Scalable architectures and services for ubiquitous Web access Michele Colajanni University of Modena colajanni@unimo.it Riccardo Lancellotti University of Modena lancellotti.riccardo@unimo.it Philip S. Yu IBM T.J. Watson Research Center psyu@us.ibm.com 1 Tutorial Overview The success of the Web in the last decade has caused an evolution of the resources that are disseminated through the Internet. The initial static contents have been enriched by an increasing amount of multimedia and dynamically generated resources. This evolution has shifted the research focus from a nowadays ma- ture content delivery scenario to a Web service generation and delivery scenario. The overall complexity is further increased by the so-called ubiquitous Web access that aims to allow the users to access Web-based services from any location through every class of devices. The key feature of the ubiquitous Web is rep- resented by the content adaptation services that tailor the Web content and the Web-based services to the characteristics of the client devices and to the preferences of the users. This feature of the ubiquitous Web introduces new performance and security problems to the infrastructure that has to generate and disseminate the resources, but it also offers a wide range of novel service opportunities. The tutorial is divided in three parts: the first related to the services for the ubiquitous Web access, the second to the architectures to build scalable services, the third to the presentation of some case studies. In the first part, we present the adaptation services by differentiating transcoding from personalization services. Transcoding services tailor Web resources to the capabilities of the client and network infrastruc- ture, while personalization requires more sophisticated services that aim to adapt the content to (a combina- tion of) user preferences, locations and behaviors. We also provide a classification of adaptation services, where the classification is based on the information used for the content adaptation service. For the design and deployment of content adaptation architectures the main difference occurs between services relying on persistent data (e.g., user profiles), and services using just volatile information. We discuss the characteris- tics of each class of content adaptation services providing examples of both research studies and commercial products. The second part of the tutorial is devoted to the analysis of the architectures for the deployment of content adaptation services with a special focus on system scalability and issues related to data consistency and privacy. We define the design of a content adaptation architecture as a mapping problem. The functions of content adaptation, data management and connection management must be assigned to the nodes composing the infrastructure. We identify different solutions: the client-based architecture that places every adaptation function directly on the client device; the server-based architecture that follows an opposite approach and utilizes a powerful system to provide ubiquitous Web access to heterogeneous clients; and the intermediary- based architecture that places adaptation services on intermediate edge servers that are closer to the clients than to the origin servers. There is a wide space of architectural, coordination and algorithmic options for the nodes of an intermediate infrastructure. We describe and evaluate the architectures having in mind the main requirements that are necessary for supporting most adaptation services, especially scalability, data consistency, identity and location management, information security and user privacy. In the third part of the tutorial, we present some interesting prototypes and systems for the support of content adaptation services. The presentation criteria is based on two parameters: whether the infrastructure