A feedback mechanism for network scheduling in LambdaGrids

Datta, Pallab; Sharma, Sanjeev; Feng, Wu-chun

doi:10.1109/ccgrid.2006.5

Cited by 4 publications

(12 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These approaches have been studied by [Banerjee et al 2006;Datta et al 2006]; however, they cannot accurately estimate the context-switch intervals, as will be proven in Section 5, and can also lead to poor circuit utilization, due to their "stop-and-go" approach.…”

Section: Problem Depiction and Approachesmentioning

confidence: 99%

“…More specifically, we rigorously expose the problems with the latest, and arguably, one of the best-performing algorithms over LambdaGrids called RBUDP + [Datta et al 2006] and propose a new end-system aware, rate-adaptive protocol called RAPID + that addresses these problems.…”

Section: Problem Depiction and Approachesmentioning

confidence: 99%

“…As part of our initial study in [Datta et al 2006], we monitored end-system performance, so as to identify forecasted periods of end-system congestion. By predicting the time at which the receiving end-system OS may allocate a large timeslice to a CPU-intensive process (and hence, not respond to packet-handling interrupts from the NIC), we can approximately estimate when end-system congestion might occur.…”

Section: Problem Depiction and Approachesmentioning

confidence: 99%

“…In [Datta et al 2006], a soft real-time (SRT) process was implemented at the receiving end-system in order to predict periods of end-system congestion and to send explicit feedback notification back to the sender to stop data transfer for a specified duration of time. This was then used to implement a feedback-based, network-scheduling protocol on the Linux 2.6 kernel.…”

Section: Problem Depiction and Approachesmentioning

confidence: 99%

See 3 more Smart Citations

End-System Aware, Rate-Adaptive Protocol for Network Transport in LambdaGrid Environments

Datta

Feng

Sharma

2006

ACM/IEEE SC 2006 Conference (SC'06)

View full text Add to dashboard Cite

Next-generation e-Science applications will require the ability to transfer information at high data rates between distributed computing centers and data repositories. A LambdaGrid offers dedicated, optical, circuit-switched, point-topoint connections that can be reserved exclusively for such applications. These dedicated high-speed connections eliminate network congestion as seen in traditional Internet, but they effectively push the network congestion to the end systems, as processing speeds cannot keep up with networking speeds. Thus, developing an efficient transport protocol over such high-speed dedicated circuits is of critical importance.We propose the idea of a end-system aware, rate-adaptive protocol for network transport, based on end-system performance monitoring. Our proposed protocol significantly improves the performance of data transfer over LambdaGrids by intelligently adapting the sending rate based on end-system constraints. We demonstrate the effectiveness of our proposed protocol and illustrate the performance gains achieved via wide-area network emulation.

show abstract

Section: Problem Depiction and Approachesmentioning

confidence: 99%

Section: Problem Depiction and Approachesmentioning

confidence: 99%

Section: Problem Depiction and Approachesmentioning

confidence: 99%

Section: Problem Depiction and Approachesmentioning

confidence: 99%

See 2 more Smart Citations

End-System Aware, Rate-Adaptive Protocol for Network Transport in LambdaGrid Environments

Datta

Feng

Sharma

2006

ACM/IEEE SC 2006 Conference (SC'06)

View full text Add to dashboard Cite

show abstract

“…Another RBUDP variant, RBUDP+ [3], uses the same scheme but a different estimate of the time and duration for which the receive process has been rescheduled. Like RAPID, the prediction of time and duration is almost impossible.…”

Section: Introductionmentioning

confidence: 99%

Making a Case for Proactive Flow Control in Optical Circuit-Switched Networks

Kumar

Chaube

Balaji

et al. 2008

High Performance Computing - HiPC 2008

View full text Add to dashboard Cite

Abstract. Optical circuit-switched networks such as National LambdaRail (NLR) offer dedicated bandwidth to support large-scale bulk data transfer. Though a dedicated circuit-switched network eliminates congestion from the network itself, it effectively "pushes" the congestion to the end hosts, resulting in lower-than-expected throughput. Previous approaches either use an ad-hoc proactive approach that does not generalize well or a sluggish reactive approach where the sending rate is only adapted based on synchronous feedback from the receiver. We address the shortcomings of such approaches using a two-step process. First, we improve the adaptivity of the reactive approach by proposing an asynchronous, fine-grained, rate-based approach. While this approach enhances performance, its limitation is that it is still reactive. Consequently, we then analyze the predictive patterns of load on the receiver and provide strong evidence that a proactive approach is not only possible, but also necessary, to achieve the best performance in dynamically varying end-host conditions.

show abstract

Grid networks and portals---End-system aware, rate-adaptive protocol for network transport in LambdaGrid environments

Datta

Feng

Sharma

2006

Proceedings of the 2006 ACM/IEEE Conference on Supercomputing - SC '06

View full text Add to dashboard Cite

Next-generation e-Science applications will require the ability to transfer information at high data rates between distributed computing centers and data repositories. A Lambda-Grid offers dedicated, optical, circuit-switched, point-topoint connections that can be reserved exclusively for such applications. These dedicated high-speed connections eliminate network congestion as seen in traditional Internet, but they effectively push the network congestion to the end systems, as processing speeds cannot keep up with networking speeds. Thus, developing an efficient transport protocol over such high-speed dedicated circuits is of critical importance.We propose the idea of a end-system aware, rate-adaptive protocol for network transport, based on end-system performance monitoring. Our proposed protocol significantly improves the performance of data transfer over LambdaGrids by intelligently adapting the sending rate based on end-system constraints. We demonstrate the effectiveness of our proposed protocol and illustrate the performance gains achieved via wide-area network emulation.

show abstract

A feedback mechanism for network scheduling in LambdaGrids

Abstract: Next-generation e-Science applications will require the ability to transfer information at high data rates between distributed computing centers and data repositories. A LambdaGrid offers dedicated, optical, circuit-switched, point-topoint

Cited by 4 publications

References 6 publications

End-System Aware, Rate-Adaptive Protocol for Network Transport in LambdaGrid Environments

End-System Aware, Rate-Adaptive Protocol for Network Transport in LambdaGrid Environments

Making a Case for Proactive Flow Control in Optical Circuit-Switched Networks

Grid networks and portals---End-system aware, rate-adaptive protocol for network transport in LambdaGrid environments

Contact Info

Product

Resources

About