Green Small Cell Network, Dissertation Example

The exponential advancements of wireless technologies has forced operators to consider the deployment of small cell networks to supplement the high demands being placed on their macrocell networks with the hopes of improving the coverage capabilities as well as the power consumption demands. However, there are a multitude of challenges to such implementation, including the problem of backhaul, and there is additional ambiguity regarding the most appropriate solutions to such issues (Bhaumik, Narlikar and Chattopadhyay). Furthermore, the requirement of designing solutions that are environmentally conscientious and sustainable places additional challenges to contend with in the establishment of a viable solution. This paper aims to help progress the path to a solution by presenting the consensus around the operators’ requirements for small cell networks with a focus on sustainability through examination of the current suggested answers to this dilemma combined in one single body of research. This analysis will examine the simulations of the small cell network configurations similar to those for macro sites, including the cost per small cell, network security options, sustainable aspects of the design, and many other facets relative to the plausibility of a solution. The Quality of Service (QoS) provided to the user is also a grave consideration and the ability to provide the same level of service will be a large factor in determining the benefits associated with a solution. Consideration of the most plausible options will also recommend that the operator plan the small cell network according to the anticipated network traffic.

Table of Contents

Acknowledgements. 2

Executive Summary. 3

Introduction. 7

Terms of Reference. 9

Scope. 13

Working Definition of Small Cell 14

Explanation of Current Network Protocols. 16

WEP (Wired Equivalent Privacy) 16

WPA (Wi-Fi Protected Access)/WPA2. 19

802.11i / Robust Secure Network (RSN) 21

VPN.. 22

Procedure. 24

Throughput (Input/Output) 25

Network Deployment Cost 28

ECR (Energy Consumption Ratio) & the Energy Consumption Gain (ECG) 30

Simulation Analysis. 32

Findings. 34

Simulation Results and Discussion. 34

Conclusions. 37

Recommendations. 38

References/Bibliography. 40

Appendices. 42

Appendix A. 42

Appendix B. 43

Appendix C.. 43

List of Figures

Figure 1: Projected BS Growth in Developing Regions, 2007-2012. 7

Figure 2: Hexagonal Network. 32

Figure 3: Cellular Geometric Path. 32

Figure 4: ECR & Capacity Density by Cell #

Figure 5: Capacity density by Cell #. 43

List of Tables

Table 1: Definitions of Terms. 9

Table 2: BS Class Power Models. 25

Table 3: Macro, micro, & Pico, CapEx & OpEx. 27

Table 4: Simulated Service Areas. 33

Table 7: Cell Capacity in MBits/s for Different Antenna Tilts Averaged over the Service Area. 35

Table 8: Simulated Energy Efficiency Rates. 36

Table 5: System Simulation Assumptions. 42

Table 6: Energy Saving by Power Method. 43

Green Small Cell Cellular Network with Multiple Cells to Reduce Power Consumption

Introduction

Technological advancements have revolutionized the way the world communicates and, in order to stay competitive, cellular phone manufacturers need to make sure their customers have access to the latest technological services available on a device capable of sustaining use for long periods of time without power or connection issues. This necessitates knowing the optimal configuration for power cells as well as how they can be produced in a sustainable manner. The current power demands include provisioning services to a myriad of remote device like personal computers (PCs), personal data assistants (PDAs), and mobile/cellular phones. There current demands for remote data access has facilitated the construction of more than 4 million base stations (BSs) throughout the nation that are serving mobile users, each consuming energy at an average rate of 25MWh per year (Hasan, Boostanimehr and Bhargava). Furthermore, according to GSMA Research, the number of BSs in developing regions was projected to nearly double between 2007-2012, as shown in fig. 1 (Hasan, Boostanimehr and Bhargava).

Figure 1: Projected BS Growth in Developing Regions, 2007-2012

Despite the phenomenal technological growth, there remains the contemporary issue of environmental conservation and sustainability in conjunction with technological developments. This presents a the global challenge of reducing greenhouse gas emissions while effectually accommodating the explosive demand of wireless data traffic through green architecture designs that satisfy this critical issue for mobile network operators (He, Zhang and Chen). Further complicating the possibility of a singular solution is the heterogeneous deployment of different cell types, although the intent was to assist in the fulfillment of the above mentioned challenges. In this capacity, a critical concern for cellular network facilitators concerns the deployment of small cells in a green manner that allows the global network to be cost-effective as well as energy-efficient. For the purpose of this discourse, we will deem the terms energy efficiency (EE) and deployment efficiency (DE) as applicable to heterogeneous wireless networks, providing appropriate consideration for realistic network power consumption models in addition to dynamic network configurations.

The main purpose of this paper is to present the various green small cellular network configurations for the purpose of presenting a comparative analysis of the different configurations. Through this analysis, the most efficient, sustainable design structure for a green cellular network will be shown, illustrating the proper density of small cells required to obtain the maximum achievable EE/DE (He, Zhang and Chen). In doing so, this study will provides useful insights for the modeling and deployment of future green wireless networks and how such can be applied to small cellular mobile networks to achieve power optimization and save energy (He, Zhang and Chen). The performance of the analysis will include calculations regarding throughput, ECR, graphs and other visual aids to achieve a comprehensive perception of green small cellular networks. The remainder of this paper will follow with an explanation of the terms to be used in the discourse, the scope of the investigation delineating the study parameters, the procedural analysis including the throughput, ECR, and other calculations, findings of the research with discussion of the determinations, and a conclusive summary of the research.

Terms of Reference

In addition to the abbreviations EE/DE, as identified above, specific industry terms of reference necessitates a basic familiarity in order to understand this discourse and the functional definitions as well as the abbreviations for such terms are presented in Table 1, as they are relevant to this paper. This table is adapted from (Badic, O’Farrell and Loskot; Bhaumik, Narlikar and Chattopadhyay; ETSI; Hasan, Boostanimehr and Bhargava; Jaloun and Guennoun; Nema, Goel and Singh; Sinha, Snai and Mitra).

