Cache-Enhanced Dynamic Movement-Based Essay Example

to allow millions of people to use cell phones simultaneously. The carrier receives approximately 800 frequencies for use across the city and each cell is usually about 10 square miles in size (26 square kilometers) [1].

Cells on a larger hexagonal grid, depicted in Figure 1.1, are typically viewed as hexagons. They consist of a base station that includes a tower and a small building containing radio equipment used for communication with Mobile Terminals on preassigned radio frequencies. Cell phones have low-power transmitters, often with two signal strengths: 0.6 watts and 3 watts [1]. The base station also operates at low power. Low-power transmitters offer two advantages: transmissions do not extend far beyond the cell, allowing for reuse of frequencies in alternate rings and non-adjacent cells; and cell phones, which are battery-operated, have relatively low power consumption, enabling the use of small batteries and the creation of handheld cellular phones. Implementing the cellular approach in a city requires numerous base stations, with hundreds of towers found in a typical large city. However, due to the high number of users, costs per user remain low.The Mobile Telephone Switching Office (MTSO) is a central office operated by every carrier in every city.

The office in charge of handling all phone connections to the traditional telephone system and managing the base stations in the region is responsible for overseeing the Mobile Switching Center (MSC) connections. These MSC connections route calls to the telephone networks. The designated area covered by an MSC is referred to as a Registration Area (RA) or Location Area (LA). Multiple RA's make up a Service Area (SA).

The Home Location Register (HLR) is responsible for supporting each

Service Area (SA). Within a wireless network, there can be multiple SAs and HLRs. For more information on the Location Registers, please refer to the Literature Review (Chapter 2).

Cell phones are equipped with distinct identification codes that serve to identify the phone itself, its owner, and the service provider. These codes consist of the Electronic Serial Number (ESN), a 32-bit number implanted in the phone during manufacturing; the Mobile Identification Number (MIN), a 10-digit number derived from the owner's phone number; and finally, the System Identification Code (SID), a unique 5-digit code assigned by the Federal Communications Commission (FCC) in the United States to each carrier.

The regulation of interstate and international communication is the responsibility of a government agency. This process involves the use of commonly used cell phone codes, such as ESN, MIN, and SID codes. The ESN code is permanently embedded in the phone, while the MIN and SID codes are programmed into the phone when activating a service plan. In order to understand how person A calls person B in 1G systems, we can examine the following procedure: When person B turns on their phone, it searches for an SID on the control channel. The control channel serves as a dedicated frequency for communication between the phone and base station regarding call establishment and channel adjustments.

If the phone cannot locate any control channels for reception, it determines that it is out of range and presents a "no service" message. When receiving the SID, the phone compares it to the programmed SID. If they match, the phone recognizes that it is within its designated network. Additionally, the phone transmits a registration request along

with the SID, which allows the MTSO to track the phone's location in a database. The MTSO uses this information to determine which cell the phone is in when making a call. When user A contacts MTSO to connect with user B, MTSO consults its database to find B's cell. MTSO then selects a frequency pair for B's phone to use in that specific cell. Using the control channel, MTSO instructs B's phone on which frequencies to use. Once B's phone and the tower switch to these frequencies, the call is established and B can communicate with A via two-way radio. As B approaches the cell's edge, the base station in B's cell detects a decrease in signal strength. Simultaneously, the base station in the adjacent cell notices an increase in signal strength from B's phone.The coordination between the two base stations occurs through the Mobile Telephone Switching Office (MTSO). During this process, B's phone receives a signal on a control channel instructing it to change frequencies. As a result, B's phone is transitioned to the new cell. This scenario is depicted in Figure 1.2 [1]. Additionally, if the SID on the control channel does not match the SID programmed into one's phone, the phone recognizes that it is currently roaming. In such cases, the MTSO of the visited cell contacts the MTSO of the home system and verifies the validity of the phone's SID using its database records.

The text discusses location management in cellular networks. The protagonist's home system connects his phone to the local MTSO, which tracks his phone's movement through its cells. This process occurs in a matter of seconds. The thesis

will focus on these concepts, which are crucial in the operation of PCS networks. The operation involves two activities: location update and paging. The second chapter of the thesis will explore the basic approaches to location update, terminal paging, and further details about the network topology.

