BROWN, A. B., AND PATTERSON, D. A. Towards availability benchmarks: A case study of software RAID systems. In USENIX '00: Proceedings of the USENIX Annual Technical Conference (San Diego, California, USA, June 2000), USENIX Association.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....benchmarking system behavior under fault loads include [20, 24] However, these works do not provide a good understanding of how one would estimate overall system availability under a given fault load. System availability studies such as [2, 25] are works in this direction. An interesting paper is [8], which outlines a methodology for benchmarking systems availability. Our work here focuses more closely on cluster based servers, and in particular the impact of faults affecting the communication system on service performability. 9 Conclusions We have studied the impact of TCP and VIA on the ....

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In Proceedings of the 2000.

....should be built to recover rapidly, in addition to being fault tolerant. While similar in viewpoint, our proposed methodology concentrates more on evaluating performability independently of the approach taken to improve performance or availability. Perhaps more similar to our work is that of [6], which outlines a methodology for benchmarking systems availability. Other works have proposed robustness [29] and reliability benchmarks [34] that quantify the degradation of system performance under faults. Our work here differs from these previous studies in that we focus on cluster based ....

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In Proceedings of the 2000.

....(FTCS 23) Toulouse, France, 1993, pp. 88 97 (IEEE CS Press) 2] T.K. Tsai, R. K. Iyer and D. Jewitt, An Approach Towards Benchmarking of Fault Tolerant Commercial Systems , in Proc. 26th Int. Syrup. on Fault Tolerant Computing (FTCS 26) Sendai, Japan, 1996, pp. 314 323 (IEEE CS Press) [3] P. Koopman and J. DeVale, Comparing the Robustness of POSIX Operating Systems , in Proc. 29th Int. Syrup. on Fault Tolerant Computing (FTCS 29) Madison, WI, USA, 1999, pp. 30 37 (IEEE CS Press) 4] A. Brown and D. A. Patterson, Towards Availability Benchmarks: A Cases Study of Software RAID ....

Brown, A. and D.A. Patterson. "Towards Availability Benchmarks: A Case Study of Software RAID Systems." Proc. 2000.

....benchmarking system behavior under fault loads include [20, 24] However, these works do not provide a good understanding of how one would estimate overall system availability under a given fault load. System availability studies such as [2, 25] are works in this direction. An interesting paper is [8], which outlines a methodology for benchmarking systems availability. Our work here focuses more closely on cluster based servers, and in particular the impact of faults affecting the communication system on service performability. 9 Conclusions We have studied the impact of TCP and VIA on the ....

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In Proceedings of the 2000.

....security) in the wide area. Additional work in the systems literature deals with evaluating the performability of a system. Wilkes and Stata [17] propose a method for describing performability using variations in quality of service under normal and degraded modes of operation. Brown and Patterson [7] describe how to measure the availability of RAID systems using a similar performability framework. In the performance space, we have successfully used the approach of declarative goal specification [16] to automate the mapping of performance and capacity requirements onto storage designs [1, 2, ....

A. Brown and D. Patterson. "Towards availability benchmarks: a case study of software RAID systems," Proc. of the 2000.

....and robustness testing, we feel that dependability benchmarking must provide a uniform, repeatable, comparable way of performing this evaluation. Dependability benchmarking is subject of a growing interest today. Both the research community and the industry are involved in research projects [6, 7] and groups meant to advance the dependability benchmarking area. Particularly, the IFIP Working Group 10.4 has created a Special Interest Group (SIG) to promote the research, practice, adoption, and dissemination of benchmarks for computer related system dependability (the authors of this paper ....

Aaron Brown and David Patterson, "Towards availability benchmark: a case study of software RAID systems", Proceedings of 2000.

....under fault loads include [25, 30] However, these works do not provide a good understanding of how one would estimate overall system availability under a given fault load. System availability studies such as [2, 33] are works in this direction. An interesting paper is that of Brown et al. [8], which outlines a methodology for benchmarking systems availability. Our work here focuses more closely on cluster based servers, and in particular the impact of faults affecting the communication system on service performability. 9 Conclusions We have studied the impact of TCP and VIA on the ....

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In Proceedings of the 2000.

....in addition to being fault tolerant [1] This view of faults is similar to our own, although our proposed methodology concentrates more on evaluating performability independent of the approach taken to improve performance or availability. Perhaps more similar to our work is that of Brown et al. [6], which outlines a methodology for benchmarking systems availability. Our work here focuses more closely on cluster based servers. Also, a significant part of our effort has been invested in the construction of Mendosus as a comprehensive faultinjection infrastructure. System availability ....

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In Proceedings of the 2000.

....in addition to being fault tolerant [1] This view of faults is similar to our own, although our proposed methodology concentrates more on evaluating performability independent of the approach taken to improve performance or availability. Perhaps more similar to our work is that of Brown et al. [6], which outlines a methodology for benchmarking systems availability. Our work here focus more closely on cluster based servers. Also, a significant part of our effort has been invested in the construction of Mendosus as a comprehensive faultinjection infrastructure. System availability studies ....

