System for managing data storage based on vector-summed size-frequency vectors for data sets, devices, and residual storage on devices

A method for managing the allocation of data sets among storage devices of a computing apparatus in such a manner that system degradation will be minimized and the storage devices will be utilized with maximum efficiency by placement of each of the data sets on a storage device whose uncommitted storage volume and access capability most nearly meets the requirement of that particular data set and by further placing data sets in such a way that access activity to the data sets will be distributed substantially uniformly across all the storage devices, said method comprising

• calculating a data storage factor of each data set, wherein the data storage factor is the vector sum, in a cartesian coordinate system, of the number of accesses per unit time to the particular data set and the volume of data in that data set;
• monitoring and recording data set access activity as a function of time;
• calculating a machine storage factor of each storage device, wherein the machine storage factor is the vector sum, in a cartesian coordinate system, of the total storage capacity of the particular storage device and the number of accesses the particular device can sustain per second;
• calculating a residual storage factor of each storage device, wherein residual storage factor is indicative of the available or uncommitted storage capacity of the particular device, with the residual storage factor being the vector difference, in a cartesian coordinate system, of the machine
• storage factor of the particular device and the vector sum, in a cartesian coordinate system, of the data storage factors of the data sets stored on that particular storage device;
• allocating the data set to a storage device that has sufficient available space and whose residual storage factor most nearly matches and exceeds the data storage factor of the data set being allocated;
• dividing each operational day of said computing apparatus into a predetermined number of time periods;
• determining the time period during the previous operational day that the storage devices as a group were least utilized;
• using the time period determined in the preceding step to move data sets among the storage devices so as to balance the residual storage factors of all the storage devices as a group for each of the time periods, starting with the time period of maximum use of the storage devices and proceeding progressively to the time period of minimum use of the storage devices.