No description available.

In recent years Peer-to-Peer Systems have gained popularity, and are best known as a convenient way of sharing content. However, even though they have existed for a considerable length of time, no method has yet been developed to measure the quality of the service they provide nor to identify cases of misbehaviour by individual peers. This thesis attempts to give to P2P systems some quality measures with the potential of giving querying peers criteria by which to judge and make predictions about the behaviour of their counterparts. The work includes the design of a reputation system from which querying peers can seek guidance before they commit to transaction with another peer. but usually as Reputation and Recommender systems have existed for years centralized services. Our innovation is the use of a distributed recommendation system which will be supported by the peers themselves. The system operates in the same manner as "word-of-mouth" in human societies does. In contrast to other reputation systems the word-of-mouth technique is itself decentralized since there is no need for central entities to exist as long as there are participants willing to be involved in the recommendation process. In order for a society to exist it is necessary that members have some way of knowing each other so that they can form relationships. The main element used to link members in an online community together is a virtual trust relationship that can be identified from the evidence that exists about their virtual partnerships. In our work we approximate the level of trust that could exist between any two parties by exploiting their similarity, constructing a network that is known as "web of trust". Using the transitivity property of trust, we make it possible for more peers to come in to contact through virtual trust relationships and thus get better results than in an ordinary system.

No description available.

No description available.

No description available.

The concept of Virtual Organisations (VOs) or Virtual Enterprises (VEs) is rapidly emerging as an important topic in many areas of computing including e-commerce, grid computing and the semantic Web. One reason for this interest is that VOs provide a means of bringing together a number of autonomous stakeholders in a dynamic fashion in order to address a specific problem or niche. These agents then work together for some period of time and then disband when it is deemed appropriate to do so. There are, however, many technical, social and economic issues associated with this VO lifecycle (i.e. creation, operation, maintenance and dissolution) that need to be addressed before VOs can be considered to be practicable. While previous technical work on VOs has concentrated on providing tools to support different aspects of the VO lifecycle, comparatively little work has focused on the mechanisms for automated VO creation, operation and maintenance. To address this shortcoming, this research aims to study and design mechanisms for the VO creation, operation and maintenance phases. In this thesis, our approach is to use combinatorial auctions and coalition formation mechanisms. In particular, novel algorithms for clearing multi-unit single-item and multi-unit combinatorial auctions have been developed as a means of tackling VO creation and one part of VO maintenance. A novel algorithm for coalition structure generation has also been developed to address VO operation and another part of VO maintenance.

Peer-to-peer (P2P) computing is primarily characterised by decentralisation, scalability, anonymity, self-organisation and ad hoc connectivity. It attracted considerable attention in open hypermedia research due to its potential for supporting collaboration among a community of people sharing similar knowledge background. The aim of this research is to investigate the feasibility and potential benefits of corporating the P2P paradigm in open hypermedia systems to support resource sharing-based collaboration. This is accomplished by utilising a distributed dynamic link service (DDLS) as a testbed, addressing issues that arise from implementing the paradigm, and demonstrating the efficiency of proposed techniques through simulation. This research begins with the development of a prototype DDLS using the open hypermedia paradigm for storing and presenting resources and a centralised P2P model which adopts a central service directory for publishing and discovering resources in a well-arranged environment. This is enhanced by an operational analysis and feature comparison between prototypes based on the traditional client-server and the centralised P2P models. Various P2P models are analysed to identify the key characteristics of and requirements for the DDLS using an unstructured P2P model which empowers collaboration in an ad hoc environment. The second phase of this research concentrates on overcoming the challenges of resource description, publishing and discovery posed by the unstructured P2P DDLS: using RDF to encode information about resources, developing a clustering technique to group resources and form the information space; and creating a semantic search mechanism to discover resources; respectively. Finally, this research proposes re-organisation techniques based on the exponential decay function and the naive estimator to enhance the performance of resource discovery in resource sharing-based collaboration.

