technology, Wuhan, China. She is the principal scientist of the
the National Grand Fundamental Research 973 Program of China “Research on the organization and key technologies of the Storage
System on the next generation Internet.” Her research interests include computer architecture, storage system, parallel I/O,
massive storage and performance evaluation.
David Swanson received a Ph.D. in physical (computational) chemistry at the University of Nebraska-Lincoln (UNL) in 1995, after which he
worked as an NSF-NATO postdoctoral fellow at the Technical University of Wroclaw, Poland, in 1996, and subsequently as a National
Research Council Research Associate at the Naval Research Laboratory in Washington, DC, from 1997–1998. In 1999 he returned
to UNL where he directs the Research Computing Facility and currently serves as an Assistant Research Professor in the Department
of Computer Science and Engineering. The Office of Naval Research, the National Science Foundation, and the State of Nebraska
have supported his research in areas such as large-scale scientific simulation and distributed systems. 相似文献
3.
Heterogeneous Disk Arrays (HDAs) allow resource sharing of their hardware by multiple RAID levels. RAID1 (mirrored disks) and RAID5 (distributed parity arrays)
are the two RAID levels considered in this study. They are both single disk failure tolerant (1DFT), but differ significantly in their efficiency in processing database workloads. The goal of the study is to maximize the
number of Virtual Array (VA) allocations in HDA. We develop an analysis to estimate the load per VA based on a few parameters: the fraction of accesses
to small versus large blocks and the fraction of updates versus reads. A VA is allocated according to the RAID level, which
minimizes the anticipated load based on input parameters. Operation in normal and degraded mode is considered for comparison
purposes, but in fact allocations are carried out using the higher load in degraded mode to ensure that single disk failures
will not result in overload. We report on parametric studies to gain insight into circumstances leading to a RAID1 or RAID5
classification. An allocation experiment with a synthetic workload is used to demonstrate the superiority of HDA with respect
to purely RAID1 or RAID5 disk arrays. This analytic study can be extended to 2DFT arrays, namely RAID6 versus 3-way replication. 相似文献
4.
Murali Vilayannur Anand Sivasubramaniam Mahmut Kandemir Rajeev Thakur Robert Ross 《Cluster computing》2006,9(1):29-44
I/O bottlenecks are already a problem in many large-scale applications that manipulate huge datasets. This problem is expected
to get worse as applications get larger, and the I/O subsystem performance lags behind processor and memory speed improvements.
At the same time, off-the-shelf clusters of workstations are becoming a popular platform for demanding applications due to
their cost-effectiveness and widespread deployment. Caching I/O blocks is one effective way of alleviating disk latencies,
and there can be multiple levels of caching on a cluster of workstations.
Previous studies have shown the benefits of caching—whether it be local to a particular node, or a shared global cache across
the cluster—for certain applications. However, we show that while caching is useful in some situations, it can hurt performance
if we are not careful about what to cache and when to bypass the cache. This paper presents compilation techniques and runtime
support to address this problem. These techniques are implemented and evaluated on an experimental Linux/Pentium cluster running
a parallel file system. Our results using a diverse set of applications (scientific and commercial) demonstrate the benefits
of a discretionary approach to caching for I/O subsystems on clusters, providing as much as 48% savings in overall execution
time over indiscriminately caching everything in some applications.
Parts of this paper have appeared in the Proceedings of the 3rd IEEE/ACM Symposium on Cluster Computing and the Grid (CCGrid'03).
This paper is an extension of these prior results, and includes a more extensive performance evaluation.
Murali Vilayannur is a Ph.D. student in the Department of Computer Science and Engineering at The Pennsylvania State University. His research
interests are in High-Performance Parallel I/O, File Systems, Virtual Memory Algorithms and Operating Systems.
Anand Sivasubramaniam received his B.Tech. in Computer Science from the Indian Institute of Technology, Madras, in 1989, and the M.S. and Ph.D.
degrees in Computer Science from the Georgia Institute of Technology in 1991 and 1995 respectively. He has been on the faculty
at The Pennsylvania State University since Fall 1995 where he is currently an Associate Professor. Anand's research interests
are in computer architecture, operating systems, performance evaluation, and applications for both high performance computer
systems and embedded systems. Anand's research has been funded by NSF through several grants, including the CAREER award,
and from industries including IBM, Microsoft and Unisys Corp. He has several publications in leading journals and conferences,
and is on the editorial board of IEEE Transactions on Computers and IEEE Transactions on Parallel and Distributed Systems.
He is a recipient of the 2002 IBM Faculty Award. Anand is a member of the IEEE, IEEE Computer Society, and ACM.
Mahmut Kandemir received the B.Sc. and M.Sc. degrees in control and computer engineering from Istanbul Technical University, Istanbul, Turkey,
in 1988 and 1992, respectively. He received the Ph.D. from Syracuse University, Syracuse, New York in electrical engineering
and computer science, in 1999. He has been an assistant professor in the Computer Science and Engineering Department at the
Pennsylvania State University since August 1999. His main research interests are optimizing compilers, I/O intensive applications,
and power-aware computing. He is a member of the IEEE and the ACM.
