Project topics and materials
Get Complete Project Material File(s) Now! »
Motivations
Once a terminal is equipped with several interfaces, it becomes possible to use simultaneously the various interfaces – and not simply to switch from one to another.
In this document, the term « multi-homing » refers to the capacity of a mobile terminal to communicate simultaneously on various network interfaces which may be of different radio access technologies, both wireless and cellular. This opens new perspectives and many benefits will be provided [Thierry]: Permanent and Ubiquitous Access This benefit targets to provide an extended coverage area via distinct access technologies. Multiple interfaces bound to distinct technologies can be used to ensure a permanent connectivity, anywhere, anytime, with anyone.
Multi-homing provides a potential means to act upon failure. Indeed, a secondary (or backup) interface or path can be used when the primary becomes unavailable. Interface or path duplication can be useful in many situations: in case of preparing a handover to reduce the packet loss, for example.
Multi-criteria Interface Selection When having multiple paths or activated interfaces, terminal may establish a strategy to select the best interface basing on multiple attributes such as the application requirements, the user preferences, the network/interface characteristics, or network operator policy, …
When having multiple available active interfaces, the mobile terminal may associate each application flow to a specific interface considering mainly the requirements of the application. Other parameters could be considered such as the terminal capacity, policy or user preferences, etc. Aggregated Bandwidth/ Load Sharing When a terminal is simultaneously connected to multiple paths, it can exploit those paths to aggregate the available bandwidths on the individual path, or out-going traffic can be split over those paths to achieve better resource aggregation, allowing better transfer rates.
In this work, we tackle the interface selection and the flow/interface association decision issues Multi-criteria interface selection decision issue The first motivation of this work is to take into account the interface selection issue where the mobile terminal equipped with several interfaces has to select at any time the best interface or the best access technology according to multiple criteria, as stated above.
Handoff techniques have been well studied and deployed in cellular systems and are of a great deal of importance in the wireless systems. Traditionally, the handover decision, especially in case of horizontal handovers, is made purely according to radio signal strength (RSS) thresholds and hysteresis values as input parameters. The vertical handover considers multiple criteria to make decision. However, the decision is mainly based on RSS, in the border region of two cells, and RSS is assumed to be the most priority criterion compared to the others.
In the multi-homing context, the mobile terminal is under different network coverage areas. The radio signal strength (RSS) attribute of the wireless networks is assumed to be higher than the necessary thresholds to make connectivity. The multi-interface mobile terminal is able to access several networks at the same time and to select the best network among the available networks basing on multiple attributes. The interface selection challenge is to determine the most favorable trade-off among all the attributes. The attribute weights depend on specific objective of the decision maker. Flow/Interface association decision issue Our second motivation is to tackle the flow/interface association issue where each application is associated to a specific interface basing mainly on the application requirements. Other parameters could be considered.
There exist three main application types such as hard-real-time, adaptive and elastic applications. Each type of application has specific characteristic. Hard real-time applications need data to arrive within a given delay bound. Examples of such applications are traditional telephony applications. Rate-adaptive or delay-adaptive applications such as video, voice, streaming are more tolerant to occasional rate, delay bound violations and dropped packets. Elastic applications are the traditional data applications such as electronic mail, remote terminal access, and file transfer, etc. which can adjust to wide changes in delay and/or throughput.
The flow/interface association challenge is to determine which network interface being suitable for which application.
Many interesting usage cases highlight the motivations of application/interface association. Consider, for example, a terminal integrates three radio interfaces: Wi-Fi, WiMAX and 3G and runs simultaneously three applications, for example, an FTP application, a streaming application (e.g. MPEG video with an average rate of 2Mbps) and a voice call. Moreover, consider that the available bandwidth is about 1Mbps on the Wi-Fi interface and about 5Mbps on the WiMAX interface, for instance. In this scenario, it would be interesting for the terminal to associate: the voice call to the 3G interface, the streaming application to the WiMAX interface since this network will provide a high streaming quality, and the FTP application to the Wi-Fi interface to economize the terminal energy consumption (considering that the energy consumption of WiMAX is higher than Wi-Fi).
