Computer systems are subject to a range of attacks that can compromise their intended operations. Conventional wisdom states that once a system has been compromised, the only way to recover is to format and reinstall. In this work, we present methods to automatically recover or self-heal from a compromise. We term the system an intrusion recovery system. The design consists of a layered architecture in which the production system and intrusion recovery system run in separate isolated virtual machines. The intrusion recovery system monitors the integrity of the production system and repairs state if a compromise is detected. A method is introduced to track the dynamic control flow graph of the production system guest kernel. A prototype of the system was built and tested against a suite of rootkit attacks. The system was able to recover from all attacks at a cost of about a 30% performance penalty.

Thesis (Ph. D.)--Oregon State University, 2010. / Printout. Includes bibliographical references (leaves 223-236). Also available on the World Wide Web.

As technology becomes ever more pervasive, the challenges of home automation are increasingly apparent. Seamless home control, home monitoring and home programming by the end user have yet to enter the mainstream. This could be attributed to the challenge of developing a fully autonomous and extensible home system that can support devices and technologies of differing protocols and functionalities. In order to offer programming facilities to the user, the underlying rule system must be fully independent, allowing support for current and future devices. Additional challenges arise from the need to detect and handle conflicts that may arise among user rules and yield undesirable results. Non-technical individuals typically struggle when faced with a programming task. It is therefore vital to encourage and ease the process of programming the home. This thesis presents Homer, a home system that has been developed to support three key features of a home system: control, monitoring and programming. Homer supports any third-party hardware or software service that can expose its functionality through Java and conform to the Homer interface. Stand-alone end user interfaces can be written by developers to offer any of Homer's functionality. Where policies (i.e. rules) for the home are concerned, Homer offers a fully independent policy system. The thesis presents a custom policy language, Homeric, that has been designed specifically for writing home rules. The Homer policy system detects overlaps and conflicts among rules using constraint satisfaction and the effect on environment variables. The thesis also introduces the notion of perspectives to ease user interactivity. These have been integrated into Homer to accommodate the range of ways in which a user may think about different aspects and features of their home. These perspectives include location, device type, time and people-oriented points of view. Design guidelines are also discussed to aid end user programming of the home. The work presented in this thesis demonstrates a system that supports control, monitoring and programming of the home. Developers can quickly and easily add functionality to the home through components. Conflicts can be detected amongst rules within the home. Finally, design guidelines and a prototype interface have been developed to allow both technically minded and non-technical people to program their home.

End users must regularly agree to lengthy software agreements prior to installing software or using software services. However, despite the fact that these agreements contain terms of direct concern to users—e.g., data collection policies—software agreements are typically read by less than 2% of the population [30]. This thesis presents techniques and heuristics for improving the presentation and visual design of software agreements, to better capture reader attention and improve comprehension. In contrast to other techniques, these techniques are applied to the full agreement content, rather than a summary, as summaries have been found to distract readers from the full content of the agreement [44,56]. This thesis introduces two techniques for improving software agreements: narrative pictograms and textured agreements. Narrative pictograms are a pictorial technique designed to improve the communication of agreement terms to non-native readers of the language of an agreement. An experimental study shows that they can successfully communicate the basic concepts of a data collection policy without words. Textured agreements are visually redesigned software agreements that highlight information relevant to users. A pair of experimental studies shows that they increase both reading time—by 30 seconds, from 7 in the first experiment and 20 in the second—and comprehension of agreement content—by 4/16 points, from 0. Finally, a solid understanding of users’ attitudes towards specific agreement content is needed to inform the design of improved software agreements. To that end, this thesis presents an analysis of EULAscan, an online community of anonymous reviewers of software agreements. An open coding is used to categorize 191 EULAscan reviews. From this analysis, functionality emerges as the most prevalent concern. The wide variety of other concerns across reviews suggests that static designs of software agreements would inadequately serve a large population of users. Instead, this thesis proposes a focus on end-user tools that identify and highlight clauses of possible interest to a given user—for example, terms that the user has not seen before.

As the Internet becomes more pervasive, the content and services are increasing in quantity as well as improving in quality. This trend is fostered by the advancement of technologies, such as RESTful services, Web 2.0, and Mashups. Service composition integrates services to fulfill specific tasks using a set of tools. The existing service composition techniques and tools are mainly designed for the Service Oriented Architecture (SOA) professionals. The business processes used in the service composition systems are primarily designed by business analysts who have extensive process knowledge. Due to the lack of process knowledge, novice business analysts and end users face challenge to identify and orchestrate service into a well-defined business process. Even for the experienced users, it is challenging to select appropriate services from a set of functionally similar services as the quality information of services may not be available. In this thesis, we propose a framework that allows a non-technical user to combine web services to achieve a goal. Our approach helps users to find the process knowledge from the web. We index web services based on the semantic concepts available in the service description documents and help users to formulate a web service search query. We use online reviews to choose a web service from a set of functionally similar web services. Our approach automatically finds the data flow between web services and generates a user interface to execute a composite service. The effectiveness of our proposed approaches is demonstrated through a series of case studies. The results of our case studies show that our approaches for process knowledge extraction, service discovery, and service selection make it easier for people with less technical knowledge to compose services. / Thesis (Ph.D, Electrical & Computer Engineering) -- Queen's University, 2014-06-30 15:18:28.155

Thesis (Ph. D.)--Computing, Georgia Institute of Technology, 2007. / Committee Chair: Dovrolis, Constantine; Committee Member: Ammar, Mostafa; Committee Member: Schwan, Karsten; Committee Member: Steenkiste, Peter; Committee Member: Zegura, Ellen.

Thesis (Ph.D.)--Drexel University, 2009. / Includes abstract and vita. Includes bibliographical references (leaves 108-114).

Thesis (M.Dev.Stud.)--Victoria University of Wellington, 2009. / Includes bibliographical references.

Thesis (Ph. D.)--Virginia Commonwealth University, 2009. / Prepared for: Dept. of Information Systems. Title from title-page of electronic thesis. Bibliography: leaves 177-188. Available online via the internet.

Thesis (M.S.) in Computer Engineering--University of Maine, 2004. / Includes vita. Includes bibliographical references (leaves 50-51).