The client/server paradigm is a common means of implementing an application over a computer network. Servers provide services, such as access to files, directories, or web pages, and clients make use of those services. The communication between the clients and servers takes the form of a network protocol. These network protocols are often rigid and inflexible due to standardization, and because they are often implemented in the operating system kernels of the clients and servers. It is difficult to add new features to existing services without having complete control of all the clients and servers in question. Virtualization is a technique that can be used to alter the properties of a network service without requiring any modifications to the clients or servers. Virtualization is typically performed on an intermediate computer that is interposed between the clients and servers, such as a programmable router. This dissertation motivates the need for virtualization and presents several different examples of successful virtualizations. These virtualizations include translation, aggregation, replication and fortification. Virtualization is demonstrated both on commodity hardware, which has the advantage of low cost, and on a specialized network processor, which offers the advantage of high performance.

NFS is the dominant network file system used to share files between UNIX-derived operating system based hosts. At the onset of this research it was found that the tested NFS implementations did not achieve data writing throughput across a Gigabit Ethernet LAN commensurate with throughput achieved with the same hosts and network for packet streams generated without NFS. A series of tests were conducted involving variation of many system parameters directed towards identification of the bottleneck responsible for the large throughput ratio between non-NFS and NFS data transfers for high speed networks. Ultimately it was found that processor, disk, and network performance are not the source of low NFS throughput but rather it is caused by an avoidable NFS behavior, the effects of which worsen with increasing network latency.

L’intégration monolithique de matériaux semi-conducteurs III-V sur substrat de Silicium est essentielle pour le développement de la photonique sur Silicium. L’objectif est de réaliser une micro-source optique à base d’un réseau ordonné de Nanofils (NFs) III-V (InAsP/InP) placés sur un guide d’onde Si. De par leur aptitude à relaxer les contraintes, les NFs sont d’un grand intérêt. C’est dans ce contexte que s’est déroulée cette thèse axée sur la croissance autocatalysée de NFs InP sur Silicium par épitaxie directe. Nous avons ainsi montré que la croissance auto-catalysée de NFs InP denses et verticaux dépend directement de la nature de l’oxyde de surface du substrat Si. Une distribution monomodale ou bimodale de NFs ont été achevées en fonction des conditions de formation des gouttelettes d’indium ou des paramètres de croissance. Une pression critique et une température critique ont permis de délimiter des domaines favorables à la croissance. Les propriétés optiques intrinsèques des NFs ont été déterminées suffisantes pour l’objectif visé. Enfin, des résultats sur la simulation optique et la polarisation de la lumière émise dans les NFs et le guide d’onde ont permis d’établir un cahier des charges pour la croissance des NFs verticaux sur SOI pour que le couplage/partage de leurs modes optiques soit le plus efficace possible aux longueurs d’onde télécom. / Monolithic integration of III-V semiconductors materials on Si substrate is essential for the Si photonic development. We aim at achieving an optical microsource based on a regular array of III-V (InAsP/InP) nanowires (NWs) standing on top a Si waveguide. Due to their ability to be fully relaxed, nanowires growth is of deep interest. This PhD thesis has been oriented towards such context especially among self-catalyzed InP NWs growth by epitaxy. Thus we have shown that highly dense and vertical self-catalyzed InP NWs accomplishment is related to Si substrate surface oxide. A monomodal or bimodal NWs distribution have been reached through a control of indium droplets formation or growth parameters. A critical pressure and a critical temperature have been found to delimit favorable growth regime. Intrinsic optical properties have been determined to be goal sufficient. Optical simulation modeling and characterization of the polarization light state in NWs and in the Si waveguide have led us to establish functional specifications to grow vertical NWs on SOI as way that their optical modes could be coupled efficiently at telecommunications wavelength.

Thesis (M.S.)--Worcester Polytechnic Institute. / Keywords: LAN; Linux; NFS. Includes bibliographical references (p. 91-92).

Att snabbt och effektivt kunna dela filer mellan system både genom WAN och LAN är något som många företag idag är beroende av. Till detta så kan fildelningsprotokoll användas, två vanliga fildelningsprotokoll är SMB/CIFS och NFS. I detta arbete så utvärderas de nyaste versionerna av protokollen i en experimentmiljö. Prestandajämförelser sker mellan två olika operativsystem och versioner av protokollen. Syftet är att besvara frågan om vilket protokoll som är mest effektivt i WAN-miljöer där olika operativsystem används. Resultatet visar att SMB är det protokoll med bäst stöd i både Linux och Windows. NFSv4.1 kan dock med en Linuxklient uppnå högre prestanda.

Performance of the I/O subsystem plays a significant role in parallel applications that need to access large amounts of data. I/O performance in such applications is expected to be scalable and balanced with respect to the communication and CPU performance. MPIIO, a part of the MPI-2 standard has many implementations. Each of the available clientside parallel architectures differ widely in their approach to achieving high performance. This thesis hypothesizes that the effectiveness of each available client-side parallel architecture differs in delivering overall parallel application performance for a given underlying file system and that increasing the performance for different workload characteristics requires different designs. This hypothesis is validated by the development of appropriate metrics and the analysis of the results, obtained from running the experiments.

