Adapting cache processing using phase libraries and real time simulators

1. A method comprising:

caching data from an input/output (I/O) stream using first cache settings;

matching the I/O stream with a first I/O pattern in a library of pre-defined caching policies, wherein the first I/O pattern is associated with second cache settings in the library;

applying the second cache settings to the I/O stream;

simulating the caching of the data from the I/O stream using a plurality of other cache settings;

creating an entry in the library as a new cache policy corresponding to third cache settings and a second I/O pattern associated with the I/O stream based on a result of simulating the caching; and

applying the third cache settings to the I/O stream based on the new cache policy.

2. The method of claim 1 , wherein matching the I/O stream with the first I/O pattern comprises matching parameters of the I/O stream to parameters of the first I/O pattern.

3. The method of claim 2 , wherein the parameters of the I/O stream include at least one item selected from a list consisting of:

a number of read operations that occurred when a portion of the I/O stream was processed;

a number of write operations that occurred when the portion of the I/O stream was processed; and

attributes of a machine learning algorithm used to process the I/O portion.

4. The method of claim 1 , wherein the first cache settings comprise default settings.

5. The method of claim 1 , wherein the result of simulating the caching indicates an item selected from a list consisting of:

that the third cache settings increase efficiency of the caching over the second cache settings;

that the third cache settings increase throughput of the caching over the second cache settings; and

that the third cache settings increase effectiveness of the caching of the second cache settings.

6. The method of claim 1 , wherein adding the entry in the library is based at least in part on determining that the second I/O pattern is not represented in the library.

7. The method of claim 1 , wherein the simulating of the caching of the data is performed in parallel with the caching.

8. The method of claim 1 , wherein the third cache settings specify at least one item selected from a list consisting of:

cache block size;

times that different data types are kept in cache before being flushed to storage devices;

the data is written to the storage devices per volume; and

the data is written to the storage devices per workload.

9. A method performed by a storage system, the method comprising:

matching an input/output (I/O) stream with a first I/O pattern in a library of pre-defined caching policies,

wherein the first I/O pattern is associated with first cache settings in the library;

applying the first cache settings to the I/O stream based on the matching;

simulating caching of data within the I/O stream using a plurality of other cache settings;

updating the library to include an entry as a new cache policy corresponding to second cache settings and a second I/O pattern associated with the I/O stream based on a result of simulating the caching; and

applying the second cache settings to the I/O stream based on the new cache policy.

10. The method of claim 9 , wherein matching the I/O stream with the first I/O pattern comprises matching parameters of the I/O stream to parameters of the first I/O pattern.

11. The method of claim 9 , wherein the result of simulating indicates an item selected from a list consisting of:

that the second cache settings increase efficiency of the caching over the first cache settings;

that the second cache settings increase throughput of the caching over the first cache settings; and

that the second cache settings increase effectiveness of the caching of the first cache settings.

12. The method of claim 9 , wherein updating the library is based at least in part on determining that the second I/O pattern is not represented in the library.

13. The method of claim 9 , wherein the simulating is performed in parallel with caching.

14. The method of claim 9 , wherein the second cache settings specify at least one item selected from a list consisting of:

cache block size;

times that different data types are kept in cache before being flushed to storage devices;

the data is written to the storage devices per volume; and

the data is written to the storage devices per workload.

15. A storage controller comprising:

a memory having stored thereon instructions for performing a method of data caching in a storage system; and

a processor coupled to the memory, the processor configured to execute the instructions to cause the processor to:

cache data from an input/output (I/O) stream using first cache settings;

simulate the caching of the data from the input/output (I/O) stream according to a plurality of different cache settings;

create an entry in a policy library as a new cache policy corresponding to second cache settings and an I/O pattern associated with the I/O stream based on a result of the simulating the caching; and

apply the second cache settings in the storage system based on the new cache policy.

16. The storage controller of claim 15 , wherein the processor is further configured to execute the instructions to cause the processor to:

identify the first cache settings, including matching parameters of the I/O stream to parameters of a policy in the policy library.

17. The storage controller of claim 15 , wherein the processor is further configured to execute the instructions to cause the processor to:

generate the new cache policy based at least in part on determining that the I/O pattern of the I/O stream is not represented in the policy library.

18. The storage controller of claim 15 , wherein the processor is configured to simulate the caching in parallel with the caching the data.

19. The storage controller of claim 15 , wherein the second cache settings specify at least one item selected from a list consisting of:

cache block size;

times that different data types are kept in cache before being flushed to storage devices;

the data is written to the storage devices per volume; and

the data is written to the storage devices per workload.

20. The storage controller of claim 15 , wherein the result of simulating indicates an item selected from a list consisting of:

that the second cache settings increase efficiency of the caching over the first cache settings;

that the second cache settings increase throughput of the caching over the first cache settings; and

that the second cache settings increase effectiveness of the caching of the first cache settings.