The technique employed for location update in this thesis is LA-based, meaning that location registration will occur when a mobile terminal (MT) moves into a new location area (LA). Location updates are typically classified into two types: static and dynamic. In static location update, the paging area is fixed and equal to the size of the LA. Each cell within the LA will initiate paging once when a call occurs.

The issue with this, however, is that if the LA (Location Area) is large, the cost of paging is very high and a mobile terminal located at the border of LAs may experience excessive location updates. Dynamic location management schemes determine the size of the location area dynamically based on changes in mobility and calling patterns of mobile terminals. Various types of dynamic location management schemes and the related concepts are discussed in Chapter 2. While commercial mobile telephone networks existed since the 1940s, many consider the analog networks of the late 1970s to be the first generation (1G) wireless networks. The details of 1G, 2G, and 3G networks, their evolutionary stages, and the involved concepts are covered in the Literature review of the second chapter. One noteworthy aspect of this thesis is that it not only focuses on implementation for 3G networks that utilize an additional register called the Gateway location register [3] (details in Chapter 3), but it also

employs a caching strategy at different registers to analyze how well the current structure can be optimized.

The caching strategy [4] is focused on users who frequently make calls to a specific region, which is common in practice. The simulation consists of four scenarios:

• Standard 3G network without any cache.
• 3G network with a cache at the GLR.
• 3G network with a cache at the VLR.
• 3G network with a cache at both GLR and VLR.

The improved scheme that utilizes caching is discussed in detail in Chapter four. A simulation model is designed to assess the performance of the proposed scheme. It assumes that the cellular network is divided into hexagonal cells of equal size. The target network includes two SAs, four G_LAs, eight LAs, 244 BSs, and 976 MTs.

In this study, we assume that there is an exponential cell residence time distribution. We also assume that each MT (Mobile Terminal) remains in a cell for a certain amount of time before moving to one of its neighboring cells with an equal probability. The incoming call arrivals are described using a Poisson process, while the service time of a phone call is described using an exponential process. To implement the simulation model, we have designed a discrete event simulator. The use of object-oriented design makes it simpler to extend or modify the existing model.

The natural behavior and structure of the objects are represented in the design. The design is

fully implemented using the Java programming language, which can be used on different platforms. Chapter 4 discusses the system design, location update, and call delivery strategies used in the improved scheme. Chapter 5 provides a detailed discussion of the simulation model and discrete event simulator.

The experiments conducted using discrete event simulator are focused on the following areas:

• Examining the relationship between Call Migrate Ratio (CMR) and the total location management cost. CMR is defined as the ratio of an MT's call arrival rate to its migration rate.
• Comparing the total location management cost among four selected schemes.
• Evaluating the impact of movement threshold value and paging delay on our proposed scheme (with cache) in comparison to the standard scheme.

The experimental results demonstrate that implementing a cache at both the GLR and VLR yields the highest benefits compared to other scenarios. However, there is always a trade-off between memory costs and performance gains. As memory costs are decreasing over time, employing a cache for most user populations is advantageous. Further details regarding the experiments can be found in Chapter 6.

In summary, this paper consists of several chapters. Chapter 1 serves as the introduction. The second chapter presents a literature review. Chapter 3 focuses on location management in 3G, specifically discussing the employed caching strategies. Chapter 4 provides the details of the enhanced location management scheme. Finally, Chapter 5 explains the simulation model in detail.

The experimental results are analyzed in Chapter 6, and Chapter 7 draws the conclusion.

justify;">
Literature Review

Location management is a critical concern in wireless network operation. The Mobile Terminals (MT's) refer to the automobile or handheld telephones, or portable computers that users utilize for call sending and receiving. For a successful delivery of an incoming call, the wireless network must constantly keep track of the location of each mobile terminal [5]. This process, known as location management [6], involves tracking and locating MTs so that incoming calls can be directed to their current location.

Location Management is distinct from location services in routing. Routing involves moving information from a source to a destination across networks, typically encountering at least one intermediate node. It includes determining the best routing paths and transporting information groups (often referred to as packets) through an internetwork. In contrast, Location Management not only tracks mobile terminals (MTs), but also focuses on managing call arrivals for call delivery.

