The current Internet consists of a large collection of Autonomous Systems (ASes) or domains, each being a network or group of networks managed by a single authority such as an Internet Service Provider (ISP) or a commercial enterprise. These ASes employ their own network policies and routing protocols. In order to support all the Internet traffic caused by everyday applications such as peer-to-peer content downloading, multimedia streaming, voice over IP, online gaming, video conferencing, etc., network providers seek to optimise their network resource usage in an efficient manner utilising Traffic Engineering (TE) techniques. These techniques control traffic routing so as to optimise operational IP network performance and can be classified into intra-AS and inter-AS. Network providers use intra-AS TE techniques to control traffic routing within their network to achieve objectives such as load balancing and/or minimising resource consumption. On the other hand, they use inter-AS TE techniques to control inbound and outbound traffic to achieve load balancing over inter-AS resources and/or minimise peering costs. Unfortunately, due to a variety of events such as malicious attacks, misconfiguration, router CPU overload, interface failure and accidental damage, various failures, and in particular link failures, occur as part of daily network operations. Research has revealed that both single intra- and inter- AS link failure are common events and transient in nature. The potential impact of a link failure can be delay, packet discard, service disruption and severe congestion due to the shifting of an excessive amount of traffic to alternative paths that are already highly utilised. Given the shortlived nature of transient failures, network operators may not have sufficient time to re-configure their networks before the failure is restored, resulting to the detrimental failure impact. In order to manage this situation, in this thesis we propose proactive network provisioning approaches that predict the impact of transient link failures and implement remedial solutions to alleviate the detrimental failure impact. The contributions of our thesis are as follows: o Improving intra-AS primary path robustness. Routing traffic flows though high availability network segments results in minimising their path failure probability and also in reducing the amount of backup resources required to protect these primary paths. We therefore propose a heuristic algorithm with four link cost functions that take link availability into account to improve primary path availability in MPLS networks while optimising network resources. o Making inter-AS outbound TE robust to inter-AS link failure for long-term and short-term network provisioning. Transient inter-AS link failures are as common as intra-AS link failures and can have similar detrimental consequences. We therefore propose an IP tunnelling approach that can provide fast rerouting and also an algorithm that provides primary and secondary egress points to achieve load balancing under both no failure and also single inter-AS link failure states for long-term network provisioning. We solve this algorithm using a tabu search heuristic. Additionally, network conditions such as traffic variations and destination prefix reachability change, these events may make the long-term outbound TE solutions inaccurate. We therefore propose a short-term network provisioning framework that provides inter-AS outbound robust TE solutions considering the network dynamics. Making both intra-AS and inter-AS outbound TE robust to intra- or inter-AS link failure. Due to the common and transient nature of both intra- and inter-AS link failures, there are significant interactions between them. As a result, the overall network performance may not be truly robust to intra- and inter-AS link failures if they are considered separately. We therefore propose a joint robust TE approach that balances the intra- and inter-AS load balancing under intra- or inter-AS link failure taking into account their interactions. In summary, in this thesis we aim to propose robust approaches for both intra-AS and inter-AS outbound TE so as to optimise IP operational network performance under both normal and also failure conditions.

Arising from the fact that the terrestrial network is becoming more IP-centric, it is crucial to be able to support IP services over the satellite component of UMTS as well to achieve close synergy with the terrestrial counterpart for seamless communication. Nevertheless, the direct utilization of the IP protocol suite in a mobile environment is not straightforward, and when coupled with the additional constraints and shortcomings that satellite-based systems pose, it imposes additional challenges to the integration of satellite systems into the IP-based T-UMTS and the global Internet. In this thesis, we study the feasibility of establishing multimedia sessions over S-UMTS based on the SIP protocol, and investigate efficient solutions to transport SIP signalling, taking into account not only the larger satellite propagation delay but also the impact of the S-UMTS radio interface arid channel impairment. The preservation of maximum commonality between T- & S-UMTS is the aim of our work to facilitate an effective integration of both segments. Given that a dedicated channel needs to be used for the transport of SIP signalling, the Random Access Channel (RACH) is first investigated. The impact of the different RACH transmission control parameters on the access performance over S-UMTS is studied with their trade-offs identified. It is further demonstrated how QoS differentiation could be provided on the RACH through Access Service Classes (ASCs). Two schemes have also been proposed to decrease the risk of code/signature collision, where it has been shown that when network-wide synchronization is already in place, these schemes perform better than the ETSI S-UMTS scheme as the access rate increases. By incorporating a link layer retransmission based on the RLC-AM mechanisms, it has been shown that SIP session establishment can work acceptably well over S-UMTS with performance comparable to the circuit-switched call setup. It is further demonstrated that with proper configuration of RLC parameters and retransmission options, the session setup performance under various channel conditions can be substantially improved. Two schemes that mitigate the inefficiencies of the existing UMTS RLC-AM procedure for session setup over S-UMTS have also been proposed. The main advantage of the first scheme lies in its ability to recover the missing last radio segments in a single round-tip whilst minimizing the feedback and avoiding redundant retransmissions. The second approach minimises redundant retransmissions when more than one feedback report trigger is configured, where improved system performances are obtained. Finally, in the joint study of SIP's own reliability mechanisms and RLC retransmission scheme, it has been shown that precious radio and network resources can be saved by making SIP and RLC 'more co-operative' through different degree of awareness of each other. Key words: S-UMTS, IP, Session Initiation Protocol, Random Access Charmel, Radio Link Control.

