Measuring and evaluating TCP splitting for cloud services

Abhinav Pathak; Y. Angela Wang; Cheng Huang; Albert Greenberg; Y. Charlie Hu; Randy Kern; Jin Li; Keith W. Ross

doi:10.1007/978-3-642-12334-4_5

Measuring and evaluating TCP splitting for cloud services

Abhinav Pathak, Y. Angela Wang, Cheng Huang, Albert Greenberg, Y. Charlie Hu, Randy Kern, Jin Li, Keith W. Ross

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

In this paper, we examine the benefits of split-TCP proxies, deployed in an operational world-wide network, for accelerating cloud services. We consider a fraction of a network consisting of a large number of satellite datacenters, which host split-TCP proxies, and a smaller number of mega datacenters, which ultimately perform computation or provide storage. Using web search as an exemplary case study, our detailed measurements reveal that a vanilla TCP splitting solution deployed at the satellite DCs reduces the 95^th percentile of latency by as much as 43% when compared to serving queries directly from the mega DCs. Through careful dissection of the measurement results, we characterize how individual components, including proxy stacks, network protocols, packet losses and network load, can impact the latency. Finally, we shed light on further optimizations that can fully realize the potential of the TCP splitting solution.

Original language	English (US)
Title of host publication	Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings
Editors	Arvind Krishnamurthy, Bernhard Plattner
Publisher	Springer Verlag
Pages	41-50
Number of pages	10
ISBN (Print)	9783642123337
DOIs	https://doi.org/10.1007/978-3-642-12334-4_5
State	Published - 2010
Event	11th International Conference on Passive and Active Network Measurement, PAM 2010 - Zurich, Switzerland Duration: Apr 7 2010 → Apr 9 2010

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	6032 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Other

Other	11th International Conference on Passive and Active Network Measurement, PAM 2010
Country/Territory	Switzerland
City	Zurich
Period	4/7/10 → 4/9/10

ASJC Scopus subject areas

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-642-12334-4_5

Cite this

Pathak, A., Wang, Y. A., Huang, C., Greenberg, A., Hu, Y. C., Kern, R., Li, J., & Ross, K. W. (2010). Measuring and evaluating TCP splitting for cloud services. In A. Krishnamurthy, & B. Plattner (Eds.), Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings (pp. 41-50). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6032 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-642-12334-4_5

Measuring and evaluating TCP splitting for cloud services. / Pathak, Abhinav; Wang, Y. Angela; Huang, Cheng et al.
Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings. ed. / Arvind Krishnamurthy; Bernhard Plattner. Springer Verlag, 2010. p. 41-50 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6032 LNCS).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Pathak, A, Wang, YA, Huang, C, Greenberg, A, Hu, YC, Kern, R, Li, J & Ross, KW 2010, Measuring and evaluating TCP splitting for cloud services. in A Krishnamurthy & B Plattner (eds), Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 6032 LNCS, Springer Verlag, pp. 41-50, 11th International Conference on Passive and Active Network Measurement, PAM 2010, Zurich, Switzerland, 4/7/10. https://doi.org/10.1007/978-3-642-12334-4_5

Pathak A, Wang YA, Huang C, Greenberg A, Hu YC, Kern R et al. Measuring and evaluating TCP splitting for cloud services. In Krishnamurthy A, Plattner B, editors, Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings. Springer Verlag. 2010. p. 41-50. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-642-12334-4_5

Pathak, Abhinav ; Wang, Y. Angela ; Huang, Cheng et al. / Measuring and evaluating TCP splitting for cloud services. Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings. editor / Arvind Krishnamurthy ; Bernhard Plattner. Springer Verlag, 2010. pp. 41-50 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{72bd0b53d1e641fb9b0a5e2a07e06627,

title = "Measuring and evaluating TCP splitting for cloud services",

abstract = "In this paper, we examine the benefits of split-TCP proxies, deployed in an operational world-wide network, for accelerating cloud services. We consider a fraction of a network consisting of a large number of satellite datacenters, which host split-TCP proxies, and a smaller number of mega datacenters, which ultimately perform computation or provide storage. Using web search as an exemplary case study, our detailed measurements reveal that a vanilla TCP splitting solution deployed at the satellite DCs reduces the 95th percentile of latency by as much as 43% when compared to serving queries directly from the mega DCs. Through careful dissection of the measurement results, we characterize how individual components, including proxy stacks, network protocols, packet losses and network load, can impact the latency. Finally, we shed light on further optimizations that can fully realize the potential of the TCP splitting solution.",

author = "Abhinav Pathak and Wang, {Y. Angela} and Cheng Huang and Albert Greenberg and Hu, {Y. Charlie} and Randy Kern and Jin Li and Ross, {Keith W.}",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2010.; 11th International Conference on Passive and Active Network Measurement, PAM 2010 ; Conference date: 07-04-2010 Through 09-04-2010",

year = "2010",

doi = "10.1007/978-3-642-12334-4_5",

language = "English (US)",

isbn = "9783642123337",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "41--50",

editor = "Arvind Krishnamurthy and Bernhard Plattner",

booktitle = "Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings",

}

TY - GEN

T1 - Measuring and evaluating TCP splitting for cloud services

AU - Pathak, Abhinav

AU - Wang, Y. Angela

AU - Huang, Cheng

AU - Greenberg, Albert

AU - Hu, Y. Charlie

AU - Kern, Randy

AU - Li, Jin

AU - Ross, Keith W.

PY - 2010

Y1 - 2010

N2 - In this paper, we examine the benefits of split-TCP proxies, deployed in an operational world-wide network, for accelerating cloud services. We consider a fraction of a network consisting of a large number of satellite datacenters, which host split-TCP proxies, and a smaller number of mega datacenters, which ultimately perform computation or provide storage. Using web search as an exemplary case study, our detailed measurements reveal that a vanilla TCP splitting solution deployed at the satellite DCs reduces the 95th percentile of latency by as much as 43% when compared to serving queries directly from the mega DCs. Through careful dissection of the measurement results, we characterize how individual components, including proxy stacks, network protocols, packet losses and network load, can impact the latency. Finally, we shed light on further optimizations that can fully realize the potential of the TCP splitting solution.

AB - In this paper, we examine the benefits of split-TCP proxies, deployed in an operational world-wide network, for accelerating cloud services. We consider a fraction of a network consisting of a large number of satellite datacenters, which host split-TCP proxies, and a smaller number of mega datacenters, which ultimately perform computation or provide storage. Using web search as an exemplary case study, our detailed measurements reveal that a vanilla TCP splitting solution deployed at the satellite DCs reduces the 95th percentile of latency by as much as 43% when compared to serving queries directly from the mega DCs. Through careful dissection of the measurement results, we characterize how individual components, including proxy stacks, network protocols, packet losses and network load, can impact the latency. Finally, we shed light on further optimizations that can fully realize the potential of the TCP splitting solution.

UR - http://www.scopus.com/inward/record.url?scp=84963791095&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84963791095&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-12334-4_5

DO - 10.1007/978-3-642-12334-4_5

M3 - Conference contribution

AN - SCOPUS:84963791095

SN - 9783642123337

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 41

EP - 50

BT - Passive and Active Measurement - 11th International Conference, PAM 2010, Proceedings

A2 - Krishnamurthy, Arvind

A2 - Plattner, Bernhard

PB - Springer Verlag

T2 - 11th International Conference on Passive and Active Network Measurement, PAM 2010

Y2 - 7 April 2010 through 9 April 2010

ER -

Measuring and evaluating TCP splitting for cloud services

Abstract

Publication series

Other

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this