Selected Topics in Communication Networks and Distributed Systems

The following sections are included:

• CONTENTS

Underlying all wireless technologies is the fundamental principle that mobile users communicate with a base station via a wireless channel. In this chapter, a basic wireless channel is explored and the effects of distance, and thereby mobility, on the wireless channel is shown. The chapter then presents a generic wireless communication network and identifies the role and function of each component i.e., mobile device, access network and core network. The merger of new technologies and data services with mobile networks has created new issues in network management. The chapter provides a tutorial on the various mobility management methods used to mitigate and solve mobility issues. Concepts such as personal mobility, session mobility and service mobility can be viewed as high level mobility management solutions while concepts such as ad hoc mobility, mode mobility and terminal mobility are viewed as low level solutions. The chapter explores these concepts and presents current protocols/methods. A guide is provided for practitioners who may implement some of the mobility management methods. The chapter highlights research directions by presenting a range of projects and standard bodies working in the mobility management field.

The Universal Mobile Telecommunications System (UMTS) constitutes the premier third generation (3G) wireless technology that is dominating the global market. Multicasting is an efficient method of supporting group communication as it allows the transmission of packets to multiple destinations, using fewer network resources. The need for broadcasting and multicasting in UMTS led to the definition of the Multimedia Broadcast/Multicast Service (MBMS) framework, which targets the efficient utilization of radio and network resources of the UMTS network. This chapter introduces the key concepts of UMTS and in particular the MBMS framework of UMTS. Moreover, it investigates the power profiles of several transport channels (common and dedicated) which could be employed for the transmission of MBMS services to mobile users. Problems regarding the high power requirements for the realization of MBMS are also presented. The reader will become familiar with these problems and the proposed techniques/solutions.

WiMAX/802.16 is a kind of network providing IP-based broadband wireless access to infrastructure networks such as the Internet. The two main envisioned applications are Web access and voice over IP. Highlights of WiMAX/802.16 are secure communications and broadband access in remote areas. This chapter covers the background, physical layer, medium access control layer, mobility support, mesh mode and multihop relay operation of WiMAX/802.16. The chapter ends with a presentation of thoughts for practitioners and a discussion of directions for future research.

Mobile Radio Frequency Identification (RFID) is a newly emerging technology which uses the mobile phone as an RFID reader with a wireless technology and provides valuable new services to the user by integrating RFID and ubiquitous sensor network infrastructure with mobile communications and wireless Internet. UHF Mobile RFID technology is based on ISO/IEC 18000-6C can share UHF RFID tags used in logistics/SCM and can avoid redundant investment expected to be integrated with other frequency band RFID. Mobile RFID enables business to provide new services to mobile customers by securing services and transactions from the end user to a company's existing e-commerce and IT systems. This chapter discusses UHF based mobile RFID technology. We begin with a look at the general of a mobile RFID system anatomy, followed by a discussion of the components that make up a typical mobile RFID system framework and the underlying sub-systems that make them work.

Wireless metropolitan networks support the mobility of subscribers over a wide range by associating them to the access points or base stations in the vicinity. The base stations together with the mobile users form an ad hoc network. Dynamic allocation of resources in such a network is challenging as the infrastructure is not fixed and the topology keeps changing fast. To make matters worse, each application, such as network gaming, file transfer, video streaming and online advertising, require different bandwidths, buffering and different levels of service quality. Maintenance of the agreed quality of service is difficult owing to the frequent handoffs, mobility of the users, and fading and synchronization issues.

Determining the location of a wireless device by measuring the signal strengths from a set of various access points (location fingerprinting) is an important issue. This work analyzes and applies the support vector machine (SVM) methodology under the framework of a statistical learning model to attain the objective of location estimation. SVMs are a set of related supervised learning methods that can be used for classification and regression. A special property of SVMs is that they simultaneously minimize the empirical classification error and also maximize the geometric margin. Received signal strength (RSS) along with the corresponding location information forms a training dataset for the experiment. As RSS is a part of the normal operating mode of wireless equipment, in particular Wi-Fi, no custom hardware is required. On a given dataset, tests are performed based on the SVM paradigm, with the results obtained being comparable to other approaches being used.

