ABSTRACT: Application service providers and web services are becoming increasingly popular despite adverse IT market conditions. New languages and protocols like XML, SOAP and UDDI provided the technical underpinnings for a global infrastructure where anybody with a networked PC has access to a large number of digital services. The input data for these services needs to be made available by the customer, i.e., the person or institution using the service. Not every potential customer, however, may feel comfortable about entrusting sensitive personal or corporate data to a service provider in an unprotected manner. Even if there is a high level of trust between customer and provider, there may be legal requirements that require a higher level of privacy. Customers may also want to be prepared for an unforeseen change of control on the provider's side - something that is a common occurence especially among start-up companies.
This talk will present several technical solutions how customers can use a provider's services without giving him access to any sensitive data. Our focus will be on privacy homomorphisms, an encryption technique originally proposed by Rivest et al. that maintains the structure of the input data while obscuring the actual content. We shall present several of our own proposals how to integrate privacy homomorphisms into existing service architectures and a comparative analysis that compares costs and benefits of the solutions currently available.
BIO: Oliver Gunther received his Ph.D. in Computer Science from the University of California at Berkeley in 1987. Since 1993 he has been Professor and Director of the Institute of Information Systems at Humboldt University in Berlin, where he also heads the Berlin-Brandenburg Graduate School of Distributed Information Systems and Humboldt's Electronic Business Forum. Since 1999, he has also been a member of the faculty of PULV in Paris. Professor Gunther has published widely in the areas of databases and information systems and served as a consultant and board member to various government agencies and high-tech companies. In 1998/99, he was Chairman of the Board of Poptel AG, Germany's first voice-over-IP company. In 2000/01, he served as Chief Technology Officer of Teamtoolz, a San Francisco-based application service provider for the marketing and advertising industry.
The adoption of high level models has been advocated by many authors as instrumental in Web site development as a strong support to both development and maintenance. We will discuss how the specification of Web sites at the logical level can greatly benefit from the introduction of specific features for the representation of time, which could interact with the support for versions and editions objects in the site. However, time can be seen as a "coordinate" of Web models, a more general notion that includes various forms of specializations, personalizations, and variations, such as those related to location, language, user preferences, device type.
BIO: Paolo Atzeni is Database Professor at Universita Roma Tre. He received his Dr. Ing. degree in Electrical Engineering from Universita di Roma "La Sapienza" in 1980. Before joining Universita Roma Tre, he was member of the research staff at IASI-CNR in Rome, Associate Professor at Universita di Napoli and Professor at Universita di Roma La Sapienza. He also had visiting appointments at the University of Toronto and at Universita dell'Aquila. He has worked on various topics in the database field, including relational database theory, conceptual models and design tools, deductive database, databases and the Web, cooperation of database systems. He published his results in major journals including ACM TODS, JCSS, SIAM Journal of Computing, Information and Computation, Information Systems and presented them at all major database conferences, including ACM-Sigmod, VLDB, ACM-PODS, ICDE, EDBT, ICDT.
He is the leader of the database group at Roma Tre, which includes three professors, a post-doctoral fellow and various students. He was general chair for the 27th International Conference on Very Large Data Bases (VLDB 2001). He was the general chair for VLDB 2001 and program chair for DBPL'95 and for the WebDB'98 and member of the program committee for various conferences.He is the President of the Executive Board of the EDBT Foundation and a Trustee of the VLDB Endowment.
ABSTRACT: In the past we talked about single information systems. In the future we expect an ever increasing number of information systems and data sources, reaching from traditional databases and large document collections, information sources contained in web pages, down to information systems in mobile devices and embedded information in mobile “smart” objects as they will occur in a pervasive computing environment. Therefore not only the immense amount of information demands new thoughts but also the number of different information sources and their coordination poses a great challenge for the development of future tools that will be suitable to access, process, and maintain information. We talk about the continuous, “infinite” information, shortly called the “ information space”. Information in this space is distributed, heterogeneous and undergoes continuous changes. So, the infrastructure for information spaces must provide convenient tools for accessing information via sophisticated search facilities and for combining or integrating search results from different sources (1), for developing applications for analyzing, mining, classifying and processing information (2), and for transactional processes that ensure consistent propagation of information changes and simultaneous invocations of several (web) services within a transactional workflow (3). For the implementation of such an infrastructure we must strive for functions including recoverability, scalability, and availability. As far as possible the infrastructure should avoid global components. Rather a peer-to-peer decentralized coordination middleware must be provided that has some self-configuration and adaptation features to various applications and their load characteristics (4). In this paper we will elaborate some of the aspects in area (3) and (4) and report on research from our hyperdatabase research framework and from experiences in ETHWorld, an ETH wide project that will establish the ETH information space. Nevertheless this paper is rather visionary and is intended to stimulate new research in this wide area.
BIO: Prof. Schek studied mathematics at the Univ. of Stuttgart (Diploma in 1968) and earned his Ph.D. in civil engineering in 1972 also from the Univ. of Stuttgart. During this time he developed – together with his dissertation advisor, Professor Linkwitz – a new method for the computation of prestressed cable-net roofs and applied this method to the analysis of roofs at the 1972 Olympic Games in Munich. He joined the IBM Scientific Centre in Heidelberg as research member (1972-1978) and as project manager (1978-1983). There he investigated end-user-oriented interfaces and transferred increasingly into database system research with special emphasis on integration with information retrieval systems. During this period he also completed habilitation and was given venia legendi for computer methods for network computation from the Univ. of Stuttgart in 1978. From 1983 to 1988 he was Full Professor (C4) at the Technical Univ. of Darmstadt. The DASDBS database project that he established there is known for its contributions to data modeling (NF2 model) to multi-level transactions and to extensible spatial databases. Since 1988 he is full Professor of Computer Science at the ETH Zurich and head of the Institute of Information Systems in the Computer Science Department. His research interests include data models, interoperability of databases and cooperation with application systems, generic, extensible object managers, and transaction management. Prof. Schek is one of the 21 members of the int. VLDB Endowment. He was a founder and editor-in-chief of the VLDB Journal. At the ETHZ he is a member of the planning commission.