Chialin Chang, Tahsin Kurc, Alan Sussman, Joel Saltz, Optimizing Retrieval and Processing of Multi-dimensional Scientific Datasets, In Proceedings of the Third Merged IPPS/SPDP Symposiums. IEEE Computer Society Press, May 2000  Home/Search   Document Details and Download   Summary   Related Articles   Check

....Division sponse to increases or decreases in client request loads. 1 Operational Data Services and Servers Server systems structured as cluster machines have become increasingly common, to drive search engines [1] operate mail servers [2] or provide scientific data or computational services [3, 4]. Our research concerns one important property of such cluster servers, which is their ability to continue to provide high levels of service even under increasing loads or when clients request behaviors vary dynamically. Operational Information Systems. In contrast to the interactive high ....

M. Beynon, T. Kurc, A. Sussman, and J. Saltz, "Optimizing Retrieval and Processing of Multi-dimensional Scientific Datasets", In Proceedings of the Third Merged IPPS/SPDP Symposiums, Cancun, Mexico, May 2000.

....are typically not designed to be coupled. Additionally, attempting to modify the internal communications scheme of these scalable codes is not only difficult but could possibly affect their validity. Existing systems like Globus[1] and Legion[2] CORBA[3] KeLP[4] and the Active Data Repository[5] provide data exchange mechanisms among their services. Systems like POLYLITH[6] Darwin[7] and Olan[8] provide a software module interconnection framework. While these systems have met with varying amounts of success, they are not currently sufficient for our purposes since implementation can ....

Chialin Chang, Kurc, T., Sussman, A., Saltz, J., Optimizing retrieval and processing of multi-dimensional scientific datasets, Proceedings of 14 th international Parallel and Distributed Symposium, 2000. pp. 405-410, May 2000

No context found.

C. Chang, T. Kurc, A. Sussman, and J. Saltz. Optimizing retrieval and processing of multi-dimensional scientific datasets. In Proceedings of the Third Merged IPPS/SPDP (14th International Parallel Processing Symposium & 11th Symposium on Parallel and Distributed Processing). IEEE Computer Society Press, May 2000.

....of filters P, F, and C. Figure 2: DataCutter stream abstraction and support for copies. gramming model. DataCutter provides support for developing applications that execute generalized reduction operations, which are common in the data processing kernel of many data analysis applications [6, 14]. This type of processing consists of retrieving the data of interest and performing user defined transformation, mapping, and aggregation operations on the data. The transformation and aggregation functions in generalized reduction operations are associative and commutative. That is, the result ....

C. Chang, T. Kurc, A. Sussman, and J. Saltz. Optimizing retrieval and processing of multi-dimensional scientific datasets. In Proceedings of the Third Merged IPPS/SPDP (14th International Parallel Processing Symposium & 11th Symposium on Parallel and Distributed Processing). IEEE Computer Society Press, May 2000.

....on a parallel machine or a cluster of workstations employing an efficient software system for providing the desired functionality. To aid in efficiently implementing storage and processing of multi perspective images, we employ an object oriented framework called the Active Data Repository (ADR) [3, 4, 7, 8], that has been developed at the University of Maryland for managing and processing large amounts of scientific data on a parallel or distributed system. In this paper we describe how to customize ADR for building an efficient and flexible framework for storing multi perspective image data and ....

C. Chang, T. Kurc, A. Sussman, and J. Saltz. Optimizing retrieval and processing of multi-dimensional scientific datasets. In Proceedings of the Third Merged IPPS/SPDP (14th International Parallel Processing Symposium & 11th Symposium on Parallel and Distributed Processing). IEEE Computer Society Press, Los Alamitos, Calif., May 2000.

....the speedup for five server processors is 3.6 compared to a one processor server. 12 3.3 Query Processing Strategies Workload partitioning and tiling have significant effects on the performance of an application implemented using the ADR framework. We have evaluated several potential strategies [22,23,43] that use different workload partitioning and tiling schemes. To simplify the presentation, we assume that the target range query involves only one input and one output dataset. Both the input and output datasets are assumed to be already partitioned into data chunks and declustered across the ....

C. Chang, T. Kurc, A. Sussman, and J. Saltz. Optimizing retrieval and processing of multi-dimensional scientific datasets. In Proceedings of the Third Merged IPPS/SPDP (14th International Parallel Processing Symposium & 11th Symposium on Parallel and Distributed Processing). IEEE Computer Society Press, May 2000.

....i i i i i i i i 2 gation functions are commutative and associative [9] That is, the correctness of the aggregation operation does not depend on the order the input data items are aggregated. In earlier work, weinvestigated three strategies for distributing the workload among processors [4,9]. The Distributed Accumulator (DA) strategy assigns the processing for an entire output element to a single processor. The output elements are partitioned across the processors and each processor carries out aggregation operations on the local output elements. Input elements are communicated to ....

....exceeding the available memory on any processor. This step is referred to as tiling.Workload partitioning is performed for each output tile. Once a partitioning of the workload is determined, processing of a tile starts with allocating in memory the output chunks assigned to the given output tile [4,9]. Output chunks that need to be replicated on multiple processors (e.g. for the RA strategies) are read bytheowner processor and forwarded to the processors that require them. After the initialization step, input chunks are read by their owner processors and aggregated with the allocated output ....

[Article contains additional citation context not shown here]

C. Chang, T. Kurc, A. Sussman, and J. Saltz. Optimizing retrieval and processing of multi-dimensional scientific datasets. In Proceedings of the ThirdMerged IPPS/SPDP Symposiums. IEEE Computer Society Press, May 2000.

No context found.

Chialin Chang, Tahsin Kurc, Alan Sussman, Joel Saltz, Optimizing Retrieval and Processing of Multi-dimensional Scientific Datasets, In Proceedings of the Third Merged IPPS/SPDP Symposiums. IEEE Computer Society Press, May 2000

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