Health signature based load shifting for part longevity

1 . A method comprising:

monitoring one or more variables associated with each of a plurality of redundant parts of a computing system, wherein one or more of the plurality of redundant parts are experiencing a load;

generating, based on the monitored one or more variables, a health signature for each of the plurality of redundant parts to obtain health signatures; and

shifting, based on the health signatures, the load between the plurality of redundant parts,

wherein shifting the load comprises:

performing, by a first redundant part of the plurality of redundant parts, a first portion of the load; and

performing, by a second redundant part of the plurality of redundant parts, a second portion of the load.

2 . The method of claim 1 , further comprising monitoring, after shifting the load, the one or more variables of the plurality of redundant parts, including updating the health signature of each of the plurality of redundant parts.

3 . The method of claim 1 , wherein the health signature is a vector, with each element of the vector corresponding to a different variable of the one or more variables.

4 . The method of claim 3 , wherein generating the health signature for each of the plurality of redundant parts includes weighting each element of the vector based on historical data associated with each of the plurality of redundant parts.

5 . The method of claim 3 , wherein shifting the load is based on comparing each of the health signatures.

6 . The method of claim 5 , wherein shifting the load is performed in response to determining that the health signature of a redundant part, of the plurality of redundant parts, experiencing the load differs from one or more of the health signatures of one or more other redundant parts, of the plurality of redundant parts, by an amount greater than a threshold.

7 . The method of claim 6 , wherein the threshold is based on an allowed Euclidian distance between the vectors.

8 . The method of claim 1 , wherein shifting the load is based on a predicted upcoming workload.

9 . The method of claim 1 , wherein shifting the load includes changing one or more operating parameters of a redundant part experiencing at least a portion of the load.

10 . The method of claim 1 , wherein shifting the load includes periodically alternating between each of the plurality of redundant parts carrying a portion of the load.

11 . A computing system comprising:

memory;

a plurality of redundant parts, wherein one or more of the plurality of redundant parts are experiencing a load; and

a processor configured to:

monitor one or more variables associated with each of the plurality of redundant parts;

generate, based on the monitored one or more variables, a health signature for each of the plurality of redundant parts to obtain the health signatures; and

shift, based on the health signatures, the load between the plurality of redundant parts,

wherein:

a first redundant part, of the plurality of redundant parts, performs a first portion of the load as a result of the load being shifted; and

a second redundant part, of the plurality of redundant parts, performs a second portion of the load as a result of the load being shifted.

12 . The computing system of claim 11 , wherein each of the plurality of redundant parts comprises a power supply unit (PSU).

13 . The computing system of claim 11 , wherein each of the plurality of redundant parts comprises a power distribution unit (PDU).

14 . The computing system of claim 11 , wherein the processor is further configured to monitor, after shifting the load, the one or more variables of the plurality of redundant parts, including updating the health signature of each of the plurality of redundant parts.

15 . The computing system of claim 11 , wherein the health signature is a vector, with each element of the vector corresponding to a different variable of the one or more variables.

16 . The computing system of claim 15 , wherein, when generating the health signature for each of the plurality of redundant parts, the processor is to weight each element of the vector based on historical data associated with each of the plurality of redundant parts.

17 . The computing system of claim 15 , wherein shifting the load is based on the processor comparing each of the health signatures.

18 . The computing system of claim 17 , wherein shifting the load is performed in response to the processor determining that the health signature of a redundant part, of the plurality of redundant parts, experiencing the load differs from one or more of the health signatures of other redundant parts, of the plurality of redundant parts, by an amount greater than a threshold.

19 . The computing system of claim 18 , wherein the threshold is based on an allowed Euclidian distance between the vectors.

20 . A computer program product comprising a computer readable storage medium and computer program instructions stored in the computer readable storage medium that, when executed, are configured to:

monitor one or more variables associated with each of a plurality of redundant parts of a computing system, wherein one or more of the plurality of redundant parts are experiencing a load;

generate, based on the monitored one or more variables, a health signature for each of the plurality of redundant parts to obtain health signatures; and

shift, based on the health signatures, the load between the plurality of redundant parts, wherein the computer program instructions, that when executed shifting the load, are configured to cause:

a first redundant part, of the plurality of redundant parts, to perform a first portion of the load; and

a second redundant part, of the plurality of redundant parts, to perform a second portion of the load.