Table 1: Definitions of Terms

Although all of these terms may not be directly discussed within the body of this work, they are all relative to the facilitation of cellular networks as well as wireless communications.

Scope

The scope of this study aims to provide a comprehensive analysis of the aggregate pool of industry knowledge regarding the requirements for small cell networks as well as the optimal configurations for a green or sustainable small cell network designed for minimal power consumption. The necessary steps involve attending to considerations such as:

• Agreement in regard to the scope of uses for small cells
• Standards specifying the capacity and Quality of Service (QoS)
• Consideration for different small cell incidences and definite requirements determining what is needed from the backhaul
• Security aspects that includes the architectures and network topologies
• Attributes of ‘sustainability’ and what challenges arise in meeting such requirements
• Total Cost of Ownership (TCO) specifications for the small cells backhauling
• Ability of small cell network to support 3G, 4G, and LTE devices

Working Definition of Small Cell

Although there is no unilateral industry definition regarding what is classified as a ‘small cell network’, this section will establish this classification for the purposes of this discourse. To achieve this, we will draw upon the researched attributes of various networks to differentiate between common aspects of all networks and those specific to a small cell network. Networks facilitating 3GPP usage have cell types that are categorized according to the minimum coupling losses between cell site and user device, which has originated four classes of cells (Bhaumik, Narlikar and Chattopadhyay). Additionally, researchers also consider the cell radius, amount of connected users, deployment options, and other relative aspects, such as the criteria considered for P-BEV project small cells, which are:

• The size of the coverage area, such that one macro cell intersects with several small cells in the same area because the coverage of an area is smaller than a macro cell
• Macro cells are deployed and managed by operators that grant open access to all users of the same operator
• Small cell networks are typically involve a lower equipment and installation cost in comparison to macro cells
• Small cell networks are typically engineered to support data services, although they can also support voice services (Bhaumik, Narlikar and Chattopadhyay)

The parameters of the definition of small cells can be further expounded and detailed through analysis of the following technical parameters that, although no specific values will be indicated, will be factored in to the consideration of the small cell network paradigm in this paper where use cases or service profiles are characterized by:

• Average and peak capacity,
• The type of data, real time data, and voice services supported,
• The mobility support, which includes S1/X2 traffic ratios, handover between macro and small cells or among small cells, and support of the X2 interface,
• The QoS requirements in terms of jitter, latency, packet loss, and availability in addition to time and frequency synchronization requirements,
• Deployment requirements,
• Power consumption associated with backhauling,
• Operational conditions, including public access and the operator’s deployed backhaul, and
• The location potential, including aspects such as indoor/outdoor, and height relative to street or rooftop levels

Another attribute for consideration involves the technique of cell breathing, which constitutes the procedure of adjusting local cell boundaries to achieve load balancing (Bhaumik, Narlikar and Chattopadhyay). Hierarchical cellular networks have been proposed with the aim of increasing capacity using smaller cells while avoiding high hand-off rates, or to integrate multiple wireless technologies. Multi-hop wireless networks that relay data to one mobile through another mobile for improved coverage and capacity have also been proposed, including schemes to save the mobile’s energy consumption. However, as far as we know, this is the first paper to propose multi-layer networks to reduce the total energy consumption of the cellular network (Bhaumik, Narlikar and Chattopadhyay). The next section highlights the pros and cons of the four principal networks involved in wireless communication and conclude with a summation of how those parameters are amalgamated to form the small cells use cases considered in this paper.

Explanation of Current Network Protocols

WEP (Wired Equivalent Privacy)

WEP is an encryption algorithm incorporated in the 802.11b standards designed to be as secure as a wired LAN, but is prone to many attacks, such as passive attacks that can decrypt traffic based on statistical analysis, active attacks to inject new traffic based on known plain text that originates from unauthorized mobile stations, active attacks to decrypt traffic that tricks the access point, and a dictionary-building attack allows real time automated decryption of traffic using the previous day’s analysis of the site traffic (Hasan, Boostanimehr and Bhargava). WEP algorithms implement a multitude of techniques, which includes:

• 2 party or 3 way handshakes
• Single static pre-shared 40/104 bit key
• Challenge response authentication mechanism using Shared Key Authentication
• Packet integrity check
• Initialization vector for creating freshness of encryption key
• Self-synchronizing
• RC4 Symmetric key cipher encryption algorithm(Boudreau, Panicker and Guo)

As the original algorithm of this nature, WEP is currently the weakest of the protocols because, as the oldest, it has the largest number of known attacks (Saunders and Argón-Zavala). The authentication process for WEP uses a keyed challenge response mechanism, where the client requests a challenge message to encrypt from the server, and if the client sends back a valid encryption of that message under the shared key, that client is considered authenticated (Jaloun and Guennoun). Additionally, WEP only uses one access point, which is the ‘server’, to authenticate a client and does not implement any of the modern EAP-based authentication mechanisms, and which leaves the network vulnerable to known MITM attacks that allow a client to pose as a server (Hoadley). The encryption strength of the WEP network uses a static key shared amongst all users for a given wireless network identifier with only three other chosen keys that a server can rotate the key. WEP uses the well-known and accepted cipher RC4 to encrypt data in linear time. Unfortunately, WEP uses this solid encryption scheme in a way that makes RC4 vulnerable to attack and thus RC4 is not a benefit, but a liability.