The IEEE 802.11 distributed coordination function standard has gained global acceptance and popularity both in wireless LANs and the wireless multi-hop ad hoc environment. In IEEE 802.11, all packets (RTS/CTS/DATA/ACK) are sent with maximum power. It has been shown that this type of handshake communication decreases spatial reuse and thus decreases capacity throughput, which additionally may yield unnecessary energy consumption. Hence, enhancing spatial reuse and conserving energy consumption have been perhaps the two most vital issues in designing IEEE 802.11-based multi-hop ad hoc networks. Transmission power control has been adopted as an efficient technique to reduce the total energy consumed in packet delivery and/or enhance network throughput by increasing the channel spatial reuse by allowing concurrent transmissions to occur at the same time. Judiciously controlling the transmit power is far from easy. Improper control may result in performance that is poorer than without such control. For instance, reducing the power too much may leave the network unconnected, or produce excessive delays. In this chapter, we present an overview of the research done in the area of transmission power control. The main theme of this chapter is to provide the reader with an understanding of the problems encountered in the exploitation of power control, approaches to solving these problems and their performance benefits.

Energy efficiency and spatial reuse have been the key design constraints for evaluating the performance of IEEE 802.11-based multi-hop ad hoc networks. Energy efficiency is crucial because mobile nodes are equipped with short-range lightweight radios operating with limited energy, while efficient spatial reuse is needed to increase the network capacity. The performance of IEEE 802.11 DCF in multi-hop networks suffers from more collisions due to the existence of hidden nodes. The increase in the number of collisions will deteriorate the network throughput and increase energy consumption. Collision avoidance MAC protocols for multi-hop ad hoc networks have addressed this challenge. Three physical attributes are of importance to avoid collisions and hence enhance throughput, namely, physical carrier sensing threshold, transmission data rate, and transmit power. In this chapter, we identify the challenges and open issues involved in designing protocols for better spatial reuse in multi-hop ad hoc networks. Specifically, we survey the research aimed at resolving these challenges and propose future work to address open issues.

Waypoint routing is a lightweight hierarchical routing model in which the route from source to destination is divided into a number of segments by waypoint nodes. Routing is then done at intersegment and intrasegment levels. A major difficulty in waypoint routing is the metric by which a route is divided into segments. There are a few ad hoc methods such as dividing the route based on the number of hops, or dividing by selecting high speed nodes as waypoints. But a better approach would be one which considers wireless environment characteristics for dividing the route into segments, i.e., considering cross layer (link and physical layer) parameters like packet loss rate, bandwidth, interference, etc. to divide the route into segments. Such a waypoint routing protocol, called cross layer DOA (dynamic source routing over ad hoc on-demand distance vector) routing protocol, is proposed and discussed in this chapter.

Usage of wireless devices give rise to interesting spontaneous networking scenarios viz. service discovery, self-organization, etc. In service discovery, devices automatically discover network services and may advertise their existence in a dynamic way. In an ad hoc network, where users are organized in groups, the management procedures are introduced to support such dynamic spontaneous groups. The proposed spontaneous group management is based on the concept of group leader, in which a terminal is responsible for the location update of a group of terminals. As energy is a scarce resource, novel schemes to reduce the search time and power consumption of all the devices in the ubiquitous environment are required. These are proposed and discussed in this chapter.

Coverage, which reflects how well a sensor field is monitored or observed, is an important performance metric for distributed sensor networks. The individual sensor coverage model is closely related to the sensing functions of different sensor types while the network coverage is a collective measure from geographically distributed sensor nodes. In the literature, the coverage problems in distributed sensor networks have been formulated in various ways with different assumptions, constraints and objectives. This chapter surveys recent research progress made to address area coverage problems. We first provide detailed discussions on the sensor coverage model and the design issues. We then classify three main types of area coverage problems in different network operation phases. Before network deployment, the main coverage problem is to determine the least number of nodes that is needed to completely cover the sensor field. After network deployment, sensing activity scheduling decides how to activate sensors alternatively in order to prolong network operation time. Network redeployment is possible if sensor nodes are equipped with a mobility unit. The node movement strategy determines how to move nodes to desired positions while guaranteeing coverage. We review representative approaches in the literature for each type of problem. We also outline some variants each with different assumptions, constraints or objectives for these problems, which we list as further reading.