Rajeev Thakur is a Computer Scientist in the Mathematics and Computer Science Division at Argonne National Laboratory. He received a B.E.
from the University of Bombay, India, in 1990, M.S. from Syracuse University in 1992, and Ph.D. from Syracuse University in
1995, all in computer engineering. His research interests are in the area of high-performance computing in general and high-performance
networking and I/O in particular. He was a member of the MPI Forum and participated actively in the definition of the I/O
part of the MPI-2 standard. He is the author of a widely used, portable implementation of MPI-IO, called ROMIO. He is also
a co-author of the book “Using MPI-2: Advanced Features of the Message Passing Interface” published by MIT Press.
Robert Ross received his Ph.D. in Computer Engineering from Clemson University in 2000. He is now an Assistant Scientist in the Mathematics
and Computer Science Division at Argonne National Laboratory. His research interests are in message passing and storage systems
for high performance computing environments. He is the primary author and lead developer for the Parallel Virtual File System
(PVFS), a parallel file system for Linux clusters. Current projects include the ROMIO MPI-IO implementation, PVFS, PVFS2,
and the MPICH2 implementation of the MPI message passing interface. 相似文献
5.
Existing hypotheses on the mode of disk formation in the photoreceptor cells of mammals appear to be incompatible: (1) plasma
membranes of adjacent evaginations form a disk which, subsequently, is internalized by a disk rim; (2) pinocytotic vesicles
are pinched off from the plasma membrane and fuse into a larger vesicle, which flattens and forms a disk. We have studied
the development of the cone outer segment and the disk formation in Tupaia
belangeri by transmission electron microscopy. During the first two postnatal weeks, the distal part of the single cilium, which is
inserted apically on the inner segment, becomes balloon-shaped. Apical to the axoneme, it contains tubular and vesicular material,
which, most probably, has been detached from the axonemal microtubules. These tubules and vesicles do not contribute to disks.
The balloon-shaped expansion, later retained as the ciliary backbone, establishes the contact with the pigment epithelium.
Formation of disks, from the 12-day-old Tupaia onwards, occurs between adjacent evaginations at the outer segment base. The initial disk rims are “hooked” to the ciliary
axonemal microtubules. The axonemal microtubules are involved in the initiation and in the alignment of the disks. Disk rim
formation and, thus, internalization of disks proceeds from the base to the apex of the outer segment, that is, from the younger
to the older disks. In the adult Tupaia, an uneven progression of disk rim formation on both sides of the axoneme is found among consecutive disks. The seemingly
incompatible hypotheses on the mode of disk formation reflect a heterochrony of the internalization of membranes and of the
disk formation among different mammals and, possibly, between cones and rods.
Received: 24 July 1997 / Accepted: 10 September 1997 相似文献
6.
Distributed redundant array of inexpensive disks (RAID) is often embedded in a cluster architecture. In a centralized RAID subsystem, the false sharing problem does not exist, because the disk array allows only mutually exclusive access by one user at a time. However, the problem does exist in a distributed RAID architecture, because multiple accesses may occur simultaneously in a distributed environment. This problem will seriously limit the effectiveness of collective I/O operations in network-based, cluster computing. Traditional accesses to disks in a RAID are done at block level. The block granularity is large, say 32 KB, often resulting in false sharing among fragments in the block. The false sharing problem becomes worse when the block size or the stripe unit becomes too large. To solve this problem, we propose an adaptive sector grouping approach to accessing a distributed RAID. Each sector has a fine grain of 512 B. Multiple sectors are grouped together to match with the data block size. The grouped sector has a variable size that can be adaptively adjusted by software. Benchmark experiments reveal the positive effects of this adaptive access scheme on the performance of a RAID. Our scheme can reduce the collective I/O access time without increasing the buffer size. Both theoretical analysis and experimental results demonstrate the performance gain in using grouped sectors for fast access of distributed RAID. 相似文献
7.
Impact of Admission and Cache Replacement Policies on Response Times of Jobs on Data Grids 总被引:1,自引:0,他引:1
Caching techniques have been used widely to improve the performance gaps of storage hierarchies in computing systems. Little
is known about the impact of policies on the response times of jobs that access and process very large files in data grids,
particularly when data and computations on the data have to be co-located on the same host. In data intensive applications
that access large data files over wide area network environment, such as data-grids, the combination of policies for job servicing
(or scheduling), caching and cache replacement can significantly impact the performance of grid jobs. We present preliminary
results of a simulation study that combines an admission policy with a cache replacement policy when servicing jobs submitted
to a storage resource manager.The results show that, in comparison to a first come first serve policy, the response times
of jobs are significantly improved, for practical limits of disk cache sizes, when the jobs that are back-logged to access
the same files are taken into consideration in scheduling the next file to be retrieved into the disk cache. Not only are
the response times of jobs improved, but also the metric measures for caching policies, such as the hit ratio and the average
cost per retrieval, are improved irrespective of the cache replacement policy used.
Ekow Otoo is research staff scientist with the scientific data management group at Lawrence Berkeley National Laboratory, University
of California, Berkeley. He received his B.Sc. degree in Electrical Engineering from the University of Science and Technology,
Kumasi, Ghana and a post graduate diploma in Computer Science from the University of Ghana, Legon. In 1977, he received his
M.Sc. degree in Computer Science from the University of Newcastle Upon Tyne in Britain and his Ph.D. degree in Computer Science
from McGill University, Montreal, Canada in 1983. He joined the faculty of the School of Computer Science, Carleton University,
in 1983 and from 1987 to 1999, he was a tenured faculty member of the School of Computer Science, Carleton University, Ottawa,
Canada. He has served as research consultant to Bell Northern Research, Ottawa, Canada, and as a research project consultant
to the GIS Division, Geomatics Canada, Natural Resources Canada, from 1990 to 1998. Ekow Otoo is a member of the ACM and IEEE.
His research interests include database management systems, data structures and algorithms, parallel I/O for high performance
computing, parallel and distributed computing.
Doron Rotem is currently a senior staff scientist and a member of the Data Management group at the Lawrence Berkeley National Lab. His
research interests include Grid Computing, Workflow, Scientific Data Management and Paralled and Distributed Computing and
Algorithms. He has published over 80 papers in international journals and conferences in these areas. Prior to that, Dr Rotem
co-founded and served as a CTO of a startup company, called CommerceRoute, that made software products in the area of workflow
and data integration and before that, he was an Associate Professor in the Department of Computer Science, University of Waterloo,
Canada. Dr. Rotem holds a B.Sc degree in Mathematics and Statistics from the Hebrew University, Jerusalem, Israel and a Ph.D.
in Computer Science from the University of the Witwatersrand, Johannesburg, South Africa.
Arie Shoshani is a senior staff scientist at Lawrence Berkeley National Laboratory. He joined LBNL in 1976. He heads the Scientific Data
Management Group. He received his Ph.D. from Princeton University in 1969. From 1969 to 1976, he was a researcher at System
Development Corporation, where he worked on the Network Control Program for the ARPAnet, distributed databases, database conversion,
and natural language interfaces to data management systems. His current areas of work include data models, query languages,
temporal data, statistical and scientific database management, storage management on tertiary storage, and grid storage middleware.
Arie is also the director of a Scientific Data Management (SDM) Integrated Software Infrastructure Center (ISIC), one of seven
centers selected by the SciDAC program at DOE in 2001. In this capacity, he is coordinating the work of collaborators from
4 DOE laboratories and 4 universities (see: http://sdmcenter.lbl.gov). Dr. Shoshani has published over 65 technical papers
in refereed journals and conferences, chaired several workshops, conferences, and panels in database management; and served
on numerous program committees for various database conferences. He also served as an associate editor for the ACM Transactions
on Database Systems. He was elected a member of the VLDB Endowment Board, served as the Publication Board Chairperson for
the VLDB Journal, and as the Vice-President of the VLDB Endowment. His home page is http://www.lbl.gov/arie. 相似文献
8.
To provide the required amount of storage and bandwidth, a video server typically comprises a large number of disks. As the
total number of disks increases, the influence of the striping algorithm that determines how video data are distributed across
the disks becomes decisive in terms of overall server cost and performance. Also introducing fault-tolerance against disk
failures becomes a must. In this paper, we first evaluate different striping algorithms in terms of throughput, buffer requirement,
and start-up latency for a non-fault-tolerant server. We then examine the impact of data striping on a fault-tolerant server
and show that the striping policy and the optimal technique to assure fault-tolerance are related: Depending on the technique
used to assure fault-tolerance (mirroring or parity), different striping techniques perform best.
This revised version was published online in July 2006 with corrections to the Cover Date. 相似文献
9.
We present gmblock, a block-level storage sharing system over Myrinet which uses an optimized I/O path to transfer data directly
between the storage medium and the network, bypassing the host CPU and main memory bus of the storage server. It is device
driver independent and retains the protection and isolation features of the OS. We evaluate the performance of a prototype
gmblock server and find that: (a) the proposed techniques eliminate memory and peripheral bus contention, increasing remote
I/O bandwidth significantly, in the order of 20–200% compared to an RDMA-based approach, (b) the impact of remote I/O to local
computation becomes negligible, (c) the performance characteristics of RAID storage combined with limited NIC resources reduce
performance. We introduce synchronized send operations to improve the degree of disk to network I/O overlapping. We deploy
the OCFS2 shared-disk filesystem over gmblock and show gains for various application benchmarks, provided I/O scheduling can
eliminate the disk bottleneck due to concurrent access. 相似文献
10.
Load balancing in a workstation-based cluster system has been investigated extensively, mainly focusing on the effective usage
of global CPU and memory resources. However, if a significant portion of applications running in the system is I/O-intensive,
traditional load balancing policies can cause system performance to decrease substantially. In this paper, two I/O-aware load-balancing
schemes, referred to as IOCM and WAL-PM, are presented to improve the overall performance of a cluster system with a general
and practical workload including I/O activities. The proposed schemes dynamically detect I/O load imbalance of nodes in a
cluster, and determine whether to migrate some I/O load from overloaded nodes to other less- or under-loaded nodes. The current
running jobs are eligible to be migrated in WAL-PM only if overall performance improves. Besides balancing I/O load, the scheme
judiciously takes into account both CPU and memory load sharing in the system, thereby maintaining the same level of performance
as existing schemes when I/O load is low or well balanced. Extensive trace-driven simulations for both synthetic and real
I/O-intensive applications show that: (1) Compared with existing schemes that only consider CPU and memory, the proposed schemes
improve the performance with respect to mean slowdown by up to a factor of 20; (2) When compared to the existing approaches
that only consider I/O with non-preemptive job migrations, the proposed schemes achieve improvements in mean slowdown by up
to a factor of 10; (3) Under CPU-memory intensive workloads, our scheme improves the performance over the existing approaches
that only consider I/O by up to 47.5%.
Xiao Qin received the BSc and MSc degrees in computer science from Huazhong University of Science and Technology in 1992 and 1999,
respectively. He received the PhD degree in computer science from the University of Nebraska-Lincoln in 2004. Currently, he
is an assistant professor in the department of computer science at the New Mexico Institute of Mining and Technology. His
research interests include parallel and distributed systems, storage systems, real-time computing, performance evaluation,
and fault-tolerance. He served on program committees of international conferences like CLUSTER, ICPP, and IPCCC. During 2000–2001,
he was on the editorial board of The IEEE Distributed System Online. He is a member of the IEEE.
Hong Jiang received the B.Sc. degree in Computer Engineering in 1982 from Huazhong University of Science and Technology, Wuhan, China;
the M.A.Sc. degree in Computer Engineering in 1987 from the University of Toronto, Toronto, Canada; and the PhD degree in
Computer Science in 1991 from the Texas A&M University, College Station, Texas, USA. Since August 1991 he has been at the
University of Nebraska-Lincoln, Lincoln, Nebraska, USA, where he is Associate Professor and Vice Chair in the Department of
Computer Science and Engineering. His present research interests are computer architecture, parallel/distributed computing,
computer storage systems and parallel I/O, performance evaluation, middleware, networking, and computational engineering.
He has over 70 publications in major journals and international Conferences in these areas and his research has been supported
by NSF, DOD and the State of Nebraska. Dr. Jiang is a Member of ACM, the IEEE Computer Society, and the ACM SIGARCH and ACM
SIGCOMM.
Yifeng Zhu received the B.E. degree in Electrical Engineering from Huazhong University of Science and Technology in 1998 and the M.S.
degree in computer science from University of Nebraska Lincoln (UNL) in 2002. Currently he is working towards his Ph.D. degree
in the department of computer science and engineering at UNL. His main fields of research interests are parallel I/O, networked
storage, parallel scheduling, and cluster computing. He is a student member of IEEE.
David Swanson received a Ph.D. in physical (computational) chemistry at the University of Nebraska-Lincoln (UNL) in 1995, after which he
worked as an NSF-NATO postdoctoral fellow at the Technical University of Wroclaw, Poland, in 1996, and subsequently as a National
Research Council Research Associate at the Naval Research Laboratory in Washington, DC, from 1997–1998. In early 1999 he returned
to UNL where he has coordinated the Research Computing Facility and currently serves as an Assistant Research Professor in
the Department of Computer Science and Engineering. The Office of Naval Research, the National Science Foundation, and the
State of Nebraska have supported his research in areas such as large-scale parallel simulation and distributed systems. 相似文献
11.
A new layout method Prime-groups is proposed to evenly distribute parity groups for declustered RAID. Prime-groups satisfies most of the layout goals for a good declustered RAID layout. For the goals that are not satisfied, it is near optimal. A new layout goal maximal write and reconstruction parallelism is also proposed. If a layout satisfies the new goal, all the surviving disks can be read in parallel and can be rewritten in parallel during reconstruction and reconfiguration. Prime-groups satisfies the new goal when the write request begins in the first disk of the array. It is also near optimal in term of declustering ratio when p is a prime. Prime-groups can compliment the layouts proposed by Alvarez et al. [2] for the criteria to obtain a good layout is quite different between Prime-groups and those proposed by Alvarez et al. [2]. 相似文献
12.