Evolution of wireless networks and mobile terminals
Wireless technologies in the last decade have attracted attention from wireless network operators, service providers, developers, vendors, and users. The breathtaking evolution of wireless technologies, services and business applications has resulted in a wide-scale deployment and usage of wireless and mobile networks.
In the first generation of mobile networks, mobile users are tied to a single national operator. The analogue and circuit-switched such as Advanced Mobile Phone System (AMPS) in USA, and Total Access Communication System (TACS), Nordic Mobile Telephony (NMT), and/or Radio Telephone Network C (C-NETZ) in different parts of European countries are considered as the first version of the mobile networks.
The Second Generation of mobile networks (2G) was digitalized, such as Global System for Mobile communications (GSM) in Europe, Personal Digital Cellular (PDC) in Japan, Digital AMPS (D-AMPS) and Code Division Multiple Access (CDMA) in the United States [Walke].
The exclusive monopoly of the operators in most countries was broken down to encourage the competition of the mobile telecommunications. The cellular network evolved to 2.5G which includes High Speed Circuit- Switched Data (HSCSD), General Packet Radio Service (GPRS), and Enhanced Data Rates for GSM Evolution (EDGE).
The 3G services are considered as the advance of the 2.5G by proposing high quality services including wide-area wireless voice telephone, video calls, and wireless data, all in a mobile environment.
3G is a family of standards for mobile telecommunications which includes GSM, EDGE, UMTS, and CDMA2000 as well as DECT and WiMAX defined by the International Telecommunication Union [itu]. Compared to 2G and 2.5G services, 3G allows simultaneous use of speech, data services. The 3G network enables network operators to offer users a wider range of advanced services while achieving greater network capacity through improved spectral efficiency. However, the wide deployment of 3G on the market introduces their shortcomings of the bandwidth capabilities. Particularly, the main drawbacks of the 3G are the low throughput rates on offer and high charges for many of advanced services that lead to a lack of attractive applications.
3.5G was born to resolve the problems of 3G with the addition of High Speed Downstream Packet Access (HSDPA) to enhance the throughput rates offered in the 3G networks. Current HSDPA deployments support downlink speeds of 1.8, 3.6, 7.2 and 14 Mbps. Further speed increases are available with Evolved High-Speed Packet Access (HSPA+), which provides speeds of up to 42 Mbps downlink and 84 Mbps with new release of the 3GPP standards [3GPP_9]. In addition, 3GPP is defining Long Term Evolution (LTE) of UMTS with objectives of increasing capacity, lowering latency and reducing costs for operators [Ekstrom][3GPPc]. The 3GPP is standardizing a Long Term Evolution Advanced as future 4G standard. At the present rates of 15-30 Mbps, 4G is capable of providing users with high-definition television streaming. At rates of 100 Mbps, 4G supports the content of a DVD-5 (e.g., a movie) [4G].
Wireless Local Area Network (WLAN) technologies (such as the Wireless Fidelity (Wi-Fi) IEEE 802.11 a/b/g family) have been widely increasing in recent years [802.11]. These wireless broadband networks are rapidly deployed in home and office for several advantages such as high speed, simplicity, and low installation cost, with very little needed wiring. The IEEE 802.11 is actually the original standard offering data rates of 1 or 2 Mbps. Revisions have been made to the standard to maximize the throughput (the case of 802.11a, 802.11b and 802.11g standards called 802.11 physical) or specify elements to ensure better security or better inter-operability.
Wireless Metropolitan Area Network (WMAN) which is also known as a Wireless Local Loop (WLL) is based on the IEEE 802.16 standard or WiMAX [802.16]. Wireless local loop can reach effective transfer speeds of 1 to 10 Mbps within a range of 4 to 10 kilometers [802.16], which makes it useful mainly for telecommunications companies. WMAN which is often cheaper and less restrictive than current wire-line options is currently widely used for small businesses. Mobile terminals are evolving into affordable high powered computing terminals (e.g., smart phones, PDAs and iPhone, …)
Beside the traditional applications such as Voice call and Short Message Service (SMS), a large number of application types are increasing, for example, the real-time services like video streaming and conferencing, interactive data services like web browsing, and non real-time background services such as video, picture, and MP3 downloads, etc. Mobile terminals are expected to have several radio interfaces (e.g., GPRS, UMTS, Wi-Fi, 3G, 4G, and WiMAX, etc.).