Evolution of Cellular Networks and the Concepts Involved:

The first generation (1G) wireless networks appeared in the late 1970's and are regarded as the analog networks at that time [7]. These networks functioned as fixed size networks, transmitting analog sound images over the air and through the networks. Both the transmitter and receiver were tuned to the same frequency, with the transmitted voice varying within a small band to generate a pattern that could be reconstructed, amplified, and sent to a speaker by the receiver. Despite being groundbreaking technology at that time, it had several drawbacks. Mobility was limited, efficiency was low due to a limited number of callers fitting into the available spectrum, and optimization techniques like coding or compression were not feasible

due to everything being analog. Moreover, the components were bulky and expensive, resulting in large-sized handsets. These limitations, combined with other factors, ultimately led to digital transmission development and 2G cellular networks emergence.

The introduction of 2G brought several advancements, including reduced power consumption, smaller equipment size, higher bandwidth requirements per call, and the crucial use of multiple access techniques - either TDMA or CDMA. In mobile systems, it is important for different users to utilize different channels to prevent traffic collisions. Cell-phone networks commonly employ three technologies - Frequency division multiple access (FDMA), Time division multiple access (TDMA), and Code division multiple access (CDMA) - for transmitting information.

FDMA:

FDMA is an analog system technology that was utilized in the initial generation cellular networks.

In this technology, each user is assigned a unique frequency for separation, meaning that the spectrum is divided into separate voice channels by dividing it into equal bandwidth sections. Although this is a straightforward technique, the problem arises when adjacent frequencies start to interfere with one another.
FIGURE 2.1. IN FDMA, EACH PHONE USES A DIFFERENT FREQUENCY.

TDMA:

Most of the current mobile data networks utilize Time Division Multiple Access (TDMA) where each conversation is allocated a frequency for a certain duration, essentially separating users by assigning distinct time slots for each channel. TDMA divides a narrow band, which is 30 kHz wide and lasts 6.7 milliseconds, into three time slots. This division is depicted in Figure 2.2.

TDMA divides a frequency into time slots. Narrow band, or "channels" as traditionally understood, is used. Each conversation is allocated one-third of the radio time. This is achievable due to the compression of digitalized voice

data, resulting in reduced transmission space requirements.

Therefore, TDMA has triple the capacity of an analog system with the same number of channels. TDMA systems operate in either the 800-MHz (IS-54) or 1900-MHz (IS-136) frequency bands [7].

CDMA:

Another approach is Code Division Multiple Access (CDMA), which is more advanced. In CDMA, different users are separated by different codes, allowing for theoretically limitless capacity. CDMA follows a completely different method than TDMA.

CDMA is a method where data is digitized and then spread out over the available bandwidth. This technique involves overlaying multiple calls on a channel, with each call being assigned a unique sequence code [1]. Spread spectrum, or the process of sending data in smaller pieces over a range of frequencies, is used in CDMA. All users transmit their signals within the same wide-band spectrum, and each signal is spread across the entire bandwidth using a unique code. At the receiver, this same code is used to recover the original signal.

CDMA systems rely on the GPS system to accurately time-stamp each signal. It can carry between eight and 10 separate calls in the same channel space as one analog AMPS call. CDMA technology is the foundation of Interim Standard 95 (IS-95) and works in both the 800-MHz and 1900-MHz frequency bands. Refer to Figure 2.3 for a clearer illustration.

FIGURE 2.3: In CDMA, each phone's data is assigned a unique code. Starting from 2G of cellular networks, either TDMA or CDMA is used, with WCDMA becoming the most popular later on. In terms of second generation, the USA employs two standards: IS-95 (CDMA) and IS-136 (D-AMPS).

The GSM (Global System for

Mobile communications) unified communication systems in Europe, while Japan uses PDC (Personal Digital Cellular). The rise of cellular radio was fueled by digital systems and advancements in chip technology. This brought about smaller and more sophisticated handsets, voice mail, call waiting, and services like SMS. GSM alone had over 360 million subscribers in Fall 2000, with other systems like cdmaOne, PDC, and TDMA also gaining users. UMTS aimed to achieve the idea of PCS data by providing services that are independent of location, network, and terminal. This led to the development of 3G cellular networks. The driving force behind 3G systems was not only capacity and global roaming, but also higher data transmission rates, standardized radio interfaces, common global frequencies, and improved quality of service.

