ROOT  is a framework for data processing, born at CERN, at the heart of the research on high-energy physics. ROOT consists of multiple components such as I/O, a C++ interpreter, GUI, and math libraries. One of its tasks is to reduce the on-disk usage of physics data; a typical approach is to use various compression techniques. ROOT I/O subsystem supports data compression using common compression algorithms such as ZLIB, LZMA and LZ4.
Zstd  is interesting fast, a lossless compression algorithm that can offer higher compression ratios comparing to most LZ-based compression algorithms, such as ZLIB or LZMA. ZSTD offers higher decompression data rates, as it is shown in standalone benchmarks  tested by Facebook. Other ZSTD functionality includes a special training mode, used to tune the algorithm for specific data patterns. The result of this training is stored in a buffer (a dictionary) which must be loaded before compression and decompression. Use of these dictionaries significantly improves the compression ratio of small data buffers, as found in ROOT-formatted files.
This project is to investigate application of ZSTD within the ROOT framework; benchmark it in comparison to the other algorithms; test it against real LHC data files; and investigate schemes to integrate dictionary-based compression in ROOT files.
Investigate application of ZSTD within the ROOT framework, benchmark it against real LHC data files and investigate schemes to integrate dictionary-based compression in ROOT files.