No description available.

No description available.

Denial of service (DoS) has become one of the most significant security threats in the Internet. The existing limitations of the Internet protocols and the available tools make DoS attack easy to launch and effective in causing damage. There are many different forms of DoS attack in terms of the attack types, attack architecture, involved machines, and attack volume. For example, the number of the machines involved in a DoS attack could vary from a single attacker machine to distributed, compromised and non-compromised machines forming what is known as distributed denial of service (DDoS). This thesis investigates DoS attacks and proposes several defending techniques, which cover detection, traceback, and mitigation. Firstly, the proposed detection technique, namely the Congestion Window Detection Technique (CWDT), provides a fast detection time and low false positive rate compared with existing techniques. The CWDT monitors the congestion window of the target in a defined time interval during the attack and uses the cumulative sum (CUSUM) statistic to detect the attack. Secondly, two different traceback techniques to identify the origin of the attack are proposed. The first traceback technique, namely the Selective Record Route Technique (SRRT), uses the Record ROute option of the IP header to selectively record the IP addresses of certain routers. The SRRT reduces the convergence time of the traceback significantly. The second technique, Uniform Probabilistic Packet Marking (UPPM), injects the IP address of the router inside the packet based on a predefined marking probability. UPPM can generalise the analysis of other marking techniques and improves the convergence time. Finally, a mitigation technique, Selective Windowed Rate Limiting (SWiRL), is presented to lessen the damage caused by the attack traffic. It significantly reduces the damage to legitimate traffic compared with other existing mitigation techniques.

No description available.

The combination of an exemplar artefact and associated search heuristics constitute a 'technological paradigm'. Such technological paradigms can emerge as industries evolve, altering the nature of innovative search from exploration to incremental improvement along a 'technological trajectory'. Disagreements exist as to the cause of standardisation and the relationship between standardisation and the related shift in innovation emphasis. The number of elements within a technology tends to increase over time. The 'constructional selection' model encapsulating this process partially explains standardisation at the individual level. However, present models of technological paradigm formation that .. feature constructional selection do not consider (1 ) whether artefact structure will induce an alteration of innovation emphasis and (2) how a standard design emerges at the population level and how the interaction of multiple agents affects the constructional selection process. This report focuses on the second point. The new models presented here introduce a heterogeneous population of firms subject to selection. Firms, in the presence of continual innovation, aim to produce superior products to their competitors. The models indicate that a technological paradigm emerges across the population (1) when firms use a range of strategies, (2) when design details of competitors' designs are freely accessible, for all rates of entry and associated competition above a low level, and (3) under conditions of imperfect imitation, provided that competition is sufficiently strong relative to the level of mutation.

Structure Activity Relationship (SAR) modelling capitalises on techniques developed within the computer science community, particularly in the fields of machine learning and data mining. These machine learning approaches are often developed for the optimisation of model accuracy which can come at the expense of the interpretation of the prediction. Highly predictive models should be the goal of any modeller, however, the intended users of the model and all factors relating to usage of the model should be considered. One such aspect is the clarity, understanding and explanation for the prediction. In some cases black box models which do not provide an interpretation can be disregarded regardless of their predictive accuracy. In this thesis the problem of model interpretation has been tackled in the context of models to predict toxicity of drug like molecules. Firstly a novel algorithm has been developed for the interpretation of binary classification models where the endpoint meets defined criteria: activity is caused by the presence of a feature and inactivity by the lack of an activating feature or the deactivation of all such activating features. This algorithm has been shown to provide a meaningful interpretation of the model’s cause(s) of both active and inactive predictions for two toxicological endpoints: mutagenicity and skin irritation. The algorithm shows benefits over other interpretation algorithms in its ability to not only identify the causes of activity mapped to fragments and physicochemical descriptors but also in its ability to account for combinatorial effects of the descriptors. The interpretation is presented to the user in the form of the impact of features and can be visualised as a concise summary or in a hierarchical network detailing the full elucidation of the models behaviour for a particular query compound. The interpretation output has been capitalised on and incorporated into a knowledge mining strategy. The knowledge mining is able to extract the learned structure activity relationship trends from a model such as a Random Forest, decision tree, k Nearest Neighbour or support vector machine. These trends can be presented to the user focused around the feature responsible for the assessment such as ACTIVATING or DEACTIVATING. Supporting examples are provided along with an estimation of the models predictive performance for a given SAR trend. Both the interpretation and knowledge mining has been applied to models built for the prediction of Ames mutagenicity and skin irritation. The performance of the developed models is strong and comparable to both academic and commercial predictors for these two toxicological activities.

