Techniques for determining workload skew

1. A method of determining and using skew for workload comprising:

performing, using a processor and for a time period in a data storage system having a first data storage configuration including N storage tiers of physical storage devices, data movement optimizations that move data portions of logical devices between the N storage tiers;

receiving, using a processor, a set of measurements for the logical devices for the time period, the set of measurements for said logical devices including N workload measurements and N capacity measurements;

determining, using a processor and said set of measurements, a set of N points denoting capacity boundaries and associated workloads for data portions of the logical devices stored on the N storage tiers of the first data storage configuration;

determining, using a processor and said set of N points and a point of origin, a curve representing a cumulative workload skew for said logical devices in the first data storage configuration; and

modeling, using a processor and the cumulative workload skew, performance of a second data storage configuration, wherein said modeling includes determining, in accordance with the cumulative workload skew, a second set of points for the second data storage configuration, wherein each point of the second set is included in the cumulative workload skew, the second set of points denoting modeled capacity boundaries and associated modeled workloads for storage tiers of the second data storage configuration, wherein said modeling performance of the second data storage configuration includes determining, in accordance with the cumulative workload skew and the second set of points, a value of a performance metric for the second data storage configuration.

2. The method of claim 1 , wherein said determining said curve includes performing curve fitting to fit said N points of the curve.

3. The method of claim 2 , wherein said curve approximates an exponential function.

4. The method of claim 1 , wherein the second data storage configuration includes M storage tiers.

5. The method of claim 4 , wherein the second data storage configuration is a revised version of the first data storage configuration wherein the second data storage configuration includes at least one of: an increase in storage capacity to at least one of the N storage tiers of the first data storage configuration, and a decrease in storage capacity to at least one of the N storage tiers of the first data storage configuration.

6. The method of claim 4 , wherein the second data storage configuration includes at least one additional storage tier not included in the first data storage configuration.

7. The method of claim 4 , wherein the first data storage configuration includes at least one storage tier not included in the second data storage configuration whereby the second data storage configuration removes the at least one storage tier.

8. The method of claim 4 , wherein N is equal to M whereby the first data storage configuration and the second data storage configuration include a same number of storage tiers.

9. The method of claim 1 , wherein said determining said curve uses only said set of N points and the point of origin to determine remaining data points of the curve.

10. The method of claim 1 , wherein the data movement optimizations are performed in accordance with workloads of the data portions and include moving a first data portion from one of the logical devices from a first of the N storage tiers to a second of the N storage tiers.

11. The method of claim 10 , wherein the first data portion is demoted due to a decrease in workload directed to the first data portion, the first storage tier having a higher performance ranking than the second storage tier.

12. The method of claim 10 , wherein the first data portion is promoted due to an increase in workload directed to the first data potion, the first storage tier having a lower performance ranking than the second storage tier.

13. The method of claim 10 , wherein the data movement optimizations place data portions having a highest relative workload on a highest ranked one of the N storage tiers.

14. The method of claim 1 , wherein the second set of points includes a number of points equal to a number of tiers in the second data storage system configuration.

15. The method of claim 1 , wherein a total capacity indicates an aggregated capacity of data portions of the logical devices stored on the N storage tiers and a total workload indicates an aggregated workload directed to said total capacity during said time period, each point in said set of N points corresponds to a single one of the N storage tiers, said each point for said single one storage tier of the N storage tiers having a first coordinate identifying a first percentage of the total capacity and a second coordinate identifying a second percentage of the total workload, said first percentage denoting a percentage of the total capacity of the data portions of the logical devices stored in said single one storage tier of the N storage tiers and all others of said plurality of storage tiers having a higher performance ranking than said single one storage tier of the N storage tiers, said second coordinate denoting an aggregated percentage of the total workload directed to said first percentage of the total capacity.

16. The method of claim 15 , wherein each point in said second set corresponds to a single storage tier of the second data storage configuration, said each point in said second set having a third coordinate identifying a third percentage of the total capacity and a fourth coordinate identifying a fourth percentage of the total workload, said third percentage denoting a percentage of the total capacity of the data portions of the logical devices stored in said single storage tier of the second data storage configuration and all other tiers of the second storage configuration having a higher performance ranking than said single storage tier, said fourth coordinate denoting an aggregated percentage of the total workload directed to said third percentage of the total capacity.

17. A non-transitory computer readable medium comprising code stored thereon that, when executed, performs a method of determining skew for workload comprising:

performing, for a time period in a data storage system having a first data storage configuration including N storage tiers of physical storage devices, data movement optimizations that move data portions of logical devices between the N storage tiers;

receiving a set of measurements for the logical devices for the time period, the set of measurements for said logical devices including N workload measurements and N capacity measurements;

determining, using said set of measurements, a set of N points denoting capacity boundaries and associated workloads for data portions of the logical devices stored on the N storage tiers of the first data storage configuration;

determining, using said set of N points and a point of origin, a curve representing a cumulative workload skew for said logical devices in the first data storage configuration; and

modeling, using the cumulative workload skew, performance of a second data storage configuration, wherein said modeling includes determining, in accordance with the cumulative workload skew, a second set of points for the second data storage configuration, wherein each point of the second set is included in the cumulative workload skew, the second set of points denoting modeled capacity boundaries and associated modeled workloads for storage tiers of the second data storage configuration, wherein said modeling performance of the second data storage configuration includes determining, in accordance with the cumulative workload skew and the second set of points, a value of a performance metric for the second data storage configuration.

18. A system comprising:

at least one processor; and

a memory comprising code stored thereon that, when executed by the at least one processor, performs a method of determining skew for workload comprising:

performing, for a time period in a data storage system having a first data storage configuration including N storage tiers of physical storage devices, data movement optimizations that move data portions of logical devices between the N storage tiers;

receiving a set of measurements for the logical devices for the time period, the set of measurements for said logical devices including N workload measurements and N capacity measurements;

determining, using said set of measurements, a set of N points denoting capacity boundaries and associated workloads for data portions of the logical devices stored on the N storage tiers of the first data storage configuration;

determining, using said set of N points and a point of origin, a curve representing a cumulative workload skew for said logical devices in the first data storage configuration; and

modeling, using the cumulative workload skew, performance of a second data storage configuration, wherein said modeling includes determining, in accordance with the cumulative workload skew, a second set of points for the second data storage configuration, wherein each point of the second set is included in the cumulative workload skew, the second set of points denoting modeled capacity boundaries and associated modeled workloads for storage tiers of the second data storage configuration, wherein said modeling performance of the second data storage configuration includes determining, in accordance with the cumulative workload skew and the second set of points, a value of a performance metric for the second data storage configuration.