Applying Database Support for Large Scale Data Driven Science in Distributed Environments
GRID '03 Proceedings of the 4th International Workshop on Grid Computing
Computer Networks: A Systems Approach, 3rd Edition
Computer Networks: A Systems Approach, 3rd Edition
The design and implementation of Grid database services in OGSA-DAI: Research Articles
Concurrency and Computation: Practice & Experience - Grid Performance
RPJ: producing fast join results on streams through rate-based optimization
Proceedings of the 2005 ACM SIGMOD international conference on Management of data
Query optimization over web services
VLDB '06 Proceedings of the 32nd international conference on Very large data bases
A critique of recently proposed buffer-sizing strategies
ACM SIGCOMM Computer Communication Review
GridDB: a data-centric overlay for scientific grids
VLDB '04 Proceedings of the Thirtieth international conference on Very large data bases - Volume 30
Robust Runtime Optimization of Data Transfer in Queries over Web Services
ICDE '08 Proceedings of the 2008 IEEE 24th International Conference on Data Engineering
Adaptive Parallelization of Queries over Dependent Web Service Calls
ICDE '09 Proceedings of the 2009 IEEE International Conference on Data Engineering
Adaptive Join Operators for Result Rate Optimization on Streaming Inputs
IEEE Transactions on Knowledge and Data Engineering
| Hi-index | 0.00 |
Modern applications of Web Services (WSs) that involve the processing of large amounts of data tend to transmit data in chunks. Several performance control techniques have been proposed to dynamically select the appropriate chunk size with a view to minimize the communication cost. However, when the data consumer is slower than the data producer, the consumer applications may suffer from memory shortage if high volumes of data arrive in the incoming buffers. To this end, we propose a specific approach to coupling performance control with congestion control features, in order to consider both performance and memory overflow issues in an integrated manner. The performance results with real data show that we can combine these controllers effectively and efficiently, so that no memory overflow occurs at the expense of negligible performance degradation.