Hierarchical Data Format 5 (HDF5): Why and How to Use It

In the last two decades the need for storage space in computational sciences has increased costantly. This necessity comes from various fields of computational sciences, such as numerical simulations (weather forecast, galaxy evolution in astrophysics, etc.) and large scale observationa surveys (e.g. LOFAR).
In addition, the expansion of the Internet has provided an extraordinary tool for sharing data around the world. This has triggered the creation of a file format that is portable, I/O efficient, standard, versatile and parallel. At the same time it has to be capable of handling very large and complex portions of data, with the guarantee of a long-term access via backward-compatible multi-platform and multi-language APIs.
The answer to these needs is the Hierarchical Data Format 5 (HDF5). In this talk I will review the main properties of the HDF5 format, why it should be used and how its usage can be tuned onto some test-cases such as numerical simulations and long-term data storage.
Finally, I will explain the main API commands that I use in my numerical project.