Link to the University of Pittsburgh Homepage
Link to the University Library System Homepage Link to the Contact Us Form

Power Management Strategies for Wired Communication Networks.

Yu, Qun (2020) Power Management Strategies for Wired Communication Networks. Doctoral Dissertation, University of Pittsburgh. (Unpublished)

PDF (Qun_PhDThesis-etd-Corrected_V2.0)
Updated Version

Download (2MB) | Preview


With the exponential traffic growth and the rapid expansion of communication infrastructures worldwide, energy expenditure of the Internet has become a major concern in IT-reliant society. This energy problem has motivated the urgent demands of new strategies to reduce the consumption of telecommunication networks, with a particular focus on IP networks. In addition to the development of a new generation of energy-efficient network equipment, a significant body of research has concentrated on incorporating power/energy-awareness into network control and management, which aims at reducing the network power/energy consumption by either dynamically scaling speeds of each active network component to make it capable of adapting to its current load or putting to sleep the lightly loaded network elements and reconfiguring the network. However, the fundamental challenge of greening the Internet is to achieve a balance between the power/energy saving and the demands of quality-of-service (QoS) performance, which is an issue that has received less attention but is becoming a major problem in future green network designs. In this dissertation, we study how energy consumption can be reduced through different power/energy- and QoS-aware strategies for wired communication networks.

To sufficiently reduce energy consumption while meeting the desire QoS requirements, we introduce several different schemes combing power management techniques with different scheduling strategies, which can be classified into experimental power management (EPM) and algorithmic power management (APM). In these proposed schemes, the power management techniques that we focus on are speed scaling and sleep mode. When the network processor is active, its speed and supply voltage can be decreased to reduce the energy consumption (speed scaling), while when the processor is idle, it can be put in a low power mode to save the energy consumption (sleep mode). The resulting problem is to determine how and when to adjust speeds for the processors, and/or to put a device into sleep mode. In this dissertation, we first discuss three families of dynamic voltage/frequency scaling (DVFS) based, QoS-aware EPM schemes, which aim to reduce the energy consumption in network equipment by using different packet scheduling strategies, while adhering to QoS requirements of supported applications. Then, we explore the problem of energy minimization under QoS constraints through a mathematical programming model, which is a DVFS-based, delay-aware APM scheme combing the speed scaling technique with the existing rate monotonic scheduling policy. Among these speed scaling based schemes, up to 26.76% dynamic power saving of the total power consumption can be achieved. In addition to speed scaling approaches, we further propose a sleep-based, traffic-aware EPM scheme, which is used to reduce power consumption by greening routing light load and putting the related network equipment into sleep mode according to twelve flow traffic density changes in 24-hour of an arbitrarily selected day. Meanwhile, a speed scaling technique without violating network QoS performance is also considered in this scheme when the traffic is rerouted. Applying this sleep-based strategy can lead to power savings of up to 62.58% of the total power consumption.


Social Networking:
Share |


Item Type: University of Pittsburgh ETD
Status: Unpublished
CreatorsEmailPitt UsernameORCID
Yu, Qunquy3@pitt.eduquy3
ETD Committee:
TitleMemberEmail AddressPitt UsernameORCID
Committee ChairZnati, Taiebznati@pitt.eduznati
Committee MemberWeiss, Martinmbw@pitt.edumbw
Committee MemberKrishnamurthy, Prashantprashk@pitt.eduprashk
Committee MemberMosse, Danielmosse@pitt.edumosse
Committee MemberPalanisamy, BalajiBPALAN@pitt.eduBPALAN
Date: 5 June 2020
Date Type: Publication
Defense Date: 13 December 2019
Approval Date: 5 June 2020
Submission Date: 20 May 2020
Access Restriction: No restriction; Release the ETD for access worldwide immediately.
Number of Pages: 152
Institution: University of Pittsburgh
Schools and Programs: School of Computing and Information > Telecommunications
Degree: PhD - Doctor of Philosophy
Thesis Type: Doctoral Dissertation
Refereed: Yes
Uncontrolled Keywords: DVFS, Energy Minimization, Energy-aware, Network Performance, QoS-aware, Sleep Mode.
Related URLs:
Date Deposited: 05 Jun 2020 21:23
Last Modified: 05 Jun 2020 21:23


Monthly Views for the past 3 years

Plum Analytics

Actions (login required)

View Item View Item