Fine-Grained Pipeline Parallelization for Network Function Programs

Seungbin Song; Heelim Choi; Hanjun Kim

doi:10.1109/CGO51591.2021.9370309

Fine-Grained Pipeline Parallelization for Network Function Programs

Seungbin Song, Heelim Choi, Hanjun Kim

Department of Electrical and Electronic Engineering

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

Abstract

Network programming languages enable programmers to implement new network functions on various hardware and software stacks in the domain of Software Defined Net-working (SDN). Although the languages extend the flexibility of network devices, existing compilers do not fully optimize the network programs due to their coarse-grained parallelization methods. The compilers consider each packet processing table that consists of match and action functions as a unit of tasks and parallelize the programs without decomposing match and action functions. This work proposes a new fine-grained pipeline parallelization compiler for network programming languages, named PSDN. First, the PSDN compiler decouples match and action functions from packet processing tables and analyzes dependencies among the matches and actions. While respecting the dependencies, the compiler efficiently schedules each match and action function into a pipeline with clock cycle estimation and fuses functions to reduce synchronization overheads. This work implements the PSDN compiler that translates a P4 network program to a Xilinx PX program, which is synthesizable to NetFPGA-SUME hardware. The proposed compiler reduces packet processing latency by 12.1 % and utilization by 3.5 % compared to previous work.

Original language	English
Title of host publication	CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization
Editors	Jae W. Lee, Mary Lou Soffa, Ayal Zaks
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	162-173
Number of pages	12
ISBN (Electronic)	9781728186139
DOIs	https://doi.org/10.1109/CGO51591.2021.9370309
Publication status	Published - 2021 Feb 27
Event	19th IEEE/ACM International Symposium on Code Generation and Optimization, CGO 2021 - Virtual, Korea, Korea, Republic of Duration: 2021 Feb 27 → 2021 Mar 3

Publication series

Name	CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization

Conference

Conference	19th IEEE/ACM International Symposium on Code Generation and Optimization, CGO 2021
Country/Territory	Korea, Republic of
City	Virtual, Korea
Period	21/2/27 → 21/3/3

Bibliographical note

Funding Information:
We thank the anonymous reviewers for their valuable feedback. We also thank the CoreLab members for their support and feedback during this work. This work is supported by IITP-2017-0-00195, IITP-2018-0-01392, and IITP-2020-0-01847 through the Institute of Information and Communication Technology Planning and Evaluation (IITP) funded by the Ministry of Science and ICT. Hanjun Kim is the corresponding author of this paper.

Publisher Copyright:
© 2021 IEEE.

All Science Journal Classification (ASJC) codes

Signal Processing
Software
Control and Optimization

Access to Document

10.1109/CGO51591.2021.9370309

Cite this

Song, S., Choi, H., & Kim, H. (2021). Fine-Grained Pipeline Parallelization for Network Function Programs. In J. W. Lee, M. L. Soffa, & A. Zaks (Eds.), CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization (pp. 162-173). [9370309] (CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/CGO51591.2021.9370309

Song, Seungbin ; Choi, Heelim ; Kim, Hanjun. / Fine-Grained Pipeline Parallelization for Network Function Programs. CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization. editor / Jae W. Lee ; Mary Lou Soffa ; Ayal Zaks. Institute of Electrical and Electronics Engineers Inc., 2021. pp. 162-173 (CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization).

@inproceedings{7d77c5d958d141ae9f3a37dd7e3fd256,

title = "Fine-Grained Pipeline Parallelization for Network Function Programs",

abstract = "Network programming languages enable programmers to implement new network functions on various hardware and software stacks in the domain of Software Defined Net-working (SDN). Although the languages extend the flexibility of network devices, existing compilers do not fully optimize the network programs due to their coarse-grained parallelization methods. The compilers consider each packet processing table that consists of match and action functions as a unit of tasks and parallelize the programs without decomposing match and action functions. This work proposes a new fine-grained pipeline parallelization compiler for network programming languages, named PSDN. First, the PSDN compiler decouples match and action functions from packet processing tables and analyzes dependencies among the matches and actions. While respecting the dependencies, the compiler efficiently schedules each match and action function into a pipeline with clock cycle estimation and fuses functions to reduce synchronization overheads. This work implements the PSDN compiler that translates a P4 network program to a Xilinx PX program, which is synthesizable to NetFPGA-SUME hardware. The proposed compiler reduces packet processing latency by 12.1 % and utilization by 3.5 % compared to previous work.",

author = "Seungbin Song and Heelim Choi and Hanjun Kim",

note = "Funding Information: We thank the anonymous reviewers for their valuable feedback. We also thank the CoreLab members for their support and feedback during this work. This work is supported by IITP-2017-0-00195, IITP-2018-0-01392, and IITP-2020-0-01847 through the Institute of Information and Communication Technology Planning and Evaluation (IITP) funded by the Ministry of Science and ICT. Hanjun Kim is the corresponding author of this paper. Publisher Copyright: {\textcopyright} 2021 IEEE.; 19th IEEE/ACM International Symposium on Code Generation and Optimization, CGO 2021 ; Conference date: 27-02-2021 Through 03-03-2021",

