LaCoDa: Layered connected topology for massive data centers

Zina Chkirbene, Sebti Foufou, Ridha Hamila, Zahir Tari, Albert Y. Zomaya

Research output: Contribution to journalArticle

Abstract

One of the fundamental challenges of existing data centers is to design a network that interconnects massive number of servers, and therefore providing an efficient and fault-tolerant routing service to upper-layer applications. Several solutions have been proposed (e.g. FatTree, DCell and BCube), however they either scale too fast (i.e., double exponentially) or too slow. This paper proposes a new data center topology, called LaCoDa, that combines the advantages of previous topologies while avoiding their limitations. LaCoDa uses a small node degree that matches physical restriction for servers, and it also interconnects a large number of servers while reducing the wiring complexity and without affecting the existing servers. LaCoDa can interconnect up to millions of servers with a small number of port switches by increasing the number of directly connected clusters per layer, and therefore avoiding the cluster connections repetitions. The theoretical and experimental results show that the average path length (APL) of a 1000-node LaCoDa is smaller than the APL of a 23-node FlatTree and 11-node BCube and DCell. Moreover, with 6-port switches and a node degree of 4, the bisection bandwidth of 4.19 Million-node LaCoDa is 7% and 14% bigger than the bisection bandwidth of 310-node DCell and BCube, respectively. Consequently, LaCoDa provides better performance than DCel, BCube, FatTree in terms of average path length, throughput, latency and bisection bandwidth.

Original languageEnglish (US)
Pages (from-to)169-180
Number of pages12
JournalJournal of Network and Computer Applications
Volume83
DOIs
StatePublished - Apr 1 2017

Fingerprint

Servers
Topology
Bandwidth
Switches
Electric wiring
Throughput

Keywords

  • Average path length
  • Bisection bandwidth
  • Data center network
  • Network topology

ASJC Scopus subject areas

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

Cite this

LaCoDa : Layered connected topology for massive data centers. / Chkirbene, Zina; Foufou, Sebti; Hamila, Ridha; Tari, Zahir; Zomaya, Albert Y.

In: Journal of Network and Computer Applications, Vol. 83, 01.04.2017, p. 169-180.

Research output: Contribution to journalArticle

Chkirbene, Zina ; Foufou, Sebti ; Hamila, Ridha ; Tari, Zahir ; Zomaya, Albert Y. / LaCoDa : Layered connected topology for massive data centers. In: Journal of Network and Computer Applications. 2017 ; Vol. 83. pp. 169-180.
@article{0628f92cf1b5487089d0f1d8ffda2df9,
title = "LaCoDa: Layered connected topology for massive data centers",
abstract = "One of the fundamental challenges of existing data centers is to design a network that interconnects massive number of servers, and therefore providing an efficient and fault-tolerant routing service to upper-layer applications. Several solutions have been proposed (e.g. FatTree, DCell and BCube), however they either scale too fast (i.e., double exponentially) or too slow. This paper proposes a new data center topology, called LaCoDa, that combines the advantages of previous topologies while avoiding their limitations. LaCoDa uses a small node degree that matches physical restriction for servers, and it also interconnects a large number of servers while reducing the wiring complexity and without affecting the existing servers. LaCoDa can interconnect up to millions of servers with a small number of port switches by increasing the number of directly connected clusters per layer, and therefore avoiding the cluster connections repetitions. The theoretical and experimental results show that the average path length (APL) of a 1000-node LaCoDa is smaller than the APL of a 23-node FlatTree and 11-node BCube and DCell. Moreover, with 6-port switches and a node degree of 4, the bisection bandwidth of 4.19 Million-node LaCoDa is 7{\%} and 14{\%} bigger than the bisection bandwidth of 310-node DCell and BCube, respectively. Consequently, LaCoDa provides better performance than DCel, BCube, FatTree in terms of average path length, throughput, latency and bisection bandwidth.",
keywords = "Average path length, Bisection bandwidth, Data center network, Network topology",
author = "Zina Chkirbene and Sebti Foufou and Ridha Hamila and Zahir Tari and Zomaya, {Albert Y.}",
year = "2017",
month = "4",
day = "1",
doi = "10.1016/j.jnca.2017.01.020",
language = "English (US)",
volume = "83",
pages = "169--180",
journal = "Journal of Network and Computer Applications",
issn = "1084-8045",
publisher = "Academic Press Inc.",

}

TY - JOUR

T1 - LaCoDa

T2 - Layered connected topology for massive data centers

AU - Chkirbene, Zina

AU - Foufou, Sebti

AU - Hamila, Ridha

AU - Tari, Zahir

AU - Zomaya, Albert Y.

PY - 2017/4/1

Y1 - 2017/4/1

N2 - One of the fundamental challenges of existing data centers is to design a network that interconnects massive number of servers, and therefore providing an efficient and fault-tolerant routing service to upper-layer applications. Several solutions have been proposed (e.g. FatTree, DCell and BCube), however they either scale too fast (i.e., double exponentially) or too slow. This paper proposes a new data center topology, called LaCoDa, that combines the advantages of previous topologies while avoiding their limitations. LaCoDa uses a small node degree that matches physical restriction for servers, and it also interconnects a large number of servers while reducing the wiring complexity and without affecting the existing servers. LaCoDa can interconnect up to millions of servers with a small number of port switches by increasing the number of directly connected clusters per layer, and therefore avoiding the cluster connections repetitions. The theoretical and experimental results show that the average path length (APL) of a 1000-node LaCoDa is smaller than the APL of a 23-node FlatTree and 11-node BCube and DCell. Moreover, with 6-port switches and a node degree of 4, the bisection bandwidth of 4.19 Million-node LaCoDa is 7% and 14% bigger than the bisection bandwidth of 310-node DCell and BCube, respectively. Consequently, LaCoDa provides better performance than DCel, BCube, FatTree in terms of average path length, throughput, latency and bisection bandwidth.

AB - One of the fundamental challenges of existing data centers is to design a network that interconnects massive number of servers, and therefore providing an efficient and fault-tolerant routing service to upper-layer applications. Several solutions have been proposed (e.g. FatTree, DCell and BCube), however they either scale too fast (i.e., double exponentially) or too slow. This paper proposes a new data center topology, called LaCoDa, that combines the advantages of previous topologies while avoiding their limitations. LaCoDa uses a small node degree that matches physical restriction for servers, and it also interconnects a large number of servers while reducing the wiring complexity and without affecting the existing servers. LaCoDa can interconnect up to millions of servers with a small number of port switches by increasing the number of directly connected clusters per layer, and therefore avoiding the cluster connections repetitions. The theoretical and experimental results show that the average path length (APL) of a 1000-node LaCoDa is smaller than the APL of a 23-node FlatTree and 11-node BCube and DCell. Moreover, with 6-port switches and a node degree of 4, the bisection bandwidth of 4.19 Million-node LaCoDa is 7% and 14% bigger than the bisection bandwidth of 310-node DCell and BCube, respectively. Consequently, LaCoDa provides better performance than DCel, BCube, FatTree in terms of average path length, throughput, latency and bisection bandwidth.

KW - Average path length

KW - Bisection bandwidth

KW - Data center network

KW - Network topology

UR - http://www.scopus.com/inward/record.url?scp=85013200586&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85013200586&partnerID=8YFLogxK

U2 - 10.1016/j.jnca.2017.01.020

DO - 10.1016/j.jnca.2017.01.020

M3 - Article

AN - SCOPUS:85013200586

VL - 83

SP - 169

EP - 180

JO - Journal of Network and Computer Applications

JF - Journal of Network and Computer Applications

SN - 1084-8045

ER -