[use case | scientific data analysis]
Problem addressed
Some scientific experiments are time consuming, hazardous, or expensive. Hence in these cases, it makes sense to prepare the experiments first "in silico": i.e. run simulations on computers, until the parameters space has been narrowed down to a small interesting set of parameters that can be used in a real experiment. Then a limited experiment can be conducted. Modeling a small angle neutron scattering experiment studying the structure of proteins is the simulation Contrail focuses on in this use case. This simulation software today however, runs on specific systems, that are not dynamically scalable.
Contrail solution
Contrail will be used to implement a complete simulation environment for neutron scattering experiments. (From the ISIS neutron scattering facility, http://www.isis.stfc.ac.uk) .)The simulation itself will be implemented in a Contrail PaaS environment. Because simulations can sometimes exist of a dynamically changing number of sub-simulations that can run in the hundreds, elasticity as provided by Contrail is important.
Expected benefits of using Contrail
The outcome will be a flexible, scalable simulation platform, which will also be applicable for other scientific simulation experiments.