No description available.

For years the Internet networking community has been struggling to develop ways to manage networks. Initial attempts brought mechanisms and protocols that focused on managing and configuring individual networking devices i.e. the Simple Network Management Protocol (SNMP). This model worked well in early deployments of IP management systems for local and metropolitan area networks but now, with the evolution of Quality of Service (QoS) models such as the Differentiated Services (DiffServ) framework, the complexity and overhead of operating and administrating networks increases enormously. There is also a need to be able to program management systems and network components to adapt to emerging requirements and subsequently be able to dynamically change the behaviour of the whole system to support modified or additional functionality. The emerging Policy-based Network Management paradigm claims to be a solution to these requirements. Policy-based Management can guide the behaviour of a network or distributed system through high-level declarative directives that are dynamically introduced, checked for consistency, refined and evaluated, resulting typically in a series of low-level actions. The objective of this thesis is to investigate the application of Policy-based Management in the context of QoS Management of IP DiffServ Networks. By using policies as a means for building extensible hierarchical management systems, we propose a novel Policy-based QoS management architecture and specify the relevant policies that can drive its behaviour dynamically, providing a holistic approach to the area of policies for QoS Management. We first present our view on policies as a means of extending the functionality of management systems dynamically, in conjunction with pre-existing "hard-wired" management logic and provide a generic framework for their application to hierarchical distributed management systems. The programmability aspect of policies is an issue which has not been properly addressed in the literature and constitutes an important contribution of this thesis. We then propose a single architecture for managing an IP DiffServ network, identifying the required functional components and their interactions addressing both service management and resource provisioning (traffic engineering) aspects of QoS Management. The design of the architecture caters for both offline and dynamic operation. Furthermore, we identify the parameters of the functional components of the architecture that are influenced by policies and present an object oriented representation of those policies based on the Policy Core Information Model (PCIM). This work differentiates from relevant work on QoS policies since it addresses the area of QoS Management in its totality defining policies related to service management and traffic engineering both at a network and element management level. Finally, we validate the proposed policy-based framework by presenting a detailed description of the design and implementation of the components of the policy management sub-system needed to be deployed in order to make our system policy-driven and present examples of QoS policies describing their transformation from their definition by the operator until their enforcement.

Grid computing is diverse and heterogeneous in nature, spanning across multiple domains where resources are not owned or managed by a single administration. This brings about many challenges to Grid resource management and exposes the user to the Grid middleware complexities. Thus this research develops a user-centric resource broker that insulates the users from the Grid complexities, alleviating them from the burden of having to know the various mechanisms of the Grid middleware. The broker is based on the SNAP (Service Negotiation and Acquisition Protocol) framework and focuses on applications that require resources on demand. It is important for applications that require resources on demand to reserve the necessary resources within the minimum time possible. Thus the work in this thesis has developed a three-phase commit protocol which enhances the traditional twophase commit protocol. Performance evaluation has been carried out to evaluate the SNAP-based resource broker using the traditional two-phase commit protocol and the newly developed three-phase commit protocol. The evaluation has been conducted on a local Grid test-bed, a distributed Grid infrastructure (the White Rose Grid) and through mathematical modelling and simulation. Throughout the evaluation, the SNAP-based resource broker using the three-phase commit protocol provides a significant performance enhancement, over the use of the traditional twophase commit protocol, in terms of the time taken between submission (to the broker) of user requirements and the job beginning execution.

No description available.

