Data storage method and method for executing an application with reduced access time to the stored data

A method for executing several processes (65 to 68) of one or more applications (83, 85), comprising:

• a first phase of observing the running of said processes (65 to 68) and the manner in which they access, over time, the stored data during said execution, during which the determination of the data file (5, 61 to 64) slices (51 to 58) likely to be accessed simultaneously by different application (83, 85) processes (65 to 68) is carried out,
• a second phase of parametrization of the storage of the data by said processes (65 to 68) on data servers (3, 4) and on the storage spaces (31 to 43) thereof, associating, with the data file (5, 61 to 64) slices (51 to 58), storage spaces (31 to 43) on the data servers (3, 4),
• a third phase of distributing the stored data file (5, 61 to 64) slices (51 to 58) over different data servers and over the storage spaces (31 to 43) thereof, this distribution being carried out such that the data file (5, 61 to 64) slices (51 to 58) likely to be subsequently simultaneously accessed by different application (83, 85) processes (65 to 68) are stored on different data servers (3, 4), where appropriate on different storage spaces (31 to 43) of these data servers (3, 4) so as to reduce the subsequent access, to each of all or part of these data servers (3, 4), where appropriate of these storage spaces (31 to 43), by too many application (83, 85) processes (65 to 68) simultaneously.

A storage method for storing, on data servers (3, 4), data from the execution of several processes (65 to 68), comprising:

• distributing the stored data over different data servers (3, 4),
• characterized in that:
• this distribution is carried out in such a way that groups of data likely to be subsequently accessed simultaneously by different processes (65 to 68) are stored on different data servers (3, 4) so as to reduce the subsequent access, to each of all or part of these data servers (3, 4) by too many processes (65 to 68) simultaneously,
• and in that:
• the determination of the groups of data likely to be accessed simultaneously by different processes (65 to 68) has been carried out, during a prior phase of executing these processes (65 to 68), by observing the behavior of these processes (65 to 68) in order to access these groups of stored data over time.