In this chapter, we study how combinations of knowledge processing techniques with networking technologies can help develop cognitive networks that involve heterogeneous devices connected through wireless/wired media which are context-aware, can sense and adapt their environment(s), make intelligent decisions and effect changes in their environment(s), can self-manage, self-heal, and self-protect. They can be profitably applied to make a difference in areas that include soil and water management in agriculture, intelligent management and optimization of power systems, reliable and smart control of fermentation processes and bio-reactors, management of environmental hazards and disasters, optimized modulation of communication antennas, precise management of hospitals and assisted living premises, and control of space and defense systems. We will study two applications: soil and water management in agriculture, and intelligent management and optimization of power systems.

We begin by providing an overview of communication networks and discussing their differences from a routing/switching perspective. Subsequently, we focus on both wired and wireless packet-switched communication networks and explain various features of routing techniques. Since routers are essentially an indispensable component in building high-performance communication networks, we briefly examine the different architectures and design issues. We then describe the fundamental routing techniques used in today's networks for unicast, multicast and broadcast communication. After that, we explore advances in routing to provide quality of service support and enforce administrative policies while making the best of the network resources. To see how theory is put into practice, we look at routing protocols for two important communication networks: the Internet and the mobile ad hoc network. We then show how traffic management techniques can help in improving the performance of routing and forwarding of network traffic while optimizing the utilization of network resources. At the end of this chapter, we highlight some directions for future research related to routing and traffic management.

Since its research and standardization started in late 1980s, network management has become more and more crucial for keeping pace with the changing management requirements of ever-developing networks. After briefly describing some background, this chapter respectively discusses network management standards (since the development of management technology is always coupled with protocol standardization) and methodology (including both techniques and strategies), and evaluates these solutions from the OSI perspective using its proposed network management model. This chapter then examines present-day challenges in the network management domain and illuminates current trends. Finally, this chapter points out directions for future research, which may be utilized to build management solutions for next-generation networks.

Efficient and cost-effective measurement of network characteristics is pivotal for distributed systems deployed on the Internet. The network characteristics are utilized by Internet-based distributed systems to provide better services to users and enhanced performance for applications. This chapter provides a detailed analysis of existing techniques for the measurement of the four important network characteristics: latency, bandwidth, path detection and loss rate. The chapter describes key concepts related to network measurements, including techniques for clock synchronization, strategies for time stamping of probes, methods for network analysis, difference between active and passive measurements, and comparison of round trip times versus oneway delay measurements. It elaborates on the usefulness of different transport and network layer protocols (i.e., Transmission Control Protocol (TCP), User Datagram Protocol (UDP) and Internet Control Message Protocol (ICMP) for obtaining network measurements and describes some important measurement tools such as Ping, Traceroue, Pathchar, Sting, Scriptroute, Spruce and Paris Traceroute. The chapter explains the effectiveness and limitations of these tools with respect to the measurement of network characteristics.

Network information flow (NIF) theory is a great breakthrough in information theory. The essence of NIF is that intermediate network nodes are allowed to take part in encoding and decoding. It is therefore also called network coding. One of its remarkable virtues compared with routing is that it can achieve maximum multicast capacity. This chapter describes the fundamental notions and principles of network coding as well as its typical applications, especially in wireless multi-hop scenarios. Future directions of NIF theory are discussed.