Third generation cellular networks included features such as integrating wired and wireless networks, a single network infrastructure, support for paging, cordless phones, wireless LANs, and sufficient bandwidth for multimedia services. The multiple access technique used could be either TDMA or CDMA, or a combination of both. These networks could be deployed to support multimedia applications. General Packet Radio Service (GPRS) and EDGE (Enhanced Datarates for Global Evolution) were packet-based network technologies used in GSM networks. The main branches of the 3G standard were WCDMA and CDMA2000. The evolution of cellular networks from analog to digital has been rapid and interesting. The different stages of this evolution are shown in Figure 2.4.

According to Rysavy Research's time estimates, the UMTS evolution and 3G systems and services are expected to experience long and steady expansion. This is due to the continuous trend of integrating wireless cellular technologies with wired networks and services, as well

as with other wireless short-range technologies like those in the IEEE 802.11 and Bluetooth series of standards.The text emphasizes the importance of upgrading networks and services in a transparent manner for users due to several factors. These factors include the expansion of the IP paradigm into wireless networks, driven by the standardization of IPv6 and its increased addressing capability, as well as the development of mobile IP protocol. Despite any significant technological advancements, it is crucial that users are not aware of these changes.

The future of profitable and successful networks relies on flexibility and responsiveness. Competition will now be based on the range of services offered rather than just tariffs, as was the case in the 2G era. The advancements in 3G wireless networks will play a significant role in shaping our future, even if most users remain oblivious to these changes. Despite their unawareness, users will enjoy the benefits of faster, more widespread, and user-friendly services [9].

Framework of a Wireless Network in 3G and Location Management in 3G

All popular Cellular/PCS Networks currently use Home Location Registers and Visitor Location Registers for location management. Location management involves tracking the movement of mobile devices as they transition from one location to another. This process comprises two main operations: location updates, which keep track of mobile device locations when not in conversation.

Mobile terminals refer to subscribers who use automobile hand-held telephones or portable computers for call transmission and reception. These terminals provide real-time location information dynamically. Paging is another operation where the system searches for mobile terminals by sending polling signals to cells in the paging area, which may consist of multiple cells. Minimizing the

cost associated with these operations is crucial for efficient system functioning. Most cellular/PCS networks follow a cellular structure, with a service area divided into cells. Each cell has a base station that communicates with mobile terminals using preassigned radio frequencies. Multiple cells are connected to a Mobile Switching Center (MSC), which routes calls to telephone networks.

MSC is a telephone exchange designed for mobile applications, serving as a link between mobile phones (via base stations) and the PSTN or PSDN, making mobile services easily accessible. MSCs host the location management system, with HLR and VLR utilized in 2G cellular networks to support this system. Services are divided into Location areas (LA's), with each LA consisting of multiple cells and being handled by a VLR, which is associated with an MSC in the networks.

The VLR holds temporary records of the profiles and location information of the MT. It is used to retrieve information for handling calls to and from a visiting mobile terminal. On the other hand, the HLR is responsible for storing permanent data, such as directory number, profile information, current location, and valid period, of mobile terminals whose primary subscription is in the area. Paging refers to the search process conducted in a Paging area (PA).

The text discusses the classification of location management schemes based on the size of the Paging Area (PA). There are two broad categories: static and dynamic location schemes. In the static scheme, if the PA size is fixed and equal to the Location Area (LA), each cell in the LA will page once when a call arrives for a Mobile Terminal (MT) registered in the LA. However, this can

result in excessive location updates if an MT is close to the boundary of an LA and moves between two LAs frequently. This problem is addressed by dynamic location management, which is the focus of this thesis. In dynamic location management, the size of a location area is determined dynamically based on changes in mobility and calling patterns of mobile terminals.

Out of the three types of dynamic location management (time-based, distance-based, and movement-based), movement-based location management is considered the most practical and effective [2]. Therefore, it is the chosen method in this thesis. The number of roaming users is increasing due to the rise in international travel. To decrease the signaling traffic for international and remote roaming, the Gateway Location Register (GLR) was proposed within the UMTS core network. This proposal was made by the Third Generation Partnership Project (3GPP), specifically the Technical Specification Group Services and Systems Aspects, Network Architecture (Release 5) in December 2000. In my thesis, I will be implementing this special feature to compare it with the analytical results proposed in "The Movement-based Location Management for 3G Cellular Networks".

