ARPACI-DUSSEAU, A. C., ARPACI-DUSSEAU, R. H., CULLER, D. E., HELLERSTEIN, J. M., AND PATTERSON, D. A. High-performance sorting on networks of workstations. In In Proceedings of the ACM International Conference on Management of Data (SIGMOD) (Tucson, AZ, May 1997), pp. 243--254.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....applications can be improved by taking into account performance heterogeneities when designing their programming environment. A combination of distributed data queues and redundancy in the layout and access of disks have been shown to achieve robustness against system performance variations [3, 4]. 8 Conclusions In the present paper, we introduced an innovative approach for running application specific computation close to stored data. We preserve the traditionally simple interface of file system accesses, and experimentally demonstrate significant performance advantages in comparison to ....

Arpaci-Dusseau, A. C., Arpaci-Dusseau, R. H., Culler, D. E., Hellerstein, J. M., and Patterson, D. High-Performance Sorting on Networks of Workstations. In ACM SIGMOD (Tucson, AZ, May 1997), pp. 243--254.

....Parallel Sort. We use Parallel Sort to show how an ac tive switch can redistribute the data according to some criteria and reduce both host CPU utilization and net work traffic. Our Parallel Sort works in the same way as a one pass parallel sort on data with a unified key distri bution [3]. Each participating host reads in a portion of the data and performs data redistribution according to a range assigned to each node. After this stage, each node sorts its local data using any sorting algorithm. In this experiment, we sort 16M data items on 4 nodes. The data format follows the ....

A. C. Arpaci-Dusseau et al. High-Performance Sort- ing on Networks of Workstations, In SIGMOD, [20] pages 243-254, May 1997.

....the reader to Graefe s survey [12] for an overview of database query processing techniques and highlight the most relevant work on parallel query processing. Early work concentrated on parallelizing individual, traditional content sensitive operators like hybrid hash join [25] and sort (e.g. [10, 20, 1]) The abstractions which inspired Flux, Exchange [11] and RiverDQ [23] were proposed to compose such operators into a dataflow. Shatdal and Naughton [27] describe how to leverage shared virtual memory across a shared nothing cluster to implement hybrid hash join. DeWitt et al. present practical ....

A. C. Arpaci-Dusseau et al. High-Performance Sorting on Networks of Workstations. In SIGMOD, 1997.

....direction. We refer the reader to [13] for an overview of database query processing techniques and highlight the most relevant work on parallel query processing. Early work concentrated on parallelizing individual, traditional content sensitive operators like hybrid hash join [25] and sort (e.g. [11, 20, 1]) The abstractions which inspired Flux, Exchange [12] and RiverDQ [23] were proposed to compose such operators into a dataflow. In [10] and [9] the authors present practical techniques for handling data skew for a hash join and external sort, respectively. These techniques rely on sampling a ....

A. C. Arpaci-Dusseau et al. High-Performance Sorting on Networks of Workstations. In SIGMOD, 1997.

....individual computations. Optimal parallel sorting algorithms are complex because they extract a maximum degree of parallelism. This is not necessary in practice since we only have a limited number of processors in the system. For clusters, HPVM MinuteSort [37] is a one pass sort based on NOW Sort [6, 7]. It uses sort nodes with more memory and CPU, and I O nodes with more disks. The I O nodes distribute records to the sort nodes which then sort and return them. Most of the work in this system is done on the sort nodes; the I O nodes are statically selected to partition the data. 8. Conclusion ....

A. C. Arpaci-Dusseau, R. H. Arpaci-Dusseau, D. E. Culler, J. M. Hellerstein, and D. A. Patterson. High-performance sorting on networks of workstations. In J. M. Peckman, editor, Proceedings, ACM SIGMOD International Conference on Management of Data: SIGMOD 1997.

No context found.

ARPACI-DUSSEAU, A. C., ARPACI-DUSSEAU, R. H., CULLER, D. E., HELLERSTEIN, J. M., AND PATTERSON, D. A. High-performance sorting on networks of workstations. In In Proceedings of the ACM International Conference on Management of Data (SIGMOD) (Tucson, AZ, May 1997), pp. 243--254.

No context found.

ARPACI-DUSSEAU, A. C., ARPACI-DUSSEAU, R. H., CULLER, D. E., HELLERSTEIN, J. M., AND PATTERSON, D. A. High-performance sorting on networks of workstations. In In Proceedings of the ACM International Conference on Management of Data (SIGMOD) (Tucson, AZ, May 1997), pp. 243--254.

No context found.

A. C. Arpaci-Dusseau, R. H. Arpaci-Dusseau, D. E. Culler, J. M. Hellerstein, and D. Patterson. HighPerformance Sorting on Networks of Workstations. In Proceedings of the 1997.

....is hidden by the storage layer from the file system. Thus, current systems must handle any disk performance heterogeneity in the storage layer the file system does not have enough information to do so itself. The research community has proposed schemes to deal with static disk heterogeneity [3, 10, 32, 52], though many of these solutions require careful tuning by an administrator. As Van Jacobsen notes, Experience shows that anything that needs to be configured will be misconfigured [18] Further complicating the issue is that the delivered performance of a device could change over time. Such ....

A. C. Arpaci-Dusseau, R. H. Arpaci-Dusseau, D. E. Culler, J. M. Hellerstein, and D. Patterson. High-Performance Sorting on Networks of Workstations. In Proceedings of the 1997.

....blocks up through the newly increased size to fetch them into the buffer cache. In the second step, Dust probes each block again, measuring the time per probe to verify if the block is still in the cache. This technique is similar to the technique used to determine available memory in NOW Sort [4]. There are two important features of this approach. First, by probing every file block in the second step, this algorithm is independent of the replacement policy used to manage the buffer cache. Second, this algorithm works even when the buffer cache is integrated with the virtual memory ....

A. C. Arpaci-Dusseau, R. H. Arpaci-Dusseau, D. E. Culler, J. M. Hellerstein, and D. A. Patterson. High-Performance Sorting on Networks of Workstations. In SIGMOD '97, Tucson, AZ, May 1997.

No context found.

A. C. Arpaci-Dusseau, R. H. Arpaci-Dusseau, D. E. Culler, J. M. Hellerstein, and D. A. Patterson. High-performance sorting on networks of workstations. In Proceedings of ACM SIGMOD'97, pages 243--254, 1997.

No context found.

A. C. Arpaci-Dusseau, R. H. Arpaci-Dusseau, D. E. Culler, J. M. Hellerstein, and D. A. Patterson. High-performance sorting on networks of workstations. In Proceedings of ACM SIGMOD'97, pages 243--254, 1997.

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