Hamid Reza Baghani Mostafa Robatjazi Seyed Rabi Mahdavi 《Radiation and environmental biophysics》2020,59(2):265-281
Radiation-shielding of healthy tissue is mandatory in breast intraoperative electron radiotherapy (IOERT). In this regard, dedicated radioprotection disks have been introduced. The aim of this study was to evaluate and compare the performance of three radioprotection disks widely used for breast IOERT. A Monte Carlo simulation approach was used for this purpose. The considered disks included Al + Pb, PMMA + Copper, and PTFE + Steel. They were stimulated by means of the MCNPX Monte Carlo code at depths around R100 and R90 of different electron energies in a water phantom, and their impact on the dosimetric properties of the therapeutic beam was evaluated in both correct and upside down disk placements. The electron energy spectrum immediately above and below each disk was calculated and analyzed. Furthermore, performance characteristics of the studied disks such as backscatter factors (BSFs) and transmission factors (TFs) at different electron energies were determined and compared. The results show that the Al + Pb disk most effectively attenuates the beam, while at the same time exhibits maximum BSF values. Employing the PMMA + Copper disk can minimize the BSF value but at the expense of an increased TF. The Al + Pb disk showed the best performance from the radiation protection viewpoint, while its highest BSF values could lead to perturbation of dose homogeneity within the target volume. PTFE + Steel disk showed an intermediate performance regarding the electron backscattering and transmission among the studied disks. The reverse placement of each disk can substantially increase the BSF value as compared to the correct situation but had less impact on the TF value. 相似文献
13.
In this paper we are concerned with the live verification of the consistency of a replicated system, an issue that has not
been addressed by the research community so far. We consider the problem of how to enable the system to detect automatically and in production whether the invariants defining the correctness of object replication are violated. This feature could greatly improve the
dependability of distributed applications and is necessary for constructing self-managing and self-healing replicated systems.
We focus on systems that enforce strongly consistent replication: all replicas of each object must be kept “continuously”
in-sync. This replication strategy is appropriate for application domains where correctness guarantees in spite of failures
are more important than performance and scalability. We present the design and implementation of a replicated web service
capable of self-checking whether all replicas are indeed kept in sync. This check occurs on-line, transparently to clients.
We also discuss the performance cost of self-checking in our prototype.
Alberto Bartoli is Associate Professor of Computer Engineering at the University of Trieste, Italy. He took a degree in Electrical Engineering
in 1989 and a doctorate in Computer Engineering in 1994, both at the University of Pisa, Italy. His research interests are
in the area of reliability and fault-tolerance in distributed systems.
Giovanni Masarin took a degree in Electronic Engineering in 2004, at the University of Trieste, Italy. He is currently involved in product
development at RadioTrevisan, a company specialized in the production of lawful interception equipments. 相似文献
14.
Dennis Gannon Sriram Krishnan Liang Fang Gopi Kandaswamy Yogesh Simmhan Aleksander Slominski 《Cluster computing》2005,8(4):271-277
Software Component Frameworks are well known in the commercial business application world and now this technology is being
explored with great interest as a way to build large-scale scientific applications on parallel computers. In the case of Grid
systems, the current architectural model is based on the emerging web services framework. In this paper we describe progress
that has been made on the Common Component Architecture model (CCA) and discuss its success and limitations when applied to
problems in Grid computing. Our primary conclusion is that a component model fits very well with a services-oriented Grid,
but the model of composition must allow for a very dynamic (both in space and in time) control of composition. We note that
this adds a new dimension to conventional service workflow and it extends the “Inversion of Control” aspects of most component
systems.
Dennis Gannon is a professor of Computer Science at Indiana University. He received his Ph.D. in Computer Science from the University of
Illinois in 1980 and his Ph.D. in Mathematics from the University of California in 1974. From 1980 to 1985, he was on the
faculty at Purdue University. His research interests include software tools for high performance distributed systems and problem
solving environments for scientific computation.
Sriram Krishnan received his Ph.D. in Computer Science from Indiana University in 2004. He is currently in the Grid Development Group at
the San Diego Supercomputer Center where he is working on designing a Web services based architecture for biomedical applications
that is secure and scalable, and is conducive to the creation of complex workflows. He received my undergraduate degree in
Computer Engineering from the University of Mumbai, India.
Liang Fang is a Ph.D. student in Computer Science at Indiana University. His research interests include Grid computing, Web services,
portals, their security and scalability issues. He is a Research Assistant in Computer Science at Indiana University, currently
responsible for investigating authorization and other security solutions to the project of Linked Environments for Atmospheric
Discovery (LEAD).
Gopi Kandaswamy is a Ph.D. student in the Computer Science Department at Indiana University where he is current a Research Assistant. His
research interests include Web services and workflow systems for the Grid.
Yogesh Simmhan received his B.E. degree in Computer Science from Madras University, India in 2000, and is a doctoral candidate in Computer
Science at Indiana University. He is currently working as a Research Assistant at Indiana University, investigating data management
issues in the LEAD project. His interests lie in data provenance for workflow systems and its use in data quality estimation.
Aleksander Slominski is a Ph.D. student in the Computer Science at Indiana University. His research interests include Grid and Web Services, streaming
XML Pull Parsing and performance, Grid security, asynchronous messaging, events, and notifications brokers, component technologies,
and workflow composition. He is currently working as a Research Assistant investigating creation and execution of dynamic
workflows using Grid Process Execution Language (GPEL) based on WS-BPEL. 相似文献
15.
