Defining HDF5 Future
Event Type
Birds of a Feather
Extreme-Scale Computing
HPC workflows
Parallel Applications
Storage Technologies
TimeWednesday, June 19th1:45pm - 2:45pm CEST
LocationSubstanz 1, 2
DescriptionHDF5 is a unique, open-source, high-performance technology suite that consists of an abstract data model, library, and file format used for storing and managing extremely large and/or complex data collections. The technology is used worldwide by government, industry, and academia in a wide range of science, engineering, and business disciplines.

We will provide a forum for the HDF5 user community to share ideas, present problems, outline possible solutions and discuss future direction of HDF5 and the HDF5 ecosystem.
Elena Pourmal will present HDF5 features under development, the HDF5 roadmap, including upcoming releases and launch a discussion for HDF5 roadmap and community involvement in HDF5 development.
Elena’s presentation and discussion will be followed by a discussion led by Steven Varga who is an active member and contributor of the HDF community. He will answer questions from the audience and survey the latest developments and challenges in the HDF5 ecosystem including (but not limited to):
● Storage devices and platforms for HDF5 (cloud, object stores, HDFS, PFS), VFDs and VOLs
● HDF5 language bindings (C++/LLVM, Python, Julia, R)
● HDF5 library development and community code contributions to HDF5
● HDF5 Performance and productivity (benchmarks, profiling, Tau plugin)
● I/O concurrency (multithreading, MPI-I/O)
● Sparse data and HDF5
● HDF5 connectors for Apache Spark and Drill
● HDF5 tools, including third-party tools and vendor support
Steven and Elena will encourage HDF5 community members to share their experiences and challenges with HDF5. The summary of discussion will be shared with the HDF5 community.