The article describes a new data schema called HDF5eis for handling large multidimensional environmental sensor time series data in high-performance computing (HPC) applications, which overpowers traditional geophysical data formats. The schema allows easy input and output protocols for data while also supporting metadata storage in UTF-8 encoded byte streams or columnar format with time series data. HDF5eis’s API enables accessing large data sets distributed across small heterogenous files conveniently. The schema outperforms conventional seismic formats by up to two orders of magnitude. HDF5eis is presented as a tool and an experimental draft that will establish the next-generation data standards in earth sciences.
https://pubs.geoscienceworld.org/geophysics/article-abstract/88/3/F29/622154/HDF5eis-A-storage-and-input-output-solution-for?redirectedFrom=fulltext