D. Carnery et al. Monitoring Streams - A New Class of Data Management Applications. In VLDB, 2002. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Flux: An Adaptive Partitioning Operator for.. - Shah, Hellerstein, .. (2002) (12 citations) (Correct)
....One thrust has been to leverage the commonality amongst queries [6, 16, 7, 2] and index queries like data in a traditional database system. Another direction has been to exploit weaker semantics and sacrifice result quality for cases in which the system cannot be scaled to match peak workloads [18, 4]. While this previous work has focused on single site implementations, we present complementary techniques for parallelizing CQ dataflows that can offer increased scalability to any CQ system. 1.1. Parallelism and Adaptive Partitioning Pipelined CQ dataflows are fundamentally limited by the ....
....utilization imbalances in a memory constrained environment, and a hybrid that is effective in both cases. While Flux can be used with a wide range of lookup based operators, in this paper we focus on its application to operators found in data stream processing systems like Telegraph [5] and Aurora [4]. In this context, we show that the imbalances outlined above can severely degrade the maximum sustainable throughput and the average response time of statically partitioned lookup based operators. By contrast, we show that Flux can achieve several factors improvement in throughput, and orders of ....
[Article contains additional citation context not shown here]
D. Carnery et al. Monitoring Streams - A New Class of Data Management Applications. In VLDB, 2002.
Flux: An Adaptive Partitioning Operator for.. - Shah, Hellerstein, .. (2002) (12 citations) (Correct)
....One thrust has been to leverage the commonality amongst queries [6, 16, 7, 2] and index queries like data in a traditional database system. Another direction has been to exploit weaker semantics and sacrifice result quality for cases in which the system cannot be scaled to match peak workloads [18, 4]. While this previous work has focused on single site implementations, we present complementary techniques for parallelizing CQ dataflows that can offer increased scalability to any CQ system. 1.1. Parallelism and Adaptive Partitioning Pipelined CQ dataflows are fundamentally limited by the ....
....utilization imbalances in a memory constrained environment, and a hybrid that is effective in both cases. While Flux can be used with a wide range of lookup based operators, in this paper we focus on its application to operators found in data stream processing systems like Telegraph [5] and Aurora [4]. In this context, we show that the imbalances outlined above can severely degrade the maximum sustainable throughput and the average response time of statically partitioned lookup based operators. By contrast, we show that Flux can achieve several factors improvement in throughput, and orders of ....
[Article contains additional citation context not shown here]
D. Carnery et al. Monitoring Streams - A New Class of Data Management Applications. In VLDB, 2002.
Highly-Available, Fault-Tolerant, Parallel Dataflows - Shah, Hellerstein, Brewer (2004) (Correct)
No context found.
D. Carnery et al. Monitoring Streams - A New Class of Data Management Applications. In VLDB, 2002.
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