With many technical advantages, mobile terminals are able to run simultaneously several applications. The mobile terminals have opportunities to take advantage of the providers’ competition when accessing a diverse range of services.
However, each technology has specific characteristics in terms of coverage area and technical characteristics (e.g., bandwidth, QoS, and bit-rate, etc.) including advantages and limitations. No radio wireless technology could always be the best choice for all applications. In each context, mobile terminals may choose specific access technologies for each application considering the user’s preference, the interface/network characteristics, the application requirements and/or the terminal capacities, etc. Always Best Connected (ABC) concept [Gazis][Fodor] states that a mobile terminal (user) can have the best service connection regardless of place and time. Moreover, a user, who wants to connect to a service, is able to choose the access in a way that suits his or her needs the best, and change the access when other better access networks become available. The users have their right to select the interface for their own applications. The challenge of mobile terminals is how to make use of the multi-technology, and multi-network to offer high quality services according to several decision metrics.
Network/Interface Selection
In cellular networks, when a mobile terminal moves away from a base station, the signal level degrades and there is a need to switch connection to another base station. The mechanism by which an ongoing connection between the mobile terminal and its correspondent is transferred from one point of access to the fixed network to another is called handover or handoff. Handoff techniques have been well studied and deployed in cellular systems and are of a great deal in the wireless systems.
Traditionally, the handover decision, especially in case of horizontal handovers, is made purely according to radio signal strength (RSS) thresholds and hysteresis values as input parameters.
A decision for vertical handoff which consists in choosing the ―best‖ interface may depend on several parameters such as network conditions, application types, power requirements, terminal conditions, user preferences, security, cost and quality of service parameters.
In traditional cellular network, each mobile terminal is attached to a single cellular radio technology. The handover procedure is used when there is a need for a cell change during a call. The handover aims at maintaining or improving the connectivity to allow the mobile terminal using efficiently the radio resources. When the mobile terminal moves within the same network technologies, the handover scheme is known as horizontal handover Handover mechanism in Global System for Mobile communications (GSM) [Brunner][Siegmund] is considered as an example of horizontal handover. Handover transfers an ongoing call from one channel to another. The voice call connection is totally controlled by the subscribers’ operator.
The handover decision is based on main attributes such as the Received Signal Strength (RSS), the Bit Error Ratio (BER). These attributes are measured by the Mobile Station (MS) and sent frequently to the Base Station Controller (BSC) via the Base Transceiver Station (BTS).
The GSM handover characteristics are summarized in Table 1. The handover procedure is controlled by the network and needs assisted information provided by the mobile station.
Table of contents :
CHAPTER 1: INTRODUCTION
1 CONTEXT
2 MOTIVATIONS
3 CONTRIBUTIONS
4 ORGANIZATION
CHAPTER 2: BACKGROUND ON INTERFACE SELECTION AND FLOW/INTERFACE ASSOCIATION ISSUES FOR MULTI-INTERFACE MOBILE TERMINALS
