QDAPS: Queueing Delay Aware Packet Spraying for Load Balancing in Data Center

Jiawei Huang; Wenjun Lv; Weihe Li; Jianxin Wang; Tian He

doi:10.1109/ICNP.2018.00017

QDAPS: Queueing Delay Aware Packet Spraying for Load Balancing in Data Center

Jiawei Huang, Wenjun Lv, Weihe Li, Jianxin Wang, Tian He

Center for Transportation Studies

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

17 Scopus citations

Abstract

Modern data center networks are usually constructed in multi-rooted tree topologies, which require the highly efficient multi-path load balancing to achieve high link utilization. Recent packet-level load balancer obtains high throughput by spraying packets to all paths, but it easily leads to the packet reordering under network asymmetry. The flow-level or flowlet-level load balancer avoids the packet reordering, while reducing the link utilization due to their inflexibility. To solve these problems, we design a Queueing Delay Aware Packet Spraying (QDAPS), that effectively mitigates the packet reordering for packet-level load balancer. QDAPS selects paths for packets according to the queueing delay of output buffer, and lets the packet arriving earlier be forwarded before the later packets to avoid packet reordering. We compare QDAPS with ECMP, LetFlow and RPS through NS2 simulation and Mininet implementation. The test results show that QDAPS reduces flow completion time (FCT) by ∼30%-50% over the state-of-the-art load balancing mechanism.

Original language	English (US)
Title of host publication	Proceedings - 26th IEEE International Conference on Network Protocols, ICNP 2018
Publisher	IEEE Computer Society
Pages	66-76
Number of pages	11
ISBN (Electronic)	9781538660430
DOIs	https://doi.org/10.1109/ICNP.2018.00017
State	Published - Nov 7 2018
Event	26th IEEE International Conference on Network Protocols, ICNP 2018 - Cambridge, United Kingdom Duration: Sep 24 2018 → Sep 27 2018

Publication series

Name	Proceedings - International Conference on Network Protocols, ICNP
Volume	2018-September
ISSN (Print)	1092-1648

Other

Other	26th IEEE International Conference on Network Protocols, ICNP 2018
Country/Territory	United Kingdom
City	Cambridge
Period	9/24/18 → 9/27/18

Bibliographical note

Funding Information:
This work is supported by the National Natural Science Foundation of China (61572530, 61629302 and 61420106009), CERNET Innovation Project (Grant No.NGII20160113) and Fundamental Research Funds for Central Universities of Central South University (2018zzts066).

Publisher Copyright:
© 2018 IEEE.

Keywords

Data center
Load balancing
Multi-path

Publisher link

10.1109/ICNP.2018.00017

Find It

Find It at U of M Libraries

Cite this

Huang, J., Lv, W., Li, W., Wang, J., & He, T. (2018). QDAPS: Queueing Delay Aware Packet Spraying for Load Balancing in Data Center. In Proceedings - 26th IEEE International Conference on Network Protocols, ICNP 2018 (pp. 66-76). Article 8526805 (Proceedings - International Conference on Network Protocols, ICNP; Vol. 2018-September). IEEE Computer Society. https://doi.org/10.1109/ICNP.2018.00017

QDAPS: Queueing Delay Aware Packet Spraying for Load Balancing in Data Center. / Huang, Jiawei; Lv, Wenjun; Li, Weihe et al.
Proceedings - 26th IEEE International Conference on Network Protocols, ICNP 2018. IEEE Computer Society, 2018. p. 66-76 8526805 (Proceedings - International Conference on Network Protocols, ICNP; Vol. 2018-September).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Huang, J, Lv, W, Li, W, Wang, J & He, T 2018, QDAPS: Queueing Delay Aware Packet Spraying for Load Balancing in Data Center. in Proceedings - 26th IEEE International Conference on Network Protocols, ICNP 2018., 8526805, Proceedings - International Conference on Network Protocols, ICNP, vol. 2018-September, IEEE Computer Society, pp. 66-76, 26th IEEE International Conference on Network Protocols, ICNP 2018, Cambridge, United Kingdom, 9/24/18. https://doi.org/10.1109/ICNP.2018.00017

