Accelerating online change-point detection algorithm using 10 GbE FPGA NIC

Takuma Iwata, Kohei Nakamura, Yuta Tokusashi, Hiroki Matsutani

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

Abstract

In statistical analysis and data mining, change-point detection that identifies the change-points which are times when the probability distribution of time series changes has been used for various purposes, such as anomaly detections on network traffic and transaction data. However, computation cost of a conventional AR (Auto-Regression) model based approach is too high and infeasible for online. In this paper, an AR model based online change-point detection algorithm, called ChangeFinder, is implemented on an FPGA (Field Programmable Gate Array) based NIC (Network Interface Card). The proposed system computes the change-point score from time series data received from 10 GbE (10 Gbit Ethernet). More specifically, it computes the change-point score at the 10 GbE NIC in advance of host applications. This paper aims to reduce the host workload and improve change-point detection performance by offloading ChangeFinder algorithm from host to the NIC. As evaluations, change-point detection in the FPGA NIC is compared with a baseline software implementation and those enhanced by two network optimization techniques using DPDK and Netfilter in terms of throughput. The result demonstrates 16.8x improvement in change-point detection throughput compared to the baseline software implementation. The throughput achieves 83.4% of the 10 GbE line rate.

Original languageEnglish
Title of host publicationEuro-Par 2018
Subtitle of host publicationParallel Processing Workshops - Euro-Par 2018 International Workshops, Revised Selected Papers
EditorsGabriele Mencagli, Dora B. Heras
PublisherSpringer Verlag
Pages506-517
Number of pages12
ISBN (Print)9783030105488
DOIs
Publication statusPublished - 2019 Jan 1
Event24th International Conference on Parallel and Distributed Computing, Euro-Par 2018 - Turin, Italy
Duration: 2018 Aug 272018 Aug 28

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11339 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference24th International Conference on Parallel and Distributed Computing, Euro-Par 2018
CountryItaly
CityTurin
Period18/8/2718/8/28

Fingerprint

Change-point Detection
Field Programmable Gate Array
Interfaces (computer)
Field programmable gate arrays (FPGA)
Ethernet
Change Point
Throughput
Autoregression
Time series
Baseline
Regression Model
Model-based
Network Optimization
Software
Probability distributions
Anomaly Detection
Data mining
Network Traffic
Statistical methods
Time Series Data

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Iwata, T., Nakamura, K., Tokusashi, Y., & Matsutani, H. (2019). Accelerating online change-point detection algorithm using 10 GbE FPGA NIC. In G. Mencagli, & D. B. Heras (Eds.), Euro-Par 2018: Parallel Processing Workshops - Euro-Par 2018 International Workshops, Revised Selected Papers (pp. 506-517). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11339 LNCS). Springer Verlag. https://doi.org/10.1007/978-3-030-10549-5_40

Accelerating online change-point detection algorithm using 10 GbE FPGA NIC. / Iwata, Takuma; Nakamura, Kohei; Tokusashi, Yuta; Matsutani, Hiroki.

Euro-Par 2018: Parallel Processing Workshops - Euro-Par 2018 International Workshops, Revised Selected Papers. ed. / Gabriele Mencagli; Dora B. Heras. Springer Verlag, 2019. p. 506-517 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11339 LNCS).