1 EVOLUTION OF WIRELESS NETWORKS AND MOBILE TERMINALS
2 NETWORK/INTERFACE SELECTION
3 FLOW/INTERFACE ASSOCIATION
4 STATE OF THE ART
4.1 Decision metrics
4.2 Decision approaches
4.2.1 MADM approach
4.2.1.1 Normalization methods
4.2.1.2 MADM methods
4.2.2 Cost function approach
4.2.3 Utility function approach
4.2.4 Profit function approach
4.2.5 Policy approach
4.2.6 Game theory approach
4.3 Related work
4.3.1 Network/Interface selection
4.3.1.1 MADM based approach
4.3.1.2 Cost function based approach
4.3.1.3 Profit function based approach
4.3.1.4 Utility function based approach
4.3.1.5 Game theory based approach
4.3.2 Flow/interface association
4.3.2.1 Utility function based approach
4.3.2.2 Policy based approach
4.3.2.3 Game theory based approach
5 SUMMARY
CHAPTER 3: THE DISTANCE TO THE IDEAL ALTERNATIVE (DIA) ALGORITHM
1 INTRODUCTION
2 COMPARATIVE STUDY OF SAW, WP AND TOPSIS
2.1 Introduction
2.2 Simulation Scenarios
3 SIMULATION RESULTS
3.1 Simulation 1
3.2 Simulation 2
3.3 Simulation 3
3.4 Discussion
4 THE DIA ALGORITHM
5 PERFORMANCE COMPARISON
5.1 Simulation 1
5.2 Simulation 2
5.3 Simulation 3
6 SUMMARY
CHAPTER 4: FLOW/INTERFACE ASSOCIATION SCHEMES
1 INTRODUCTION
2 SINGLE FLOW/INTERFACE ASSOCIATION SCHEME
2.1 Motivation usage case
2.2 Related work
2.3 Interface Utility Function
2.3.1 The Application Utility Function
2.3.2 The Battery Consumption Function
2.3.3 The Interface Utility Function
2.4 Utility-based flow/interface association scheme
2.5 Performance Evaluation
2.5.1 Simulation scenarios
2.5.2 Simulation cases:
2.5.2.1 Case 1:
2.5.2.2 Case 2:
2.5.2.3 Case 3:
2.5.2.4 Case 4:
2.5.2.5 Case 5:
3 IMPLEMENTATION CONSIDERATIONS
3.1 Basic concept of IEEE 802.21
3.2 System implementation
3.2.1 Architecture
3.2.1.1 Multi-interface mobile terminal
3.2.1.2 User preference
3.2.1.3 IEEE 802.21 client
3.2.1.4 Network manager library
3.2.1.5 Bandwidth estimation module
3.2.1.6 Interface selection decision
3.3 Parameters Fetching
3.4 Conclusions
4 MULTIPLE FLOW/INTERFACE ASSOCIATION SCHEME
4.1 Model Description
4.2 Basic concepts of stochastic heuristic problems
4.2.1 Notion of Neighborhood
4.2.2 The stochastic heuristic algorithms
4.2.2.1 Local Search
4.2.2.2 Tabu Search
4.2.2.3 Simulated Annealing Algorithm
4.2.2.4 Genetic Algorithm
4.3 Performance evaluation comparison
4.3.1 Simulation set up
4.3.2 Simulation scenarios
4.3.3 Simulation results
4.3.3.1 Local search
4.3.3.2 Tabu search
4.3.3.3 Simulated annealing
4.3.4 Discussions
4.4 Oriented diversification of Tabu search for the multiple flow/interface association
4.5 Performance evaluation
5 SUMMARY
CHAPTER 5: STRATEGY GAME FOR FLOW/INTERFACE ASSOCIATION IN MULTI-INTERFACE MOBILE TERMINALS
1 INTRODUCTION
2 INTRODUCTION TO GAME THEORY
2.1 Definition of games
2.2 Examples of Games
2.3 Example of Nash equilibria
2.4 Equilibrium strategies
2.4.1 Nash Equilibria for pure strategies
2.4.2 Nash Equilibria for mixed strategies
2.5 Introduction to Evolutionary Games
2.5.1 Potential game
2.5.2 Evolutionary dynamics
2.5.2.1 Replicator dynamics
2.5.2.2 Brown-von Neumann-Nash (BNN) dynamics
2.5.3 Positive correlation (PC)
2.5.4 Equilibrium
2.5.5 Nash learning algorithm
2.6 Related work
3 FRAMEWORK AND MODEL
3.1 Application-based model
3.2 Mixed strategy and equilibrium
3.3 Replicator Dynamic
3.4 Efficiency of the equilibrium points
3.5 Nash learning algorithm
4 IMPLEMENTATION AND VALIDATION
4.1 Implementation
4.1.1 Utility function
4.1.2 Bandwidth allocation
4.2 Simulations
4.2.1 Scenario 1 (Simple scenario)
4.2.2 Scenario 2
4.2.3 Discussion
5 SUMMARY
CHAPTER 6: CONCLUSION
1 NOTION OF NEIGHBORHOOD
2 LOCAL SEARCH
3 TABU SEARCH (TS)
4 SIMULATED ANNEALING ALGORITHM (SA)
5 GENETIC ALGORITHM (GA)
BIBLIOGRAPHY
