In-switch approximate processing

Delayed tasks management for MapReduce applications

Koya Mitsuzuka, Ami Hayashi, Michihiro Koibuchi, Hideharu Amano, Hiroki Matsutani

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

1 Citation (Scopus)

Abstract

In MapReduce, the parallel processing performance is often limited by only a few compute nodes that delay to complete given tasks. Although various techniques have been invented to handle such stragglers, these techniques mostly impose a burden on master node to monitor the progress of all the compute nodes, resulting in a new bottleneck as the number of compute nodes increases. As an alternative approach, in this paper, we propose to move such straggler management burden from master node to network switch that connects the master and compute nodes, because all the information goes through the switch. More specifically, the proposed network switch monitors output packets from Map tasks to detect stragglers. When detected, the proposed switch generates a response instead of the straggler based on the outputs of the other normal Map tasks, so that Reduce tasks can be started without delay. We introduce some approximate techniques for the proxy computation and response at the switch; thus our switch is called 'ApproxSW.' We implement ApproxSW on NetFPGA-SUME board that has four 10Gbit Ethernet (10GbE) interfaces and a Virtex-7 FPGA. An experiment shows that the ApproxSW functions do not degrade the original 10GbE switch performance. We also analyze the accuracy of the proxy computation and response for stragglers and show that the proposed approximation based on task similarity achieves the best accuracy.

Original languageEnglish
Title of host publication2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9789090304281
DOIs
Publication statusPublished - 2017 Oct 2
Event27th International Conference on Field Programmable Logic and Applications, FPL 2017 - Gent, Belgium
Duration: 2017 Sep 42017 Sep 6

Other

Other27th International Conference on Field Programmable Logic and Applications, FPL 2017
CountryBelgium
CityGent
Period17/9/417/9/6

Fingerprint

Switches
Processing
Ethernet
Field programmable gate arrays (FPGA)
Experiments

ASJC Scopus subject areas

  • Computer Networks and Communications
  • Computer Science Applications
  • Hardware and Architecture
  • Software

Cite this

Mitsuzuka, K., Hayashi, A., Koibuchi, M., Amano, H., & Matsutani, H. (2017). In-switch approximate processing: Delayed tasks management for MapReduce applications. In 2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017 [8056802] Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.23919/FPL.2017.8056802

In-switch approximate processing : Delayed tasks management for MapReduce applications. / Mitsuzuka, Koya; Hayashi, Ami; Koibuchi, Michihiro; Amano, Hideharu; Matsutani, Hiroki.

2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017. Institute of Electrical and Electronics Engineers Inc., 2017. 8056802.

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

Mitsuzuka, K, Hayashi, A, Koibuchi, M, Amano, H & Matsutani, H 2017, In-switch approximate processing: Delayed tasks management for MapReduce applications. in 2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017., 8056802, Institute of Electrical and Electronics Engineers Inc., 27th International Conference on Field Programmable Logic and Applications, FPL 2017, Gent, Belgium, 17/9/4. https://doi.org/10.23919/FPL.2017.8056802
Mitsuzuka K, Hayashi A, Koibuchi M, Amano H, Matsutani H. In-switch approximate processing: Delayed tasks management for MapReduce applications. In 2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017. Institute of Electrical and Electronics Engineers Inc. 2017. 8056802 https://doi.org/10.23919/FPL.2017.8056802
Mitsuzuka, Koya ; Hayashi, Ami ; Koibuchi, Michihiro ; Amano, Hideharu ; Matsutani, Hiroki. / In-switch approximate processing : Delayed tasks management for MapReduce applications. 2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017. Institute of Electrical and Electronics Engineers Inc., 2017.
@inproceedings{e7dbe442ba434b7a8dcb27aa0a4b32b5,
title = "In-switch approximate processing: Delayed tasks management for MapReduce applications",
abstract = "In MapReduce, the parallel processing performance is often limited by only a few compute nodes that delay to complete given tasks. Although various techniques have been invented to handle such stragglers, these techniques mostly impose a burden on master node to monitor the progress of all the compute nodes, resulting in a new bottleneck as the number of compute nodes increases. As an alternative approach, in this paper, we propose to move such straggler management burden from master node to network switch that connects the master and compute nodes, because all the information goes through the switch. More specifically, the proposed network switch monitors output packets from Map tasks to detect stragglers. When detected, the proposed switch generates a response instead of the straggler based on the outputs of the other normal Map tasks, so that Reduce tasks can be started without delay. We introduce some approximate techniques for the proxy computation and response at the switch; thus our switch is called 'ApproxSW.' We implement ApproxSW on NetFPGA-SUME board that has four 10Gbit Ethernet (10GbE) interfaces and a Virtex-7 FPGA. An experiment shows that the ApproxSW functions do not degrade the original 10GbE switch performance. We also analyze the accuracy of the proxy computation and response for stragglers and show that the proposed approximation based on task similarity achieves the best accuracy.",
author = "Koya Mitsuzuka and Ami Hayashi and Michihiro Koibuchi and Hideharu Amano and Hiroki Matsutani",
year = "2017",
month = "10",
day = "2",
doi = "10.23919/FPL.2017.8056802",
language = "English",
booktitle = "2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

