Title: Advancing Exascale Data Management with Trust-Driven Lossy Compression

Abstract:

Today’s extreme-scale scientific simulations and experiments are generating more data than that can be stored, transmitted, and analyzed. The upcoming exscale systems and high-resolution scientific instruments are going to exacerbate this problem with the imbalanced growth between storage systems and scientific data. In this talk, we will see how scientists address those data challenges via effective lossy data compression. This includes design and development of trust-driven data compression algorithms, optimizations on high-end computing clusters, and their interactions with storage and I/O systems. The ultimate goal is to provide effective scientific data management solutions to reduce time to insights on the cutting-edge computing systems for mission-critical applications.