O principal objetivo deste trabalho é apresentar os aspectos de especificação e implementação dos serviços de transferência de arquivos desenvolvidos para duas redes locais, usadas em ambientes de automação industrial e de automação de escritório, respectivamente. São apresentados, inicialmente, alguns conceitos e os requisitos de especificação que devem ser satisfeitos por estes serviços. À luz desta caracterização, algumas especificações de serviços de transferência de arquivos para redes de computadores são estudadas e analisadas. De acordo com esta análise e com as tendências de padronização de redes de computadores usadas nas aplicações acima mencionadas, é escolhida uma especificação a ser tomada como base na implementação destes serviços. Estes serviços são especificados detalhadamente, considerando as características das redes locais utilizadas. Alguns aspectos de implementação da máquina de protocolo que presta os serviços especificados são discutidos, descrevendo-se o conjunto de processos que constituem tal máquina. Em seguida, são especificados os programas aplicativos desenvolvidos a partir destes serviços de transferência de arquivos, considerando-se os requisitos das aplicações envolvidas. Para avaliar o sistema implementado, alguns testes são definidos, visando verificar a operacionalidade, conformidade e desempenho deste sistema. Os resultados destes test6es são apresentados e discutidos. Finalmente, são feitas considerações sobre o sistema implementado e a especificação tomada como base. Como continuidade do presente trabalho, é sugerida a implementação de outros serviços utilizando a infra-estrutura criada pelo mesmo. / The main purpose of this work is to present aspects on the specification and the implementation of the file transfer services, developed for two local area networks used in manufacturing and office automation. Some concepts and the specification requirements for such services are initially presented. In the light of this introduction some specifications of file transfer services for computer networks are studied and analyzed. Based on this analysis and on the trends in computer networks standardization for the above mentioned applications, a specification is selected as the basis for the file transfer services. These services are then fully specified taking into account the characteristics of the target local area networks. Some implementation aspects of protocol machine which provides the specified services are discussed describes the processes that make up such machine. The application programs developed using these file transfer services are then described. Some tests are conceived the verify the functionalism, conformity and performance of the implemented system. The results of these tests are presented and discussed in some detail. Finally, some considerations on the implemented system and the specification used as basis are presented. As future work the implementation of other services using the framework created is suggested.

O principal objetivo deste trabalho é apresentar os aspectos de especificação e implementação dos serviços de transferência de arquivos desenvolvidos para duas redes locais, usadas em ambientes de automação industrial e de automação de escritório, respectivamente. São apresentados, inicialmente, alguns conceitos e os requisitos de especificação que devem ser satisfeitos por estes serviços. À luz desta caracterização, algumas especificações de serviços de transferência de arquivos para redes de computadores são estudadas e analisadas. De acordo com esta análise e com as tendências de padronização de redes de computadores usadas nas aplicações acima mencionadas, é escolhida uma especificação a ser tomada como base na implementação destes serviços. Estes serviços são especificados detalhadamente, considerando as características das redes locais utilizadas. Alguns aspectos de implementação da máquina de protocolo que presta os serviços especificados são discutidos, descrevendo-se o conjunto de processos que constituem tal máquina. Em seguida, são especificados os programas aplicativos desenvolvidos a partir destes serviços de transferência de arquivos, considerando-se os requisitos das aplicações envolvidas. Para avaliar o sistema implementado, alguns testes são definidos, visando verificar a operacionalidade, conformidade e desempenho deste sistema. Os resultados destes test6es são apresentados e discutidos. Finalmente, são feitas considerações sobre o sistema implementado e a especificação tomada como base. Como continuidade do presente trabalho, é sugerida a implementação de outros serviços utilizando a infra-estrutura criada pelo mesmo. / The main purpose of this work is to present aspects on the specification and the implementation of the file transfer services, developed for two local area networks used in manufacturing and office automation. Some concepts and the specification requirements for such services are initially presented. In the light of this introduction some specifications of file transfer services for computer networks are studied and analyzed. Based on this analysis and on the trends in computer networks standardization for the above mentioned applications, a specification is selected as the basis for the file transfer services. These services are then fully specified taking into account the characteristics of the target local area networks. Some implementation aspects of protocol machine which provides the specified services are discussed describes the processes that make up such machine. The application programs developed using these file transfer services are then described. Some tests are conceived the verify the functionalism, conformity and performance of the implemented system. The results of these tests are presented and discussed in some detail. Finally, some considerations on the implemented system and the specification used as basis are presented. As future work the implementation of other services using the framework created is suggested.

Die SFS-Benchmarksuite (Release 1.1) wurde auf mehreren Rechnerarchitekturen der TU Chemnitz-Zwickau installiert und zu einer Leistungs- bewertung dieser Architekturen genutzt. Eine speziell untersuchte Bedingung war die Hardwareausstattung und Rechnerkonfiguraion von NFS-Servern.

Le stockage de données utilise souvent des systèmes se caractérisant par une grande intrusivité : ceux-ci requièrent de nombreuses modifications logicielles, voire parfois même matérielles, pour être déployés et utilisés. Notre solution consiste à offrir un stockage distribué logiciel pour architectures de type grappes de nature faiblement intrusive dans la mesure où le protocole standard omni-présent du monde Unix, NFS, est utilisé. L'approche retenue se caractérise par une séparation de la gestion des méta-données et des données permettant ainsi de répartir la charge d'entrées/sorties et d'obtenir de meilleures performances. L'ajout de redondance permet aussi de disposer à moindre coût de stockage distribué encore plus performant et plus sûr. Le développement d'outils de transfert efficace inter-grappes et d'un système distribué de fichiers à plus grande échelle a permis de valider notre approche.