No silver bullet: Extending SDN to the data plane

Anirudh Sivaraman, Keith Winstein, Suvinay Subramanian, Hari Balakrishnan

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

The data plane is in a continuous state of flux. Every few months, researchers publish the design of a new highperformance queueing or scheduling scheme that runs inside the network fabric. Many such schemes have been queen for a day, only to be surpassed soon after as methods - or evaluation metrics - evolve. The lesson, in our view: there will never be a conclusive victor to govern queue management and scheduling inside network hardware. We provide quantitative evidence by demonstrating bidirectional cyclic preferences among three popular contemporary AQM and scheduling configurations. We argue that the way forward requires carefully extending Software-Defined Networking to control the fast-path scheduling and queueing behavior of a switch. To this end, we propose adding a small FPGA to switches. We have synthesized, placed, and routed hardware implementations of CoDel and RED. These schemes require only a few thousand FPGA "slices" to run at 10 Gbps or more - a minuscule fraction of current low-end FPGAs - demonstrating the feasibility and economy of our approach.

Original languageEnglish (US)
Title of host publicationProceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013
DOIs
StatePublished - 2013
Event12th ACM Workshop on Hot Topics in Networks, HotNets 2013 - College Park, MD, United States
Duration: Nov 21 2013Nov 22 2013

Other

Other12th ACM Workshop on Hot Topics in Networks, HotNets 2013
CountryUnited States
CityCollege Park, MD
Period11/21/1311/22/13

Fingerprint

Scheduling
Field programmable gate arrays (FPGA)
Switches
Hardware
Fluxes
Software defined networking

ASJC Scopus subject areas

  • Computer Networks and Communications

Cite this

Sivaraman, A., Winstein, K., Subramanian, S., & Balakrishnan, H. (2013). No silver bullet: Extending SDN to the data plane. In Proceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013 [19] https://doi.org/10.1145/2535771.2535796

No silver bullet : Extending SDN to the data plane. / Sivaraman, Anirudh; Winstein, Keith; Subramanian, Suvinay; Balakrishnan, Hari.

Proceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013. 2013. 19.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Sivaraman, A, Winstein, K, Subramanian, S & Balakrishnan, H 2013, No silver bullet: Extending SDN to the data plane. in Proceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013., 19, 12th ACM Workshop on Hot Topics in Networks, HotNets 2013, College Park, MD, United States, 11/21/13. https://doi.org/10.1145/2535771.2535796
Sivaraman A, Winstein K, Subramanian S, Balakrishnan H. No silver bullet: Extending SDN to the data plane. In Proceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013. 2013. 19 https://doi.org/10.1145/2535771.2535796
Sivaraman, Anirudh ; Winstein, Keith ; Subramanian, Suvinay ; Balakrishnan, Hari. / No silver bullet : Extending SDN to the data plane. Proceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013. 2013.
@inproceedings{aa40d7e2a5414931b97b01148283284a,
title = "No silver bullet: Extending SDN to the data plane",
abstract = "The data plane is in a continuous state of flux. Every few months, researchers publish the design of a new highperformance queueing or scheduling scheme that runs inside the network fabric. Many such schemes have been queen for a day, only to be surpassed soon after as methods - or evaluation metrics - evolve. The lesson, in our view: there will never be a conclusive victor to govern queue management and scheduling inside network hardware. We provide quantitative evidence by demonstrating bidirectional cyclic preferences among three popular contemporary AQM and scheduling configurations. We argue that the way forward requires carefully extending Software-Defined Networking to control the fast-path scheduling and queueing behavior of a switch. To this end, we propose adding a small FPGA to switches. We have synthesized, placed, and routed hardware implementations of CoDel and RED. These schemes require only a few thousand FPGA {"}slices{"} to run at 10 Gbps or more - a minuscule fraction of current low-end FPGAs - demonstrating the feasibility and economy of our approach.",
author = "Anirudh Sivaraman and Keith Winstein and Suvinay Subramanian and Hari Balakrishnan",
year = "2013",
doi = "10.1145/2535771.2535796",
language = "English (US)",
isbn = "9781450325967",
booktitle = "Proceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013",

}

TY - GEN

T1 - No silver bullet

T2 - Extending SDN to the data plane

AU - Sivaraman, Anirudh

AU - Winstein, Keith

AU - Subramanian, Suvinay

AU - Balakrishnan, Hari

PY - 2013

Y1 - 2013

N2 - The data plane is in a continuous state of flux. Every few months, researchers publish the design of a new highperformance queueing or scheduling scheme that runs inside the network fabric. Many such schemes have been queen for a day, only to be surpassed soon after as methods - or evaluation metrics - evolve. The lesson, in our view: there will never be a conclusive victor to govern queue management and scheduling inside network hardware. We provide quantitative evidence by demonstrating bidirectional cyclic preferences among three popular contemporary AQM and scheduling configurations. We argue that the way forward requires carefully extending Software-Defined Networking to control the fast-path scheduling and queueing behavior of a switch. To this end, we propose adding a small FPGA to switches. We have synthesized, placed, and routed hardware implementations of CoDel and RED. These schemes require only a few thousand FPGA "slices" to run at 10 Gbps or more - a minuscule fraction of current low-end FPGAs - demonstrating the feasibility and economy of our approach.

AB - The data plane is in a continuous state of flux. Every few months, researchers publish the design of a new highperformance queueing or scheduling scheme that runs inside the network fabric. Many such schemes have been queen for a day, only to be surpassed soon after as methods - or evaluation metrics - evolve. The lesson, in our view: there will never be a conclusive victor to govern queue management and scheduling inside network hardware. We provide quantitative evidence by demonstrating bidirectional cyclic preferences among three popular contemporary AQM and scheduling configurations. We argue that the way forward requires carefully extending Software-Defined Networking to control the fast-path scheduling and queueing behavior of a switch. To this end, we propose adding a small FPGA to switches. We have synthesized, placed, and routed hardware implementations of CoDel and RED. These schemes require only a few thousand FPGA "slices" to run at 10 Gbps or more - a minuscule fraction of current low-end FPGAs - demonstrating the feasibility and economy of our approach.

UR - http://www.scopus.com/inward/record.url?scp=84893330906&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84893330906&partnerID=8YFLogxK

U2 - 10.1145/2535771.2535796

DO - 10.1145/2535771.2535796

M3 - Conference contribution

AN - SCOPUS:84893330906

SN - 9781450325967

BT - Proceedings of the 12th ACM Workshop on Hot Topics in Networks, HotNets 2013

ER -