The GLR is a node that sits between the VLR/SGSN and the HLR. Its purpose is to manage the location of roaming subscribers in a visited network without constantly involving the HLR for every change of location. This helps reduce signaling traffic between the visited and home mobile systems, while optimizing location updates and the handling of user profile data across network boundaries. The GLR is specifically located within the visited network.

It stores the information of the roamer and manages location within the network. The interface between the HLR and the GLR is

the same as that between the HLR and VLR because the presence of the GLR is not visible from the home network.

Teletraffic Modeling

Teletraffic models serve as crucial tools for network analysis and design. Before delving into the model to be used in this thesis, this section briefly discusses the different modeling techniques employed in the performance analysis of location update and paging schemes in general.

Teletraffic models can be categorized into four main groups: Network Topology Model, Mobile Residence Model, Mobility Model, and Call Model. In the following section, each type of model will be discussed in detail.

• Network Topology Model
• Mobile Residence Model
• Mobility Model
• Call Model

Network Topology Model

The network topology model defines the connectivity between base stations or cells. Regular cell topologies are commonly used to represent the coverage area of cellular networks [11]. There are three types of regular topology models: Linear Model (for one-dimensional networks), Mesh Model and Hexagonal Model (for two-dimensional networks). Although these models simplify analytical computation, they do not accurately depict a realistic cellular network topology. This is because cell sizes depend on factors such as transmit power, receiver sensitivity, antenna radiation pattern, propagation environment, and the number of neighboring cells can vary from cell to cell [11].

In a paper [12] by S. K. Sen et al., they introduced a graph model for the cellular network. This model applies to a zone or LA-based cellular

mobile system. The network is represented by a connected graph G = (V, E), where V represents the LAs (local areas) and E represents the access paths between pairs of LAs. The graph has a bounded degree.

The Mobile Residence Model

The residence time at a location is the duration a mobile user stays there before moving to another place. Some location update schemes, like selective LA and profile-based update algorithms, rely on accurate estimations of residence time at various Las [11]. Typically, research on location management assumes a geometric cell residence time distribution for performance analysis. This assumption assumes that the distribution is Independent and Identically Distributed (IID) across all cells. However, the IID geometric residence time assumption does not accurately represent individual user mobility patterns. In reality, users may spend longer periods of time at certain locations, such as their homes or offices.

In another article that is quite interesting [13], the authors presented a model for residence that allows an IID general cell residence time distribution. However, this model is only applicable to a hexagonal cell configuration.

Mobility Models

The term mobility refers to the ability of a person to move and is typically described by the average velocity and distance covered by the mobile user [14]. Mobility models play a crucial role in analyzing various aspects of wireless networks, such as resource allocation, handoff, and location management. Generally, the mobility models are dependent on factors like speed, direction, or movement history of the mobile user [15].

Mobility models are divided into two categories. One category includes models that are based on individual movement behavior such as the Gaussian Model and the Markov

Model. The other category includes models that are based on aggregate movement behavior such as the Fluid Flow Model and Gravity Model. In the next section, we will discuss each of these models in detail.

The Gaussian Mobility Model is established on the concept of isotropic random user motion with drift, which is defined as the average velocity in a given direction.

The probability distribution function in a one-dimensional model is given by (Equation 2-1).

Equation 2-1

where x is the location variable, t is the time elapsed since last contact with the mobile terminal, v is the mean drift velocity, and D is a constant. When the time is partitioned into small intervals and v = 0, Gaussian user location distribution has been used to study the timer and state-based update schemes [11].

The Markov Model is a model that aims to analyze a user's movement pattern by assigning probabilities to neighboring cells. When a mobile user is in cell i, they have the option to either stay in that cell with probability Pr(i|i) or move to a neighboring cell j with probability Pr(j|i). This probability function represents the user's preference. The residence time in each cell follows a geometric distribution. The Markov Model has been utilized to analyze the performance of selective LA and threshold-based update schemes.

One limitation of this approach is the lack of a concept for the movement history or trip of a specific mobile user [11]. The Fluid Flow Model describes aggregate movement behavior as the flow of a fluid. Mobile users are assumed to move at an average velocity of v, with their direction uniformly distributed over [0, 2?].

Assuming a uniform density of ? for the mobile users and a location area boundary length of L, the rate at which users move out of LA C can be determined using Equation 2-2.

Equation 2-2

This model accurately represents the boundary crossing rate in a symmetric grid of streets, such as the Manhattan-style. It has been utilized to study the pr.