Cooperation towards public goods relies on credible threats of punishment to deter cheats. However, punishing is costly, so
it remains unclear who incurred the costs of enforcement in our evolutionary past. Theoretical work suggests that human cooperation
may be promoted if people believe in supernatural punishment for moral transgressions. This theory is supported by new work in cognitive psychology and by anecdotal ethnographic
evidence, but formal quantitative tests remain to be done. Using data from 186 societies around the globe, I test whether
the likelihood of supernatural punishment—indexed by the importance of moralizing “high gods”—is associated with cooperation.
Dominic Johnson is a fellow in the Princeton University Society of Fellows. He holds a D.Phil. in biology from Oxford University
and a Ph.D. in political science from Geneva University. His research revolves around the evolutionary biology of human behavior
and how this impacts on conflict, cooperation, politics, and religion. His recent book, Overconfidence and War: The Havoc and Glory of Positive Illusions, was published in 2004 by Harvard University Press. 相似文献
16.
Spine degeneration is a pathology that will affect 80% of the population. Since the intervertebral disks play an important role in transmitting loads through the spine, the aim of this study was to evaluate the biomechanical impact of disk properties on the load carried by healthy (Thompson grade I) and degenerated (Thompson grades III and IV) disks. A three-dimensional parametric poroelastic finite element model of the L4/L5 motion segment was developed. Grade I, grade II, and grade IV disks were modeled by altering the biomechanical properties of both the annulus and nucleus. Models were validated using published creep experiments, in which a constant compressive axial stress of 0.35 MPa was applied for 4 h. Pore pressure (PP) and effective stress (S(E)) were analyzed as a function of time following loading application (1 min, 5 min, 45 min, 125 min, and 245 min) and discal region along the midsagittal profile for each disk grade. A design of experiments was further implemented to analyze the influence of six disk parameters (disk height (H), fiber proportion (%F), drained Young's modulus (E(a),E(n)), and initial permeability (k(a),k(n)) of both the annulus and nucleus) on load-sharing for disk grades I and IV. Simulations of grade I, grade III, and grade IV disks agreed well with the available published experimental data. Disk height (H) had a significant influence (p<0.05) on the PP and S(E) during the entire loading history for both healthy and degenerated disk models. Young's modulus of the annulus (E(a)) significantly affected not only S(E) in the annular region for both disk grades in the initial creep response but also S(E) in the nucleus zone for degenerated disks with further creep response. The nucleus and annulus permeabilities had a significant influence on the PP distribution for both disk grades, but this effect occurred at earlier stages of loading for degenerated than for healthy disk models. This is the first study that investigates the biomechanical influence of both geometrical and material disk properties on the load transfer of healthy and degenerated disks. Disk height is a significant parameter for both healthy and degenerated disks during the entire loading. Changes in the annulus stiffness, as well as in the annulus and nucleus permeability, control load-sharing in different ways for healthy and degenerated disks. 相似文献
17.
High-performance computing increasingly occurs on “computational grids” composed of heterogeneous and geographically distributed
systems of computers, networks, and storage devices that collectively act as a single “virtual” computer. A key challenge
in this environment is to provide efficient access to data distributed across remote data servers. Our parallel I/O framework,
called Armada, allows application and data-set providers to flexibly compose graphs of processing modules that describe the
distribution, application interfaces, and processing required of the dataset before computation. Although the framework provides
a simple programming model for the application programmer and the data-set provider, the resulting graph may contain bottlenecks
that prevent efficient data access. In this paper, we present an algorithm used to restructure Armada graphs that distributes
computation and data flow to improve performance in the context of a wide-area computational grid.
This work was supported by Sandia National Laboratories under DOE contract DOE-AV6184.
Ron A. Oldfield is a senior member of the technical staff at Sandia National Laboratories in Albuquerque, NM. He received the B.Sc. in computer
science from the University of New Mexico in 1993. From 1993 to 1997, he worked in the computational sciences department of
Sandia National Laboratories, where he specialized in seismic research and parallel I/O. He was the primary developer for
the GONII-SSD (Gas and Oil National Information Infrastructure–Synthetic Seismic Dataset) project and a co-developer for the
R&D 100 award winning project “Salvo”, a project to develop a 3D finite-difference prestack-depth migration algorithm for
massively parallel architectures. From 1997 to 2003 he attended graduate school at Dartmouth college and received his Ph.D.
in June, 2003. In September of 2003, he returned to Sandia to work in the Scalable Computing Systems department. His research
interests include parallel and distributed computing, parallel I/O, and mobile computing.