Cette thèse se positionne dans le contexte des réseaux sans fil multi-sauts (réseaux de capteurs/actionneurs/robots mobiles). Ces réseaux sont composés d’entités indépendantes à la puissance limitée et fonctionnant sur batteries qui communiquent exclusivement par voie radio. Pour pouvoir relayer les messages d’un robot à une station de base, on utilise des protocoles dits « de routage» qui ont en charge de déterminer quel robot doit relayer le message. Nous nous sommes basés sur le protocole CoMNet, qui adapte la topologie du réseau à son trafic lors du routage afin d’économiser de l’énergie en déplaçant les robots. Mais modifier la topologie c'est aussi modifier les possibilités de routage. Nous proposons donc MobileR (Mobile Recursivity) qui choisit le prochain noeud en ayant anticipé par le calcul les conséquences de tous les changements de topologie possibles. Un autre problème vient du fait qu’il y a souvent plusieurs nœuds qui détectent un même événement et vont émettre des messages à router vers la station de base. Ces messages vont finir par se croiser, et le nœud de croisement va sans cesse être relocalisé sur chacun des chemins. Le protocole PAMAL (PAth Merging ALgorithm) détecte ces intersections : il va provoquer une fusion des chemins de routage en amont du nœud de croisement et une agrégation de paquets en aval. Enfin, le protocole GRR (Greedy Routing Recovery) propose un mécanisme de récupération pour augmenter le taux de délivrance des messages dans les réseaux de capteurs/actionneurs avec obstacle(s). En effet, les protocoles de routage actuels échouent face à un obstacle. GRR va permettre de contourner l’obstacle en relocalisant des nœuds tout autour. / This thesis is about wireless multi-hop networks such as sensor/actuator networks and actuator networks. Those networks are composed of independent entities which have limited computing and memory capabilities and are battery powered. They communicate through the radio medium and do not require any static infrastructure. In order to relay messages between actuators up to the base station, we use what is called "routing protocols". My works rely on CoMNet, the first geographic routing protocol which aims to adapt the network topology to the routed traffic in order to save energy. Nevertheless, CoMNet does not consider the consequences of those relocations more than in a one-hop way. We proposed MobileR (Mobile Recursivity), which anticipates the routing in a multi-hop manner through computations over its one-hop neighbors. Hence it can select the “best” next forwarding node according to its knowledge. Another important topic is that events are likely to be detected by multiple sensors and all of them transmit message toward the destination. But those messages are likely to cross over an intersection node. This crossing provokes useless oscillation for it and premature node death. The PAMAL (PAth Merging ALgorithm) routing algorithm detects those routing path crossing and provokes a path merging upstream and uses a packet aggregation downstream. Finally, the Greedy Routing Recovery (GRR) protocol takes controlled mobility into account in order to increase delivery rate on topology with holes or obstacles. GRR includes a dedicated relocation pattern which will make it circumvent routing holes and create a routing path.

In recent years the growth in demand for multimedia applications delivered over a network has been extensive illustrated by the phenomenal expansion of the World Wide Web in today's Internet. Increasing amounts of electronic information are presented in a multimedia format and delivered over networks, which generates increased traffic demands. In short, future networks will not only comprise multiple, interconnected data transport platforms but will also need to handle an increasing range of distributed multimedia services, delivered to a diverse user community that is increasingly becoming mobile.

