Thesis (M.S.)--Baylor University, 2006. / Includes bibliographical references (p. 50-52).

This dissertation is concerned with the management and performance issues of those register-insertion (R-I) ring type local area networks (LANs) which employ the message removal by destination node protocol. A typical example of such a network is the Distributed Loop Computer Network (DLCN). developed by Liu et al for fundamental research in the field of distributed computing. After considering the management issues of low cost R-J LANs. the research deals with performance analysis of the dynamically reconfigurable register-insertion (DRR) network. In the first part of this research. a set of management functions is identified which are desirable and can be provided economically by a low cost LAN. A unique feature of the DLCN is that messages are removed from the network by the destination node. that is. messages do not travel whole of the loop. Therefore. it is not possible for a special control node to monitor the data traffic on the network without providing support functionality in each network access unit (NAU). The minimum functionality which must be provided in each NAU is identified in the thesis. A skeleton network was implemented to verify the feasibility of the proposed scheme. A paper describing the findings of this research was published and is reproduced as appendix A. In the second part of this research. a new feature of the DLCN network is introduced. namely. that the performance of a network employing removal by destination protocol can be improved by reconfiguring the network in a particular way. A methodology to find the optimal configuration is developed and is shown. by worked examples. to lead to improved performance. The findings of this research are particularly applicable to the dynamically reconfigurable register-insertion (DRR) network. A paper dealing with the optimisation of a hypothetical fully connected DRR network has been accepted for publication. Another paper. which considers the general case of less than fully connected DRR networks. is to be published. Both papers are reproduced as append ices Band C. Finally. a performance study of the ORR network is undertaken. As there seems to be no published attempt at formal analysis or simulation of a ORR network. a survey of literature dealing with performance study of the basic OLeN is performed. A simulation model of the DRR was then developed and implemented to verify the results arrived at in the previous section. Later. a queueing model of the DRR network. based on the work of Bux and Schlatter [7] is developed and analysed. Both simulation and analysis support the claim that the performance of a DRR network can be improved by adopting the configuration strategy developed in this thesis.

published_or_final_version / abstract / Electrical and Electronic Engineering / Doctoral / Doctor of Philosophy

Network applications and Internet services fail all too frequently. However, end users cannot effectively identify the root cause using traditional troubleshooting techniques due to the limited capability to distinguish failures caused by local network elements from failures caused by elements located outside the local area network. To overcome these limitations, we propose a new approach, one that leverages collaboration of user machines to assist end users in diagnosing various failures related to Internet connectivity and poor network performance. First, we present DYSWIS ("Do You See What I See?"), an automatic network fault detection and diagnosis system for end users. DYSWIS identifies the root cause(s) of network faults using diagnostic rules that consider diverse information from multiple nodes. In addition, the DYSWIS rule system is specially designed to support crowdsourced and distributed probes. We also describe the architecture of DYSWIS and compare its performance with other tools. Finally, we demonstrate that the system successfully detects and diagnoses network failures which are difficult to diagnose using a single-user probe. Failures in lower layers of the protocol stack also have the potential to disrupt Internet access; for example, slow Internet connectivity is often caused by poor Wi-Fi performance. Channel contention and non-Wi-Fi interference are the primary reasons for this performance degradation. We investigate the characteristics of non-Wi-Fi interference that can severely degrade Wi-Fi performance and present WiSlow ("Why is my Wi-Fi slow?"), a software tool that diagnoses the root causes of poor Wi-Fi performance. WiSlow employs user-level network probes and leverages peer collaboration to identify the physical location of these causes. The software includes two principal methods: packet loss analysis and 802.11 ACK number analysis. When the issue is located near Wi-Fi devices, the accuracy of WiSlow exceeds 90%. Finally, we expand our collaborative approach to the Internet of Things (IoT) and propose a platform for network-troubleshooting on home devices. This platform takes advantage of built-in technology common to modern devices --- multiple communication interfaces. For example, when a home device has a problem with an interface it sends a probe request to other devices using an alternative interface. The system then exploits cooperation of both internal devices and remote machines. We show that this approach is useful in home networks by demonstrating an application that contains actual diagnostic algorithms.

