Dissertation (PhD)--University of Arizona, Tucson, Arizona, 2006.

Orientador: Shusaburo Motoyama / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Elétrica e de Computação / Made available in DSpace on 2018-08-20T23:50:39Z (GMT). No. of bitstreams: 1 Pazeto_TatianaAnnoni_D.pdf: 1711898 bytes, checksum: 8065d763289ade9949c23d195aa6fce9 (MD5) Previous issue date: 2012 / Resumo: As redes atuais, baseadas em tecnologia IP, transportam uma variedade de trafegos, tais como voz, dados e videos (trafego multimidia) e sao denominadas de redes convergentes. Outras redes estao em desenvolvimento para aplicacoes especificas como Redes de Sensores Corporais Sem Fio (RSC). O problema de prover a qualidade de servico (QoS - Quality of Service) de cada tipo de trafego em redes convergentes e em RSC, e essencial, pois os trafegos exigem diferentes requisitos de qualidade. Um dos principais parametros para prover QoS nessas redes e o escalonador de trafegos. O objetivo principal desta tese e analisar e propor escalonadores de trafego para a rede convergente e para RSC. O escalonador de trafegos proporciona uma utilizacao mais equitativa da banda disponivel. Como a rede baseada em IP foi originalmente projetada para transportar somente trafegos de dados, e estudada, nesta tese, a influencia do trafego multimidia no desempenho e no projeto do escalonador. O estudo e realizado atraves do desenvolvimento de varias plataformas de simulacoes que contem os varios tipos de trafegos, um buffer de armazenamento de pacotes, um link de saida e os varios tipos de escalonamento. Como o escalonador FIFO foi e continua sendo o mais utilizado, esse escalonador foi tomado como referencia para confrontar os resultados obtidos com o escalonador DRR e com o outro modelo inedito de escalonamento que usa em seu calculo de distribuicao de quotas, o conceito de banda efetiva. O escalonador DRR e o de banda efetiva podem fazer distincao entre os trafegos, de modo que a alocacao de bandas entre os varios trafegos se torne mais justa e atenda aos requisitos de QoS. Pelos resultados obtidos, pode-se constatar que a solucao de escalonamento proposta consegue controlar a perda de pacotes, mas o atraso deve ser melhor investigado. Alem disso, o escalonador DRR e mais indicado para trafego multimidia se comparado ao escalonador FIFO. Na maioria das propostas examinadas na literatura, o escalonador mais utilizado para RSC e aquele baseado em tecnica TDM. Existem poucos estudos em que o escalonador e baseado em servico ciclico, conhecido na literatura como polling. Nesta tese, estuda-se a conveniencia da utilizacao do escalonador baseado em polling para RSC. O estudo e, tambem, realizado atraves de uma plataforma de simulacoes que contempla fontes apropriadas desenvolvidas para sensores corporais, o escalonador polling e um buffer. As fontes desenvolvidas sao ineditas na literatura. Os resultados mostram que o escalonador polling pode ser uma boa alternativa para coletar dados dos sensores sobre ou subcutaneos implantados no corpo humano / Abstract: Today's networks are based on IP technology and carry a variety of traffic, such as voice, data and video (multimedia traffic) and are called convergent networks. Other networks are being developed for specific applications such as Wireless Body Sensor Networks (WBSN). The problem of providing Quality of Service (QoS) of each type of traffic in convergent networks and WBSN is essential, since the traffics demand different quality requirements. One of the main providers of QoS in these networks is the traffic scheduler. The main objective of this thesis is to analyze and propose traffic schedulers for convergent network and WBSN. The traffic scheduler provides a fairer use of the available bandwidth. As IP based network was originally designed to carry only data traffic, it is studied in this thesis, the influence of multimedia traffic on the performance and design of the scheduler. The study is carried out by developing multiple simulation platforms that contain various types of traffics, a buffer for packet storing, an output link and the various types of scheduling. As the FIFO scheduler was and remains the most widely used, it was taken as a reference to compare the results obtained with the DRR scheduler and with another unprecedented model of scheduling that uses in its calculation of the distribution of quotas, the concept of effective bandwidth. The DRR and the effective bandwidth schedulers can separate the traffics, so the allocation of bandwidths among the various traffics becomes fair and requirements of QoS can be met. From the results, it can be seen that the proposed scheduling solution can control packet loss, but the delay should be better investigated. Moreover, the DRR scheduler is best suited for multimedia traffic compared to the FIFO scheduler. In most of the proposals examined in the literature, the scheduler most used on WBSN is based on TDM technique. There are few studies in which the scheduler is based on cyclic service, known in the literature as polling. In this thesis, the convenience of using the scheduler based on polling for WBSN is studied. The study is also undertaken through a platform of simulations that include appropriate fonts developed for body sensors, the polling scheduler and a buffer. The sources developed are unprecedented in literature. The results show that the polling scheduler can be a good alternative to collect data from sensors implanted in subcutaneous or on the human body / Doutorado / Telecomunicações e Telemática / Doutora em Engenharia Elétrica

