Relative fitness modeling

Mesnier, Michael P.; Wachs, Matthew; Sambasivan, Raja R.; Zheng, Alice X.; Ganger, Gregory R.

doi:10.1145/1498765.1498789

Cited by 6 publications

(4 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…More elaborate techniques are possible, but we leave them for future work. For example, we can use techniques such as those described in [21] to evolve models collected for one system and database configuration so that these models are accurate for another system and database configuration.…”

Section: Discussion and Extensionsmentioning

confidence: 99%

Predicting completion times of batch query workloads using interaction-aware models and simulation

Ahmad

Duan

Aboulnaga

et al. 2011

Proceedings of the 14th International Conference on Extending Database Technology

View full text Add to dashboard Cite

A question that database administrators (DBAs) routinely need to answer is how long a batch query workload will take to complete. This question arises, for example, while planning the execution of different report-generation workloads to fit within available time windows. To answer this question accurately, we need to take into account that the typical workload in a database system consists of mixes of concurrent queries. Interactions among different queries in these mixes need to be modeled, rather than the conventional approach of considering each query separately. This paper presents a new approach for estimating workload completion times that takes the significant impact of query interactions into account. This approach builds performance models using an experimentdriven technique, by sampling the space of possible query mixes and fitting statistical models to the observed performance at these samples. No prior assumptions are made about the internal workings of the database system or the cause of query interactions, making the models robust and portable. We show that a careful choice of sampling and statistical modeling strategies can result in accurate models, and we present a novel interaction-aware workload simulator that uses these models to estimate workload completion times. An experimental evaluation with complex TPC-H queries on IBM DB2 shows that this approach consistently predicts workload completion times with less than 20% error.

show abstract

Section: Discussion and Extensionsmentioning

confidence: 99%

Predicting completion times of batch query workloads using interaction-aware models and simulation

Ahmad

Duan

Aboulnaga

et al. 2011

Proceedings of the 14th International Conference on Extending Database Technology

View full text Add to dashboard Cite

show abstract

“…on I/O trace generation and replay [26]. Neither of these scale as they do not perform any compression.…”

Section: Related Workmentioning

confidence: 99%

Preserving time in large-scale communication traces

Ratn

Mueller

Supinski

et al. 2008

Proceedings of the 22nd Annual International Conference on Supercomputing

View full text Add to dashboard Cite

“…These closed performance correlations of different factors can inspire us to develop some relative performance model like Ref. [27] [28].…”

Section: ) Relational Analysismentioning

confidence: 99%

Performance Evaluation of Parallel File System Based on Lustre and Grey Theory

Zhao

2010

2010 Ninth International Conference on Grid and Cloud Computing

View full text Add to dashboard Cite

High Performance Computing (HPC) system need to be coupled with efficient parallel file systems, such as Lustre file system, that can deliver commensurate IO throughput to scientific applications. It is important to gain insights into the deliverable parallel file system IO efficiency. In order to gain a good understanding on what and how to impact the performance of parallel file systems. This paper presents a study on performance evaluation of parallel file systems using Lustre file system. We conduct an in-depth survey on the basic performance factors of Lustre. Based on this survey, a series of test cases are designed to validate the performance of Lustre and we adopt relational analysis model and grey prediction model to analyze and predict the performance changes. In our relational analysis, we find that the performance of Lustre has a more closed correlation when performance factors change. Our prediction results indicate that our prediction model can obtain better prediction precision and could be further applied to performance evaluation of other parallel file systems.

show abstract

Relative fitness modeling

Cited by 6 publications

References 8 publications

Predicting completion times of batch query workloads using interaction-aware models and simulation

Predicting completion times of batch query workloads using interaction-aware models and simulation

Preserving time in large-scale communication traces

Performance Evaluation of Parallel File System Based on Lustre and Grey Theory

Contact Info

Product

Resources

About