With the advent and development of modern information technology, such as the Internet, the difficulty in transmitting data has been reduced significantly. This makes it easier for entities to share their experience to a larger extent than before. In this thesis, we study the design and analysis of feedback mechanisms, which are the information systems that enable entities to learn information from others' experience. We provide a framework for feedback mechanisms. We first provide an abstract model of a feedback mechanism which defines the scope of our concept and identifies the necessary components of a feedback mechanism. We then provide a framework for feedback mechanisms. This provides a global and systematic view of feedback mechanisms. We also use our model and framework to decompose and analyse several existing feedback mechanisms. We propose an electronic marketplace which can be used for trading online services such as computational resources and digital storage. This marketplace incorporates a dispute prevention and resolution mechanism that is explicitly designed to encourage the good conduct of marketplace users, as well as providing important security features and being cost-effective. We also show how to incorporate the marketplace into Grid computing for exchanging computational resources. We propose a novel feedback mechanism for electronic marketplaces. In this setting, the role of feedback is no longer a “shadow of the future”, but a “shadow of the present”. In other words, feedback directly impacts on the seller's payoff for the current transaction instead of future transactions. This changes the fundamental functionality of feedback, which solves many inherent problems of reputation systems that are commonly applied in electronic marketplaces. We provide a novel announcement scheme for vehicular ad-hoc networks (VANETs) based on a reputation system in order to evaluate message reliability. This scheme features robustness against adversaries, efficiency and fault tolerance to temporary unavailability of the central server.

Drive-by download is a sort of network attack which uses different techniques to plant malicious codes in their computers. It makes the traditional intrusion detection systems and firewalls nonfunctional in the reason that those devices could not detect web-based threats. The Crawler-based approach has been proposed by many studies to discover drive-by download sites. However, the Crawler-based approach could not simulate the real user behavior of web browsing when drive-by download attack happens. Therefore, this study proposes a new approach to detect drive-by download by sniffing HTTP flow. This study uses reputation system to improve the efficiency of client honeypots, and adjusts client honeypots to process the raw data of HTTP flow. In the experiment conducted in real network environment, this study show the performance of a single client honeypot could reach average 560,000 HTTP success access log per day. Even in the peak traffic, this mechanism reduced the process time to 22 hours, and detected drive-by download sites that users were actually browsing. Reputation system in this study is applicable to varieties of domain names because it does not refer to online WHOIS database. It established classification model on machine learning in 12 features. The correct classification rate of the reputation system applied in this study is 90.9%. Compared with other Reputation System studies, this study not only extract features from DNS A-Type but also extract features from DNS NS-Type. The experiment results show the Error Rate of the new features from DNS NS-Type is only 19.03%.

Reputation systems have the potential of improving the quality of on-line markets by identifying fraudulent users and subsequently dealing with these users can be prevented. The behaviour of participants involved in e-commerce can be recorded and then this information made available to potential transaction partners to make decisions to choose a suitable counterpart. Unfortunately current reputation systems suffer from various vulnerabilities. Solutions for many of these problems will be discussed. One of the major threats is that of unfair feedback. A large number of negative or positive feedbacks could be submitted to a particular user with the aim to either downgrade or upgrade the user's reputation. As a result the produced reputation does not reflect the user's true trustworthiness. To overcome this threat a variation of Bayesian Reputation system is proposed. The proposed scheme is based on the subjective logic framework proposed Josang et al. [65]. The impact of unfair feedback is countered through some systematic approaches proposed in the scheme. Lack of anonymity for participants leads to reluctance to provide negative feedback. A novel solution for anonymity of feedback providers is proposed to allow participants to provide negative feedback when appropriate without fear of retaliation. The solution is based on several primitive cryptographic mechanisms; e-cash, designated verifier proof and knowledge proof. In some settings it is desirable for the reputation owner to control the distribution of its own reputation and to disclose this at its discretion to the intended parties. To realize this, a solution based on a certificate mechanism is proposed. This solution allows the reputation owner to keep the certificate and to distribute its reputation while not being able to alter that information without detection. The proposed solutions cater for two modes of reputation systems: centralised and decentralised. The provision of an off-line reputation system is discussed by proposing a new solution using certificates. This is achieved through the delegation concept and a variant of digital signature schemes known as proxy signatures. The thesis presents a security architecture of reputation systems which consists of different elements to safeguard reputation systems from malicious activities. Elements incorporated into this architecture include privacy, verifiability and availability. The architecture also introduces Bayesian approach to counter security threat posed by reputation systems. This means the proposed security architecture in the thesis is a combination of two prominent approaches, namely, Bayesian and cryptographic, to provide security for reputation systems. The proposed security architecture can be used as a basic framework for further development in identifying and incorporating required elements so that a total security solution for reputation systems can be achieved.