TY - GEN

T1 - In-switch approximate processing

T2 - Delayed tasks management for MapReduce applications

AU - Mitsuzuka, Koya

AU - Hayashi, Ami

AU - Koibuchi, Michihiro

AU - Amano, Hideharu

AU - Matsutani, Hiroki

PY - 2017/10/2

Y1 - 2017/10/2

N2 - In MapReduce, the parallel processing performance is often limited by only a few compute nodes that delay to complete given tasks. Although various techniques have been invented to handle such stragglers, these techniques mostly impose a burden on master node to monitor the progress of all the compute nodes, resulting in a new bottleneck as the number of compute nodes increases. As an alternative approach, in this paper, we propose to move such straggler management burden from master node to network switch that connects the master and compute nodes, because all the information goes through the switch. More specifically, the proposed network switch monitors output packets from Map tasks to detect stragglers. When detected, the proposed switch generates a response instead of the straggler based on the outputs of the other normal Map tasks, so that Reduce tasks can be started without delay. We introduce some approximate techniques for the proxy computation and response at the switch; thus our switch is called 'ApproxSW.' We implement ApproxSW on NetFPGA-SUME board that has four 10Gbit Ethernet (10GbE) interfaces and a Virtex-7 FPGA. An experiment shows that the ApproxSW functions do not degrade the original 10GbE switch performance. We also analyze the accuracy of the proxy computation and response for stragglers and show that the proposed approximation based on task similarity achieves the best accuracy.

AB - In MapReduce, the parallel processing performance is often limited by only a few compute nodes that delay to complete given tasks. Although various techniques have been invented to handle such stragglers, these techniques mostly impose a burden on master node to monitor the progress of all the compute nodes, resulting in a new bottleneck as the number of compute nodes increases. As an alternative approach, in this paper, we propose to move such straggler management burden from master node to network switch that connects the master and compute nodes, because all the information goes through the switch. More specifically, the proposed network switch monitors output packets from Map tasks to detect stragglers. When detected, the proposed switch generates a response instead of the straggler based on the outputs of the other normal Map tasks, so that Reduce tasks can be started without delay. We introduce some approximate techniques for the proxy computation and response at the switch; thus our switch is called 'ApproxSW.' We implement ApproxSW on NetFPGA-SUME board that has four 10Gbit Ethernet (10GbE) interfaces and a Virtex-7 FPGA. An experiment shows that the ApproxSW functions do not degrade the original 10GbE switch performance. We also analyze the accuracy of the proxy computation and response for stragglers and show that the proposed approximation based on task similarity achieves the best accuracy.

UR - http://www.scopus.com/inward/record.url?scp=85034443526&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85034443526&partnerID=8YFLogxK

U2 - 10.23919/FPL.2017.8056802

DO - 10.23919/FPL.2017.8056802

M3 - Conference contribution

BT - 2017 27th International Conference on Field Programmable Logic and Applications, FPL 2017

PB - Institute of Electrical and Electronics Engineers Inc.

ER -