High throughput filtering using FPGA-acceleration

Wim Vanderbauwhede, Anton Frolov, Leif Azzopardi, Sai Rahul Chalamalasetti, Martin Margala

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

Abstract

With the rise in the amount information of being streamed across networks, there is a growing demand to vet the quality, type and content itself for various purposes such as spam, security and search. In this paper, we develop an energy-efficient high performance information filtering system that is capable of classifying a stream of incoming document at high speed. The prototype parses a stream of documents using a multicore CPU and then performs classification using Field-Programmable Gate Arrays (FPGAs). On a large TREC data collection, we implemented a Naive Bayes classifier on our prototype and compared it to an optimized CPU based-baseline. Our empirical findings show that we can classify documents at 10Gb/s which is up to 94 times faster than the CPU baseline (and up to 5 times faster than previous FPGA based implementations). In future work, we aim to increase the throughput by another order of magnitude by implementing both the parser and filter on the FPGA.
Original languageEnglish
Title of host publicationCIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management
Place of PublicationNew York, NY, USA
Pages1245-1248
Number of pages4
DOIs
Publication statusPublished - 27 Oct 2013

Fingerprint

information system
energy
demand
performance

Keywords

  • classification
  • parsing
  • fpga
  • filtering

Cite this

Vanderbauwhede, W., Frolov, A., Azzopardi, L., Chalamalasetti, S. R., & Margala, M. (2013). High throughput filtering using FPGA-acceleration. In CIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management (pp. 1245-1248). New York, NY, USA. https://doi.org/10.1145/2505515.2507866
Vanderbauwhede, Wim ; Frolov, Anton ; Azzopardi, Leif ; Chalamalasetti, Sai Rahul ; Margala, Martin. / High throughput filtering using FPGA-acceleration. CIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management. New York, NY, USA, 2013. pp. 1245-1248
@inproceedings{877a2514536c4c53a49b5cca691efa09,
title = "High throughput filtering using FPGA-acceleration",
abstract = "With the rise in the amount information of being streamed across networks, there is a growing demand to vet the quality, type and content itself for various purposes such as spam, security and search. In this paper, we develop an energy-efficient high performance information filtering system that is capable of classifying a stream of incoming document at high speed. The prototype parses a stream of documents using a multicore CPU and then performs classification using Field-Programmable Gate Arrays (FPGAs). On a large TREC data collection, we implemented a Naive Bayes classifier on our prototype and compared it to an optimized CPU based-baseline. Our empirical findings show that we can classify documents at 10Gb/s which is up to 94 times faster than the CPU baseline (and up to 5 times faster than previous FPGA based implementations). In future work, we aim to increase the throughput by another order of magnitude by implementing both the parser and filter on the FPGA.",
keywords = "classification, parsing, fpga, filtering",
author = "Wim Vanderbauwhede and Anton Frolov and Leif Azzopardi and Chalamalasetti, {Sai Rahul} and Martin Margala",
year = "2013",
month = "10",
day = "27",
doi = "10.1145/2505515.2507866",
language = "English",
isbn = "978-1-4503-2263-8",
pages = "1245--1248",
booktitle = "CIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management",

}

Vanderbauwhede, W, Frolov, A, Azzopardi, L, Chalamalasetti, SR & Margala, M 2013, High throughput filtering using FPGA-acceleration. in CIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management. New York, NY, USA, pp. 1245-1248. https://doi.org/10.1145/2505515.2507866

High throughput filtering using FPGA-acceleration. / Vanderbauwhede, Wim; Frolov, Anton; Azzopardi, Leif; Chalamalasetti, Sai Rahul; Margala, Martin.

CIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management. New York, NY, USA, 2013. p. 1245-1248.

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

TY - GEN

T1 - High throughput filtering using FPGA-acceleration

AU - Vanderbauwhede, Wim

AU - Frolov, Anton

AU - Azzopardi, Leif

AU - Chalamalasetti, Sai Rahul

AU - Margala, Martin

PY - 2013/10/27

Y1 - 2013/10/27

N2 - With the rise in the amount information of being streamed across networks, there is a growing demand to vet the quality, type and content itself for various purposes such as spam, security and search. In this paper, we develop an energy-efficient high performance information filtering system that is capable of classifying a stream of incoming document at high speed. The prototype parses a stream of documents using a multicore CPU and then performs classification using Field-Programmable Gate Arrays (FPGAs). On a large TREC data collection, we implemented a Naive Bayes classifier on our prototype and compared it to an optimized CPU based-baseline. Our empirical findings show that we can classify documents at 10Gb/s which is up to 94 times faster than the CPU baseline (and up to 5 times faster than previous FPGA based implementations). In future work, we aim to increase the throughput by another order of magnitude by implementing both the parser and filter on the FPGA.

AB - With the rise in the amount information of being streamed across networks, there is a growing demand to vet the quality, type and content itself for various purposes such as spam, security and search. In this paper, we develop an energy-efficient high performance information filtering system that is capable of classifying a stream of incoming document at high speed. The prototype parses a stream of documents using a multicore CPU and then performs classification using Field-Programmable Gate Arrays (FPGAs). On a large TREC data collection, we implemented a Naive Bayes classifier on our prototype and compared it to an optimized CPU based-baseline. Our empirical findings show that we can classify documents at 10Gb/s which is up to 94 times faster than the CPU baseline (and up to 5 times faster than previous FPGA based implementations). In future work, we aim to increase the throughput by another order of magnitude by implementing both the parser and filter on the FPGA.

KW - classification

KW - parsing

KW - fpga

KW - filtering

U2 - 10.1145/2505515.2507866

DO - 10.1145/2505515.2507866

M3 - Conference contribution book

SN - 978-1-4503-2263-8

SP - 1245

EP - 1248

BT - CIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management

CY - New York, NY, USA

ER -

Vanderbauwhede W, Frolov A, Azzopardi L, Chalamalasetti SR, Margala M. High throughput filtering using FPGA-acceleration. In CIKM '13 Proceedings of the 22nd ACM International Conference on Information & Knowledge Management. New York, NY, USA. 2013. p. 1245-1248 https://doi.org/10.1145/2505515.2507866