This paper describes SARAVÁ, a research project that aims at investigating new challenges in P2P data sharing for online communities. The major advantage of P2P is a completely decentralized approach to data sharing which does not require centralized administration. Users may be in high numbers and interested in different kinds of collaboration and sharing their knowledge, ideas, experiences, etc. Data sources can be in high numbers, fairly autonomous, i.e. locally owned and controlled, and highly heterogeneous with different semantics and structures. Our project deals with new, decentralized data management techniques that scale up while addressing the autonomy, dynamic behavior and heterogeneity of both users and data sources. In this context, we focus on two major problems: query processing with uncertain data and management of scientific workflows.