Simulating Energy-Aware Networks in Large-Scale Distributed Systems

No Thumbnail Available

URL

Journal Title

Journal ISSN

Volume Title

Perustieteiden korkeakoulu | Master's thesis

Date

2017-08-28

Department

Major/Subject

Cloud Computing and Services

Mcode

SCIXXXX

Degree programme

Master's Programme in ICT Innovation

Language

en

Pages

58

Series

Abstract

Due to the emergence of new technologies such as IoT, the number of devices connected to the data center infrastructures is rapidly increasing. In response to this growing trend, the data-centers are also expanding. This expansion led to a major energy consumption issue. To tackle this issue researchers are designing and running experiments on real network, testbed environments and simulation software. The first two experiments environments are limited in terms of availability, flexibility and cost for conducting large-scale energy consumption estimation experiments. Simulation is the viable alternative in the context of large-scale networks. Currently proposed simulation software that are capable of estimating energy consumption of a network are in the category of packet-level simulators. The main problem of this kind of simulators is that they fail to scale well in the area of large-scale networks. In this paper we have proposed a reasonably accurate and scalable flow-level simulator alternative in the context of SimGrid, a large-scale distributed network simulator. Our flow-level simulator registered relative error less than 1% in accuracy compared to a packet-level simulator and it is also at least 243 times more faster.

Description

Supervisor

Heljanko, Keijo

Thesis advisor

Quinson, Martin
Orgerie, Anne-Cécile

Keywords

flow-level, energy, SimGrid, ECOFEN

Other note

Citation