Skip to main navigation Skip to search Skip to main content

Network-aware end-to-end data throughput optimization

  • SUNY Buffalo

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

17 Scopus citations

Abstract

The rapidly advancing optical networking technology allows us high-bandwidth connectivity up to 100Gbps these days. However, the end-users and their applications can only observe a fraction of this available bandwidth capacity due to inefficient transport protocols and other end-system bottlenecks such as disk I/O limitations, processor speed, and NIC restrictions. In this paper, we present a novel network-aware end-to-end throughput prediction and optimization framework which provides us with the best parameter combination (i.e. parallel stream, disk, and CPU numbers) to achieve the highest end-to-end throughput between two end-systems (i.e. clusters, data centers, parallel disk systems) possible. Our experiments show that the model and algorithm we have developed enable us to achieve close-to-optimal end-to-end throughput performance with negligible sampling and prediction overhead.

Original languageEnglish
Title of host publicationNDM'11 - Proceedings of the 2011 International Workshop on Network-Aware Data Management, Co-located with SC'11
Pages21-30
Number of pages10
DOIs
StatePublished - 2011
EventInternational Workshop on Network-Aware Data Management, NDM'11, Held in Conjunction with the International Conference for High Performance Computing, Networking, Storage and Analysis, SC'11 - Seattle, WA, United States
Duration: Nov 14 2011Nov 14 2011

Publication series

NameNDM'11 - Proceedings of the 2011 International Workshop on Network-Aware Data Management, Co-located with SC'11

Conference

ConferenceInternational Workshop on Network-Aware Data Management, NDM'11, Held in Conjunction with the International Conference for High Performance Computing, Networking, Storage and Analysis, SC'11
Country/TerritoryUnited States
CitySeattle, WA
Period11/14/1111/14/11

Keywords

  • Data parallelism
  • GridFTP
  • Optimization
  • Parallel TCP streams
  • Prediction
  • Striping

Fingerprint

Dive into the research topics of 'Network-aware end-to-end data throughput optimization'. Together they form a unique fingerprint.

Cite this