WEP uses a relatively weak 40 bit initialization vector (IV) version in comparison to the 104 bit version and other 128 bit keys in other protocols, since the 40 bit version was created to meet exportation laws (Jaloun and Guennoun). The 2²⁴ space presents frequent opportunities for collisions, which, when combined with a short key on a high-traffic network makes it easy to access the network with as little as four million packets and the WEP network is also vulnerable to automatic revocation of the key from the network via a management interface not built-in to WEP (Jaloun and Guennoun). These WEP network uses a simple checksum to guarantee the integrity of the data itself and does nothing to guarantee the header. The WEP network is essentially vulnerable to attacks on all the network layers it uses and has a DoS weakness in its normal operation that requires the basic challenge response authentication mechanism to remember what challenge message it sent the client to prevent replay attacks, which helps the network protect the host’s identity since the only identity revealed is its association with a group that knows the key to the wireless station (ETSI). However, WEP specifies no use of cookies, although there is a instance of fresh encryption for each packet by using IVs that uses the small space of 2²⁴.

WEP is easy and relatively inexpensive to implement since it consists of two simple security mechanisms, which are the authentication phase and the communication under encryption phase. Both of these phases are remarkably simple. Authentication occurs with a simple challenge response exchange. Encryption occurs with an XOR which is linear in time relative to the length of the message. The computational cost of doing an XOR is quite low when compared to other encryption standards like AES. The number of actors needed to coordinate authentication and encryption are just the client and access point. Since WEP is already a mainstream product, products that support WEP versus no encryption are equally cheap. In sum, WEP is a great protocol when we examine just the cost in hardware,

In general, WEP does not consume much power. This is mainly due to the computational simplicity of the encryption scheme, but it can also be credited to the sleep mode standard with most implementations. Since WEP does not have a method for detecting attacks and automatically sleeping for a while. WEP neither provides a way to determine physical location nor provides something creative of its own. So, while WEP was the first approach at a wireless security protocol, something novel in and of itself, that novelty does not translate to measurable merits.

WPA (Wi-Fi Protected Access)/WPA2

WPA is basically a corrected version of WEP with additional mechanisms to improve security and performance. There are two modes of WPA, Pre-Shared Key mode which is quite similar to WEP and Enterprise mode, which will be considered since it is more secure than WEP and has additional mechanisms (ETSI). The enterprise mode is not used in a home wireless network because it requires a RADIUS server which is too expensive for the home user (ETSI). WPA networks use similar techniques to WEP, but have several features that are upgrades, as follows:

• 128 bit keys,
• 802.1X uses the standard for Extensible Authentication Protocol (EAP) to dynamically assigns and distributes keys per session/user/packet,
• MIC Packet Integrity Check is used to protect the header as well as the payload,
• 3 party, multi way handshake,
• Temporal Key Integrity Protocol (TKIP) provides a fresh key, and
• RC4 is used although some vendors also implement WPA2 which uses AES instead of RC4 (Saunders and Argón-Zavala)

WPA implements fixes for the multitude of issues apparent in the WEP network configuration to provide a much more solid protocol, starting with the improved authentication requirements (Saunders and Argón-Zavala). WPA employs a complex authentication mechanism involving four parties, which are the client, the access point, a RADIUS authentication server, and a certificate authority (ETSI). WPA allows four to five main variants of an Extensible Authentication Protocol (EAP). Included in this list are LEAP, EAP-TLS, EAP-TTLS, and PEAP and they always involve server certificates for verifying server identity and helping derive the key. The client can be authenticated using legacy methods like CHAP or more modern methods like EAP-MD5. There are no known MITM attacks for WPA, and while it is credentials based system, it uses only one authentication server. The encryption scheme employed by WPA is still RC4. However, WPA uses RC4 in a way that allows it to maintain its strength. For starters, the initialization vector is twice as long at 48 bits, and the key length is a standard 128 bits. Additionally, WPA has a built-in mechanism for providing fresh keys to clients. Every time a client authenticates, it and the server derive a new pair-wise key. While WEP had a later modification to allow a rotation of keys, called TKIP, WPA signifies the completion of this idea, in that it provides a fresh key every time. The larger space of IVs (2^48) ensures that there will be less collisions for the time that one key is used, and that trillions, not millions of packets must be collected before cryptanalysis can be done. One potential future vulnerability could be in how a server and client compute a fresh key. Each key should be perfectly independent of the previously derived key. WPA still uses the older encryption scheme RC4 to maintain backwards compatibility with older hardware. A stronger encryption scheme is still desired.

WPA uses the integrity method called Michael (MIC). MIC is described under Explanation of Terms. Our methodology could not be clearer, in that MIC is a technique more complex than a checksum, but still simpler than other signature methods. To prevent attacks, WPA uses per packet 48-bit IVs that provide good prevention of replay attacks. There is no cookie, so WPA is still vulnerable to DoS, and now there are more parties to deny service – the access point and the RADIUS server. There are no known attacks on WPA, but this may change with time and it does not use any cookies but asks revealing information online, so identity issues are prevalent.

The ease and cost of implementation for WPA is significantly more difficult to implement than WEP. The addition of two other parties, the RADIUS server and the certificate authority, require many more lines of code to implement the protocol correctly. While it is significantly more complex, the basic encryption scheme is still relatively simple, so WPA can still run on legacy hardware. In general, WPA does not consume much power, which is mainly due to the computational simplicity of the encryption scheme, but it can also be credited to the sleep mode standard with most implementations. Since WPA does not have a method for detecting attack and automatically sleeping for a while, it receives only a score of 1/4 exactly like WEP.

802.11i / Robust Secure Network (RSN)

While WPA implements the improvements possible on legacy hardware, RSN includes a multitude of hardware upgrades not present in WPA networks due to hardware restrictions. Many of the same techniques used in WPA networks are present in 802.11/RRSN networks, in addition to the many improvements, as follows:

• EAP
• 3 party, multi way handshake
• Advanced Encryption Standard
• Symmetric cipher
• 128 bit keys
• Requires more computationally powerful hardware
• Packet Integrity Check (Counter Mode Encryption)

The authentication methods and encryption techniques used in WPA are also used in RSN networks. AES provides a stronger encryption scheme over WEP’s RC4 and provides solid strength that accounts for the integrity, identity protection, and attack prevention issues present in WPA networks. However, it is not as easy to implement and requires additional hardware, which makes it more expensive than WPA and WEP network implementation. With more expensive hardware, AES can use the integrity method called CCM. Put shortly, CCM guarantees more than MIC. It is quite similar to WPA, except it is even more difficult to implement and upgrade since it requires new client hardware. When AES is implemented in hardware, its power consumption is drastically less than when in software. As a result, we assume that in the long run power consumption of AES will be acceptable. Much like WPA, since RSN does not have a method for detecting attack and automatically sleeping for a while, it receives only a score of (2/4), exactly like WEP and WPA. Much like WEP and WPA, RSN neither addresses our own idea of determining physical location nor creates something completely new that our methodology does not cover.

