Distributed Architectures for High Performance and Privacy-Aware Content Generation and Delivery

The current trend in the evolution of the Web is towards an every increasing demand for personalized Web contents. Tailoring Web resources to the characteristics of heterogeneous client devices and user preferences opens two main novel issues in the research area of content generation and delivery. First, the high computational cost characterizing most adaptation services requires efficient and scalable support systems. Second, adaptation services typically rely on user information that may include sensitive data. When system scalability is achieved through architectures that are distributed over a geographical area, privacy issues concerning the user profiles management become particularly critical. In this paper, we propose a distributed architecture for the ubiquitous Web access that guarantees high performance and privacy of sensitive user information by splitting the adaptation services over the nodes of a two-level topology. We evaluate the performance and the overheads of the proposed architecture in order to understand to which extent it is convenient to distribute adaptation services over multiple nodes that are geographically distributed