This research work presents a dynamic algorithm for improving energy saving in Long Term Evolution (LTE) mobile access networks through off mode, sleep mode and multi-cell cooperation utilization at the eNodeBs. The LTE mobile access network environment and the eNodeBs energy saving models were developed with a view to implementing a dynamic energy saving algorithm. The dynamic energy saving algorithm is an integration of two algorithms, namely: energy estimation algorithm and load/traffic sharing algorithm. The energy estimation algorithm is use to estimates the energy consumption of the eNodeBs when they are powered on, irrespective of the traffic loading. The load/traffic sharing algorithm transfers traffic between eNodeBs which enabled the off mode, sleep mode and multi-cell cooperation of the eNodeBs. The dynamic energy saving algorithm wasimplemented in MATLAB 2013b environment. The performance of the dynamic energy saving algorithm was carried out by simulation using the developed MATLAB graphical user interface (GUI) program called the LTE network energy saving analysis software based on dynamic scheduling. Energy savings were analysed for call blocking probabilitiesof,,,, and , while varying the energy load proportionality constant between and in steps ofAn optimum energy saving for the network was achieved when maintaining a call blocking probability of which corresponded to , , , , , , , , , and for the energy-load proportionality constant of and respectively. Validation of the proposed dynamic energy saving algorithm was carried out by comparison with the “always-on” algorithm by Chiaraviglio et al., (2012) and the “sleep-wake” algorithm by Hossain et al., (2013). The result showed that the proposed dynamic energy saving algorithm achieved the highest energy saving of and as compared to the “always-on” algorithm by Chiaraviglio et al., (2012)and the“sleep-wake” algorithm by Hossain et al., (2013)which achieve an energy saving of 0% and 40% respectively while guaranteeing a call blocking probability of at an energy-load proportionality constant of 1.
TABLE OF CONTENTS
TABLE OF CONTENTS vii
LIST OF FIGURES ix
LIST OF TABLES x
LIST OF ABBREVIATIONS xi
CHAPTER ONE: INTRODUCTION
1.1 BACKGROUND 1
1.2 STATEMENT OF PROBLEM 6
1.3 AIM AND OBJECTIVES 7
1.4 METHODOLOGY 7
1.5 SCOPE 8
1.6 SIGNIFICANT CONTRIBUTIONS 8
1.7 DISSERTATION ORGANIZATION 9
CHAPTER TWO: LITERATURE REVIEW
2.1 INTRODUCTION 10
2.2 REVIEW OF FUNDAMENTAL CONCEPTS 10
2.2.1 LTE Radio Access Scheme 10
2.2.2 Load Utilization Factor 12
2.2.3 Architecture and Power Consumption of eNodeBs 14
2.2.4 Multi-Cell Cooperation in Cellular Networks 15
2.2.5 Self-Organizing Networks 16
2.2.6 Power Consumption Model 16
2.2.7 Energy-Load Proportionality Constant 17
2.2.8 Power Consumption Parameters 17
2.2.9 Blocking Probability 18
2.2.10 Load Curve 19
2.2.11 Review of Existing Algorithms 20
184.108.40.206 Always-On Algorithm 20
220.127.116.11 Sleep-Wake Algorithm 21
2.3 REVIEW OF SIMILAR RESEARCH WORKS 23
CHAPTER THREE: MATERIALS AND METHODS
3.1 INTRODUCTION 30
3.2 MODELING THE LTE CELLULAR ENVIRONMENT 30
3.2.1 Cell Structure 31
3.2.2 Location of eNodeBs 32
3.2.3 Mobile Stations 35
3.2.4 Adjacent eNodeBs 38
3.3 DEVELOPMENT OF THE ENERGY SAVING MODEL 40
3.3.1 Quality of Service Constraint 44
3.4 PROPOSED DYNAMIC ENERGY SAVING ALGORITHM 45
3.4.1 Energy Estimation Algorithm 45
3.4.2 Load/Traffic Sharing Algorithm 48
3.4.3 Dynamic Energy Saving Algorithm 51
3.4.4 The Developed MATLAB GUI for Energy Saving Analysis 52
3.5 SIMULATION SETUP 55
CHAPTER FOUR: RESULTS AND DISCUSSIONS
4.1 INTRODUCTION 57
4.2 MOBILE STATION DISTRIBUTION FACTOR 57
4.3 INSTANTANEOUS POWER CONSUMPTION 58
4.4 ENERGY CONSUMPTION 59
4.5 DAILY ENERGY SAVING 62
4.6 VALIDATION 65
CHAPTER FIVE: SUMMARY, CONCLUSION AND RECOMMENDATIONS
5.1 INTRODUCTION 68
5.2 SUMMARY 68
5.3 CONCLUSION 69
5.4 LIMITATIONS 69