VPN

The Virtual Private Network is constructed by using public wires to privately connect nodes, such as with the number of systems that enable you to create networks using the Internet as the medium for data communication. These systems use encryption and other security mechanisms to ensure that only authorized users can access the network and that the data cannot be decrypted upon interception. The mechanism used in VPN differs from the other approach by “Tunneling,” which is the process of placing an entire packet within another packet and sending it over a network. The protocol of the outer packet is understood by the network and both end points, called tunnel interfaces for where the packet enters and exits the network. To implement tunneling, we require 3 additional protocols (Carrier protocol, Encapsulating protocol, and Passenger protocol)

VPN uses several methods for keeping the connection and data secure. Some mechanisms are firewalls, encryption, IPSec, and the AAA server:

• A firewall provides a strong barrier between your private network and the Internet. You can set firewalls to restrict the number of open ports, what types of packets are forwarded and which protocols are allowed.
• IPSec (Internet Protocol Security Protocol) provides enhanced security features such as better encryption algorithms and more comprehensive authentication. IPSec has two encryption modes: tunnel and transport. Tunnel mode encrypts the header and the payload of each packet while transport mode only encrypts the payload. Only systems that are IPSec compliant can take advantage of this protocol. Also, all devices must use a common key and the firewalls of each network must have very similar security policies set up. IPSec can encrypt data between various devices, such as: router to router, firewall to router, PC to router, and PC to server
• AAA (Authentication, Authorization and Accounting) servers are used for more secure access in a remote-access VPN environment. When a request to establish a session comes in from a dial-up client, the request is proxied to the AAA server. AAA then checks the following: who you are (authentication), what you are allowed to do (authorization) and what you actually do (accounting).

VPN seems to be powerful but VPN is also expensive. First of all, tunneling requires additional protocols to manage it (Carrier protocol, Encapsulating protocol, and Passenger protocol). Moreover there are other requirements for administrating the VPN – the administrator must know how much the VPN will be used and what type of data will be traveling through it. VPN is a type of overlay for WEP. Since it is not in and of itself a wireless protocol, it is not directly comparable. We can dream up many overlays for any of the technologies discussed, but we are focusing on securing the lowest layer possible. We include discussion of it above as an example of one solution to the WEP problem that industry has adopted but stop short of measuring it, because it is not the lowest layer of security.

Procedure

Most devices that use wireless connections run from a battery-powered power source. As a result, it is only fair to include power consumption in our evaluation of wireless protocols. Power consumption is best measured in a relative manner between protocols. For example, a protocol which uses AES will use more power than one that uses RC4. Additionally, when a client receives attack-like behavior, it would be good for a protocol to specify a means to detect the attack and conserve power. This is especially important for networks that are seldom recharged, like sensor networks. A wireless protocol evaluation methodology would be incomplete without considering power. It should be noted that implementing AES in hardware instead of running it from ROM software cuts the power cost significantly. In this section, we will introduce the concept and methodology of calculating the network deployment cost and network energy consumption, based on which we will then derive the tradeoff between DE and EE.

Addressing the challenge of the limited spectrum availability, coupled with the increasing consumer demand for stronger bandwidths capable of handling higher traffic volume, requires innovation so that consumer demand can be accommodated while the carrier’s business models continues to perform effectively (Hoadley). New solutions must be developed that: use available spectrum with the utmost efficiency to allow higher data throughput over the wireless link; support a greater number of users within individual cells and significantly enhance the user experience; and reduce the carrier cost of transporting megabit-rate traffic and carry that lower carrier cost through to the consumer (Hoadley). High capacity and variable bit rate information transmission with high bandwidth efficiency are the key requirements that the modern transceivers have to meet in order to provide a variety of new high quality services to be delivered to the customers (Nema, Goel and Singh). For this analysis, the bandwidth will be calculated as follows:

[Equation 1]

Throughput (Input/Output)

Here, we give the definition of the main metrics for the design of green wireless networks, such as deployment efficiency and energy efficiency. We first give the definition of network throughput which is the key measure of the wireless network performance. The network throughput is defined as the summation of the average throughput of all BS sites within the considered network area, where the average throughput of a BS site is the delivered information bits averaged over all mobile terminals served by that BS site, with the different BS class power models being illustrated comprehensively in Table 2, adapted from (He, Zhang and Chen).

Table 2: BS Class Power Models

When the small cells are deployed in the macro cell in a sparse manner, such that there is no significant interference between small cells, and given the frequency division spectrum scheme, the throughput of small cells T_s and throughput of macro cell T_m are given by:

[Equation 2] (He, Zhang and Chen)

[Equation 3] (He, Zhang and Chen)

where P_s and P_m are the transmit power of small cell BS and macro BS, respectively. Additionally, g_s = ?1r –? 1 and g_m = ?0r?? 0 are the channel gains of macro user and small cell users, correspondingly, while ?0 and ?1 represent fading coefficients (He, Zhang and Chen). In this instance, deployment efficiency and energy efficiency adhere to the following meanings:

• Deployment efficiency (DE) is defined as the ratio of the network throughput over the network deployment cost within a certain period. The network deployment cost includes two parts: Capital Expenditure (CapEx) and Operational Expenditure (OpEx). In short, CapEx is the money to buy fixed assets, e.g., BS equipments. OpEx is the ongoing cost for running BS, e.g., power and maintenance cost(He, Zhang and Chen).
• Energy efficiency (EE) is defined as the ratio of the network throughput over the network energy consumption within a certain period(He, Zhang and Chen).