Privacy-invasive software is an increasingly common problem for today's computer users, one to which there is no absolute cure. Most of the privacy-invasive software are positioned in a legal gray zone, as the user accepts the malicious behaviour when agreeing to the End User License Agreement. This thesis proposes the use of a specialized reputation system to gather and share information regarding software behaviour between community users. A client application helps guide the user at the point of executing software on the local computer, displaying other users' feedback about the expected behaviour of the software. We discuss important aspects to consider when constructing such a system, and propose possible solutions. Based on the observations made, we implemented a client/server based proof-of-concept tool, which helped us discover other issues such as the effect on system stability. We also compare this solution to other, more conventional, protection methods such as anti-virus and anti-spyware software.

Content distribution is the primary function of the Internet today. Technologies like multicast and peer-to-peer networks hold the potential to serve content to large populations in a scalable manner. While multicast provides an efficient transport mechanism for one-to-many and many-to-many delivery of data in an Internet environment, the peer-to-peer networks allow scalable content location and retrieval among large groups of users in the Internet. Incorporating quality-consciousness in these technologies is necessary to enhance the overall experience of clients. This dissertation focuses on the architectures and mechanisms to enhance multicast and peer-to-peer content distribution through quality-consciousness. In particular, the following aspects of quality-consciousness are addressed: 1) client latency, 2) service differentiation, and 3) content quality. Data analysis shows that the existing multicast scheduling algorithms behave unfairly when the access conditions for the popular files changes. They favor the popular files while penalizing the files whose access conditions have not changed. To maintain the client latency for all files under dynamic access conditions we develop a novel multicast scheduling algorithm that requires no change in server provisioning. Service differentiation is a desirable functionality for both multicast and peer-to-peer networks. For multicast, we design a scalable and low overhead service differentiation architecture. For peer-to-peer networks, we focus on a protocol to provide different levels of service to peers based on their contributions in the system. The ability to associate reliable reputations with peers in a peer-to-peer network is a useful feature of these networks. Reliable reputations can help establish trust in these networks and hence improve content quality. They can also be used as a substrate for a service differentiation scheme for these networks. This dissertation develops two methods of tracking peer reputations with varying degrees of reliability and overheads.