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

Iwata, T, Nakamura, K, Tokusashi, Y & Matsutani, H 2019, Accelerating online change-point detection algorithm using 10 GbE FPGA NIC. in G Mencagli & DB Heras (eds), Euro-Par 2018: Parallel Processing Workshops - Euro-Par 2018 International Workshops, Revised Selected Papers. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11339 LNCS, Springer Verlag, pp. 506-517, 24th International Conference on Parallel and Distributed Computing, Euro-Par 2018, Turin, Italy, 18/8/27. https://doi.org/10.1007/978-3-030-10549-5_40
Iwata T, Nakamura K, Tokusashi Y, Matsutani H. Accelerating online change-point detection algorithm using 10 GbE FPGA NIC. In Mencagli G, Heras DB, editors, Euro-Par 2018: Parallel Processing Workshops - Euro-Par 2018 International Workshops, Revised Selected Papers. Springer Verlag. 2019. p. 506-517. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-030-10549-5_40
Iwata, Takuma ; Nakamura, Kohei ; Tokusashi, Yuta ; Matsutani, Hiroki. / Accelerating online change-point detection algorithm using 10 GbE FPGA NIC. Euro-Par 2018: Parallel Processing Workshops - Euro-Par 2018 International Workshops, Revised Selected Papers. editor / Gabriele Mencagli ; Dora B. Heras. Springer Verlag, 2019. pp. 506-517 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{c17f4d562e2f495987125732aefdda82,
title = "Accelerating online change-point detection algorithm using 10 GbE FPGA NIC",
abstract = "In statistical analysis and data mining, change-point detection that identifies the change-points which are times when the probability distribution of time series changes has been used for various purposes, such as anomaly detections on network traffic and transaction data. However, computation cost of a conventional AR (Auto-Regression) model based approach is too high and infeasible for online. In this paper, an AR model based online change-point detection algorithm, called ChangeFinder, is implemented on an FPGA (Field Programmable Gate Array) based NIC (Network Interface Card). The proposed system computes the change-point score from time series data received from 10 GbE (10 Gbit Ethernet). More specifically, it computes the change-point score at the 10 GbE NIC in advance of host applications. This paper aims to reduce the host workload and improve change-point detection performance by offloading ChangeFinder algorithm from host to the NIC. As evaluations, change-point detection in the FPGA NIC is compared with a baseline software implementation and those enhanced by two network optimization techniques using DPDK and Netfilter in terms of throughput. The result demonstrates 16.8x improvement in change-point detection throughput compared to the baseline software implementation. The throughput achieves 83.4{\%} of the 10 GbE line rate.",
author = "Takuma Iwata and Kohei Nakamura and Yuta Tokusashi and Hiroki Matsutani",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-030-10549-5_40",
language = "English",
isbn = "9783030105488",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "506--517",
editor = "Gabriele Mencagli and Heras, {Dora B.}",
booktitle = "Euro-Par 2018",
address = "Germany",

}

TY - GEN

T1 - Accelerating online change-point detection algorithm using 10 GbE FPGA NIC

AU - Iwata, Takuma

AU - Nakamura, Kohei

AU - Tokusashi, Yuta

AU - Matsutani, Hiroki

PY - 2019/1/1

Y1 - 2019/1/1

N2 - In statistical analysis and data mining, change-point detection that identifies the change-points which are times when the probability distribution of time series changes has been used for various purposes, such as anomaly detections on network traffic and transaction data. However, computation cost of a conventional AR (Auto-Regression) model based approach is too high and infeasible for online. In this paper, an AR model based online change-point detection algorithm, called ChangeFinder, is implemented on an FPGA (Field Programmable Gate Array) based NIC (Network Interface Card). The proposed system computes the change-point score from time series data received from 10 GbE (10 Gbit Ethernet). More specifically, it computes the change-point score at the 10 GbE NIC in advance of host applications. This paper aims to reduce the host workload and improve change-point detection performance by offloading ChangeFinder algorithm from host to the NIC. As evaluations, change-point detection in the FPGA NIC is compared with a baseline software implementation and those enhanced by two network optimization techniques using DPDK and Netfilter in terms of throughput. The result demonstrates 16.8x improvement in change-point detection throughput compared to the baseline software implementation. The throughput achieves 83.4% of the 10 GbE line rate.

AB - In statistical analysis and data mining, change-point detection that identifies the change-points which are times when the probability distribution of time series changes has been used for various purposes, such as anomaly detections on network traffic and transaction data. However, computation cost of a conventional AR (Auto-Regression) model based approach is too high and infeasible for online. In this paper, an AR model based online change-point detection algorithm, called ChangeFinder, is implemented on an FPGA (Field Programmable Gate Array) based NIC (Network Interface Card). The proposed system computes the change-point score from time series data received from 10 GbE (10 Gbit Ethernet). More specifically, it computes the change-point score at the 10 GbE NIC in advance of host applications. This paper aims to reduce the host workload and improve change-point detection performance by offloading ChangeFinder algorithm from host to the NIC. As evaluations, change-point detection in the FPGA NIC is compared with a baseline software implementation and those enhanced by two network optimization techniques using DPDK and Netfilter in terms of throughput. The result demonstrates 16.8x improvement in change-point detection throughput compared to the baseline software implementation. The throughput achieves 83.4% of the 10 GbE line rate.

UR - http://www.scopus.com/inward/record.url?scp=85061694776&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85061694776&partnerID=8YFLogxK

U2 - 10.1007/978-3-030-10549-5_40

DO - 10.1007/978-3-030-10549-5_40

M3 - Conference contribution

AN - SCOPUS:85061694776

SN - 9783030105488

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

SP - 506

EP - 517

BT - Euro-Par 2018

A2 - Mencagli, Gabriele

A2 - Heras, Dora B.

PB - Springer Verlag

ER -