We first analyze the energy model for the heterogeneous wireless networks. In general, the energy consumption of a heterogeneous wireless network can be considered as the summation of the energy consumption of different classes of BS sites, such that

[Equation 4] (He, Zhang and Chen)

where i is the class index, Ni and E_site i are the number of BS sites and site average power consumption in class i, respectively. The latter is the accumulation of power consumption P_site i over a certain time duration T (He, Zhang and Chen). Note that in this framework the power consumption of mobile terminals is not taken into consideration. Normally, the power consumption of a BS site includes power losses from circuit power of signal processing, radio frequency, A/D D/A converter, power supply, battery backup, antenna feeder, site cooling consumption, etc. It is shown in that the relationships between BS transmit power and BS site power consumption is nearly linear for LTE systems (He, Zhang and Chen). Thus, the BS site power consumption can be approximated using the following linear model:

[Equation 5] (He, Zhang and Chen)

where P^T i is the BS transmit power, P^base i is the power consumption when BS transmits at the minimum non-zero power, ?i is the slope of the traffic-dependent power consumption which depends mostly on the power amplifier efficiency, i.e., BS transmit power and traffic have a near-linear relation, Psleep i is the sleep mode power consumption that is normally smaller than Pbase i . Dynamically switching BS into sleep mode enables the deactivation of components inside BS when there is nothing to transmit, which is believed to be a promising solution for energy saving (He, Zhang and Chen). In this paper, we will analyze the network performance (throughput, EE and DE) taking into account the realistic power model (3) and the impact of dynamic BS sleeping. Some typical power model parameters for different BS classes are listed in Table 3 for reference.

Table 3: Macro, micro, & Pico, CapEx & OpEx

Network Deployment Cost

In general, the total cost of a heterogeneous wireless network can be consider as the summation of the cost from different classes of BS sites csite where CCapEx i and COpEx i are BS site’s CapEx and OpEx, which can be further specified as

[Equation 6] (He, Zhang and Chen)

where C^CapEx i and C^OpEx i are BS site’s CapEx and OpEx, which can be further specified as

[Equation 7] (He, Zhang and Chen)

[Equation 8] (He, Zhang and Chen)

where cBSE i , cRNE i and cSB i represent the cost of BS equipments (BSE), radio network equipments (RNE), and BS site buildout of class i, cpower i , ctrans i and clease i are the annual expense of electric power and man power, backhaul transmission lease, and BS site lease of class i (He, Zhang and Chen). Table 3 gives some typical CapEx and OpEx values for macro, micro, and pico BS (He, Zhang and Chen).

Based on the concepts of network energy consumption and deployment cost introduced above, we could then derive the concept of network energy efficiency and deployment efficiency (He, Zhang and Chen). Suppose that the transmit power of small cell BS ps and the transmit power of macro BS pm in equations (2) and (3) shall support the rate requirement of the corresponding users. Therefore, when a small cell is serving k users in its coverage, the throughput of the small cell is equal to the sum rate of the k users, such that Ts = kR and the transmit power of the small cell BS is specified by

[Equation 9] (He, Zhang and Chen)

where ?_SE = KR/B symbolizes the network spectrum efficiency in bps/Hz, such that k can be 1 to K with the probability ?k 1, as indicated in the equation adapted from (He, Zhang and Chen).

[Equation 10]

such that in this paper, a user is assumed to be served by small cell if it is located in a small cell’s coverage. Otherwise, the user will be served by macro cell. Similarly, the throughput of macro cell is the sum rate of all the macro users, i.e., , where n is the number of deployed small cells, is the average number of users served by a small cell BS^{2 (He, Zhang and Chen)}. Therefore, the transmit power of macro BS² is calculated by

[Equation 11] (He, Zhang and Chen) whereas the network energy efficiency can be expressed as

[Equation 12] (He, Zhang and Chen) where

[Equation 13] (He, Zhang and Chen)

[Equation 14] (He, Zhang and Chen)

where is the average in terms of user traffic power consumption of a small cell BS, p_m is the power consumption of the macro BS, ? is defined as the probability that a small cell is in the active mode, such as serving at least one user (He, Zhang and Chen).

ECR (Energy Consumption Ratio) & the Energy Consumption Gain (ECG)

Large cell networks or macrocells are effective in providing area coverage for voice and low-speed data traffic, but due to their typically large coverage macrocells are generally limited in providing high data rates per unit area. On the other hand, operating expenses of a macrocell are high, especially when subscriber revenue does not match the increased data rate demand. An increase of data rate has a direct impact on the required number of macrocells and as a result, the energy consumption of such a network can increase significantly. Energy metrics for realistic configurations and realistic architectures have to be defined hand in hand with the corresponding minimum energy optimization problems. The total system wide energy should include embodied energy as well as operational energy for services delivery. Particularly, the total energy

[Equation 15] (He, Zhang and Chen)

where the operational energy EOP is a function of the radio access architecture, which includes:

• the cell size,
• the height of the base station antenna,
• the radiation pattern of the antenna,
• the transmitting and receiving antenna distances,
• interference,
• multipath fading,
• shadowing,
• radio resource management (RRM),
• user density,
• user mobility, and
• traffic scenarios among other considerations(He, Zhang and Chen)

The energy is the Energy Budget Model for calculating the overall energy consumption of a cellular network over a given period of time T. This energy model takes into account the total energy consumption EOP due to the actual information transmission activity and also the embodied energy EEM over the equipment’s total lifetime. One energy metric to be considered is the energy consumption ratio (ECR) such as the energy per delivered information bit, such that

[Equation 16] (He, Zhang and Chen)

where E is the energy required to deliver M bits over time T, and D = M/T is the data rate in bits per seconds. This energy metric provides energy consumption in Joules consumed for transportation of one information bit (He, Zhang and Chen). An additional energy metric is the Energy Consumption Gain (ECG) which is defined by

[Equation 17] (He, Zhang and Chen)

where P_RAN is the RAN power consumption (He, Zhang and Chen).