We live in a world which is becoming increasingly rich in technology, with a wide array of portable and embedded devices being readily available and surrounding us in everyday use. Similarly, advances in communications technologies and the explosive growth of data being published on the Internet have provided access to information on an unparalleled scale. However, device interoperability is often poor at best, and accessing data which is relevant to any given situation can be difficult due to the sheer quantity of information which is available. A contextually aware environment is envisioned as one in which integrated computer systems have an understanding or representation of not only the physical space and the resources within it, but also the activities, interests, actions and intent of the human occupants at any given time. Given such knowledge, a contextually aware and technology rich pervasive environment may offer services and applications which attempt to adapt the surroundings in a manner which assists its users, such as by configuring devices or assimilating information which is relevant to activities currently being undertaken. The research presented in this thesis combines the fields of knowledge management, semantic technologies, logic and reasoning with those from the predominantly hardware and communications oriented field of pervasive computing, in order to facilitate the creation of contextually aware environments. Requirements for such a system are discussed in detail, resulting in the development of a generic framework of components and data representations from which domain specific deployments can be created. To demonstrate and test the proposed framework, experimentation has been conducted in the example domain of an academic environment, including the development of two contextually aware applications. The experiences and lessons learned during this research are documented throughout, and have influenced the proposed avenues for future related research in this area.

The current generation of networks is referred as the passive or conventional networks. Their functionality is limited to transfer packets from one point to another. Passive nodes such as routers and switches are vertically integrated devices that allow small or no modification on the operations they perform. Their functionality is limited to the network and (sometimes) transport protocol layers. They perform no computation on the payload data of the packets they handle. Due to their nature, passive networks impose limitations such as the difficulties of introducing new protocols and services, and limited performance. There is a different approach in the computer networks field, called Active Networks. Active Networks consist of routers and switches (Active Nodes) that not only forward packets from one point to another but also perform customised computations on them. The applications that execute on the Active Nodes can be user-driven. This allows end-users to program the network and tailor its services to their needs. Users could also inject their own code, which programs intermediate Active Nodes. This thesis presents an Active Router architecture using programmable hardware. The router consists of two hosts running Linux. The separation of the router functionality into two hosts is performed for safety reasons. A PCI-based board that hosts a Field Programmable Gate Array (FPGA) device comprises the programmable hardware element of the router. The motivation for using an FPGA device is that it is reprogrammable, so Active Applications can be installed on the fly, and it provides a safer execution environment. The software architecture of the Active Router comprises several modules that implement tasks such as the safe download of applications from dedicated servers, resource management, fault detection and isolation. The performance evaluation of the Active Router reveals several bottlenecks and limitations such as the PCI bus, the interrupt-driven nature of the Linux operating system.

This thesis studies concurrency control and composition of transactions in computing environments with long living transactions where local data autonomy of transactions is indispensable. This kind of computing architecture is referred to as a Disconnected System where reads are segregated -disconnected- from writes enabling local data autonomy. Disconnecting reads from writes is inspired by Bertrand Meyer's "Command Query Separation" pattern. This thesis provides a simple yet precise definition for a Disconnected System with a focus on transaction management. Concerning concurrency control, transaction management frameworks implement a 'one concurrency control mechanism fits all needs strategy'. This strategy, however, does not consider specific characteristics of data access. The thesis shows the limitations of this strategy if transaction load increases, transactions are long lived, local data autonomy is required, and serializability is aimed at isolation level. For example, in optimistic mechanisms the number of aborts suddenly increases if load increases. In pessimistic mechanisms locking causes long blocking times and is prone to deadlocks. These findings are not new and a common solution used by database vendors is to reduce the isolation. This thesis proposes the usage of a novel approach. It suggests choosing the concurrency control mechanism according to the semantics of data access of a certain data item. As a result a transaction may execute under several concurrency control mechanisms. The idea is to introduce lanes similar to a motorway where each lane is dedicated to a certain class of vehicle with the same characteristics. Whereas disconnecting reads and writes sets the traffic's direction, the semantics of data access defines the lanes. This thesis introduces four concurrency control classes capturing the semantics of data access and each of them has an associated tailored concurrency control mechanism. Class O (the optimistic class) implements a first-committer-wins strategy, class R (the reconciliation class) implements a first-n-committers-win strategy, class P (the pessimistic class) implements a first-reader-wins strategy, and class E (the escrow class) implements a first-n-readers-win strategy. In contrast to solutions that adapt the concurrency control mechanism during runtime, the idea is to classify data during the design phase of the application and adapt the classification only in certain cases at runtime. The result of the thesis is a transaction management framework called O|R|P|E. A performance study based on the TPC-C benchmark shows that O|R|P|E has a better performance and a considerably higher commit rate than other solutions. Moreover, the thesis shows that in O|R|P|E aborts are due to application specific limitations, i.e., constraint violations and not due to serialization conflicts. This is a result of considering the semantics