M.Comm. / The problem of establishing and controlling remote access to corporate networks has become one of the most difficult issues facing network administrators and information security professionals (Tipton & Krause (b), 2001:99). As the connection devices become more powerful and less expensive, more and more personally owned devices make their way to the connected edge of the network. The rise of the Internet as an accepted vehicle for business use has dramatically increased the points of entry to the corporate network. As employees and third parties gain mobility through the use of laptops, smart phones and Personal Digital Assistants (PDAs), the threat to organisational security grows (Cartwright, 2001). There are three main reasons why remote access connections are often insecure: Lack of awareness regarding security risks associated with remote access; Protecting a remote access connection is a complex task which is unlikely to be done well unless a methodical approach is taken; and Remote locations are not under the direct control of the organisation and are inherently insecure (Information Security Forum (a), 1999). The purpose of this short dissertation is to: Provide an overview of remote access, including concepts and definitions; Identify the risks associated with remote access to corporate networks; Identify possible controls to mitigate these risks; and highlight possible future trends regarding remote access.

The Rhodes Computer Music Network (RHOCMN) is a network which allows main studio resources to be shared. RHOCMN is growing into a multi-workstation environment and additional devices are being incorporated into the system. A star configuration is used for transmitting MIDI from a MIDI patch bay to the workstations and MIDI devices. This imposes several disadvantages on the use of the studio, such as wiring problems. In a quest to avoid problems related to MIDI in RHOCMN, the MIDINet system was developed. The idea was to acquire a viable solution to MIDI's main problems which does not involve a redefinition of the MIDI specification.

Grid computing technologies facilitate complex scientific collaborations between globally dispersed parties, which make use of heterogeneous technologies and computing systems. However, in recent years the commercial sector has developed a growing interest in Grid technologies. Prominent Grid researchers have predicted Grids will grow into the commercial mainstream, even though its origins were in scientific research. This is much the same way as the Internet started as a vehicle for research collaboration between universities and government institutions, and grew into a technology with large commercial applications. Grids facilitate complex trust relationships between globally dispersed business partners, research groups, and non-profit organizations. Almost any dispersed “virtual organization” willing to share computing resources can make use of Grid technologies. Grid computing facilitates the networking of shared services; the inter-connection of a potentially unlimited number of computing resources within a “Grid” is possible. Grid technologies leverage a range of open standards and technologies to provide interoperability between heterogeneous computing systems. Newer Grids build on key capabilities of Web-Service technologies to provide easy and dynamic publishing and discovery of Grid resources. Due to the inter-organisational nature of Grid systems, there is a need to provide adequate security to Grid users and to Grid resources. This research proposes a framework, using a specific brokered pattern, which addresses several common Grid security challenges, which include: Providing secure and consistent cross-site Authentication and Authorization; Single-sign on capabilities to Grid users; Abstract iii; Underlying platform and runtime security, and; Grid network communications and messaging security. These Grid security challenges can be viewed as comprising two (proposed) logical layers of a Grid. These layers are: a Common Grid Layer (higher level Grid interactions), and a Local Resource Layer (Lower level technology security concerns). This research is concerned with providing a generic and holistic security framework to secure both layers. This research makes extensive use of STRIDE - an acronym for Microsoft approach to addressing security threats - as part of a holistic Grid security framework. STRIDE and key Grid related standards, such as Open Grid Service Architecture (OGSA), Web-Service Resource Framework (WS-RF), and the Globus Toolkit are used to formulate the proposed framework.

No description available.

A method of performance improvement of token ring networks is presented, based on the use of stations with two latency states. Station latency is defined as the time delay introduced in passing data through a station. Most token ring protocol standards (e.g. IEEE 802.5 or ANSI X3T9.5) require incoming data to be decoded and encoded in the station before transmission onto the ring. These encoding and decoding operations add significantly to the station latency. The bypassing of the encoding and decoding steps is proposed, which improves the mean message waiting time. A detailed evaluation and comparison of the networks is based on both analytical and simulation results. The performance of identical stations and symmetric traffic is obtained analytically. A discrete event simulation model for a token ring network is written in GPSS for general traffic. Results show a significant reduction in mean waiting time for the dual latency ring, with performance approaching or exceeding that of gated and exhaustive service, for certain ranges of network utilization. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

A dissertation submitted to the Faculty of Engineering, University of the Witwatersrand, Johannesburg, in fulfilment of the requirements for the degree of Master of Science in Engineering / The MC2/64 system is a configureable multi-user transputer- based system which was designed using a modular approach. The MC2/64 consists of MC2 Clusters which are connected using a modified Clos network. The MC2 Clusters were designed and realised as completely configurable modules using and extending an algorithm based on Eulerian cycles through a requested graph. This dissertation discusses the configuration algorithm and the extensions made to the algorithm for the MC2 Clusters. The total MC2/64 system is not completely configurable as a MC2 Cluster releases only a limited number of links for inter-cluster connections. This dissertation analyses the configurability of MC2/64, but also presents algorithms which enhance the usability of the system from the user's point of view. The design and the implementation of the network control software are also submitted as topics in this dissertation. The network control software must allow multiple users to use the system, but without them influencing each other's transputer domains. This dissertation therefore seeks to give an overview of network control problems and the solutions implemented in current MC2/64 systems. The results of the research done for this dissertation will hopefully aid in the design of future MC2 systems which will provide South Africa with much needed, low cost, high performance computing power. / Andrew Chakane 2018

No description available.

No description available.