David Kotz is a Professor of Computer Science at Dartmouth College in Hanover NH. After receiving his A.B. in Computer Science and Physics
from Dartmouth in 1986, he completed his Ph.D in Computer Science from Duke University in 1991. He returned to Dartmouth to
join the faculty in 1991, where he is now Professor of Computer Science, Director of the Center for Mobile Computing, and
Executive Director of the Institute for Security Technology Studies. His research interests include context-aware mobile computing,
pervasive computing, wireless networks, and intrusion detection. He is a member of the ACM, IEEE Computer Society, and USENIX
associations, and of Computer Professionals for Social Responsibility. For more information see http://www.cs.dartmouth.edu/dfk/. 相似文献
18.
Jiannong Cao Alvin T. S. Chan Yudong Sun Sajal K. Das Minyi Guo 《Cluster computing》2006,9(3):355-371
Application scheduling plays an important role in high-performance cluster computing. Application scheduling can be classified
as job scheduling and task scheduling. This paper presents a survey on the software tools for the graph-based scheduling on
cluster systems with the focus on task scheduling. The tasks of a parallel or distributed application can be properly scheduled
onto multi-processors in order to optimize the performance of the program (e.g., execution time or resource utilization).
In general, scheduling algorithms are designed based on the notion of task graph that represents the relationship of parallel
tasks. The scheduling algorithms map the nodes of a graph to the processors in order to minimize overall execution time. Although
many scheduling algorithms have been proposed in the literature, surprisingly not many practical tools can be found in practical
use. After discussing the fundamental scheduling techniques, we propose a framework and taxonomy for the scheduling tools
on clusters. Using this framework, the features of existing scheduling tools are analyzed and compared. We also discuss the
important issues in improving the usability of the scheduling tools.
This work is supported by the Hong Kong Polytechnic University under grant H-ZJ80 and by NASA Ames Research Center by a cooperative
grant agreement with the University of Texas at Arlington.
Jiannong Cao received the BSc degree in computer science from Nanjing University, Nanjing, China in 1982, and the MSc and the Ph.D degrees
in computer science from Washington State University, Pullman, WA, USA, in 1986 and 1990 respectively. He is currently an
associate professor in Department of Computing at the Hong Kong Polytechnic University, Hong Kong. He is also the director
of the Internet and Mobile Computing Lab in the department. He was on the faculty of computer science at James Cook University
and University of Adelaide in Australia, and City University of Hong Kong. His research interests include parallel and distributed
computing, networking, mobile computing, fault tolerance, and distributed software architecture and tools. He has published
over 120 technical papers in the above areas. He has served as a member of editorial boards of several international journals,
a reviewer for international journals/conference proceedings, and also as an organizing/programme committee member for many
international conferences. Dr. Cao is a member of the IEEE Computer Society, the IEEE Communication Society, IEEE, and ACM.
He is also a member of the IEEE Technical Committee on Distributed Processing, IEEE Technical Committee on Parallel Processing,
IEEE Technical Committee on Fault Tolerant Computing, and Computer Architecture Professional Committee of the China Computer
Federation.
Alvin Chan is currently an assistant professor at the Hong Kong Polytechnic University. He graduated from the University of New South
Wales with a Ph.D. degree in 1995 and was subsequently employed as a Research Scientist by the CSIRO, Australia. From 1997
to 1998, he was employed by the Centre for Wireless Communications, National University of Singapore as a Program Manager.
Dr. Chan is one of the founding members and director of a university spin-off company, Information Access Technology Limited.
He is an active consultant and has been providing consultancy services to both local and overseas companies. His research
interests include mobile computing, context-aware computing and smart card applications.
Yudong Sun received the B.S. and M.S. degrees from Shanghai Jiao Tong University, China. He received Ph.D. degree from the University
of Hong Kong in 2002, all in computer science. From 1988 to 1996, he was among the teaching staff in Department of Computer
Science and Engineering at Shanghai Jiao Tong University. From 2002 to 2003, he held a research position at the Hong Kong
Polytechnic University. At present, he is a Research Associate in School of Computing Science at University of Newcastle upon
Tyne, UK. His research interests include parallel and distributed computing, Web services, Grid computing, and bioinformatics.
Sajal K. Das is currently a Professor of Computer Science and Engineering and the Founding Director of the Center for Research in Wireless
Mobility and Networking (CReWMaN) at the University of Texas at Arlington. His current research interests include resource
and mobility management in wireless networks, mobile and pervasive computing, sensor networks, mobile internet, parallel processing,
and grid computing. He has published over 250 research papers, and holds four US patents in wireless mobile networks. He received
the Best Paper Awards in ACM MobiCom’99, ICOIN-16, ACM, MSWiM’00 and ACM/IEEE PADS’97. Dr. Das serves on the Editorial Boards
of IEEE Transactions on Mobile Computing, ACM/Kluwer Wireless Networks, Parallel Processing Letters, Journal of Parallel Algorithms
and Applications. He served as General Chair of IEEE PerCom’04, IWDC’04, MASCOTS’02 ACM WoWMoM’00-02; General Vice Chair of
IEEE PerCom’03, ACM MobiCom’00 and IEEE HiPC’00-01; Program Chair of IWDC’02, WoWMoM’98-99; TPC Vice Chair of ICPADS’02; and
as TPC member of numerous IEEE and ACM conferences.
