Scientific Data Management in the Exascale Era Poster Jan 2015

This project aims to improve scientific productivity for DOE mission critical applications with innovative data processing tools and techniques. These techniques could run orders of magnitude faster or scale much better than existing approaches.  The overarching goal in the next few years is to develop an autonomous scientific data services framework to serve as the basis of an exascale data ecosystem.  More specifically, this project has been working on (1) designing a data storage system interface that enables dynamic performance optimization, (2) developing fundamental algorithms and data structures for users to select the most relevant data records based on their applications, and (3) taking advantage of emerging many-core architectures and expanding the capability of in situ data processing systems.  The work is primarily fundamental computer science research.  The research results are delivered to early adopters through close collaborations for immediate uses.  Techniques with wider applicability are distributed to the user community as publications, open-source software packages, and community standards.