The number of internet users is growing by the second and with an increase of internet users, there is also an increase of User-generated content. Web application that bases their buisness on User-generated content is often called by the name Collaborative Web Applications. There is a challenge for those types of applications to maintain high quality of content, since the amount of content that flows in is massive. Previous studies in the area propose that this challenge might be answered by a Web Reputation System. Previous studies has also proposed to divide Web Reputation System into various simple models. One of these models is karma, which is the focus in this report. Karma as a model aims to reward users who contribute content that is considered by other users as rewarding or entertaining. Through literature studies, guidelines on karmabased systems could not be found, which suggets that its has not previously been investigated. This made it a relevant problem for this report. The work has been carried out as a case study and used documents and a questionnaire to answer the research questions. Through data acquisition from both users of existing karma-based systems and a comparison of these existing system, it has resulted in a number of guidlines which are presented in this report. / Antalet internetanvändare ökar hela tiden och med en ökning av användare, tillkommer även en ökning av användargenererat innehåll. Webbapplikationer som baserar sig på användargenererat innehåll benämns som en ’Collaborative Web Application’ och det är en utmaning för dessa att bibehålla bra informationskvalité med tanke på mängden information som strömmar in. Tidigare studier inom området föreslår att ett ’Web Reputation System’ kan vara lösningen för att hantera detta. Tidigare studier har även valt att dela upp ’Web Reputation System’ i olika enkla modeller, där fokuset i denna studie har varit på karma. Karma syftar till att belöna användare som bidrar med innehåll som anses av användare som givande. Via litteraturstudier så kunde inte riktlinjer gällande karmabaserade system hittas, vilket föreslår att det inte tidigare har efterforskats. Detta gjorde att det blir ett relevant problem för detta arbete. Arbetet har genomförts som en fallstudie och använt sig av dokument och en enkätundersökning för att svara på frågeställningarna. Genom datainsamling både från användare av befintliga karma-baserade system och en jämförelse av befintliga karma-baserade system, så har det resulterat i ett antal riktlinjer som presenteras i detta arbete.

Online auction, as one way of the online shopping has become increasingly popular. During the process of online auction, there are many factors influencing the the final auction price. This dissertation will focus on the factors influencing the online auction price. Thus, the sellers and buyers in online auction process will be much more clear how to make decision in future auction process.

Trust plays an important role in our daily life, both implicitly and explicitly. Our decisions are based on our estimation of how trustworthy a person is or how reliable a service is. Consequently, there has been a rise in trust systems that model human trust in a virtual or computing environment. These trust systems or trust models help to bridge the gap of human feelings and intuition in an unfamiliar environment. Trust models collect information regarding the participants' activities and give a trust rating based on observed activities. In a network environment, a plethora of network devices are in constant communication as data packets are transported from source to destination. The autonomous nature of network environments and devices make it difficult to monitor the services and devices from a central point. Security mechanisms, such as IPSec, exist in routing protocols to safeguard network packets travelling in a network, however routing devices that act as service providers are not protected by malicious attacks. For example, an attack aimed at the routing architecture of a network involves a routing device advertising itself as another routing device in order to divert network traffic away from its intended destination. This dissertation investigates trust models in network environments as a possible approach to predict and ultimately eliminate at- tacks on routing devices. To accomplish this, the role of routing devices as service providers and requesters must be stated explicitly. Activities on a routing device must be collected and used to determine the trust level of the routing device. This dissertation presents the TSONE - Trust in Service-Oriented Net- work Environment - model. The model incorporates traditional service- oriented architecture (SOA) principles to define a service-oriented network environment. Services in this environment are then defined. Furthermore the characteristics of this environment are adapted from SOA principles. An approach is defined to collect and measure activities on routing devices. This is later used to determine the trust level of the routing device. Finally, a pro- to type illustrates that incorporation of trust models is a possible option in assessing availability and reliability of routing devices. / Dissertation (MSc)--University of Pretoria, 2011. / Computer Science / unrestricted

This work describes security mechanisms of wireless network based on 802.11 standard and security enhancement 802.11i of these networks known as WPA2, where the analysis of vulnerabilities and attacks on these networks were performed. The work discusses two major security issues. The first is unsecure management frames responsible for vulnerability with direct impact on availability and the other is the vulnerability that allows executing the impersonalize type of attacks. The system for generation attacks was designed to realize any attack very fast and efficient. The core of the thesis is the design of a system for attack analysis using the principle of trust and reputation computation. The conclusion of the work is devoted to experimenting with the proposed system, especially with the selection of suitable metrics for calculating the trust value.

京都大学 / 0048 / 新制・課程博士 / 博士(情報学) / 甲第20029号 / 情博第624号 / 新制||情||108(附属図書館) / 33125 / 京都大学大学院情報学研究科社会情報学専攻 / (主査)教授 石田 亨, 教授 矢守 克也, 教授 喜多 一 / 学位規則第4条第1項該当 / Doctor of Informatics / Kyoto University / DFAM