Overview of Green Wireless Networks

Overview of Green Wireless Networks Abstract: Traditional mobile networks largely focus on availability, variety, stability and large capacity. Due to the rapid development of Information and Communication Technology (ICT) industry whose major constituent are the mobile networks, CO2 emissions have been increasing rapidly. This shows the need for energy efficient wireless networks or green wireless networks which will put emphasis on saving the energy and environmental protection. The current wireless networks concentrates mainly on non-energy related factors such as Quality of Service (QoS), throughput and reliability. So these factors have to kept in mind while transitioning to green wireless networks. The techniques that need to be implemented are aimed at improving energy efficiency but not compromising the QoS, throughput and reliability.   In this paper the various metrics which help in evaluating performance of wireless networks are reviewed. Also different approaches to improve energy efficiency in wireless networks an d how to combine them for higher energy efficiency are discussed. Introduction: The latest mobile phones provide multiple services which led to rise of ICT traffic. ICT constitutes for 2% of total Green House Gas (GHG) and CO2 emissions worldwide. Within the ICT sector, mobile sector was responsible for 43% of emissions until 2002 while studies suggest that this number would go up to 51% by 2020 [1]. The predominant energy consuming part in a wireless network is the Radio Access Network (RAN). This comes from the fact the RF power amplifier within the RAN consumes a lot of input power for operation and releases a lot of heat contributing to energy wastage. In addition to this, the inconsistent distribution of real world mobile traffic among the BSs leads to underutilization of supplied energy [1]. These two reasons give us an idea of where the energy is being wasted or not utilized, helping us in formulating new techniques for energy efficient wireless networks. While discussing about various techniques for energy efficiency, we have to keep in mind that the QoS is not compromised at all. Because if an operator uses a technique, they should be able to serve the users by utilizing less energy but not by compromising users service. The various parts of a mobile network that consume power are data centers in backhaul, macro cells, femtocells, mobile stations or end hosts and their services. But the major part that consumes the highest power is the power amplifier section and Base station or RAN section. Hence the various techniques presented in this paper are aimed at energy efficiency in these sections only. Section II of the paper outlines various metrics which can be used to evaluate the energy performance of systems. Section III discusses cell layout adaptation techniques for reducing energy consumption and is divided into 3-subsections that outline various cell shaping algorithms. Section IV explains some challenges and research directions for energy efficient networks like Cognitive Radio (CR), M2M communication etc. Metrics for measuring energy performance: Energy efficiency can be achieved by employing better techniques. But in order to measure the energy consumption or utilization, metrics are needed. Energy efficiency metric can be defined as ratio of output to the input power supplied [1]. The output here may correspond to how much the distance of transmission is, number of bits transmitted or output power etc. The metrics for energy efficiency are broadly categorized into 3 levels: Component level metrics, Access node level metrics and network level metrics. Component level metrics mainly focus on power amplifier section, Access node level metrics focus on RAN or Base station, and network level metrics focus on performance of RAN [1]. These metrics help to quantify energy efficiency of various devices and therefore it becomes easier to compare which technique is better. Firstly, at the component level, where we focus on power amplifier section, there are 2 possible types of metric categories. One is Analog and the other is digital. The two important metrics of analog RF transmission are Power Amplifier efficiency (PA efficiency) and Peak to average power ratio (PAPR). PA efficiency is the ratio of PA output power to the input power supplied to it. Higher value of PA efficiency is desired, but in reality this is the part where most of the input power is wasted. PAPR, as the name suggests is the ratio of Peak power to average power. Lower value of PAPR is desired, as higher values tend to reduce the amplifier efficiency. The significant digital metrics in component level are millions of instructions per second per watt (MIPS/W) and millions of floating point operations per second per watt (MFLOPS/W). Higher value of MIPS/W and MFLOPS/W are desired as they refer to digital output generated for a given power input [1]. Secondly, at access node level there are 2 major metrics. Power efficiency and radio efficiency. Power efficiency refers to transmitted data rate over a given bandwidth and input power supplied. It is measured in bits per second per hertz per watt (b/S/Hz/W). Radio efficiency refers to transmitted data rate and transmitted distance over a given bandwidth and input power supplied. It is measured in bits meters per second per hertz per watt (b-m/S/Hz/W) [1]. Higher values of power and radio efficiency are desired as they measure the data rate and distance of transmission which are always desired to be a higher value for a given power input. Finally, at the network level also there are 2 metrics which measure the number of subscribers served during peak hours and coverage area respectively. The first metric measures the number of subscribers served during peak hour to the supplied input power and is measured in number of subscribers per watt (Subscribers/W) and the second metric measures the coverage area of the radio signal to the supplied input power and is measured in square meters per watt (m2/W) [1]. Higher values for both these metrics are desired as they signify serving more number of subscribers or a