Autonomous workload homing in a multi-tenant environment
A methods for identifying a multi-tenant storage array for an application workload includes identifying workload parameters and defining a plurality of groups for each parameter and a plurality of “bins” corresponding to tuples of the groups. Exemplary workload parameters include a percent read parameter and an I/O size parameter. A bin mix of the workload is determined based on historical data wherein the bin mix indicates bins associated with workload activity exceeding a specified threshold. The bin mix is used to define at least some inputs for a supervised learning model of a process for homing application workloads in a multi-tenant storage array. After appropriate training of the model with a generative adversarial network, the model may be invoked to infer or predict attributes of a suitable storage array. The workload may be associated with a scaling factor that influences the determination of a suitable storage array.
1. A method for homing a source workload within a multitenant storage environment, the method comprising:
identifying workload parameters, including a first workload parameter and a second workload parameter, associated with an application workload;
defining a plurality of value ranges for each of the workload parameters including a first plurality of values ranges for the first workload parameter and a second plurality of ranges for the second workload parameter;
defining a plurality of bins wherein each bin is associated with a distinct tuple of the value ranges, wherein each of the plurality of bins includes a first tuple component indicative of one of the first plurality of value ranges and a second tuple component indicative of one of the second plurality of value ranges;
determining a bin mix of the workload, wherein the bin mix indicates which of the plurality of bins correspond to workload activity exceeding a specific threshold;
specifying inputs and labels for a supervised learning model of a process for homing the application workload in a multi-tenant storage array, wherein the inputs include inputs associated with the bin mix and the labels correspond to attributes of the multi-tenant storage array;
training the model based at least in part on at least one of: historical data and one or more test cases;
invoking the model to predict attributes of the multi-tenant array for a new workload mix; and
identifying a suitable multi-tenant array based on the attributes.
2. The method of claim 1 , further comprising: migrating the workload application to the suitable multi-tenant array.
3. The method of claim 1 , wherein the workload parameters include at least one parameter selected from a percent read parameter and an I/O size parameter.
4. The method of claim 3 , wherein the workload parameters include a percent read parameter and an I/O size parameter and wherein the plurality of bins include a bin corresponding to each combination of percent read value range and I/O size value range.
5. The method of claim 1 , further comprising associating the application workload with a scaling factor, wherein identifying a suitable multi-tenant array includes identifying a suitable multi-tenant array based on the scaling factor.
6. The method of claim 1 , wherein the features for the supervised learning model include, in addition to the bin mix features, one or more features corresponding to a storage resource attribute associated with the workload application.
7. The method of claim 6 , wherein the workload application is associated with block storage and wherein the features for the supervised learning model include a block storage attribute.
8. The method of claim 6 , wherein the workload application is associated with file storage and wherein the features for the supervised learning model include a file storage attribute.
9. The method of claim 1 , wherein training the model includes generating one or more test cases with a long short term memory (LSTM) generator and discriminating between the one or more test cases and real data.
10. The method of claim 9 , wherein the inputs include at least one numerical input and at least one categorical input and wherein generating the test cases includes employing a tanh activation function for the numerical inputs and a softmax function for the categorical inputs.
11. An information handling system, comprising: a central processor unit; and non-transitory computer readable storage, accessible to the processor, including processor executable instructions that, when executed by the processor, result in operations comprising:
identifying workload parameters, including a first workload parameter and a second workload parameter, associated with an application workload;
defining a plurality of value ranges for each of the workload parameters including a first plurality of values ranges for the first workload parameter and a second plurality of ranges for the second workload parameter;
defining a plurality of bins wherein each bin is associated with a distinct tuple of the value ranges, wherein each of the plurality of bins includes a first tuple component indicative of one of the first plurality of value ranges and a second tuple component indicative of one of the second plurality of value ranges;
determining a bin mix of the workload, wherein the bin mix indicates which of the plurality of bins correspond to workload activity exceeding a specific threshold;
specifying inputs and labels for a supervised learning model of a process for homing the application workload in a multi-tenant storage array, wherein the inputs include inputs associated with the bin mix and the labels correspond to attributes of the multi-tenant storage array;
training the model based at least in part on at least one of: historical data and one or more test cases;
invoking the model to predict attributes of the multi-tenant array for a new workload mix; and
identifying a suitable multi-tenant array based on the attributes.
12. The information handling system of claim 11 , further comprising: migrating the workload application to the suitable multi-tenant array.
13. The information handling system of claim 11 , wherein the workload parameters include at least one parameter selected from a percent read parameter and an I/O size parameter.
14. The information handling system of claim 13 , wherein the workload parameters include a percent read parameter and an I/O size parameter and wherein the plurality of bins include a bin corresponding to each combination of percent read value range and I/O size value range.
15. The information handling system of claim 11 , further comprising associating the application workload with a scaling factor, wherein identifying a suitable multi-tenant array includes identifying a suitable multi-tenant array based on the scaling factor.
16. The information handling system of claim 11 , wherein the features for the supervised learning model include, in addition to the bin mix features, one or more features corresponding to a storage resource attribute associated with the workload application.
17. The information handling system of claim 16 , wherein the workload application is associated with block storage and wherein the features for the supervised learning model include a block storage attribute.
18. The information handling system of claim 16 , wherein the workload application is associated with file storage and wherein the features for the supervised learning model include a file storage attribute.
19. The information handling system of claim 11 , wherein training the model includes generating one or more test cases with a long short term memory (LSTM) generator and discriminating between the one or more test cases and real data.
20. The information handling system of claim 19 , wherein the inputs include at least one numerical input and at least one categorical input and wherein generating the test cases includes employing a tanh activation function for the numerical inputs and a softmax function for the categorical inputs.