Social tagging is considered one of the important tools used in the Web2.0 applications. It represents an economical way of supporting systems by employing users to provide tags. A number of functionalities can be provided using tagging such as retrieval, categorisation, and expression of opinion just to name a few. One of the functions that can be provided using tagging is the support of cross-lingual information retrieval (CUR) through providing tags in different languages. This approach promises a solution that may not include the complexities associated with the traditional approaches, especially automatic translation problems. One of the main problems that exist with tagging is that users are free to tag. This can result in tags that do not support the system as needed. Directing users' tagging behaviour towards supporting the system might be a challenge while maintaining users' freedom. This problem was linked to the lack of a design process that can be followed to construct tagging systems. The problem was approached by investigating ways to identify the interacting components of tagging systems. A design process was identified to construct tagging systems that directs tagging behaviour. The dynamic of the tagging system was interpreted based on two theories: organisational semiotics and activity theory. The design process was evaluated using three experiments that tested three behaviours: self organisation, categorisation for others and CUR support. Based on these studies, it was found that users can be directed into providing desired types of tagging behaviour. The design process can be extended to construct systems that are based on user collaboration, though further research needs to be conducted to verify this.

No description available.

The study of Combinatorics on words started at the beginning of the 20th century with the work of the Norwegian Mathematician Axel Thue, who published several articles in a relatively unknown journal. His work had primarily theoretical objectives, but ever since many of his results have been rediscovered independently by other researchers in relation to other problems. Although many questions have been studied and solved in the area, there are yet many open questions left to be studied. Among the basic discoveries of Thue are the existence of infinite words with no occurrence of squares (words of the form uu for a nonempty word u) on an alphabet of at least three symbols, and with no occurrence of cubes (and even overlaps) on a binary alphabet. The constraints on repetitions in infinite words have been raised to optimality after Dejean’s conjecture on the repetitive threshold associated with the alphabet size, which last cases have been proved recently by Rao after the works of Carpi, Pansiot, Moulin-Ollagnier, Mohammad-Noori and Currie, Currie and Rampersad. The first case says that the repetitive threshold of the binary alphabet is 2 (infinite binary words can avoid factor of exponent larger than 2 but cannot do more) and the second case, proved by Dejean, states that it is 7/4 for the three-letter alphabet. The constraint studied later on by Fraenkel and Simpson is somewhat orthogonal to the previous notion. Their parameter to the complexity of binary infinite words is the number of squares occurring in them without any restriction on the number of occurrences. The analysis of repetitions in strings is primarily of combinatorial interest in relation to the entropy of sequences. But repetitions or repeats are also of main concerns in the domains of text compression and of pattern matching. The knowledge of extreme situations or strongest constraints on words help analyse the behaviour of the corresponding algorithms. In this document, we provide a new proof for the Fraenkel and Simpson result, we give a proof that there exists an infinite binary word which contains finitely many squares and simultaneously avoids words of exponent larger than 7/3, which leads us to the concept introduced hereafter. A chapter is dedicated to new notion of Finite-Repetition threshold and some results about it. We give some new results on the trade-off between the number of squares and the number of maximal-exponent powers in infinite binary words. This is done in three cases where the maximal exponent is 7/3, 5/2, and 3, that is the only cases of interest. We show that there exists no infinite 3+-free binary word avoiding squares of odd period. This study also reveals there exists no infinite binary word, simultaneously avoiding cubes and squares of even period. Moreover, we proof that there exists an infinite 3+-free binary word avoiding squares of even-period length. We investigate the trade-off between the maximal period length of repetitions contained and their number. Similarly we exhibit a trade-off between number of cubes and number of squares occurring in an infinite word avoiding even-period squares. All bounds provided in these cases are shown to be optimal. Repetitions or repeats are also of main concern in the domains of text compression and of pattern matching. The knowledge of extreme situations or strongest constraints on words helps analyse the behaviour of the corresponding algorithms. In this document we mostly deal with the combinatorial aspects of the question. The algorithmic part is strongly linked and it is used to explore the words satisfying constraints on the repetitions they contain.