larger coverage area for a given power input. Hence when evaluating various techniques for wireless energy efficiency, it is better to know at whether energy efficiency is augmented in component level or access node level or network level. That way it would become easier to compare the efficiency in terms of various levels individual metrics. Reducing Energy consumption through Cell Layout Adaptation: Cell layout adaptation (CLA) techniques focus on energy efficiency at network level. But for these techniques to improve energy efficiency, it is important to improve efficiency in component level and access node level as well, because all these 3 levels are inter-related to each other and one works on the basis of another. Power is first supplied from power amplifier and then to RAN and at last to the network level, that means it is possible to save more energy in component level and access node level and the remaining energy that is used by the network can be efficiently utilized by implementing these cell layout adaptation techniques. CLA techniques are basically divided into 3 major categories. First part consists of cell shaping techniques like Base Stations (BSs) turning off and cell breathing, second part consists of hybrid macro femtocell deployments and the final part consists of relaying techniques [1]. A. Cell Shaping Techniques: As mentioned earlier, base stations turning off and cell breathing techniques encompass cell shaping techniques. The basic idea behind the former is turning off BSs and redistributing the remaining traffic to neighboring base stations. Here we need to make sure that we are turning off BSs which are idle or the ones which have very less traffic that can be taken up by neighboring cells. This way energy consumption is reduced and only the BSs that have traffic will operate and consume energy. Cell breathing scheme goes one step further by not actually turning off BSs, but by reducing the power consumption of a cell. This can be achieved by covering a low distance depending on the traffic. That means BSs experiencing higher traffic operate in full power mode while the BSs with medium traffic operate in medium power mode and cells with very less traffic operate at low power mode, thereby reducing the coverage area depending on subscriber traffic. This is analogous to a cell breathing acc ording to traffic patterns. As these cell shaping techniques are based on network level, number of subscribers served and coverage area metrics should be maintained in order to ensure good QoS and less call drop rate when implementing these techniques. The broader explanation of cell shaping techniques is mentioned above, but to implement those techniques there are 2 major algorithms: switching-off network planning algorithm and cell breathing coordination algorithm [1]. Firstly, switching-off network planning algorithm works on the basis of 3 factors, number of BSs to turn off, number of BSs to operate, and time period for which BSs are turned off. The ratio of number of BSs to turn off and BSs to operate and a specific time period for which turn-off is implemented based on the low traffic pattern is calculated. Once these values are calculated, it is made sure that the blocking probability limit is not exceeded, which means definite QoS is maintained. Cell breathing coordination algorithm works on the basis of a central node called a cell zooming server. The cell zooming server analyzes the incoming traffic and tries to turn of the BSs which do not have any traffic at all. Then it tries to distribute the traffic from less active BSs to busy BSs. It also makes sure to distribute traffic based on input traffic and turns on the sleep mode BSs when required. This centralized approach works good in smaller networks and when it comes to large scale networks, it would be very ineffective. The same applies to switching off network planning algorithm because there is no centralized node to turn on the BSs if needed, as the turn -off time if fixed based on traffic patterns [1]. The cell shaping techniques also bring up a new trade-off, i.e. SE-EE tradeoff (spectral efficiency-energy efficiency) [3]. SE-EE trade-off focuses on network level characteristics like number of subscribers served and coverage area for input power supplied. By implementing these cell shaping techniques although energy efficiency is obtained, there is always chance where coverage area is reduced and some subscribers are ignored. Ideally, higher the energy efficiency lower is the spectral efficiency. But in reality, because of component level energy issues, transmission distances, coding schemes the relationship between SE and EE is not inversely proportional, but it is of the form of a bell curve. So it is better to apply cell shaping techniques until the point where spectral efficiency is not compromised. B. Hybrid macro femtocell deployment: Femtocell deployment in combination with macro cells is a second method under cell layout adaptation. Femtocell deployments provide coverage in the order of 10 meters and utilize a small BS, which requires less power to operate. Femtocell deployment is advantageous as it provides good coverage and QoS to a set of users within its range with less operating expenses when compared to a macro BS [1]. Although femtocell deployment is a good concept, it is not desirable to have too many femtocells as it increases the power consumption and utilizes more network resources for a lesser coverage area. A better way of deployment is having hybrid macro and femtocell deployment. In the case of hybrid deployment, the macro BS provides coverage to users who are evenly spread over a long distance and the femtocell provides coverage to users who are located in a dense region. This way the energy is utilized efficiently, as a new macro BS is not being deployed to provide coverage to those dense set of users. The hybrid macro cell and -femtocell deployment poses a new challenge for handoffs, as macro BS and femtocell BS might have same signal strength in the others coverage region. The handoffs issue can be solved by defining a clear boundary between the macro and femtocell BS. Within