The need to accurately model and measure complex phenomena is driving a continual increase in the size of scientific data. In practically all disciplines, multi-gigabyte, terabyte, or even petabytesized datasets are becoming more common. While advances have accelerated our ability to create large data, the ability to share and analyze these datasets has struggled to keep pace. Through the technological advances of this project, the simple, intuitive, and wide-reaching use of large scientific data will be possible. Currently, sharing and analyzing even a single large dataset can hold overwhelming challenges for researchers. To overcome these challenges, this project will develop technologies that provide search-by-example queries of scientific datasets, intelligent data reduction and reconstruction, and formal models for large data systems.