rDAN: Toward robust demand-aware network designs

Chen Avin; Alexandr Hercules; Andreas Loukas; Stefan Schmid

doi:10.1016/j.ipl.2017.12.008

rDAN: Toward robust demand-aware network designs

Chen Avin, Alexandr Hercules, Andreas Loukas, Stefan Schmid

Research output: Contribution to journal › Article › peer-review

Abstract

We currently witness the emergence of interesting new network topologies optimized towards the traffic matrices they serve, such as demand-aware datacenter interconnects (e.g., ProjecToR) and demand-aware peer-to-peer overlay networks (e.g., SplayNets). This paper introduces a formal framework and approach to reason about and design robust demand-aware networks (DAN). In particular, we establish a connection between the communication frequency of two nodes and the path length between them in the network, and show that this relationship depends on the entropy of the communication matrix. Our main contribution is a novel robust, yet sparse, family of networks, short rDANs, which guarantee an expected path length that is proportional to the entropy of the communication patterns.

Original language	American English
Pages (from-to)	5-9
Number of pages	5
Journal	Information Processing Letters
Volume	133
DOIs	https://doi.org/10.1016/j.ipl.2017.12.008
State	Published - 1 May 2018

Keywords

Algorithms
Coding theory
Datacenter interconnects
Network design
Peer-to-peer topologies

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Signal Processing
Information Systems
Computer Science Applications

Access to Document

10.1016/j.ipl.2017.12.008

Cite this

@article{311c549bd3444fd994af8d52f053f928,

title = "rDAN: Toward robust demand-aware network designs",

abstract = "We currently witness the emergence of interesting new network topologies optimized towards the traffic matrices they serve, such as demand-aware datacenter interconnects (e.g., ProjecToR) and demand-aware peer-to-peer overlay networks (e.g., SplayNets). This paper introduces a formal framework and approach to reason about and design robust demand-aware networks (DAN). In particular, we establish a connection between the communication frequency of two nodes and the path length between them in the network, and show that this relationship depends on the entropy of the communication matrix. Our main contribution is a novel robust, yet sparse, family of networks, short rDANs, which guarantee an expected path length that is proportional to the entropy of the communication patterns.",

keywords = "Algorithms, Coding theory, Datacenter interconnects, Network design, Peer-to-peer topologies",

author = "Chen Avin and Alexandr Hercules and Andreas Loukas and Stefan Schmid",

note = "Publisher Copyright: {\textcopyright} 2017 Elsevier B.V.",

year = "2018",

month = may,

day = "1",

doi = "10.1016/j.ipl.2017.12.008",

language = "American English",

volume = "133",

pages = "5--9",

journal = "Information Processing Letters",

issn = "0020-0190",

publisher = "Elsevier B.V.",

}

TY - JOUR

T1 - rDAN

T2 - Toward robust demand-aware network designs

AU - Avin, Chen

AU - Hercules, Alexandr

AU - Loukas, Andreas

AU - Schmid, Stefan

PY - 2018/5/1

Y1 - 2018/5/1

N2 - We currently witness the emergence of interesting new network topologies optimized towards the traffic matrices they serve, such as demand-aware datacenter interconnects (e.g., ProjecToR) and demand-aware peer-to-peer overlay networks (e.g., SplayNets). This paper introduces a formal framework and approach to reason about and design robust demand-aware networks (DAN). In particular, we establish a connection between the communication frequency of two nodes and the path length between them in the network, and show that this relationship depends on the entropy of the communication matrix. Our main contribution is a novel robust, yet sparse, family of networks, short rDANs, which guarantee an expected path length that is proportional to the entropy of the communication patterns.

AB - We currently witness the emergence of interesting new network topologies optimized towards the traffic matrices they serve, such as demand-aware datacenter interconnects (e.g., ProjecToR) and demand-aware peer-to-peer overlay networks (e.g., SplayNets). This paper introduces a formal framework and approach to reason about and design robust demand-aware networks (DAN). In particular, we establish a connection between the communication frequency of two nodes and the path length between them in the network, and show that this relationship depends on the entropy of the communication matrix. Our main contribution is a novel robust, yet sparse, family of networks, short rDANs, which guarantee an expected path length that is proportional to the entropy of the communication patterns.

KW - Algorithms

KW - Coding theory

KW - Datacenter interconnects

KW - Network design

KW - Peer-to-peer topologies

UR - http://www.scopus.com/inward/record.url?scp=85040126190&partnerID=8YFLogxK

U2 - 10.1016/j.ipl.2017.12.008

DO - 10.1016/j.ipl.2017.12.008

M3 - Article

SN - 0020-0190

VL - 133

SP - 5

EP - 9

JO - Information Processing Letters

JF - Information Processing Letters

ER -

rDAN: Toward robust demand-aware network designs

Abstract

Keywords

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Cite this