Simulation Analysis

The case study in this paper is based on an outdoor wide area cellular network scenario (Badic, O’Farrell and Loskot). A classical hexagonal deployment is considered using three sectored cell sites as shown in (Fig. 2). The cell sites and cells are numbered from the center cells outwards in a clockwise direction as shown in Fig. 2.

Figure 2: Hexagonal Network

The radius of a cell-site denoted by R is fully adjustable with the inter-site distance = 1.5R for an hexagonal geometry (Fig. 3) (Badic, O’Farrell and Loskot).

Figure 3: Cellular Geometric Path

Base stations are configured with three sector antennas, with directions of 00, 1200, and 2400. A multi- tier approach has been used with number of tiers ? 3. The area to be served is defined when N = 57 cells each of radius R_cell = R/2 = 500m, and average cell transmission power P_cell is set to 20W (i.e. the power per sector). One objective of this paper is to investigate the ECR and ECG for this service area when reducing both R and P_cell while keeping the total power throughout the whole service area and the number of users constant. Equations (4), (5) and (6) give expression for the cell area, A_cell, service area, A_RAN, and total average transmitted power, P_RAN, respectively (Badic, O’Farrell and Loskot).

[Equation 18] (Badic, O’Farrell and Loskot)

[Equation 19] (Badic, O’Farrell and Loskot)

[Equation 20] (Badic, O’Farrell and Loskot)

Various cell sizes with R < 1000m have been investigated and compared with the initial large cell scenario with R = 1000m (Badic, O’Farrell and Loskot).

Using (4) and (6) the new total number of cells and P_cell for a new radius can be derived. The number of cells in the area increases by a factor of n when R changes from R_large to R_small as shown in the equation: (Badic, O’Farrell and Loskot). In order to fill the original coverage area with smaller cells while maintaining the same overall RAN power the new cell transmission power needs to be reduced by the same factor n. Table 4 summaries the small cell scenarios studied and related parameters used in the investigation.

Table 4: Simulated Service Areas

The Sleep Mode was also simulated for analysis of the consequences of the capacity density exceeding the demand to determine whether the architecture is unnecessarily consuming energy by keeping unused base station switched on (Badic, O’Farrell and Loskot). In this capacity, a sleep mode has been introduced where the cells that are not populated with users are turned off (Badic, O’Farrell and Loskot). In circumstances that consist of a constant user density, the number of users in a small cell scenario is the same as in a large cell scenario and therefore for comparison only 57 small cells need to be active to serve the same number of users while the remaining (N ?57) cells can be switched off (Badic, O’Farrell and Loskot). Thus, when (N ?57) cells are in sleep mode the RAN power consumption P_RAN can be reduced to: [Equation 21] (Badic, O’Farrell and Loskot)

Findings

Simulation Results

In this section, simulation results for the evaluation of the proposed cellular scenarios are presented. The analysis operated under the assumptions indicated in Table 5: System Simulation Assumptions, shown in Appendix A and adapted from (He, Zhang and Chen), which specifies the values for various attributes relative to the functionality of a small cell network. Once the network simulation was completed, it became more apparent that there were specific benefits that separated some small cell network configurations over others, which is illustrated in Table 6: Energy Saving by Power Method, which is shown in Appendix B and adapted from (He, Zhang and Chen). The average cell transmission power P_cell is cell size dependent and only the carrier is assumed in simulations with power allocations given in Table 5 where a bin size of 50m is used throughout the simulation and the main simulation parameters used are presented in Table 5.

The data in Table 7 presents a summary of the average cell capacity (S_cell) rates obtained through the simulation for different cell sizes in addition to antenna heights, showing that the average cell capacity is obtained by averaging the throughput by the overall bin locations in all active cells (He, Zhang and Chen). This simulation was repeated for different antenna down-tilt values to find out the optimum down-tilt providing maximum capacity as well as how the best performance is achieved with antenna down-tilts, such as E7 M0 in Table 7, where the average throughput is between 2.44 and 2.56 MBits/s (He, Zhang and Chen).

Table 7: Cell Capacity in MBits/s for Different Antenna Tilts Averaged over the Service Area

The energy per delivered information bit given in (2) is calculated using the best achieved throughput performance. The ECR values for different cell sizes are summarized in Table 8, which also indicates cell capacity densities calculated according to the equation C_d = S_cell/A_cell and the cell power densities calculated by P_dcell = P_cell/A_cell expressed in W/m².

Table 8: Simulated Energy Efficiency Rates

The results for the Sleep Mode Simulation indicated that the sleep mode functionality being introduced into the model as a method of compensating for the increasing cell capacity density through the reduced cell size in order to diminish the RAN energy consumption rates without corrupting the cell capacity (Badic, O’Farrell and Loskot). When in the sleep mode, the cells that do host an active user are powered-off, which allows the energy consumption to decrease and rise according to actual usage demands as a benefit derived from sleep mode (Bhaumik, Narlikar and Chattopadhyay). This attribute has been characterized with respect to a reference macro-cell deployment corresponding to R = 1000m using at least one active user per cell, which should be calculated according to the assumption that all 57 cells are active (Badic, O’Farrell and Loskot). The graphs in Fig. 4, shown in Appendix C, plots the ECR against the number of cells and Fig. 5, also shown in Appendix C, graphs the cell capacity density against the number of cells. From the results it can be observed that the cell ECR decreases with increasing number of cells, meaning that they decrease in cell size while the cell capacity density increases linearly with the number of cells. However, the cell power density Pdcell ? 31 [?W/m2] remains constant (Badic, O’Farrell and Loskot).