Grid, cluster or Internet computing is essentially concurrent. There are many different models of concurrent processes, which are inherent for distributed systems. The goal of this chapter is to introduce a common formalized framework aimed at grid, cluster and Internet computing for current research in this area and to eliminate the shortcomings of existing models of concurrency. Following on from previous research, here we build a high-level metamodel event-action-process (EAP) for concurrent processes. This metamodel comprises a variety of other models of concurrent processes. We shape these mathematical models and study events, actions and processes in relation to important practical problems, such as communication in networks, concurrent programming, and distributed computations. In the third section of the chapter, a three-level algebra of events, actions and processes is constructed and studied as a new stage of algebra for concurrent processes. Relations between EAP process algebra and other models of concurrency are considered in the fourth section of this chapter.

This chapter presents the design, implementation and evaluation of a dataflow system, including a macro-dataflow programming model, runtime system and an online scheduling algorithm, to simplify the development and deployment of distributed applications. The model provides users with a simple interface for programming applications with complex parallel patterns. The associated runtime system dispatches tasks onto distributed resources through a proposed online algorithm, called localized heterogeneous earliest-finish-time (L-HEFT), and manages failures and load balancing in a transparent manner. The system has been implemented in a .NET-based enterprise grid software platform, called Aneka. Evaluation of the scalability and fault tolerance properties of the system has been carried out. The results demonstrate that our L-HEFT scheduling algorithm is efficient compared to existing techniques as it introduces a low overhead while making mapping decisions.

With the vast amount of computing resources available across the Internet, basic communication and programming mechanisms have been evolving from simple client/server socket communication model to object-oriented middleware systems, and lately, to the Grid computing service system platform. Nowadays, computers can be interconnected with networks to create high-performance parallel, distributed, and clustered computing systems. From a networking design perspective, the system model has moved from a basic multiple-to-one client/server system to a peer-to-peer model. In this chapter, we review the foundation technologies that have been developed into the Grid computing. Different algorithmic designs for overlay networks are also discussed. We then investigate the design of Grid computing operating platform for serving largescale real-time online applications. In order to validate the proposed design, which we call a massively multi-user online platform (MMOP), we design an online game to carry out thorough experiments for evaluating the overall system performance.

The Internet is a heterogeneous network environment and the network resources that are available to multimedia applications can be modified very quickly. Multimedia applications must have the capability to adapt their operation to network changes. In order to add adaptation characteristics to multimedia applications, we can use techniques both at the network and application layers. Adaptive multimedia transmission techniques have the capability to transmit multimedia data over heterogeneous networks and adapt media transmission to network changes. In order to provide adaptive multimedia transmission, mechanisms to monitor the network conditions and mechanisms to adapt the transmission of the data to the network changes must be implemented. This chapter concentrates on the architecture and design issues of adaptive multimedia techniques that provide the capability to transmit multimedia data over heterogeneous networks and adapt the transmission of the multimedia data to the network changes.

Theoretically packet inter-arrival patterns will follow a Poisson distribution. In practice they do not. Further, in a distributed processing environment, there are often different protocols involved which may exacerbate the inter-arrival pattern variability. This chapter uses four examples to provide thoughts for practitioners. In the first example, there is evidence to suggest that the efficiency of the inter-processor communication decayes as the complexity of the matrix increases. The next example uses a better designed algorithm. The improved algorithm allows for a decrease in the time it takes to solve the problem as additional nodes are added; however, the decrease is not a linear one. In the third example, a performance gain is seen with up to 12 processors. In the final example, performance of a password cracking experiment increases with each additional processor. The low cost and performance of distributed computing makes it a popular option to solve complex problems. Perhaps this environment will replace the supercomputer at some point in the future.

Detecting intrusions in networks and applications has become one of the most critical tasks to prevent the misuse of network resources by attackers. The cost involved in protecting these valuable resources is often negligible when compared with the actual cost of a successful intrusion. This is one of the motivating factors for developing more powerful intrusion detection systems. In this chapter, we describe the problem of intrusion detection in detail and discuss various methods that have been used for building intrusion detection systems. We also discuss some of the emerging methods, such as conditional random fields, which can detect a wide variety of attacks with relatively higher accuracy.