@inproceedings{9a66d34b8b334d82a6411225a08c5bbf,

title = "QDAPS: Queueing Delay Aware Packet Spraying for Load Balancing in Data Center",

abstract = "Modern data center networks are usually constructed in multi-rooted tree topologies, which require the highly efficient multi-path load balancing to achieve high link utilization. Recent packet-level load balancer obtains high throughput by spraying packets to all paths, but it easily leads to the packet reordering under network asymmetry. The flow-level or flowlet-level load balancer avoids the packet reordering, while reducing the link utilization due to their inflexibility. To solve these problems, we design a Queueing Delay Aware Packet Spraying (QDAPS), that effectively mitigates the packet reordering for packet-level load balancer. QDAPS selects paths for packets according to the queueing delay of output buffer, and lets the packet arriving earlier be forwarded before the later packets to avoid packet reordering. We compare QDAPS with ECMP, LetFlow and RPS through NS2 simulation and Mininet implementation. The test results show that QDAPS reduces flow completion time (FCT) by ∼30%-50% over the state-of-the-art load balancing mechanism.",

keywords = "Data center, Load balancing, Multi-path",

author = "Jiawei Huang and Wenjun Lv and Weihe Li and Jianxin Wang and Tian He",

note = "Funding Information: This work is supported by the National Natural Science Foundation of China (61572530, 61629302 and 61420106009), CERNET Innovation Project (Grant No.NGII20160113) and Fundamental Research Funds for Central Universities of Central South University (2018zzts066). Publisher Copyright: {\textcopyright} 2018 IEEE.; 26th IEEE International Conference on Network Protocols, ICNP 2018 ; Conference date: 24-09-2018 Through 27-09-2018",

year = "2018",

month = nov,

day = "7",

doi = "10.1109/ICNP.2018.00017",

language = "English (US)",

series = "Proceedings - International Conference on Network Protocols, ICNP",

publisher = "IEEE Computer Society",

pages = "66--76",

booktitle = "Proceedings - 26th IEEE International Conference on Network Protocols, ICNP 2018",

}

TY - GEN

T1 - QDAPS

T2 - 26th IEEE International Conference on Network Protocols, ICNP 2018

AU - Huang, Jiawei

AU - Lv, Wenjun

AU - Li, Weihe

AU - Wang, Jianxin

AU - He, Tian

N1 - Funding Information: This work is supported by the National Natural Science Foundation of China (61572530, 61629302 and 61420106009), CERNET Innovation Project (Grant No.NGII20160113) and Fundamental Research Funds for Central Universities of Central South University (2018zzts066). Publisher Copyright: © 2018 IEEE.

PY - 2018/11/7

Y1 - 2018/11/7

N2 - Modern data center networks are usually constructed in multi-rooted tree topologies, which require the highly efficient multi-path load balancing to achieve high link utilization. Recent packet-level load balancer obtains high throughput by spraying packets to all paths, but it easily leads to the packet reordering under network asymmetry. The flow-level or flowlet-level load balancer avoids the packet reordering, while reducing the link utilization due to their inflexibility. To solve these problems, we design a Queueing Delay Aware Packet Spraying (QDAPS), that effectively mitigates the packet reordering for packet-level load balancer. QDAPS selects paths for packets according to the queueing delay of output buffer, and lets the packet arriving earlier be forwarded before the later packets to avoid packet reordering. We compare QDAPS with ECMP, LetFlow and RPS through NS2 simulation and Mininet implementation. The test results show that QDAPS reduces flow completion time (FCT) by ∼30%-50% over the state-of-the-art load balancing mechanism.

AB - Modern data center networks are usually constructed in multi-rooted tree topologies, which require the highly efficient multi-path load balancing to achieve high link utilization. Recent packet-level load balancer obtains high throughput by spraying packets to all paths, but it easily leads to the packet reordering under network asymmetry. The flow-level or flowlet-level load balancer avoids the packet reordering, while reducing the link utilization due to their inflexibility. To solve these problems, we design a Queueing Delay Aware Packet Spraying (QDAPS), that effectively mitigates the packet reordering for packet-level load balancer. QDAPS selects paths for packets according to the queueing delay of output buffer, and lets the packet arriving earlier be forwarded before the later packets to avoid packet reordering. We compare QDAPS with ECMP, LetFlow and RPS through NS2 simulation and Mininet implementation. The test results show that QDAPS reduces flow completion time (FCT) by ∼30%-50% over the state-of-the-art load balancing mechanism.

KW - Data center

KW - Load balancing

KW - Multi-path

UR - http://www.scopus.com/inward/record.url?scp=85058087376&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85058087376&partnerID=8YFLogxK

U2 - 10.1109/ICNP.2018.00017

DO - 10.1109/ICNP.2018.00017

M3 - Conference contribution

AN - SCOPUS:85058087376

T3 - Proceedings - International Conference on Network Protocols, ICNP

SP - 66

EP - 76

BT - Proceedings - 26th IEEE International Conference on Network Protocols, ICNP 2018

PB - IEEE Computer Society

Y2 - 24 September 2018 through 27 September 2018

ER -

QDAPS: Queueing Delay Aware Packet Spraying for Load Balancing in Data Center

Abstract

Publication series

Other

Bibliographical note

Keywords

Publisher link

Find It

Other files and links

Fingerprint

Cite this