Buffer sizing to reduce interference and increase throughput of real-time stream processing applications

Philip Sebastian Kurtin, S.J. Geuns, J.P.H.M. Hausmans, Marco Jan Gerrit Bekooij

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

6 Citations (Scopus)
41 Downloads (Pure)

Abstract

Existing temporal analysis and buffer sizing techniques for real-time stream processing applications ignore that FIFO buffers bound interference between tasks on the same processor. By considering this effect it can be shown that a reduction of buffer capacities can result in a higher throughput. However, the relation between buffer capacities and throughput is non-monotone in general, which makes an exploitation of the effect challenging. In this paper a buffer sizing approach is presented which exploits that FIFO buffers bound interference between tasks on shared processors. The approach combines temporal analysis using a cyclic dataflow model with computation of buffer capacities in an iterative manner and thereby enables higher throughput guarantees at smaller buffer capacities. It is shown that convergence of the proposed analysis flow is guaranteed. The benefits of the presented approach are demonstrated using a WLAN 802.11p transceiver application executed on a multiprocessor system with shared processors. If buffers without blocking writes are used an up to 25% higher guaranteeable throughput and up to 23% smaller buffer capacities can be determined compared to existing approaches. For systems using buffers with blocking writes the guaranteeable throughput is even up to 43% higher and buffer capacities up to 11% smaller.
Original languageUndefined
Title of host publication2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015
Place of PublicationUSA
PublisherIEEE Computer Society
Pages9-18
Number of pages10
ISBN (Print)1555-0885
DOIs
Publication statusPublished - 13 Apr 2015
Event18th IEEE International Symposium on Object-Oriented Real-Time Distributed Computing, ISORC 2015 - Auckland, New Zealand
Duration: 13 Apr 201517 Apr 2015
Conference number: 18

Publication series

Name
PublisherIEEE Computer Society
ISSN (Print)1555-0885

Conference

Conference18th IEEE International Symposium on Object-Oriented Real-Time Distributed Computing, ISORC 2015
Abbreviated titleISORC
CountryNew Zealand
CityAuckland
Period13/04/1517/04/15

Keywords

  • EWI-25068
  • IR-95740
  • METIS-312448

Cite this

Kurtin, P. S., Geuns, S. J., Hausmans, J. P. H. M., & Bekooij, M. J. G. (2015). Buffer sizing to reduce interference and increase throughput of real-time stream processing applications. In 2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015 (pp. 9-18). USA: IEEE Computer Society. https://doi.org/10.1109/ISORC.2015.14
Kurtin, Philip Sebastian ; Geuns, S.J. ; Hausmans, J.P.H.M. ; Bekooij, Marco Jan Gerrit. / Buffer sizing to reduce interference and increase throughput of real-time stream processing applications. 2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015. USA : IEEE Computer Society, 2015. pp. 9-18
@inproceedings{3fc6a32bd7ce4fb4bdbe0aa1cb1356d4,
title = "Buffer sizing to reduce interference and increase throughput of real-time stream processing applications",
abstract = "Existing temporal analysis and buffer sizing techniques for real-time stream processing applications ignore that FIFO buffers bound interference between tasks on the same processor. By considering this effect it can be shown that a reduction of buffer capacities can result in a higher throughput. However, the relation between buffer capacities and throughput is non-monotone in general, which makes an exploitation of the effect challenging. In this paper a buffer sizing approach is presented which exploits that FIFO buffers bound interference between tasks on shared processors. The approach combines temporal analysis using a cyclic dataflow model with computation of buffer capacities in an iterative manner and thereby enables higher throughput guarantees at smaller buffer capacities. It is shown that convergence of the proposed analysis flow is guaranteed. The benefits of the presented approach are demonstrated using a WLAN 802.11p transceiver application executed on a multiprocessor system with shared processors. If buffers without blocking writes are used an up to 25{\%} higher guaranteeable throughput and up to 23{\%} smaller buffer capacities can be determined compared to existing approaches. For systems using buffers with blocking writes the guaranteeable throughput is even up to 43{\%} higher and buffer capacities up to 11{\%} smaller.",
keywords = "EWI-25068, IR-95740, METIS-312448",
author = "Kurtin, {Philip Sebastian} and S.J. Geuns and J.P.H.M. Hausmans and Bekooij, {Marco Jan Gerrit}",
note = "eemcs-eprint-25068",
year = "2015",
month = "4",
day = "13",
doi = "10.1109/ISORC.2015.14",
language = "Undefined",
isbn = "1555-0885",
publisher = "IEEE Computer Society",
pages = "9--18",
booktitle = "2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015",
address = "United States",