Minyi Guo received his Ph.D. degree in information science from University of Tsukuba, Japan in 1998. From 1998 to 2000, Dr. Guo had
been a research scientist of NEC Soft, Ltd. Japan. He is currently a professor at the Department of Computer Software, The
University of Aizu, Japan. From 2001 to 2003, he was a visiting professor of Georgia State University, USA, Hong Kong Polytechnic
University, Hong Kong. Dr. Guo has served as general chair, program committee or organizing committee chair for many international
conferences, and delivered more than 20 invited talks in USA, Australia, China, and Japan. He is the editor-in-chief of the
Journal of Embedded Systems. He is also in editorial board of International Journal of High Performance Computing and Networking,
Journal of Embedded Computing, Journal of Parallel and Distributed Scientific and Engineering Computing, and International
Journal of Computer and Applications.
Dr. Guo’s research interests include parallel and distributed processing, parallelizing compilers, data parallel languages,
data mining, molecular computing and software engineering. He is a member of the ACM, IEEE, IEEE Computer Society, and IEICE.
He is listed in Marquis Who’s Who in Science and Engineering. 相似文献
19.
Differential membrane protein phosphorylation in bovine retinal rod outer segment disk membranes as a function of disk age 总被引:1,自引:0,他引:1
Kathleen Boesze-Battaglia Arlene D. Albert James S. Frye Philip L. Yeagle 《Bioscience reports》1996,16(4):289-297
The outer segment portion of photoreceptor rod cells is composed of a stacked array of disk membranes. Newly formed disks are found at the base of the rod outer segment (ROS) and are relatively high in membrane cholesterol. Older disks are found at the apical tip of the ROS and are low in membrane cholesterol. Disk membranes were separated based on their membrane cholesterol content and the extent of membrane protein phosphorylation determined. Light induced phosphorylation of ROS disk membrane proteins was investigated using magic angle spinning31P NMR. When intact rod outer segment preparations were stimulated by light, in the presence of endogenously available kinases, membrane proteins located in disks at the base of the ROS were more heavily phosphorylated than those at the tip. SDS-gel electrophoresis of the phosphorylated disk membranes subpopulations identified a phosphoprotein species with a molecular weight of approximately 68–72 kDa that was more heavily phosphorylated in newly formed disks than in old disks. The identity of this phosphoprotein is presently under investigation. When the phosphorylation reaction was carried out in isolated disk membrane preparations with exogenously added co-factors and kinases, there was no preferential protein phosphorylation. Taken collectively, these results suggest that within the ROS there is a protein phosphorylation gradient that maybe indicative of co-factor or kinase heterogeneity. 相似文献
20.
An attacker’s connection can propagate quickly to different parts of a transparent All-Optical Network. Such attacks affect
the normal traffic and cause a quality of service degradation or outright service denial. Attack monitors can collect the
information of each link and each node to help diagnose the attacker’s exact location.
Quick detection and localization of an attack source helps avoid losing large amounts of data in an all-optical network. However,
to detect attack sources, it is not necessary to put monitors on all nodes. Since not every wavelength on every link is being
used all the time, we propose to use the idle wavelengths to setup diagnostic connections and obtain the necessary information
needed for diagnosis purposes. We show that placing a relatively small number of monitors at some key nodes in a network is
sufficient to achieve level of performance. However, the monitor placement policy, routing policy, and diagnosis method are
challenging problems.
We, in this paper, first develop a monitor placement policy, a test connection policy, and a routing policy based on our definition
of crosstalk attack and monitor node models. With these policies, we show that we can always detect and localize the malicious
connections as long as there is no more than one malicious connection on each wavelength in the whole network. After this,
we develop a scalable diagnosis method, which can localize the sources of the such malicious attacks in a fast manner.
Arun K. Somani is currently Jerry R. Junkins Chair Professor of Electrical and Computer Engineering at Iowa State University. He earned
his MSEE and Ph.D. degrees in electrical engineering from the McGill University, Montreal, Canada, in 1983 and 1985, respectively.
He worked as Scientific Officer for Govt. of India, New Delhi from 1974 to 1982. From 1985 to 1997, he was a faculty member
at the University of Washington, Seattle, WA, where he was a Professor of EE and CSE from 1995 onwards. From 1997 to 2002,
he served as David C. Nicholas Professor of Electrical and Computer Engineering at Iowa State University. Professor Somani’s
research interests are in the area of fault tolerant computing, computer communication and networks, wireless and optical
networking, computer architecture, and parallel computer systems.
Tao Wu received the B.S. and M.S.E.E. degrees in telecommunication engineering from the University of Electronic Science and Technology
of China, Sichuan, China, in 1993 and 1996, respectively, and the Ph.D. degree in computer and electrical engineering from
Iowa State University, Ames, in 2003.
He is currently a Software Engineer with Microsoft Corporation. His research interests are in the area of WDM-based optical
networking, network security, and image processing. 相似文献