Mousika: Enable General In-Network Intelligence in Programmable Switches by Knowledge Distillation

Guorui Xie1; Qing Li2; Yutao Dong1; Guanglin Duan1; Yong Jiang1; Jingpu Duan3

10.1109/INFOCOM48880.2022.9796936

2022

41st IEEE Conference on Computer Communications (IEEE INFOCOM)

0743-166X

978-1-6654-5823-8

IEEE INFOCOM 2022 - IEEE Conference on Computer Communications

2022-May

1938-1947

2-5 May 2022

London, United Kingdom

345 E 47TH ST, NEW YORK, NY 10017 USA

IEEE

Given the power efficiency and Tbps throughput of packet processing, several works are proposed to offload the decision tree (DT) to programmable switches, i.e., in-network intelligence. Though the DT is suitable for the switches' match-action paradigm, it has several limitations. E.g., its range match rules may not be supported well due to the hardware diversity; and its implementation also consumes lots of switch resources (e.g., stages and memory). Moreover, as learning algorithms (particularly deep learning) have shown their superior performance, some more complicated learning models are emerging for networking. However, their high computational complexity and large storage requirement are cause challenges in the deployment on switches. Therefore, we propose Mousika, an in-network intelligence framework that addresses these drawbacks successfully. First, we modify the DT to the Binary Decision Tree (BDT). Compared with the DT, our BDT supports faster training, generates fewer rules, and satisfies switch constraints better. Second, we introduce the teacher-student knowledge distillation in Mousika, which enables the general translation from other learning models to the BDT. Through the translation, we can not only utilize the super learning capabilities of complicated models, but also avoid the computation/memory constraints when deploying them on switches directly for line-speed processing.

In-network Intelligence Decision Tree Programmable Switch

其他

英语

EI ; CPCI-S

National Key Research and Development Program of China[2020YFB1804704] ; National Natural Science Foundation of China["61972189","61902171"] ; Shenzhen Key Lab of Software Defined Networking[ZDSYS20140509172959989]

Computer Science ; Engineering ; Telecommunications

Computer Science, Hardware & Architecture ; Engineering, Electrical & Electronic ; Telecommunications

WOS:000936344400195

20222712316904

Binary trees ; Deep learning ; Distillation ; Learning algorithms ; Learning systems ; Personnel training

Ergonomics and Human Factors Engineering:461.4 ; Machine Learning:723.4.2 ; Chemical Operations:802.3 ; Personnel:912.4 ; Combinatorial Mathematics, Includes Graph Theory, Set Theory:921.4 ; Systems Science:961

IEEE

1.Tsinghua University,International Graduate School,Shenzhen,China
2.Peng Cheng Laboratory,Shenzhen,China
3.Southern University of Science and Technology,Shenzhen,China

Guorui Xie,Qing Li,Yutao Dong,et al. Mousika: Enable General In-Network Intelligence in Programmable Switches by Knowledge Distillation[C]. 345 E 47TH ST, NEW YORK, NY 10017 USA:IEEE,2022:1938-1947.