Mobile ad hoc networks (MANETs) have become increasingly important in view of their promise of ubiquitous connectivity beyond traditional fixed infrastructure networks. Such networks, consisting of potentially highly mobile nodes, have provided new challenges by introducing special consideration stemming from the unique characteristics of the wireless medium and the dynamic nature of the network topology. The TCP protocol, which has been widely deployed on a multitude of internetworks including the Internet, is naturally viewed as the de facto reliable transport protocol for use in MANETs. However, assumptions made at TCP’s inception reflected characteristics of the prevalent wired infrastructure of networks at the time and could subsequently lead to sub-optimal performance when used in wireless ad hoc environments. The basic presupposition underlying TCP congestion control is that packet losses are predominantly an indication of congestion in the network. The detrimental effect of such an assumption on TCP’s performance in MANET environments has been a long-standing research problem. Hence, previous work has focused on addressing the ambiguity behind the cause of packet loss as perceived by TCP by proposing changes at various levels across the network protocol stack, such as at the MAC mechanism of the transceiver or via coupling with the routing protocol at the network layer. The main challenge addressed by the current work is to propose new methods to ameliorate the illness-effects of TCP’s misinterpretation of the causes of packet loss in MANETs. An assumed restriction on any proposed modifications is that resulting performance increases should be achievable by introducing limited changes confined to the transport layer. Such a restriction aids incremental adoption and ease of deployment by requiring minimal implementation effort. Further, the issue of packet loss ambiguity, from a transport layer perspective, has, by definition, to be dealt with in an end-to-end fashion. As such, a proposed solution may involve implementation at the sender, the receiver or both to address TCP shortcomings. Some attempts at describing TCP behaviour in MANETs have been previously reported in the literature. However, a thorough enquiry into the performance of those TCP agents popular in terms of research and adoption has been lacking. Specifically, very little work has been performed on an exhaustive analysis of TCP variants across different MANET routing protocols and under various mobility conditions. The first part of the dissertation addresses this shortcoming through extensive simulation evaluation in order to ascertain the relative performance merits of each TCP variant in terms of achieved goodput over dynamic topologies. Careful examination reveals sub-par performance of TCP Reno, the largely equivalent performance of NewReno and SACK, whilst the effectiveness of a proactive TCP variant (Vegas) is explicitly stated and justified for the first time in a dynamic MANET environment. Examination of the literature reveals that in addition to losses caused by route breakages, the hidden terminal effect contributes significantly to non-congestion induced packet losses in MANETs, which in turn has noticeably negative impact on TCP goodput. By adapting the conservative slow start mechanism of TCP Vegas into a form suitable for reactive TCP agents, like Reno, NewReno and SACK, the second part of the dissertation proposes a new Reno-based congestion avoidance mechanism which increases TCP goodput considerably across long paths by mitigating the negative effects of hidden terminals and alleviating some of the ambiguity of non-congestion related packet loss in MANETs. The proposed changes maintain intact the end-to-end semantics of TCP and are solely applicable to the sender. The new mechanism is further contrasted with an existing transport layer-focused solution and is shown to perform significantly better in a range of dynamic scenarios. As solution from an end-to-end perspective may be applicable to either or both communicating ends, the idea of implementing receiver-side alterations is also explored. Previous work has been primarily concerned with reducing receiver-generated cumulative ACK responses by “bundling” them into as few packets as possible thereby reducing misinterpretations of packet loss due to hidden terminals. However, a thorough evaluation of such receiver-side solutions reveals limitations in common evaluation practices and the solutions themselves. In an effort to address this shortcoming, the third part of this research work first specifies a tighter problem domain, identifying the circumstances under which the problem may be tackled by an end-to-end solution. Subsequent original analysis reveals that by taking into account optimisations possible in wireless communications, namely the partial or complete omission of the RTS/CTS handshake, noticeable improvements in TCP goodput are achievable especially over long paths. This novel modification is activated in a variety of topologies and is assessed using new metrics to more accurately gauge its effectiveness in a wireless multihop environment.

Στο παρόν κείμενο μελετώνται οι τεχνικές υλοποίησης L2 MPLS VPNs Cicruit Cross Connect (CCC) και Any Trasfer over MPLS (AToM) που παρέχουν οι δύο κύριοι κατασκευαστές δικτυακού υλικού Juniper και Cisco αντίστοιχα. Πρόθεση της μελέτης είναι να διερευνηθούν τα ιδιαίτερα τεχνικά χαρακτηριστικά της κάθε τεχνικής και στην συνέχεια να μελετηθούν τρόποι επίτευξης της διαλειτουργικότητας μεταξύ των δύο. Το τελευταίο είναι χρήσιμο στην περίπτωση που μια υπηρεσία παροχής L2 MPLS VPNs επεκτείνεται σε ένα δίκτυο που αποτελείται από εξοπλισμό και των δύο κατασκευαστών. Στην ιδιαίτερη περίπτωση του ΕΔΕΤ το οποίο είναι ένα αμιγώς «Cisco» δίκτυο το παραπάνω έχει εφαρμογή στο όριο διασύνδεσης με το GEANT το οποίο αντίστοιχα είναι ένα αμιγώς «Juniper» δίκτυο. Συνεπώς η παροχή L2 MPLS VPN μεταξύ οποιουδήποτε φορέα που συνδέεται στο ΕΔΕΤ( πχ ενός ακαδημαϊκού ιδρύματος) και ενός άλλου φορέα που είναι πίσω από το Geant ( φυσικά εκτός ΕΔΕΤ) προϋποθέτει την επίλυση των προβλημάτων διαλειτουργικότητας μεταξύ των δύο τεχνικών. / The subject of this paper is the implementation techniques of L2 MPLS VPNs Cicruit Cross Connect (CCC) and Any Trasfer over MPLS (AtoM) that are provided by the two main manufactures of network connection (Juniper and Cisco). Furthermore this paper searches the special technical features of each technique and tries to find ways of interoperability between those two. This is useful when a providing L2 MPLS VPNs service expands in a network that consists of equipment of these two main manufactures . In the special occasion of EDET, which is a completely «Cisco» network the upper has implementation in the connection limits with GEANT, which is a completely «Juniper» network. Therefore the providing of L2 MPLS VPN between any carrier that is connected with EDET (eg. Academic Institute) and another carrier that is behind GEANT (of course out of EDET) presupposes the solution of the interoperability problems between these techniques.