Dynamic application migration across storage platforms
Embodiments of the present disclosure relate to load balancing application processing between storage platforms. Input/output (I/O) workloads can be anticipated during one or more time-windows. Each I/O workload can comprise one or more I/O operations corresponding to one or more applications. Processing I/O operations of each application can be dynamically migrated to one or more storage platforms of a plurality of storage platforms based on the anticipated workload.
1. A method comprising:
anticipating input/output (I/O) workloads during one or more time-windows, each I/O workload comprising one or more I/O operations corresponding to one or more applications, wherein anticipating the IO workloads includes:
determining an amount of additional workload each storage platform can process based on one or more response time goals associated with each Application's I/O operations, and
determining each storage array's breaking workload, wherein a breaking workload corresponds to an amount of time each storage array exceeds the one or more response time goals; and
dynamically migrating processing of each Application's corresponding I/O operations to one or more storage platforms of a plurality of storage platforms based on the anticipated workload.
2. The method of claim 1 further comprising determining the one or more response time goals associated with each Application.
3. The method of claim 2 further comprising determining loads of each of the storage platforms.
4. The method of claim 3 further comprising predicting I/O operation patterns of the anticipated I/O workloads.
5. The method of claim 4 , wherein predicting the I/O operation patterns includes determining amounts of each type of I/O operation in each anticipated I/O workload, wherein each I/O operation type includes at least a read or write (R/W) operation.
6. The method of claim 5 further comprising predicting each storage platform's response times for processing each Application's I/O operations.
7. The method of claim 6 , wherein predicting each response time includes determining one or more of each storage platform's configuration, I/O operation types in each anticipated workload, and each I/O operation's size.
8. The method of claim 6 , wherein predicting each storage platform's response times for processing each Application's I/O operations includes:
determining each storage platform's current workload; and
monitoring each storage platform's response times for processing I/O operations of the current workload.
9. The method of claim 8 , wherein dynamically migrating processing of each Application's I/O operations to one or more storage platforms includes selecting the one or more storage platforms based on one or more of the amounts of additional workload each storage platform can process and each storage array's breaking workload.
10. An apparatus comprising a memory and at least one processor configured to:
anticipate input/output (I/O) workloads during one or more time-windows, each I/O workload comprising one or more I/O operations corresponding to one or more applications, wherein anticipating the IO workloads includes:
determining an amount of additional workload each storage platform can process based on one or more response time goals associated with each Application's I/O operations, and
determining each storage array's breaking workload, wherein a breaking workload corresponds to an amount of time each storage array exceeds the one or more response time goals; and
dynamically migrate processing of each Application's corresponding I/O operations to one or more storage platforms of a plurality of storage platforms based on the anticipated workload.
11. The apparatus of claim 10 further configured to determine the one or more response time goals associated with each Application.
12. The apparatus of claim 11 further configured to determine loads of each of the storage platforms.
13. The apparatus of claim 12 further configured to predict I/O operation patterns of the anticipated I/O workloads.
14. The apparatus of claim 13 further configured to predict the I/O operation patterns by determining amounts of each type of I/O operation in each anticipated I/O workload, wherein each I/O operation type includes at least a read or write (R/W) operation.
15. The apparatus of claim 14 further configured to predict each storage platform's response times for processing each Application's I/O operations.
16. The apparatus of claim 15 further configured to predict each response time by determining one or more of: each storage platform's configuration, I/O operation types in each anticipated workload, and each I/O operation's size.
17. The apparatus of claim 15 , wherein predicting each storage platform's response times for processing each Application's I/O operations includes:
determining each storage platform's current workload; and
monitoring each storage platform's response times for processing I/O operations of the current workload.
18. The apparatus of claim 17 , wherein dynamically migrating processing of each Application's I/O operations to one or more storage platforms includes selecting the one or more storage platforms based on one or more of the amounts of additional workload each storage platform can process and each storage array's breaking workload.