Discussion

The cell ECR continues to decrease with decreasing cell size while the cell capacity density continues to increase. Thus the cell power density P_dcell ? 31 ?W/m2 still remains constant, although, from a RAN perspective, the results demonstrate that the sleep mode functionality and the hence energy and overall RAN power consumption decreases with diminishing cell size, as does the RAN power density (Badic, O’Farrell and Loskot). However, the RAN capacity density defined by C_dRAN = S_RAN/A_cellN now remains constant. Thus with sleep mode, the ECG increases with decreasing cell size. This latter characteristic is such that the ECG is directly proportional to the quantity of cells and confirms that by reducing the cell size from R_large to R_small, accordingly (Badic, O’Farrell and Loskot). The EGG is increased by the factor n, which has been previously defined, and reducing the cell size using sleep mode reduces both the cell ECR and RAN energy consumption rates without taxing the power capacity (Badic, O’Farrell and Loskot).

The ideal green configuration for the small cell network demonstrates energy behavior that expresses power management strategies that regulates consumption throughout the whole system (Capone, Santos and Filippini). In the ideal configuration, the power distribution is linearly dependent on the traffic load with rates that range from very low with no traffic to maximum value with a full load (Boudreau, Panicker and Guo). If we were able to achieve this ideal behavior, addition of any technology improvement to those of the energy management would be seamless or at the very least possible (Lim, Lee and Choi). However, there are some limiting constraints of the traditional cellular architecture that has been used so far for all wireless access technologies that prevent an optimal power management and, more in general, to reach very high reductions of the energy consumption (Capone, Santos and Filippini). Full coverage of the service area is the concept that ensures the cellular architecture of wireless access networks has its foundation on user terminals in any point within the network area that can be accessed at all times.

The maximum coverage area of a base station depends on several issues, including the transmission power and the propagation conditions (Saunders and Argón-Zavala). In this aspect, when traffic density is low, adopted cell layouts are usually based on the full coverage capacity and coverage these ranges of base stations are typically almost fully exploited by limiting the overlap among neighboring cells to necessary capacities for mobility management (Capone, Santos and Filippini). In areas where traffic is much higher, base station density is exceedingly augmented in order to reduce cell size and increase the available access capacity per unit area, which results in the cellular layout being characterized by redundant coverage such that each point in the area is supported by several base stations (Capone, Santos and Filippini).

It is quite evident that energy management procedures in traditional cellular architectures can only exploit the redundant coverage of the network used to provide high capacity traffic when traffic is low (Bhaumik, Narlikar and Chattopadhyay). Nonetheless, the insurance of since full coverage at all times is non-negligible aspect of the network that can never be switched off even if there is no active user. It has been shown that potential energy savings achievable with energy management strategies in current cellular technologies are in the range of 20%-40%, depending on the considered traffic profiles and network layouts (Capone, Santos and Filippini). Paradoxically, this savings may even reduce (in percentage) in the future with the new cellular architectures based on micro cells. Indeed, it is commonly accepted that these micro-cellular layouts can reduce the nominal consumption since the energy per covered area of micro base stations is lower than that of macro base stations. Unfortunately, since micro cells provide high capacity with limited coverage overlap, they leave little room for energy management since basically all cells are essential for guaranteeing full coverage (Capone, Santos and Filippini).

The basic concept of Heterogeneous data networks integrates the new architecture is independent from the wireless technology adopted for the data network even if only future generation low-power wireless technologies will allow to achieve very high energy efficiency that is main target of the new architecture (He, Zhang and Chen). Additional aspects warranting mention explores the fact that the new architecture enables a flexible management of a set of heterogeneous data networks using different wireless technologies since their lifecycles are much longer than expected (He, Zhang and Chen). The new technologies incorporate the use of multiple radio interfaces with different technologies and some mobile voice calls is still handled by the old GSM technology in some developing and developed countries (He, Zhang and Chen). Most modern mobile handhelds use base station equipment that is currently being replaced with multi-technology devices attached to a single radio access network, incorporating the coexistence of heterogeneous wireless technologies, which is unavoidable being that it is commonly recognized as a key element for the design of future generation mobile network architectures (He, Zhang and Chen).

The new architecture is particularly suitable for managing heterogeneous wireless technologies as it reverses the classical approach to network selection (Lim, Lee and Choi). Since the access to communication service is mediated by the signaling network, it is no longer the user terminal that selects the access point, but this is basically delegated to the network (Capone, Santos and Filippini). This allows a more flexible and intelligent management of traffic with the set of technologies and radio resources available through algorithms that are able to take into account the status of the whole system. The design of these new algorithms is an interesting technical challenge since several issues can be considered including specific mobility and resource management policies of mobile operators (Capone, Santos and Filippini).

Conclusions

Cost reduction and upgrades in available technologies, along with the development of inexpensive digital network infrastructures have propelled forward a phenomenal trend known as technological convergence. Technological convergence is the merging of singular technological entities into new technologies that bring together a myriad of media. Technology originally handled one medium or accomplished one or two tasks. Technological convergence allows devices to present and interact with a wide array of media. The increase in the variations of media present in homes due to the continued advancement in available technologies has also greatly contributed to the convergence of technologies.

The most common example of technological convergence is the modern cell phone. In its original conception, cell phones were simply mobile telephones. Due to technological convergence, today’s models are complex instruments, able to perform many functions of a desktop PC, a telephone, a music player, and even a digital video camera, all in small, pocket-sized devices. Other examples include Apple IPhones that can play videos, various music formats, and play the radio, transmitting or receiving sound and can be combined with other air-interface solutions that facilitate communication through next generation wireless local area networks (WLANs) and 4G mobile communication systems, such as Multiple-input multiple-output (MIMO) wireless technology amalgamated with orthogonal frequency division multiplexing (MIMO-OFDM) (Chen and Lee; Nema, Goel and Singh). The recent advances in communications applications such as OFDM and MIMO have presented opportunities for improvement in 4G networks, making them more reliable transmitters of data for wireless applications.