the dense region, the femtocell should have higher signal strength and it should properly handoff at the bounda ry of macro BS. Also within the coverage area of macro BS, the femtocell BS should have very less signal strength [1]. This would give a clear idea to define a boundary. A better way of implementing this hybrid deployment is by utilizing the cell shaping techniques like BS turning off and cell breathing coordination. If there are a set of femtocells, and one of the coverage area is totally idle, then that femtocell BS can be turned off and basic coverage is provided by the macro BS at that location. Similarly, if incoming traffic is analyzed, femtocells and macro cells can use the cell breathing techniques to lower their power utilization [1]. Also the hybrid macro and femtocell deployment leads to a DE-EE tradeoff (deployment efficiency-energy efficiency) [3]. Ideally energy efficiency increases when more femtocells are deployed and deployment efficiency goes down because of increase in deployment expenses, network utilization and energy consumption. In a practical scenario, the relationship between DE and EE is more like a bell curve, with a peak point where deployment and energy efficiency are in good standing. Hence it is a good idea to use hybrid deployment until the point where it does not degrade the deployment efficiency and energy efficiency. C. Relaying techniques: Energy efficiency can be achieved through 2 types of relaying techniques. The first technique uses repeater stations or green antennas for relaying and the second technique uses mobile stations for relaying. In the first technique, a repeater station or a green antenna with receiver capability is connected to the macro BS through a coax cable or optical fiber, with the latter utilizing less energy. These green antennas are placed very near to the mobile stations, which helps to reduce the energy consumption in uplink by the mobile stations. Although this technique improves energy efficiency for mobile stations, it increases operating expenses for the service provider. In the second technique, the mobile stations work in coordination and perform the relaying operation. This way the transmission distance for the macro BS is reduced and it consumes less energy. Although this technique assumes mobile stations as relays which work selflessly. Practically, the mobile stations may not work in coordination which would break the link for relaying. One more drawback of this technique is that for maintaining coordination between the mobile stations, more energy is consumed [1]. Challenges and directions for energy efficient wireless networks: Cognitive Radio (CR) and M2M (Machine to Machine) communication systems provide new opportunities in the field of green wireless networks, but also pose significant challenges at the same time. Cognitive Radio can be defined as a RF transceiver that is used to switch users from a very busy spectrum to an unused one and vice versa if needed. The origin for this topic came from the fact that many RF spectrums are congested with several users and some other spectrums are underutilized. Hence the concept of CR would efficiently manage users in various spectrums and help to deliver better QoS. Indirectly this switching of spectrums or utilizing unused spectrums is resulting in energy efficiency as spectrums with more users will not utilize additional energy as users are transferred to other spectrum. Also underutilized spectrums which were consuming energy for operation, now serve the new users efficiently resulting in energy and spectrum utilization. The only disadvantage of CR technique is that monitoring various RF spectrums and switching users from one spectrum to another requires significant energy. Hence this technique would be energy efficient only if more energy is saved by intelligently switching users than that is utilized for monitoring spectrums or users [2]. M2M wireless communication systems are aimed at connecting various wireless devices directly. This approach also helps in reducing energy consumption from the point of view of a mobile station. M2M helps to reduce the computation required by various physical devices and also tries to offload them to the network itself. This way the mobile stations consume less energy as the number of computations is reduced. The major disadvantage with this approach is that if more computation is offloaded to the main network, it might consume more energy that that is being saved by mobile stations by utilizing this approach. Hence this technique would be energy efficient if the main network does not consume a lot of energy for some additional computations [2]. Conclusion and Future Scope: The rise in carbon footprint, especially the contribution to it from the ICT sector and consequently mobile sector led to interest in energy efficient wireless networks. Energy efficiency can be achieved at various levels such as power amplifier, RAN and network. The techniques proposed in the paper focus on energy efficiency in RAN and network levels. But they also have trade-offs like DE-EE and SE-EE, which can be vanquished by emerging techniques like CR and M2M communications. These emerging techniques can be improved in a way where they consume less energy for monitoring in comparison with the prevailing levels. Alongside that, at the power amplifier level, the current solution for energy efficiency is to use expensive components which would trade off the gains achieved by energy savings. Hence a future research direction would be addressing energy efficiency at power amplifier level and improving CR and M2M techniques. VI. References: [1] Luis .S, Nuaymi .L, and Bonnin .J, An overview and classification of research approaches in green wireless networks. Eurasip journal on wireless communications and networking 2012.1 (2012): pp.1-18. [2] Xiaofei .W, et al. A survey of green mobile networks: Opportunities and challenges. Mobile Networks and Applications 17.1 (2012): pp.4-20. [3] Yan Chen; Shunqing Zhang; Shugong Xu; Li, G.Y., Fundamental trade-offs on green wireless networks, in Communications Magazine, IEEE , vol.49, no.6, pp.30-37, June 2011.