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In Proceedings of the 2000.

....operator errors will soon become the dominant factor in component failures. More recent work in the Windows NT context validates these arguments [17, 18, 25] Expanding on this theme, work has emerged using human factor studies on to determine how actual humans behave in the context of RAID repair [3]. While our choice of workstation reboot metric capture human factor effects to a large degree, our work leaves open the question if the human factor is dominant cause of the observed reboot distribution. 7. CONCLUSION We have proposed the application of a general three step framework to the ....

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In 2000.

No context found.

BROWN, A. B., AND PATTERSON, D. A. Towards availability benchmarks: A case study of software RAID systems. In USENIX '00: Proceedings of the USENIX Annual Technical Conference (San Diego, California, USA, June 2000), USENIX Association.

No context found.

A. B. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. 2000.

No context found.

A. Brown and D.A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. Proceedings of the 2000.

....work was initially published, there had been several studies of RAID reliability and availability [25] 26] but these had focused on simulation studies of hardware RAID, and none had examined RAID in the context of a general availability benchmark. Following on this work s initial publication [10], researchers at EMC, a major commercial RAID system vendor, adapted our availability benchmarking techniques to measure the availability of the centralized cache component in an undisclosed EMC storage system, and found the techniques useful in identifying availability weaknesses as well as ....

A. Brown and D.A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. Proceedings of the 2000.

.... it is not unheard of to find shipping software products with significant bugs in their recovery code; an illustrative example is the software RAID 5 driver shipped with Solaris x86, which improperly handles a double disk failure by returning fabricated garbage data in response to I O requests [7]. Often times the bugs are in performance, not functionality; again choosing an example from software RAID, the Linux software RAID 5 driver puts such low priority on disk reconstruction that repair can take hours even with small volumes and moderate user loads [7] Repair centric systems must ....

.... data in response to I O requests [7] Often times the bugs are in performance, not functionality; again choosing an example from software RAID, the Linux software RAID 5 driver puts such low priority on disk reconstruction that repair can take hours even with small volumes and moderate user loads [7]. Repair centric systems must address the problem of buggy and ineffective maintenance code. The only way to do this is to exercise the maintenance code to expose bugs and allow them to be corrected before they are needed in an emergency. This must be done frequently and in the context of a ....

[Article contains additional citation context not shown here]

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. Proceedings of the

....which can be used to improve software and thereby stretch its MTTF. Progress on performance was so quick in part because we had a common yardstickbenchmarksto measure success. To make such rapid progress on recovery, we need similar incentives. Prior work has successfully benchmarked availability [Brown00] Brown02] and so we do not cover the topic in this paper. With any benchmark, one of the first questions is whether it is realistic. Rather than guess why systems fail, we need facts on which to base fault workloads. Section 2 presents the type of data that we need to drive the benchmarks. Although ....

A. Brown and D.A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. Proc

....issue by examining the practical, delivered availability and reliability of a real database system in the context of realistic disk faults and failures. Our approach is based on the availability benchmarking technique introduced in our previous work measuring the availability of RAID systems [3]. In essence, this technique quantifies availability behavior by examining the variations in delivered quality of service as the system is subjected to targeted fault injection. For this study of DBMSs, we adapted the general methodology defined in our earlier work by defining quality of service ....

....are presented in Section 4 and discussed in Section 5. We discuss related work in Section 6, and wrap up with conclusions and future work in Section 7. 2 Availability Benchmarking Methodology This section provides a brief recap of the general availability benchmarking methodology defined in [3], then describes how that methodology can be specialized to the case of database availability benchmarking. 2.1 Overview The general availability benchmarking methodology laid out in [3] consists of four parts: a set of qualityof service metrics that measure the test system s behavior, a ....

[Article contains additional citation context not shown here]

A. Brown and D. A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. Proceedings of the

No context found.

Aaron Brown and David Patterson. Towards availability benchmarks: A case study of software RAID systems. In Proceedings of the 2000.

No context found.

A. Brown and D. Patterson. Towards availability benchmarks: A case study of software RAID systems. In Proceedings of the 2000.

No context found.

Aaron Brown and David A. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In Proceedings of the 2000.

No context found.

A. Brown and D. Patterson. Towards availability benchmarks: a case study of software RAID systems. Proc. 2000 USENIX Annual Technical Conference, pp. 263--276, June 2000.

No context found.

A. Brown and D. Patterson. Towards availability benchmarks: a case study of software RAID systems. Proceedings of the 2000.

No context found.

A. Brown and D. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In USENIX Annual Technical Conference, San Diego, CA, June 2000.

No context found.

A. Brown and D. Patterson. Towards Availability Benchmarks: A Case Study of Software RAID Systems. In USENIX Annual Technical Conference, San Diego, CA, June 2000.

No context found.

A. Brown and D. A. Patterson, "Towards Availability Benchmarks: A Cases Study of Software RAID Systems", in Proc. 2000.

First 50 documents

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC