The download duration of peer-to-peer overlay networks is highly dependent upon the client's selection of candidate node-servers and the algorithms used in that process. Recent findings suggest that as node-server network capacity increases the deviation from the average total download time can vary as much as 300 percent between selection algorithms. This work investigated the current selection algorithms based upon chunk size, parallel connections, permanent connection, and time based switching. The time based switching algorithm is a variation of the chunk based algorithm. Time based switching enables a client to randomly select a new node-server regardless of connection speed at predetermined time slots. Simulations indicate a 41% percent decrease in download time when compared to chunk based switching. The effects of inserting chokepoints in the time based switching algorithm were investigated. This work investigated improving a client's download performance by preemptively releasing a client from a poor performing node-server. To achieve this, the client will gather a peer-to-peer network overlay capacity from a global catalog. This information will be used to seed a client choke algorithm. Clients will then be able to continually update a local capacity average based upon past download sessions. This local average will be used to make a comparison between the current download session and the previously calculated average. A margin has been introduced to allow the client to vary from the average calculated capacity. The client will perform comparisons against chokepoints and make performance decisions to depart a node-server that does not meet minimum capacity standards. Experimental results in this research demonstrated the effectiveness of applying a choking algorithm to improve upon client download duration as well as increasing the accuracy of download duration estimates. In the single downloader scenario, the choke based algorithm improved performance up to 44% in extreme congestion and a more modest 13% under normal conditions. The multiple client scenarios yielded on average a 1% decrease in client download duration along with a 44% increase download homogeneity. Furthermore, the results indicate that a client based choking algorithm can decrease overall peer-to-peer network congestion buy improving upon client selection of node-servers.

As grades computacionais e as redes peer-to-peer (P2P) surgiram como áreas distintas, com diferentes propósitos, modelos e ferramentas. No decorrer dos últimos anos, estas áreas foram convergindo, uma vez que a infraestrutura e o modelo de execução descentralizada das redes P2P provaram ser uma alternativa adequada para o tratamento de questões relacionadas à manutenção de grades de larga escala, tais como escalabilidade, descoberta, alocação e monitoramento de recursos. O modelo MultiCluster trata a convergência entre grades computacionais e redes peer-to-peer de uma forma mais restrita: os problemas de escalabilidade, de descoberta e alocação de recursos são minimizados considerando-se apenas recursos localmente disponíveis para a construção de uma grade, a qual pode ser usada para a execução de aplicações com diferentes características de acoplamento e comunicação. Esse trabalho apresenta a arquitetura do modelo e seus aspectos funcionais, bem como um primeira implementação do modelo, realizada através da adaptação da biblioteca de programação DECK sobre os protocolos do projeto JXTA. A avaliação do funcionamento dessa implementação é apresentada e discutida, com base em algumas aplicações com diferentes características. / Grid computing and peer-to-peer computing emerged as distinct areas with different purposes, models and tools. Over the last years, these areas has been converging since the infrastructure and the execution model used in peer-to-peer networks have proven to be a suitable way to treat some problems related to the maintenance of large scale grids, such as scalability, monitoring, and resource discovery and allocation. The MultiCluster model addresses the convergence of grids and peer-to-peer networks in a more restricted way: the problems related to scalability, resource allocation and discovery are minimized by considering only local resources for the conception of a small scale grid, which can be used to run applications with different characteristics of granularity and communication. This work presents the MultiCluster architecture and its functional aspects, as well as a first implementation carried out by adapting the DECK programming library to use JXTA protocols and its consequent evaluation, based on applications with different characteristics.

With the extremely rapid development of the Internet in recent years, emerging peer-to-peer network overlays are meeting the requirements of a more sophisticated communications environment, providing a useful substrate for applications such as scalable file sharing, data storage, large-scale multicast, web-cache, and publish-subscribe services. Due to its design flexibility, peer-to-peer networks can offer features including self-organization, fault-tolerance, scalability, load-balancing, locality and anonymity. As the Internet grows, there is an urgent requirement to understand real-time network performance degradation. Measurement tools currently used are ping, traceroute and variations of these. SNMP (Simple Network Management Protocol) is also used by network administrators to monitor local networks. However, ping and traceroute can only be used temporarily, SNMP can only be deployed at certain points in networks and these tools are incapable of sharing network measurements among end-users. Due to the distributed nature of networking performance data, peer-to-peer overlay networks present an attractive platform to distribute this information among Internet users. This thesis aims at investigating the desirable locality property of peer-to-peer overlays to create an application to share Internet measurement performance. When measurement data are distributed amongst users, it needs to be localized in the network allowing users to retrieve it when external Internet links fail. Thus, network locality and robustness are the most desirable properties. Although some unstructured overlays also integrate locality in design, they fail to reach rarely located data items. Consequently, structured overlays are chosen because they can locate a rare data item deterministically and they can perform well during network failures. In structured peer-to-peer overlays, Tapestry, Pastry and Chord with proximity neighbour selection, were studied due to their explicit notion of locality. To differentiate the level of locality and resiliency in these protocols, P2Psim simulations were performed. The results show that Tapestry is the more suitable peer-to-peer substrate to build such an application due to its superior localizing data performance. Furthermore, due to the routing similarity between Tapestry and Pastry, an implementation that shares network measurement information was developed on freepastry, verifying the application feasibility. This project also contributes to the extension of P2Psim to integrate with GT-ITM and link failures.

<p>Concepts like peer-to-peer networks and distributed computing are not new. They have been available in different forms for a long time. This thesis examines the possibility of merging these concepts. The assumption is that non-centralized peer-to-peer networks can be used for effective sharing of computing resources. While most peer-to-peer systems today concentrate on sharing of data in various forms, this thesis concentrates on sharing of clock cycles instead of files.</p>

<p>Advances in peer-to-peer overlay networks and Semantic Web technology will have a substantial influence on the design and implementation of future digital libraries. However, it remains unclear how best to combine their advantages in constructing digital library systems. This thesis is devoted for investigating, proposing and evaluating possible solutions to advance developments in this field.</p><p>The main research goal of this work is to combine the strengths of both peer-to-peer overlay networks and Semantic Web for facilitating semantic searches in large-scale distributed digital library systems. The approach has been conducted in a sequential and progressive manner. Firstly, we recognize system infrastructure and metadata heterogeneity as two major challenges in conducting semantic searching across distributed digital libraries. Next, we investigate the strengths and weaknesses of both peer-to-peer and Semantic Web technology and justify that these two fields are complementary and can be combined in conducting semantic searches in a large-scale distributed environment. Thirdly, due to various topologies, functionalities and limitations different peer-to-peer infrastructures may possess, we survey current classical peer-to-peer systems so as to facilitate determinating appropriate infrastructure for specific application scenario. Fourthly, we probe into approaches in generating ontology-enriched metadata records for semantic search purpose. Finally and most importantly, we will propose a semantic search process for interoperation among heterogeneous resources, basing on ontology mapping mechanism.</p><p>A major contribution expected in our work is, in a broader term, proposing and investigating possible solutions in combining the strengths of both peer-to-peer overlay networks and Semantic Web for facilitating semantic search among highly distributed digital libraries. From a specific perspective, we provide an appropriate benchmark for facilitating decision making in choosing appropriate peer-to-peer networks for digital library construction; especially, we consider in this work no global schema exists and further justify the feasibility and advantages of ontology engineering method in semantic enriched metadata management; to support federated search in such a distributed environment, we also propose an extended super-peer network model, emphasizing in load-balancing and self-organizing capabilities; Based on semantic enriched metadata management, we propose also direct ontology mapping method to enable runtime semantic search process. Evaluation results have illustrated the feasibility and robustness of our approaches.</p><p>The future direction of this work includes studies on user authentication,efficient ontology parsing and real-life applications.</p>

Advances in peer-to-peer overlay networks and Semantic Web technology will have a substantial influence on the design and implementation of future digital libraries. However, it remains unclear how best to combine their advantages in constructing digital library systems. This thesis is devoted for investigating, proposing and evaluating possible solutions to advance developments in this field. The main research goal of this work is to combine the strengths of both peer-to-peer overlay networks and Semantic Web for facilitating semantic searches in large-scale distributed digital library systems. The approach has been conducted in a sequential and progressive manner. Firstly, we recognize system infrastructure and metadata heterogeneity as two major challenges in conducting semantic searching across distributed digital libraries. Next, we investigate the strengths and weaknesses of both peer-to-peer and Semantic Web technology and justify that these two fields are complementary and can be combined in conducting semantic searches in a large-scale distributed environment. Thirdly, due to various topologies, functionalities and limitations different peer-to-peer infrastructures may possess, we survey current classical peer-to-peer systems so as to facilitate determinating appropriate infrastructure for specific application scenario. Fourthly, we probe into approaches in generating ontology-enriched metadata records for semantic search purpose. Finally and most importantly, we will propose a semantic search process for interoperation among heterogeneous resources, basing on ontology mapping mechanism. A major contribution expected in our work is, in a broader term, proposing and investigating possible solutions in combining the strengths of both peer-to-peer overlay networks and Semantic Web for facilitating semantic search among highly distributed digital libraries. From a specific perspective, we provide an appropriate benchmark for facilitating decision making in choosing appropriate peer-to-peer networks for digital library construction; especially, we consider in this work no global schema exists and further justify the feasibility and advantages of ontology engineering method in semantic enriched metadata management; to support federated search in such a distributed environment, we also propose an extended super-peer network model, emphasizing in load-balancing and self-organizing capabilities; Based on semantic enriched metadata management, we propose also direct ontology mapping method to enable runtime semantic search process. Evaluation results have illustrated the feasibility and robustness of our approaches. The future direction of this work includes studies on user authentication,efficient ontology parsing and real-life applications.

Peer-to-peer (P2P) streaming has recently received much research attention, with successful commercial systems showing its viability in the Internet. Despite the remarkable popularity in real-world systems, the fundamental properties and limitations are not yet well understood from a theoretical perspective, as there exists a significant gap between the fundamental limits and the performance achieved in practice. In this thesis, we seek to provide an in-depth analytical understanding of fundamental properties and limitations of P2P streaming systems, with a particular spotlight on the performance gap. We first identify the major problem in existing streaming protocols and show that this problem accounts for most of the gap separating the actual and optimal performances of the streaming systems. We then propose a remedy based on network coding to address this problem and show that the gap to the fundamental limits can be significantly reduced.

The growth of wireless telecommunications has stipulated the interest for anywhere-anytime computing. The synergy between networking and mobility will engender new collaborative applications with mobile devices on heterogeneous platforms. One such middleware is “SYSTEM ON MOBILE DEVICES”, SYD developed by the Yamacraw Embedded Systems research team. This type of middleware is an opening step towards Peer-to-Peer mobile networks. This project envisioned collaborative applications among mobile devices and PDAs were used as servers. This thesis studies various existing architectures in mobile computing and their scalability issues. We also proposed new scalable flexible thick client proxy system FTCPS, an architecture suitable for mobile Peer-to-Peer networks. Our empirical study showed that FTCPS has low response time compared to other architectures.

Peer-to-peer (P2P) streaming has recently received much research attention, with successful commercial systems showing its viability in the Internet. Despite the remarkable popularity in real-world systems, the fundamental properties and limitations are not yet well understood from a theoretical perspective, as there exists a significant gap between the fundamental limits and the performance achieved in practice. In this thesis, we seek to provide an in-depth analytical understanding of fundamental properties and limitations of P2P streaming systems, with a particular spotlight on the performance gap. We first identify the major problem in existing streaming protocols and show that this problem accounts for most of the gap separating the actual and optimal performances of the streaming systems. We then propose a remedy based on network coding to address this problem and show that the gap to the fundamental limits can be significantly reduced.

No description available.