by Wong Leung-Chung Chris. / Thesis (M.Phil.)--Chinese University of Hong Kong, 1997. / Includes bibliographical references (leaves 86-[92]). / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- Routing Overview --- p.2 / Chapter 1.2 --- Routing in the Internet --- p.5 / Chapter 1.2.1 --- Inter-Domain Routing --- p.6 / Chapter 1.2.2 --- Intra-Domain Routing --- p.7 / Chapter 1.2.3 --- The Future Trend --- p.7 / Chapter 2 --- Inter-Domain Routing --- p.9 / Chapter 2.1 --- Inter-Domain Routing Protocols --- p.9 / Chapter 2.1.1 --- Exterior Gateway Protocol (EGP) --- p.10 / Chapter 2.1.2 --- Border Gateway Protocol (BGP) --- p.11 / Chapter 2.1.3 --- Inter-Domain Policy Routing (IDPR) --- p.12 / Chapter 2.1.4 --- Other Protocols --- p.13 / Chapter 2.2 --- The Need for Pricing on Inter-Domain Routing Protocols --- p.13 / Chapter 2.3 --- Pricing Scheme on the Inter-Domain level --- p.15 / Chapter 2.4 --- Routing Protocols to Support Pricing on the Internet --- p.16 / Chapter 2.4.1 --- Routing Towards Multiple-Additive Metrics --- p.16 / Chapter 2.4.2 --- "Network Model, Notations and Assumptions" --- p.16 / Chapter 2.4.3 --- The Problem Statement --- p.18 / Chapter 3 --- Application of Neural Nets in Route Selection --- p.20 / Chapter 3.1 --- Neural Network (NN) Overview --- p.20 / Chapter 3.1.1 --- Brief History on Neural Network Research --- p.20 / Chapter 3.1.2 --- Definition of Neural Network --- p.21 / Chapter 3.1.3 --- Neural Network Architectures --- p.22 / Chapter 3.1.4 --- Transfer Fucntion of a Neuron --- p.24 / Chapter 3.1.5 --- Learning Methods --- p.25 / Chapter 3.1.6 --- Applications in Telecommunications --- p.26 / Chapter 3.2 --- Review on the Applications of Neural Networks in Packet Routing --- p.27 / Chapter 3.2.1 --- The JEB Branch --- p.27 / Chapter 3.2.2 --- The Hopfield/Energy Minimization Branch (HEM) --- p.29 / Chapter 3.2.3 --- Supervised Learning (SL) --- p.34 / Chapter 3.3 --- Discussions --- p.35 / Chapter 4 --- Route Selection as “Link-state´ح Classification --- p.36 / Chapter 4.1 --- Multi-Layer Feedforward Network (MLFN) --- p.37 / Chapter 4.1.1 --- Function Approximation Power of MLFN --- p.38 / Chapter 4.1.2 --- Choosing MLFN parameters..........´ب --- p.40 / Chapter 4.1.3 --- Trailing a MLFN --- p.41 / Chapter 4.2 --- The Utility Function --- p.43 / Chapter 4.3 --- The Neural Network Architecture --- p.46 / Chapter 4.3.1 --- Routing Graph Representation with Successor Sequence Table (SST) --- p.46 / Chapter 4.3.2 --- The Neural Network Layout --- p.52 / Chapter 4.3.3 --- How the Neural Network Controller Works --- p.55 / Chapter 4.3.4 --- Training --- p.56 / Chapter 4.4 --- Simulation --- p.56 / Chapter 4.4.1 --- Performance Parameters --- p.56 / Chapter 4.4.2 --- Simulation Results --- p.57 / Chapter 4.5 --- Conclusions and Discussions --- p.70 / Chapter 5 --- Route Selection as Energy Minimization - A Theoretical Study --- p.73 / Chapter 5.1 --- The Hopfield/Tank NN Model --- p.73 / Chapter 5.2 --- Boltzman's Machine --- p.76 / Chapter 5.3 --- Boltzman's Machine Model for Multiple-Metrices Routing --- p.79 / Chapter 5.4 --- Conclusions --- p.82 / Chapter 6 --- Conclusions --- p.84 / Bibliography --- p.86

Thesis (Ph. D.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

Thesis (M.S.)--State University of New York at Binghamton, Dept. of Electrical & Computer Engineering, 2006. / Includes bibliographical references.

As the role of the Internet has been steadily gaining in importance, overlays are increasingly being used to provide new services and to deploy older ones. Some of the services for which overlays have been proposed include multicast, quality of service (QoS), search, and resilient networks. The use of overlays, in turn, has led to more interest in improving their performance. The performance of an overlay network depends significantly on how the network is structured, i.e., the placement of the nodes in the underlying network topology, the links between the overlay nodes and the access links of these nodes. This thesis focuses on algorithms for improving the performance of

Thesis (M.S. in Computer Science)--Naval Postgraduate School, March 2004. / Thesis advisor(s): Wen Su, John Gibson. Includes bibliographical references (p. 189). Also available online.

Thesis (Ph.D)--University of Wollongong, 2005. / Typescript. Includes bibliographical references: leaf 172-181.