5.5 RECOMMENDATIONS FOR FURTHER WORK 70
The information and communication technology (ICT) systems consume up to 10% of the world‟s energy accounting for about 2% of global emissions(Marsan et al., 2009). The telecommunications network is one of the main energy consumer of the information and communication technology sector(Yigitel et al., 2014). About 37% of the total emissionsfrom ICT devices and systems are due to the telecommunication infrastructure and devices (Oh and Krishnamachari, 2010), where about a tenth of the estimate is due to cellularmobile communication networks (Son et al., 2013). This accounts for about 0.2% of the global emissions and 1% of the world energy consumption(Richter et al., 2009). The mobile cellular communications sector alone consumes approximately 60 billion kWh per year(Dufková et al., 2010).Correspondingly, energy consumption as well as footprint of mobile cellular networks are increasing at an alarming rate due to the exponential growth in mobile data traffic (Wu et al., 2015). A projection showing the exponential growth in global mobile data trafficis illustrated in Figure 1.1(“Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2012-2017,” 2013).
Figure 1.1: A Projection of Mobile Data Traffic Growth(“Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2012-2017,” 2013)
This leads to high network operating costs and a considerable contribution to the worsening global warming phenomenon respectively(Son and Krishnamachari, 2012).A typical power consumption of a mobile cellular network is shown in Figure 1.2 (Han et al., 2011).
Figure 1.2: Power Consumption of a Mobile Cellular Network(Han et al., 2011)
As can been seen from Figure 1.2, in amobile cellular network, the base station consumes the highest amount of energy which represents about 58% of the total energy utilization, whereas the accumulated energy requirement for mobile stations is approximately 3% of the total energy utilization. However, current mobile cellular networks are typically designed and operated to meet a given coverage and capacity level by considering the peak traffic demand, while energy efficiency takes a minor (or no role at all) at the design and operation stages(Oh et al., 2013).
Consequently, minimization of energy consumption at base stations will considerably enhance the energy efficiency of cellular networks(Zhang and Wang, 2013). Remarkably, contemporary base stations have a high-degree of non-load-energy proportional consumption characteristic and consume significant amount of energy even atno-load condition (Xiang et al., 2013). For example, a typical active base station consumes 800-1500W, while its power
amplifier output needed to transmit from the antennas amounts to 40-80W only during the high-traffic hours (Han et al., 2013). This implies that a base station consumes more than 90% of the operational power when there is no transmission (Han et al., 2013).
On the other hand, cellular mobile network traffic exhibits a high-degreeof temporal-spatial diversity, which means that traffic demand varies both in time and space (Peng et al., 2011). This variation is directly related to the random call making behaviour and mobility pattern of the mobile users(Paul et al., 2011). A typical normalized traffic profile of a mobile cellular network for a week is as shown in Figure 1.3(Oh and Krishnamachari, 2010).
Figure 1.3: A Normalised Traffic Profile of a Cellular Mobile Network(Oh and Krishnamachari, 2010)
As shown in Figure 1.3, the traffic load decreases dramatically during the late night hours. There is also low traffic all day long during weekends or holidays in particular places such as government (or private) offices, schools etc., which operate mostly on week days. Therefore, infrastructures of the cellular access networks are underutilized during the low traffic periods(Yigitel et al., 2014).
