Abstract: When database researchers talk about integrating or exchanging data, they generally assume the end goal is to give everyone a consistent, unified view of whatever heterogeneous data exists within a system. However, in many settings among collaborators, there is no consensus as to what that view should look like -- neither in terms of its schema *nor* even its data contents! Particularly in the life sciences, this tends to be the case, as many data items may be controversial, speculative, or based on a local laboratory's specialized expertise.

We are building a new type of system, the *collaborative data sharing system* or CDSS. Our implementation, Orchestra, is focused on enabling data sharing where there is heterogeneity among schemas and disagreement in data. In contrast to previous work, we enable a 'publish first, resolve conflicts later' model where individual peers determine how to resolve conflicts according to their own policies. I will discuss the three main pillars of our work -- an extended data model that captures provenance and trust; a decentralized approach to reconciling different participants' database instances based on trust policies and the transactions applied to each database; and our work-in-progress on translating updates from one schema to another.

Joint work with Nicholas Taylor, Grigoris Karvounarakis, Todd Green, Val Tannen.

Short Bio: Zachary Ives is an Assistant Professor at the University of Pennsylvania and an Associated Faculty Member at the Penn Center for Bioinformatics. He received his B.S. from Sonoma State University and his PhD from the University of Washington. His research interests include data integration, peer-to-peer models of data sharing, processing and security of heterogeneous sensor streams, and data exchange between autonomous systems. He is a recipient of the NSF CAREER award and a member of the DARPA Computer Science Study Panel.