Load-balancing multipath switching system with flow slice

Lei Shi, Bin Liu, Changhua Sun, Zhengyu Yin, Laxmi N. Bhuyan, H. Jonathan Chao

Research output: Contribution to journalArticle

Abstract

Multipath Switching systems (MPS) are intensely used in state-of-the-art core routers to provide terabit or even petabit switching capacity. One of the most intractable issues in designing MPS is how to load balance traffic across its multiple paths while not disturbing the intraflow packet orders. Previous packet-based solutions either suffer from delay penalties or lead to O(N 2 ) hardware complexity, hence do not scale. Flow-based hashing algorithms also perform badly due to the heavy-tailed flow-size distribution. In this paper, we develop a novel scheme, namely, Flow Slice (FS) that cuts off each flow into flow slices at every intraflow interval larger than a slicing threshold and balances the load on a finer granularity. Based on the studies of tens of real Internet traces, we show that setting a slicing threshold of 1-4 ms, the FS scheme achieves comparative load-balancing performance to the optimal one. It also limits the probability of out-of-order packets to a negligible level (10 -6) on three popular MPSes at the cost of little hardware complexity and an internal speedup up to two. These results are proven by theoretical analyses and also validated through trace-driven prototype simulations.

Original languageEnglish (US)
Article number5674025
Pages (from-to)350-365
Number of pages16
JournalIEEE Transactions on Computers
Volume61
Issue number3
DOIs
StatePublished - 2012

Fingerprint

Switching Systems
Switching systems
Multipath
Load Balancing
Slice
Resource allocation
Hardware
Routers
Telecommunication traffic
Slicing
Internet
Trace
Load Balance
Hashing
Router
Granularity
Penalty
Speedup
Traffic
Prototype

Keywords

  • Load balancing
  • switching theory
  • traffic measurement

ASJC Scopus subject areas

  • Hardware and Architecture
  • Software
  • Computational Theory and Mathematics
  • Theoretical Computer Science

Cite this

Load-balancing multipath switching system with flow slice. / Shi, Lei; Liu, Bin; Sun, Changhua; Yin, Zhengyu; Bhuyan, Laxmi N.; Chao, H. Jonathan.

In: IEEE Transactions on Computers, Vol. 61, No. 3, 5674025, 2012, p. 350-365.

Research output: Contribution to journalArticle

Shi, L, Liu, B, Sun, C, Yin, Z, Bhuyan, LN & Chao, HJ 2012, 'Load-balancing multipath switching system with flow slice', IEEE Transactions on Computers, vol. 61, no. 3, 5674025, pp. 350-365. https://doi.org/10.1109/TC.2010.279
Shi, Lei ; Liu, Bin ; Sun, Changhua ; Yin, Zhengyu ; Bhuyan, Laxmi N. ; Chao, H. Jonathan. / Load-balancing multipath switching system with flow slice. In: IEEE Transactions on Computers. 2012 ; Vol. 61, No. 3. pp. 350-365.
@article{ea7040c93d8d43bd8225229c5b44172f,
title = "Load-balancing multipath switching system with flow slice",
abstract = "Multipath Switching systems (MPS) are intensely used in state-of-the-art core routers to provide terabit or even petabit switching capacity. One of the most intractable issues in designing MPS is how to load balance traffic across its multiple paths while not disturbing the intraflow packet orders. Previous packet-based solutions either suffer from delay penalties or lead to O(N 2 ) hardware complexity, hence do not scale. Flow-based hashing algorithms also perform badly due to the heavy-tailed flow-size distribution. In this paper, we develop a novel scheme, namely, Flow Slice (FS) that cuts off each flow into flow slices at every intraflow interval larger than a slicing threshold and balances the load on a finer granularity. Based on the studies of tens of real Internet traces, we show that setting a slicing threshold of 1-4 ms, the FS scheme achieves comparative load-balancing performance to the optimal one. It also limits the probability of out-of-order packets to a negligible level (10 -6) on three popular MPSes at the cost of little hardware complexity and an internal speedup up to two. These results are proven by theoretical analyses and also validated through trace-driven prototype simulations.",
keywords = "Load balancing, switching theory, traffic measurement",
author = "Lei Shi and Bin Liu and Changhua Sun and Zhengyu Yin and Bhuyan, {Laxmi N.} and Chao, {H. Jonathan}",
year = "2012",
doi = "10.1109/TC.2010.279",
language = "English (US)",
volume = "61",
pages = "350--365",
journal = "IEEE Transactions on Computers",
issn = "0018-9340",
publisher = "IEEE Computer Society",
number = "3",

}

TY - JOUR

T1 - Load-balancing multipath switching system with flow slice

AU - Shi, Lei

AU - Liu, Bin

AU - Sun, Changhua

AU - Yin, Zhengyu

AU - Bhuyan, Laxmi N.

AU - Chao, H. Jonathan

PY - 2012

Y1 - 2012

N2 - Multipath Switching systems (MPS) are intensely used in state-of-the-art core routers to provide terabit or even petabit switching capacity. One of the most intractable issues in designing MPS is how to load balance traffic across its multiple paths while not disturbing the intraflow packet orders. Previous packet-based solutions either suffer from delay penalties or lead to O(N 2 ) hardware complexity, hence do not scale. Flow-based hashing algorithms also perform badly due to the heavy-tailed flow-size distribution. In this paper, we develop a novel scheme, namely, Flow Slice (FS) that cuts off each flow into flow slices at every intraflow interval larger than a slicing threshold and balances the load on a finer granularity. Based on the studies of tens of real Internet traces, we show that setting a slicing threshold of 1-4 ms, the FS scheme achieves comparative load-balancing performance to the optimal one. It also limits the probability of out-of-order packets to a negligible level (10 -6) on three popular MPSes at the cost of little hardware complexity and an internal speedup up to two. These results are proven by theoretical analyses and also validated through trace-driven prototype simulations.

AB - Multipath Switching systems (MPS) are intensely used in state-of-the-art core routers to provide terabit or even petabit switching capacity. One of the most intractable issues in designing MPS is how to load balance traffic across its multiple paths while not disturbing the intraflow packet orders. Previous packet-based solutions either suffer from delay penalties or lead to O(N 2 ) hardware complexity, hence do not scale. Flow-based hashing algorithms also perform badly due to the heavy-tailed flow-size distribution. In this paper, we develop a novel scheme, namely, Flow Slice (FS) that cuts off each flow into flow slices at every intraflow interval larger than a slicing threshold and balances the load on a finer granularity. Based on the studies of tens of real Internet traces, we show that setting a slicing threshold of 1-4 ms, the FS scheme achieves comparative load-balancing performance to the optimal one. It also limits the probability of out-of-order packets to a negligible level (10 -6) on three popular MPSes at the cost of little hardware complexity and an internal speedup up to two. These results are proven by theoretical analyses and also validated through trace-driven prototype simulations.

KW - Load balancing

KW - switching theory

KW - traffic measurement

UR - http://www.scopus.com/inward/record.url?scp=84863078190&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84863078190&partnerID=8YFLogxK

U2 - 10.1109/TC.2010.279

DO - 10.1109/TC.2010.279

M3 - Article

AN - SCOPUS:84863078190

VL - 61

SP - 350

EP - 365

JO - IEEE Transactions on Computers

JF - IEEE Transactions on Computers

SN - 0018-9340

IS - 3

M1 - 5674025

ER -