However, under the current network operation approach, all base stations are keptpowered on irrespective of traffic load (Wu and Niu, 2012).This traditional network operation and the aforementioned non-load-energy proportional utilization at base stationsare the major causes
of the substantial amount of energy wastage in existing cellular networks(Chiaraviglio et al., 2012).
Correspondingly, a cellular network is a radio network distributed over geographical areas called cells, each served by at least one fixedlocation transceiver known as a cell site or base station. (Murthy and Kavitha, 2012). Cellular networks run on technology platforms which evolved from the first generation (1G) to the current fourth generation (4G) Long Term Evolution (LTE) mobile networks.The LTE standard was developed by the 3rd Generation Partnership Project (3GPP) to cope with the rapid increase of mobile data usage (Sesia et al., 2009).
The LTE is marketed as 4G LTE. LTE is a standard for wireless data communications technology and a development from the Global System for Mobile communication (GSM ) and Universal Mobile Terrestrial System (UMTS) standards(Sesia et al., 2009). The main goal of LTE is to provide a high data rate, low latency and packet optimised radio access technology supporting flexible bandwidth deployments. The network architecture of an LTE mobile network is as shown in Figure 1.4.
Figure 1.4: LTE Mobile Network Architecture(Sesia et al., 2009)
As illustrated in Figure 1.4, the network architecture of LTE consists of two network domains – an access network and a core network. This research work focuses on the access network which contains the base stations that consumes the highest amount of energy in the LTE mobile network. The network architecture of LTE comprises the following three main components (Sesia et al., 2009):
i. The Mobile Station (MS)
ii. The Evolved Universal Terrestrial Radio Access Network (E-UTRAN)
iii. The Evolved Packet Core (EPC)
The E-UTRAN comprises of the evolved base stations called eNodeBs which handles the radio communicationbetween the MS and the EPC.The eNodeBs are connected to EPCvia serving gateways (SGWs) and mobility management entities (MMEs) using the “SI” interface. The eNodeBs are interconnected with each other using an “X2” interface, which is introduced for inter-eNodeB message exchange in various coordination and cooperation phases (Sesia et al., 2009). The X2 interface is used for signalling and packet forwarding during handover.
The E-UTRAN which consists the eNodeBs is the major energy consumer of the LTE mobile network. Thus, it is imperative to exploit the non-load-energy proportional utilization ofeNodeBs to devise techniques that manage the energy consumption of LTE mobile access networks more efficiently (Hasan et al., 2011).Switching-off eNodeBs during low traffic situations has been proposed for future LTE systems, however the standard so far does not specify implementation schemes (Hossain et al., 2011). This is the primary focus of green cellular mobile communication. It finds radio networking solutions that can greatly improve energy saving and resource efficiency without compromising the quality of service of the mobile stations (Han et al., 2014).Given the nature of energy wastage, natural traffic diversity
and the projection of traffic growth trend, it is crucial to develop dynamic algorithms to improveenergy savingin LTE mobile access networks.Many algorithms have been developed by researchers in the area of green wireless communication for minimizing the energy consumption of base stations using sleep mode utilization. Most of the researchers validate the performance of their algorithms against standard algorithms which are the “always-on” algorithm by Chiaraviglio et al., (2012) and the “sleep-wake” algorithm by Hossain et al., (2013)(Bousia et al., 2014).
