GPU-accelerated hash and wildcard hybrid flow switching for tackling massive flow entries

Nobutaka Matsumoto; Michiaki Hayashi; Itsuro Morita

doi:10.1109/HPSR.2013.6602315

GPU-accelerated hash and wildcard hybrid flow switching for tackling massive flow entries

Nobutaka Matsumoto, Michiaki Hayashi, Itsuro Morita

Research output: Contribution to journal › Conference article › peer-review

4 Citations (Scopus)

Abstract

GPU-accelerated high performance switching design using hybrid flow tables allocated in host and GPU memories is proposed based on LightFlow architecture. In this work, careful placement of hash and wildcard tables is proposed considering overhead of data copy during lookup process, in order to tackle massive flow entries using commodity hardware. Experimental results show that the proposal achieves 14.6 Mpps which is the fastest record of the software flow switch operated under a large scale reaching 4 million flow entries. With this proposal, it is also demonstrated that the average processing delay is about one forth compared with Open vSwitch.

Original language	English
Article number	6602315
Pages (from-to)	213-214
Number of pages	2
Journal	IEEE International Conference on High Performance Switching and Routing, HPSR
DOIs	https://doi.org/10.1109/HPSR.2013.6602315
Publication status	Published - 2013
Externally published	Yes
Event	2013 IEEE 14th International Conference on High Performance Switching and Routing, HPSR 2013 - Taipei, Taiwan, Province of China Duration: 2013 Jul 8 → 2013 Jul 11

Keywords

flow switching
GPU
look-up
software switch

ASJC Scopus subject areas

Hardware and Architecture
Electrical and Electronic Engineering

Access to Document

10.1109/HPSR.2013.6602315

Cite this

@article{48c427b2622d48dba5b8e4572ca135b8,

title = "GPU-accelerated hash and wildcard hybrid flow switching for tackling massive flow entries",

abstract = "GPU-accelerated high performance switching design using hybrid flow tables allocated in host and GPU memories is proposed based on LightFlow architecture. In this work, careful placement of hash and wildcard tables is proposed considering overhead of data copy during lookup process, in order to tackle massive flow entries using commodity hardware. Experimental results show that the proposal achieves 14.6 Mpps which is the fastest record of the software flow switch operated under a large scale reaching 4 million flow entries. With this proposal, it is also demonstrated that the average processing delay is about one forth compared with Open vSwitch.",

keywords = "flow switching, GPU, look-up, software switch",

author = "Nobutaka Matsumoto and Michiaki Hayashi and Itsuro Morita",

year = "2013",

doi = "10.1109/HPSR.2013.6602315",

language = "English",

pages = "213--214",

journal = "IEEE International Conference on High Performance Switching and Routing, HPSR",

issn = "2325-5595",

note = "2013 IEEE 14th International Conference on High Performance Switching and Routing, HPSR 2013 ; Conference date: 08-07-2013 Through 11-07-2013",

}

TY - JOUR

T1 - GPU-accelerated hash and wildcard hybrid flow switching for tackling massive flow entries

AU - Matsumoto, Nobutaka

AU - Hayashi, Michiaki

AU - Morita, Itsuro

PY - 2013

Y1 - 2013

N2 - GPU-accelerated high performance switching design using hybrid flow tables allocated in host and GPU memories is proposed based on LightFlow architecture. In this work, careful placement of hash and wildcard tables is proposed considering overhead of data copy during lookup process, in order to tackle massive flow entries using commodity hardware. Experimental results show that the proposal achieves 14.6 Mpps which is the fastest record of the software flow switch operated under a large scale reaching 4 million flow entries. With this proposal, it is also demonstrated that the average processing delay is about one forth compared with Open vSwitch.

AB - GPU-accelerated high performance switching design using hybrid flow tables allocated in host and GPU memories is proposed based on LightFlow architecture. In this work, careful placement of hash and wildcard tables is proposed considering overhead of data copy during lookup process, in order to tackle massive flow entries using commodity hardware. Experimental results show that the proposal achieves 14.6 Mpps which is the fastest record of the software flow switch operated under a large scale reaching 4 million flow entries. With this proposal, it is also demonstrated that the average processing delay is about one forth compared with Open vSwitch.

KW - flow switching

KW - GPU

KW - look-up

KW - software switch

UR - http://www.scopus.com/inward/record.url?scp=84889053892&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84889053892&partnerID=8YFLogxK

U2 - 10.1109/HPSR.2013.6602315

DO - 10.1109/HPSR.2013.6602315

M3 - Conference article

AN - SCOPUS:84889053892

SN - 2325-5595

SP - 213

EP - 214

JO - IEEE International Conference on High Performance Switching and Routing, HPSR

JF - IEEE International Conference on High Performance Switching and Routing, HPSR

M1 - 6602315

T2 - 2013 IEEE 14th International Conference on High Performance Switching and Routing, HPSR 2013

Y2 - 8 July 2013 through 11 July 2013

ER -

GPU-accelerated hash and wildcard hybrid flow switching for tackling massive flow entries

Abstract

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this