}

Kurtin, PS, Geuns, SJ, Hausmans, JPHM & Bekooij, MJG 2015, Buffer sizing to reduce interference and increase throughput of real-time stream processing applications. in 2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015. IEEE Computer Society, USA, pp. 9-18, 18th IEEE International Symposium on Object-Oriented Real-Time Distributed Computing, ISORC 2015, Auckland, New Zealand, 13/04/15. https://doi.org/10.1109/ISORC.2015.14

Buffer sizing to reduce interference and increase throughput of real-time stream processing applications. / Kurtin, Philip Sebastian; Geuns, S.J.; Hausmans, J.P.H.M.; Bekooij, Marco Jan Gerrit.

2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015. USA : IEEE Computer Society, 2015. p. 9-18.

Research output: Chapter in Book/Report/Conference proceedingConference contributionAcademicpeer-review

TY - GEN

T1 - Buffer sizing to reduce interference and increase throughput of real-time stream processing applications

AU - Kurtin, Philip Sebastian

AU - Geuns, S.J.

AU - Hausmans, J.P.H.M.

AU - Bekooij, Marco Jan Gerrit

N1 - eemcs-eprint-25068

PY - 2015/4/13

Y1 - 2015/4/13

N2 - Existing temporal analysis and buffer sizing techniques for real-time stream processing applications ignore that FIFO buffers bound interference between tasks on the same processor. By considering this effect it can be shown that a reduction of buffer capacities can result in a higher throughput. However, the relation between buffer capacities and throughput is non-monotone in general, which makes an exploitation of the effect challenging. In this paper a buffer sizing approach is presented which exploits that FIFO buffers bound interference between tasks on shared processors. The approach combines temporal analysis using a cyclic dataflow model with computation of buffer capacities in an iterative manner and thereby enables higher throughput guarantees at smaller buffer capacities. It is shown that convergence of the proposed analysis flow is guaranteed. The benefits of the presented approach are demonstrated using a WLAN 802.11p transceiver application executed on a multiprocessor system with shared processors. If buffers without blocking writes are used an up to 25% higher guaranteeable throughput and up to 23% smaller buffer capacities can be determined compared to existing approaches. For systems using buffers with blocking writes the guaranteeable throughput is even up to 43% higher and buffer capacities up to 11% smaller.

AB - Existing temporal analysis and buffer sizing techniques for real-time stream processing applications ignore that FIFO buffers bound interference between tasks on the same processor. By considering this effect it can be shown that a reduction of buffer capacities can result in a higher throughput. However, the relation between buffer capacities and throughput is non-monotone in general, which makes an exploitation of the effect challenging. In this paper a buffer sizing approach is presented which exploits that FIFO buffers bound interference between tasks on shared processors. The approach combines temporal analysis using a cyclic dataflow model with computation of buffer capacities in an iterative manner and thereby enables higher throughput guarantees at smaller buffer capacities. It is shown that convergence of the proposed analysis flow is guaranteed. The benefits of the presented approach are demonstrated using a WLAN 802.11p transceiver application executed on a multiprocessor system with shared processors. If buffers without blocking writes are used an up to 25% higher guaranteeable throughput and up to 23% smaller buffer capacities can be determined compared to existing approaches. For systems using buffers with blocking writes the guaranteeable throughput is even up to 43% higher and buffer capacities up to 11% smaller.

KW - EWI-25068

KW - IR-95740

KW - METIS-312448

U2 - 10.1109/ISORC.2015.14

DO - 10.1109/ISORC.2015.14

M3 - Conference contribution

SN - 1555-0885

SP - 9

EP - 18

BT - 2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015

PB - IEEE Computer Society

CY - USA

ER -

Kurtin PS, Geuns SJ, Hausmans JPHM, Bekooij MJG. Buffer sizing to reduce interference and increase throughput of real-time stream processing applications. In 2015 IEEE 18th International Symposium on Real-Time Distributed Computing, ISORC 2015. USA: IEEE Computer Society. 2015. p. 9-18 https://doi.org/10.1109/ISORC.2015.14