Reducing Storage Footprint and Bandwidth Requirements to a Minimum: Compressing Sets of Simulation Results
Simulations are a key component of research and development in many industries. Due to the increasing number of simulations and size of their generated output, managing simulation output has become an important task. Consequently, simulation data management (SDM) systems were widely deployed. Efficient storage and distribution of large numbers of simulation results is one of the major challenges many SDM-systems face nowadays. This challenge represents an opportunity to deploy advanced compression techniques and transfer protocols. SDMZIP is a new approach to compress results of LS-DYNA simulations by aggregating related simulation files into a set, thus reducing redundant information in the whole set. This is achieved by storing common information between simulations in a central database, which is complemented by simulation-specific detail files for each simulation result. The database can be extended incrementally to add new simulations to the set, possibly resulting in an update file for each extension. This modular storage strategy enables component-wise transfer and deletion protocols. Thereby the needed bandwidth and storage requirements are further reduced significantly. For a set of 30 simulations as a part of a parameter variation study, a compression rate of over 100 was achieved by the new software solution. This correlates to an improvement of a factor 4 compared to current industry standards. Moreover, it suggests that there is a further potential for reduction in the data already available in many application areas today.
https://www.dynamore.de/de/download/papers/2016-ls-dyna-forum/Papers%202016/dienstag-11.10.16/simulation-data-management-and-compression/reducing-storage-footprint-and-bandwidth-requirements-to-a-minimum-compressing-sets-of-simulation-results/view
https://www.dynamore.de/@@site-logo/DYNAmore_Logo_Ansys.svg
Reducing Storage Footprint and Bandwidth Requirements to a Minimum: Compressing Sets of Simulation Results
Simulations are a key component of research and development in many industries. Due to the increasing number of simulations and size of their generated output, managing simulation output has become an important task. Consequently, simulation data management (SDM) systems were widely deployed. Efficient storage and distribution of large numbers of simulation results is one of the major challenges many SDM-systems face nowadays. This challenge represents an opportunity to deploy advanced compression techniques and transfer protocols. SDMZIP is a new approach to compress results of LS-DYNA simulations by aggregating related simulation files into a set, thus reducing redundant information in the whole set. This is achieved by storing common information between simulations in a central database, which is complemented by simulation-specific detail files for each simulation result. The database can be extended incrementally to add new simulations to the set, possibly resulting in an update file for each extension. This modular storage strategy enables component-wise transfer and deletion protocols. Thereby the needed bandwidth and storage requirements are further reduced significantly. For a set of 30 simulations as a part of a parameter variation study, a compression rate of over 100 was achieved by the new software solution. This correlates to an improvement of a factor 4 compared to current industry standards. Moreover, it suggests that there is a further potential for reduction in the data already available in many application areas today.