1.2 STATEMENT OF PROBLEM
The non-load-energy consumption proportionality of eNodeBs are the major causes behind the substantial amount of energy wastage in LTE cellular access networks, especially in the low traffic periods (Chiaraviglio et al., 2012). The existing dynamic switching off/on energy saving algorithms of LTE cellular access network eNodeBs was developed using the constant eNodeB power consumption model (Oh et al., 2013). This leads to the non-load-energy consumption proportionality of the eNodeBs which results to an increased energy consumption at low traffic period. Consequently, there is a need to develop a robust dynamic energy saving algorithm for LTE mobile access networks that will incorporate the load-proportional power consumption model of eNodeBs and allowing all eNodeBs under a cluster coordinate with their neighbouring eNodeBs to turn off/sleep the least loaded eNodeBs through load sharing with any moderately loaded eNodeBs.The algorithm will incorporate the inherent temporal -time traffic diversity of cellular access networks and the load-proportional power consumption model of eNodeBs. The proposed algorithm is expected to improve the energy savings by at least 50% while guaranteeing the quality of service offered to the mobile stations using off mode, sleep mode and active mode utilization of eNodeBs in a muti-cell cooperation and self-organizing manner.
1.3 AIM AND OBJECTIVES
The aim of this research is the development of a dynamic algorithm to improve energy saving in LTE mobile access networks while guaranteeing the quality of service offered to mobile stations. The objectives of this research work are as follows:
i. Development of a dynamic algorithm for improving energy saving of LTE access network.
ii. Development of an energy saving analysis MATLAB graphical user interface (GUI).
iii. Evaluation of the energy savings resulting from the developed algorithm.
iv. Validation of the developed algorithm by comparing its performance in terms of the energy saving and blocking probability with the“always-on” algorithm by Chiaraviglio et al., (2012) and the “sleep-wake” algorithmby Hossain et al., (2013).
The step by step approach of the proposed methodology is itemized as:
i. Development of mathematical model for the LTE cellular environment to represent the cell structure, location of an eNodeB, mobile station distribution factor and adjacent eNodeBs.
ii. Development of mathematical model for the energy saving of the LTE mobile access networks and the quality of service constraint (blocking probability).
iii. Development of an energy estimation algorithm, a load/traffic sharing algorithm and the integration of the two algorithms to form the dynamic energy saving algorithm for the LTE mobile access networks.
iv. Development of a MATLAB GUI for the simulation and analysis of the LTE access network energy saving.
v. Simulation and analysis of the energy consumption of the LTE access network resulting from the energy estimation algorithm.
vi. Validating the developed dynamic energy saving algorithm by comparing its performance in terms of the energy saving and blocking probability with the “always-on” algorithm and the “sleep-wake” algorithm‟.
This dissertation entails the energy saving in homogeneous LTE mobile access network by implementing dynamic algorithm at the eNodeBs to enable inter-base station cooperation while guarantee the quality of service offered to the mobile station using the downlink communication (that is from eNodeB to mobile station).
1.6 SIGNIFICANT CONTRIBUTIONS
The significant contributions of this research work are as follows:
i. Development of a distributed sleep-wake, self-organizing and dynamic energy saving algorithm for LTE mobile access networks that uses three different modes of operation: off mode, sleep mode and active mode.
ii. Development of an LTE network energy saving analysis software (MATLAB graphical user interface) based on dynamic scheduling for energy saving analysis.
iii. The proposed dynamic energy saving algorithm achieved a maximum energy saving of and with respect to the “always-on” algorithm by Chiaraviglio et al., (2012) and “sleep-wake” algorithm by Hossain et al., (2013)which achieve an energy saving of 0% and 40% respectivelywhile guaranteeing a call blocking probability of .
1.7 DISSERTATION ORGANIZATION
The general introduction has been presented in Chapter One. The rest of the chapters are presented as follows: A review of the fundamental concepts of LTE, energy saving in LTE mobile access networks and a review of similar research works are presented in Chapter Two. Modeling of the LTE cellular environment, development of the energy saving model in the LTE mobile access networks, the quality of service constraint model as well as the proposed dynamic energy saving algorithm are presented in Chapter Three. Analysis and discussion of the results are presented in Chapter Four. Conclusion, Limitation and recommendation for further works are discussed in Chapter Five.
[email protected].[email protected].