year = "2021",

month = feb,

day = "27",

doi = "10.1109/CGO51591.2021.9370309",

language = "English",

series = "CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "162--173",

editor = "Lee, {Jae W.} and Soffa, {Mary Lou} and Ayal Zaks",

booktitle = "CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization",

address = "United States",

}

Song, S, Choi, H & Kim, H 2021, Fine-Grained Pipeline Parallelization for Network Function Programs. in JW Lee, ML Soffa & A Zaks (eds), CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization., 9370309, CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization, Institute of Electrical and Electronics Engineers Inc., pp. 162-173, 19th IEEE/ACM International Symposium on Code Generation and Optimization, CGO 2021, Virtual, Korea, Korea, Republic of, 21/2/27. https://doi.org/10.1109/CGO51591.2021.9370309

Fine-Grained Pipeline Parallelization for Network Function Programs. / Song, Seungbin; Choi, Heelim; Kim, Hanjun.

CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization. ed. / Jae W. Lee; Mary Lou Soffa; Ayal Zaks. Institute of Electrical and Electronics Engineers Inc., 2021. p. 162-173 9370309 (CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization).

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

TY - GEN

T1 - Fine-Grained Pipeline Parallelization for Network Function Programs

AU - Song, Seungbin

AU - Choi, Heelim

AU - Kim, Hanjun

N1 - Funding Information: We thank the anonymous reviewers for their valuable feedback. We also thank the CoreLab members for their support and feedback during this work. This work is supported by IITP-2017-0-00195, IITP-2018-0-01392, and IITP-2020-0-01847 through the Institute of Information and Communication Technology Planning and Evaluation (IITP) funded by the Ministry of Science and ICT. Hanjun Kim is the corresponding author of this paper. Publisher Copyright: © 2021 IEEE.

PY - 2021/2/27

Y1 - 2021/2/27

N2 - Network programming languages enable programmers to implement new network functions on various hardware and software stacks in the domain of Software Defined Net-working (SDN). Although the languages extend the flexibility of network devices, existing compilers do not fully optimize the network programs due to their coarse-grained parallelization methods. The compilers consider each packet processing table that consists of match and action functions as a unit of tasks and parallelize the programs without decomposing match and action functions. This work proposes a new fine-grained pipeline parallelization compiler for network programming languages, named PSDN. First, the PSDN compiler decouples match and action functions from packet processing tables and analyzes dependencies among the matches and actions. While respecting the dependencies, the compiler efficiently schedules each match and action function into a pipeline with clock cycle estimation and fuses functions to reduce synchronization overheads. This work implements the PSDN compiler that translates a P4 network program to a Xilinx PX program, which is synthesizable to NetFPGA-SUME hardware. The proposed compiler reduces packet processing latency by 12.1 % and utilization by 3.5 % compared to previous work.

AB - Network programming languages enable programmers to implement new network functions on various hardware and software stacks in the domain of Software Defined Net-working (SDN). Although the languages extend the flexibility of network devices, existing compilers do not fully optimize the network programs due to their coarse-grained parallelization methods. The compilers consider each packet processing table that consists of match and action functions as a unit of tasks and parallelize the programs without decomposing match and action functions. This work proposes a new fine-grained pipeline parallelization compiler for network programming languages, named PSDN. First, the PSDN compiler decouples match and action functions from packet processing tables and analyzes dependencies among the matches and actions. While respecting the dependencies, the compiler efficiently schedules each match and action function into a pipeline with clock cycle estimation and fuses functions to reduce synchronization overheads. This work implements the PSDN compiler that translates a P4 network program to a Xilinx PX program, which is synthesizable to NetFPGA-SUME hardware. The proposed compiler reduces packet processing latency by 12.1 % and utilization by 3.5 % compared to previous work.

UR - http://www.scopus.com/inward/record.url?scp=85103181354&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85103181354&partnerID=8YFLogxK

U2 - 10.1109/CGO51591.2021.9370309

DO - 10.1109/CGO51591.2021.9370309

M3 - Conference contribution

AN - SCOPUS:85103181354

T3 - CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization

SP - 162

EP - 173

BT - CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization

A2 - Lee, Jae W.

A2 - Soffa, Mary Lou

A2 - Zaks, Ayal

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 19th IEEE/ACM International Symposium on Code Generation and Optimization, CGO 2021

Y2 - 27 February 2021 through 3 March 2021

ER -

Song S, Choi H, Kim H. Fine-Grained Pipeline Parallelization for Network Function Programs. In Lee JW, Soffa ML, Zaks A, editors, CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization. Institute of Electrical and Electronics Engineers Inc. 2021. p. 162-173. 9370309. (CGO 2021 - Proceedings of the 2021 IEEE/ACM International Symposium on Code Generation and Optimization). doi: 10.1109/CGO51591.2021.9370309

Fine-Grained Pipeline Parallelization for Network Function Programs

Abstract

Publication series

Conference

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this