Security is paramount in mobile ad hoc networks (MANETs) since a MANET is neither conducive to centralized authorities nor suitable for inheriting the solutions that have been proposed for wired networks. Given that end-to-end communication between applications relies on the self-organized characteristics of MANETs, most if not all the proposed security solutions concentrate on securing communication through multi-hop trustworthy nodes. In this chapter, we present state-of-the-art security in MANETs and the survey comprises MANET-based secure routing, key management, and trust management systems. However, we confine ourselves to a few well-regarded proposals due to the exhaustive list of proposals available in each of the above-mentioned categories. First, we discuss the features inherent in MANETs and their impact on the design of security mechanisms, in addition to the threats and attacks that are common in MANETs. Second, we describe a few well-known solutions in the area of secure routing and key management to demonstrate their role as a prevention system. We then discuss the limitations of those systems such as their inability to react to dynamically changing attack patterns and their assumption that nodes will cooperate for routing and network management. Finally, we address the recent advancements in security systems, where a defense-in-depth approach is adopted to incorporate trust management systems as the second layer of defense to prevention systems. Trust management systems complement prevention systems by measuring the trustworthiness of nodes and promptly react to dynamically changing attack patterns. We then detail the limitations of trust management systems and discuss possible research directions to address those limitations.

Identity management in mobile communication systems has the advantage that it can be bound not only to something that the user knows, but also to something that he possesses, i.e., it provides two-factor authentication. In converging communication networks the smart card can support a generic user identity management. These considerations led to the development of the Generic Bootstrapping Architecture (GBA) by the mobile standardization body 3rd Generation Partnership Project (3GPP) presented in this chapter. We describe existing mobile identity management approaches, introduce the Liberty Alliance Framework and outline how the mobile GBA can be utilized together with the Liberty Alliance Identity Federation Framework and the Web Services Framework for single sign-on or authentication services. Finally, we highlight the latest network convergence aspects in the area of identity management and outline the latest research questions and activities for meta-identity management systems.

The enormous growth of the Internet has had a huge impact on users' daily lives, providing us with a wealth of information easily and conveniently. At the same time, this accessibility has made the Internet an easy target for users who wish to disrupt the flow of information or exploit the Web for personal gain. Those who compromise the flow of information or access information in an unauthorized manner are called attackers. Many tools are now available for attackers to test the vulnerability of Internet.

One form of attack is denial-of-service (DoS). A DoS attack is launched by sending a stream of packets to a victim (affected system or destination system). This consumes some key resources of the victim and renders them unavailable to legitimate users, or it provides an attacker with unlimited access to the victim so that the attacker can inflict arbitrary damage. If DoS attacks are generated from different places then it is called distributed denial-of-service (DDoS). Attackers deploy multiple systems to attain this goal. DoS/DDoS attacks are one of the major threats to the Internet. The normal functions of popular servers on the Internet are disturbed due to this. In both cases, the attackers hide their identity by spoofing their Internet Protocol (IP) addresses. The mitigation of a DoS attack is a reactive countermeasure that is actually initiated by the victim after detecting an attack. The main challenges in DoS mitigation is to accurately identify attack packets and filter them without causing collateral damage to legitimate traffic destined for the victim. Two models have been proposed for mitigating DoS/DDoS attacks: secure overlay services and server hopping using distributed firewalls. In this chapter, the effectiveness of these approaches is assessed for different types of DoS/DDoS attacks. The simulation results of DoS/DDoS that without any defense models shows the packet delivery time increases and the packet delivery to the server is delayed. In the secure overlay services defense model for DoS/DDoS attacks, the variation in packet delivery time remains almost constant with the actual packet delivery time. The server hopping using distributed firewalls model also maintains a constant packet delivery time. However, the amount of variation in packet delivery time in secure overlay services is greater than with server hopping using distributed firewalls. Through simple analytical models, it is shown that DoS/DDoS attacks directed against any part of the secure overlay services infrastructure have a negligible probability of disrupting the communication between two parties. Furthermore, the resistance of a secure overlay services network to DoS/DDoS attacks increases greatly with the number of nodes that participate in the overlay. The secure overlay services and server hopping using distributed firewalls architectures provide a range of defenses that can severely limit the damage caused by DDoS attacks.

The following sections are included:

• SOLUTIONS