Extent group workload forecasts
Forecasting workload activity for data stored on a data storage device includes selecting at least one metric for measuring workload activity, providing at least one grouping of portions of the data according to a workload affinity determination provided for each of the portions at a subset of a plurality of time steps, where the workload affinity determination is based on each of the data portions in the group experiencing above-average workload activity during same ones of the subset of the plurality of time steps, the subset corresponding to at least one business cycle for accessing the data, and forecasting workload activity for all of the portions of data in the group based on forecasting workload activity for a subset of the data portions that is less than all of the data portions.
1. A method of forecasting workload activity for data stored on a data storage device, comprising:
selecting at least one metric for measuring workload activity;
grouping a plurality of portions of the data into at least one grouping of portions of the data according to a workload affinity determination provided for each of the portions for a subset of a plurality of time steps, wherein the workload affinity determination is based on a similarity of values determined for the metric for each of the data portions, including each of the data portions in the at least one grouping experiencing above-average workload activity relative to a reference level of workload activity for the respective data portion only, independent of workload activity levels of others of the plurality of data portions, during same ones of the subset of the plurality of time steps, the subset corresponding to at least one business cycle for accessing the data;
for a first of the plurality of portions, for each time step of the subset of time steps:
determining a first workload activity ratio indicative of a workload activity of the first portion during the time step relative to a reference level of workload activity for the first portion;
determining a group workload activity ratio indicative of a workload activity of the at least one grouping collectively during the time step relative to a reference level of workload activity for the at least one grouping collectively;
determining whether to adjust a first workload affinity determination of the first portion based at least in part on the first workload activity ratio and the group workload activity ratio; and
if it is determined to adjust the first workload affinity determination, adjusting the first workload affinity determination; and
forecasting workload activity for all of the portions of data in the at least one grouping based on forecasting workload activity for a subset of the data portions that is less than all of the data portions.
2. A method, according to claim 1 , wherein the at least one business cycle for accessing the data corresponds to particular days of a week.
3. A method, according to claim 1 , wherein the at least one business cycle for accessing the data corresponds to a pre-planned event.
4. A method, according to claim 1 , wherein there are at least two business cycles and different groupings of data portions are maintained for each of the business cycles.
5. A method, according to claim 1 , wherein the at least one metric is selected from a group consisting of: a read miss metric, a serial read metric, and a write metric.
6. A method, according to claim 1 , wherein the at least one business cycle for accessing the data corresponds to detecting repeatable patterns in workload activity coherence of the at least one grouping of data.
7. A method, according to claim 6 , wherein the workload activity coherence is based on a mean weighted activity ratio of each data portion and a weighted variance of activity ratios of the data portions at each time step, wherein weighting used for each particular one of the data portions is an activity workload for the particular one of the data portions at each time step.
8. A method, according to claim 1 , wherein the workload affinity determination for each data portion is a time-weighted average.
9. A non-transitory computer readable medium containing software that forecasts workload activity for data stored on a data storage device, the software comprising:
executable code that groups a plurality of portions of the data into at least one grouping of portions of the data according to a workload affinity determination provided for each of the portions for a subset of a plurality of time steps, wherein the workload affinity determination is based on a similarity of values determined for the metric for each of the data portions, including each of the data portions in the at least one grouping experiencing above-average workload activity relative to a reference level of workload activity for the respective data portion only, independent of workload activity levels of others of the plurality of data portions, during same ones of the subset of the plurality of time steps, the subset corresponding to at least one business cycle for accessing the data and wherein the workload activity is based on at least one metric selected for measuring workload activity;
executable code that, for a first of the plurality of portions, for each time step of the subset of time steps:
determines a first workload activity ratio indicative of a workload activity of the first portion during the time step relative to a reference level of workload activity for the first portion;
determines a group workload activity ratio indicative of a workload activity of the at least one grouping collectively during the time step relative to a reference level of workload activity for the at least one grouping collectively;
determines whether to adjust a first workload affinity determination of the first portion based at least in part on the first workload activity ratio and the group workload activity ratio; and
if it is determined to adjust the first workload affinity determination, adjusts the the first workload affinity determination; and
executable code that forecasts workload activity for all of the portions of data in the at least one grouping based on forecasting workload activity for a subset of the data portions that is less than all of the data portions.
10. A non-transitory computer readable medium, according to claim 9 , wherein the at least one business cycle for accessing the data corresponds to particular days of a week.
11. A non-transitory computer readable medium, according to claim 9 , wherein the at least one business cycle for accessing the data corresponds to a pre-planned event.
12. A non-transitory computer readable medium, according to claim 9 , wherein there are at least two business cycles and different groupings of data portions are maintained for each of the business cycles.
13. A non-transitory computer readable medium, according to claim 9 , wherein the at least one metric is selected from a group consisting of: a read miss metric, a serial read metric, and a write metric.
14. A non-transitory computer readable medium, according to claim 9 , wherein the at least one business cycle for accessing the data corresponds to detecting repeatable patterns in workload activity coherence of the at least one grouping of data.
15. A non-transitory computer readable medium, according to claim 14 , wherein the workload activity coherence is based on a mean weighted activity ratio of each data portion and a weighted variance of activity ratios of the data portions at each time step, wherein weighting used for each particular one of the data portions is an activity workload for the particular one of the data portions at each time step.
16. A non-transitory computer readable medium, according to claim 9 , wherein the workload affinity determination for each data portion is a time-weighted average.