The remarkable advances of mobile technologies and the prevalence of mobile devices have profoundly transformed telephony systems. They provide functionalities which surpass telephony needs, and which motivate the development of value-added mobile services and functions. The number of mobile phones in use far exceeds any other technical devices that could be used to market, sell, produce, or deliver products and services to consumers. These developments open lucrative opportunities to retailers and service providers. The literature highlights the fact that mobile transactions (m-transactions) are one of the most critical incentives for successful mobile commerce (m-commerce). However, the successfulness of m-transaction systems in Saudi Arabia requires a strong acceptance of the Saudi consumers. The importance of this study is accentuated by the fact that mobile commerce and its services are still in their infancy and there is still an apparent lack of acceptance of mobile transactions amongst Saudi users. Research needs to address the issue of acceptance of m-transactions from the user’s perspective, particularly within developing countries as they suffer from a noticeable lack of studies in this field. This research initially starts with a comprehensive literature review about the critical factors affecting the acceptance and use of electronic commerce (e-commerce), m-commerce and focuses on m-transactions, including different technology acceptance models and theories, helping to investigate whether there exists an effective comprehensive framework for adopting m-transactions within the context of Saudi Arabia and, more specifically, from a consumer’s perspective. Furthermore, to emphasize the true value measure of m-transaction, we must comprehend and evaluate the potency and limitations of mobile purchasing and the key factors affecting the m-transaction use decision. M-transactions hold a huge potential for online business and sales, but merely having an m-transaction service “hosted” on the World Wide Web (WWW) should not lead us to believe that customers will rush into mobile commercial websites for their desired products. Recognising that fact and realizing that there are different important factors and concerns over m-transactions playing a significant role, highlight the need for investigating and developing a framework that encompasses the critical factors affecting the intention to use m-transaction within the context of a Saudi consumer’s perspective. In order to achieve this goal, this study evolved in several stages aiming to reach a satisfactory level of maturity. These stages can be divided into three main phases: (1) exploratory phase which contains three exploratory studies which helped to add the cultural qualities as a further dimension that would play a significant role in such a unique cultural region. Consequently, a holistic framework is integrated that includes the key factors affecting the intention to use m-transactions. This framework is empirically validated in (2) a further study using a statistically representative sample size of 1008 Saudi users from different demographic backgrounds. The empirical analysis revealed that security, ease of use, usefulness, social influence, navigational structure, telecommunication infrastructure, individualism, masculinity, power distance, uncertainty avoidance, and visual appeal have a significant impact on the intention to use m-transactions. Amongst those factors, ease of use was the most significant influential factor. Therefore, this led to (3) another study aiming to empirically investigate the level of ease of use (usability) of conducting m-transactions within the Saudi context. In total, this research comprised five different empirical studies to extend our understanding of the phenomenon of m-transactions. The ultimate product of this research is to develop a consolidated framework for the intention to use m-transactions, combined with a set of recommendations for mobile websites and application developers, designers, government, and other organizations that intend to extend their business in the mobile commerce area, and this should eventually benefit the users.