Automated SmartNIC Offloading Insights for Network Functions

被引:21
作者
Qiu, Yiming [1 ]
Xing, Jiarong [1 ]
Hsu, Kuo-Feng [1 ]
Kang, Qiao [1 ]
Liu, Ming [2 ]
Narayana, Srinivas [3 ]
Chen, Ang [1 ]
机构
[1] Rice Univ, Houston, TX 77005 USA
[2] UW Madison, VMware, Madison, WI USA
[3] Rutgers State Univ, New Brunswick, NJ USA
来源
PROCEEDINGS OF THE 28TH ACM SYMPOSIUM ON OPERATING SYSTEMS PRINCIPLES, SOSP 2021 | 2021年
关键词
Network function; SmartNIC; Machine learning; MODEL;
D O I
10.1145/3477132.3483583
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
The gap between CPU and networking speeds has motivated the development of SmartNICs for NF (network functions) offloading. However, offloading performance is predicated upon intricate knowledge about SmartNIC hardware and careful hand-tuning of the ported programs. Today, developers cannot easily reason about the offloading performance or the effectiveness of different porting strategies without resorting to a trial-and-error approach. Clara is an automated tool that improves the productivity of this workflow by generating offloading insights. Our tool can a) analyze a legacy NF in its unported form, predicting its performance characteristics on a SmartNIC (e.g., compute vs. memory intensity); and b) explore and suggest porting strategies for the given NF to achieve higher performance. To achieve these goals, Clara uses program analysis techniques to extract NF features, and combines them with machine learning techniques to handle opaque SmartNIC details. Our evaluation using Click NF programs on a Netronome Smart-NIC shows that Clara achieves high accuracy in its analysis, and that its suggested porting strategies lead to significant performance improvements.
引用
收藏
页码:772 / 787
页数:16
相关论文
共 71 条
  • [31] Gables: A Roofline Model for Mobile SoCs
    Hill, Mark D.
    Reddi, Vijay Janapa
    [J]. 2019 25TH IEEE INTERNATIONAL SYMPOSIUM ON HIGH PERFORMANCE COMPUTER ARCHITECTURE (HPCA), 2019, : 317 - 330
  • [32] Hong S, 2010, CONF PROC INT SYMP C, P280, DOI 10.1145/1816038.1815998
  • [33] Hong S, 2009, CONF PROC INT SYMP C, P152, DOI 10.1145/1555815.1555775
  • [34] ieee802, 2018, IEEE P802.3bs 400 GbE Task Force. Adopted Timeline
  • [35] Iyer R, 2019, PROCEEDINGS OF THE 16TH USENIX SYMPOSIUM ON NETWORKED SYSTEMS DESIGN AND IMPLEMENTATION, P517
  • [36] Katsikas GP, 2018, PROCEEDINGS OF THE 15TH USENIX SYMPOSIUM ON NETWORKED SYSTEMS DESIGN AND IMPLEMENTATION (NSDI'18), P171
  • [37] High Performance Packet Processing with FlexNIC
    Kaufmann, Antoine
    Peter, Simon
    Sharma, Naveen Kr.
    Anderson, Thomas
    Krishnamurthy, Arvind
    [J]. ACM SIGPLAN NOTICES, 2016, 51 (04) : 67 - 81
  • [38] Lattner C, 2004, INT SYM CODE GENER, P75, DOI 10.1109/CGO.2004.1281665
  • [39] KV-Direct: High-Performance In-Memory Key-Value Store with Programmable NIC
    Li, Bojie
    Ruan, Zhenyuan
    Xiao, Wencong
    Lu, Yuanwei
    Xiong, Yongqiang
    Putnam, Andrew
    Chen, Enhong
    Zhang, Lintao
    [J]. PROCEEDINGS OF THE TWENTY-SIXTH ACM SYMPOSIUM ON OPERATING SYSTEMS PRINCIPLES (SOSP '17), 2017, : 137 - 152
  • [40] Liu M, 2019, PROCEEDINGS OF THE 2019 USENIX ANNUAL TECHNICAL CONFERENCE, P363