Recommendations

Although, as of 2008, broadband service was available in 182 markets and mobile networks exceeded one billion by the start of 2009, more than half the world nations still do not have an Internet Exchange Point, (IXP) primarily where local traffic can be transmitted and pay exorbitant charges, as much as $2,000-$5,000 USD per megabyte (Mb) per month, for transiting traffic information IXP within an improved world (Lim, Lee and Choi). Data transmission is recognized by technologies that appears in normal everyday life including cell phones and blue tooth technology, Wi-Fi, infra-red, and wired technologies, all of which operate as part of each network. Although originally coined for airborne multichannel moving target indicator (MTI) radar space-time adaptive processing (STAP) has been adopted in many disciplines in which joint adaptive sensor temporal and spatial processing are performed, such as multidimensional adaptive filtering (Stüber, Barry and McLaughlin).

With these considerations in mind, it is recommended that for network operators to first decide the total number of small cells n_total based on the constraint of maximum traffic demand, the deployment budget, and the average DE (in terms of traffic distribution), such that:

[Equation 22] (He, Zhang and Chen)

This will allow operators to activate small cells during the times when the traffic T_net is low for energy saving purposes (He, Zhang and Chen). Overall, technological convergence currently does and will remain a large aspect of future devices and these instruments will undoubtedly present higher power demands than modern electronics. Small cell networks will have to be sustainable if they are to be profitable since configurations that are not green are fast becoming archaic. When considering the requirements of the small cell network implementation, framing the decision with strictest regard for the needs of the business ensures a comprehensive understanding of the appropriate business landscape as well as the requirements of the organizations that wish to make use of the new technologies. Both of these aspects must be considered in account of their planning for the different business requirements that apply to individual users’ computing environments. To be sure the small cell network will sustain the needs of the customers, the authentication protocol must not be prone to man in the middle attacks and all exchanged passwords must be securely transmitted. Also, in the event that an intruder can capture an authentication server, the greater the redundancy between servers and synchronized decisions between them, the better. A good protocol must choose an encryption scheme that is secure under a probabilistic polynomial time model. Additionally, the protocol must apply the encryption scheme in a way that does not open a good encryption scheme to vulnerability. A good protocol should have a key management mechanism so the user will not have to worry about the manually generating and installing new keys. 

References/Bibliography

Badic, B., et al. “Energy Efficient Radio Access Architectures for Green Radio: Large versus Small Cell Size Deployment, .” IEEE (2009): 1-5.

Bhaumik, Sourjya, et al. “Breathe to Stay Cool: Adjusting Cell Sizes to Reduce Energy Consumption.” Green Networking (2010): 41-46.

Boudreau, Gary, et al. “Interference Coordination and Cancellation for 4G Networks.” IEEE Communications Magazine (2009).

Capone, Antonio, et al. “Looking Beyond Green Cellular Networks.” International Conference on Wireless On-demand Network Systems and Services (WONS) (2012): 127-130. Online.

ETSI. LTE; Evolved Universal Terrestrial Radio Access (E-UTRA); Physical layer procedures (3GPP TS 36.213 version 8.8.0 Release 8). ETSI TS 136 213 V8.8.0 (2009-10). Sophia Antipolis Cedex, France: European Telecommunications Standards Institute, 2009. Onine.

Hasan, Ziaul, Hamidreza Boostanimehr and Vijay K. Bhargava. “Green Cellular Networks: A Survey, Some Research Issues and Challenges.” (2011): 1-16. Online.

He, Gaoning, et al. “Energy Efficiency and Deployment Efficiency Tradeoff for Heterogeneous Wireless Networks.” Globecom 2012 – Symposium on Selected Areas in Communications. IEEE, 2012. 3213-3218.

Hoadley, John. “Building future networks with MIMO and OFDM.” 19 September 2005. 27 February 2013. <http://www.printthis.clickability.com/pt/cpt?expire=&title=Building+future+networks+with+MIMO+and+OFDM&urlID=418712567&action=cpt&partnerID=97671&fb=Y>.

Jaloun, Mohammed and Zouhair Guennoun. “Wireless Mobile Evolution to 4G Network.” Wireless Sensor Network 2 (2010): 309-317.

Lim, Youngmi, Joo Hyung Lee and Jun Kyun Choi. “The Effects of Cell Size on Total Power Consumption, Handover, User Density of a Base Station, and Outage Probability.” ICNS 2011: The Seventh International Conference on Networking and Services. Daejeon, Republic of Korea: IARIA, 2011. 1-4.

Nema, Shikha, Aditya Goel and R P Singh. “Integrated DWDM and MIMO-OFDM System for 4G High Capacity Mobile Communication.” Signal Processing: An International Journal 3.5 (n.d.): 132-143. <http://www.cscjournals.org/csc/manuscript/Journals/SPIJ/volume3/Issue5/SPIJ-38.pdf>.

Saunders, Simon R. and Alejandro Argón-Zavala. Antennas and Propogation for wireless communication systems. 2nd. Chichester,West Sussex: JohnWiley & Sons Ltd, 2007.

Sinha, Nirmalendu Bikas, Makar Chand Snai and M. Mitra. “Per formance Enhancement of MIMO-OFDM Technology for High Data Rate Wireless Networks.” International Journal of Computer Science and Application 2010 (2010): 122-128.

Stüber, Gordon L., et al. Broadband MIMO-OFDM Wireless Communications. Atlanta, GA: Georgia Institute of Technology, n.d.

Appendices

Appendix A

Table 5: System Simulation Assumptions

Appendix B

Table 6: Energy Saving by Power Method

Appendix C

Figure 4: ECR & Capacity Density by Cell # Figure 5: Capacity density by Cell #