Techniques for storing data
Described are techniques for storing data. A write operation is received that is directed to a target logical address of a logical device where physical storage has not been allocated for the target logical address. The logical device has a service level objective denoting a performance goal to be maintained with respect to I/O operations for the logical device. A storage tier is selected from a plurality of storage tiers from which to allocate storage for storing write data of the write operation. The selecting is performed in accordance with determining a match between an expected level of performance of the storage tier and the performance goal. First physical storage is allocated from the storage tier. A logical address subrange of the logical device is mapped to the first physical storage. The write data is stored in the first physical storage at a location mapped to the target logical address.
1. A method of storing data comprising:
receiving a write operation directed to a target logical address of a logical device where physical storage has not been allocated for the target logical address, said logical device having a service level objective denoting a performance goal to be maintained with respect to I/O operations for the logical device;
selecting a storage tier from a plurality of storage tiers from which to allocate storage for storing write data of the write operation, wherein said selecting is performed in accordance with determining a match between an expected level of performance of said storage tier of physical storage devices having a same physical device type and said performance goal of the logical device as denoted by the service level objective, wherein said selecting includes:
determining whether the logical device has any associated service level objective denoting a performance goal to be maintained with respect to I/O operations for the logical device; and
responsive to determining that the logical device does not have any associated service level objective, using a default ranking of the plurality of storage tiers to select the storage tier as a highest ranked one of the plurality of storage tiers having available capacity, wherein said plurality of tiers are configured into RAID groups having write performance penalties, wherein said default ranking of the plurality of storage tiers is ranked in accordance with one or more primary criteria including media type characteristics and one or more secondary criteria including the write performance penalties associated with different RAID types and configurations, wherein each of the different RAID types with an associated configuration has an associated one of the write performance penalties denoting a number of additional back-end I/O operations incurred when issuing a single write operation that stores data in a RAID group configured with said each RAID type and associated configuration;
allocating first physical storage from the storage tier;
mapping a logical address subrange of the logical device to the first physical storage; and
storing the write data in the first physical storage at a physical storage location mapped to the target logical address.
2. The method of claim 1 , wherein each of the plurality of storage tiers has characteristics denoting a media type of physical storage devices of said each tier and denoting a RAID type and configuration of physical storage devices.
3. The method of claim 2 , wherein the media type associated with said each storage tier denotes an expected level of performance with respect to I/O operations directed to said each tier.
4. The method of claim 1 , wherein the performance goal of the service level objective and the expected level of performance of said storage tier are response times.
5. The method of claim 2 , wherein said plurality of storage tiers includes a first storage tier having first characteristics identifying a media type with an expected level of performance matching the performance goal of the service level objective and identifying a first RAID type and configuration, and a second storage tier having second characteristics identifying a media type with an expected level of performance matching the performance goal of the service level objective and identifying a second RAID type and configuration, and wherein said selecting further comprises:
determining a first write penalty associated with the first RAID type and configuration;
determining a second write penalty associated with the second RAID type and configuration; and
selecting the first storage tier rather than the second storage tier from which to allocate storage for storing the write data if the first write penalty is less than said second write penalty, otherwise selecting the second storage tier rather than the first storage tier.
6. The method of claim 1 , wherein the storage tier is characterized by a first media type denoting that the storage tier includes physical devices of the first media type and wherein the storage tier is characterized by a first RAID type and configuration denoting that physical devices of the storage tier are configured into one or more RAID groups of the first RAID type and configuration, and the method further comprising:
collecting I/O workload information for a data portion including the target logical address and the logical address subrange of the logical device;
determining, based on the I/O workload information, a preferred media type and a preferred RAID type and configuration for the data portion;
determining, for first data of the data portion currently stored in the storage tier, whether said first RAID type and configuration matches the preferred RAID type and configuration, and whether said first media type matches said preferred media type;
responsive to determining that any of said first RAID type and configuration does not match the preferred RAID type and configuration, and said first media type does not match said preferred media type, selecting the first data of the data portion to be moved from the storage tier to a target storage tier.
7. The method of claim 6 , wherein the target storage tier has said preferred RAID type and configuration and said preferred media type.
8. The method of claim 6 , wherein the I/O workload information includes a write probability for said data portion.
9. The method of claim 6 , wherein the I/O workload information includes an I/O density for the data portion.
10. The method of claim 6 , wherein the I/O workload information includes a first metric based on I/O throughput for the data portion.
11. The method of claim 6 , wherein the I/O workload information includes a second metric based on an amount of data read in connection with I/Os directed to the data portion and based on an amount of data written in connection with I/Os directed to the data portion.
12. A system comprising:
a processor; and
a memory comprising code stored therein that, when executed, performs a method comprising:
receiving a write operation directed to a target logical address of a logical device where physical storage has not been allocated for the target logical address, said logical device having a service level objective denoting a performance goal to be maintained with respect to I/O operations for the logical device;
selecting a storage tier from a plurality of storage tiers from which to allocate storage for storing write data of the write operation, wherein said selecting is performed in accordance with determining a match between an expected level of performance of said storage tier of physical storage devices having a same physical device type and said performance goal of the logical device as denoted by the service level objective, wherein said selecting includes:
determining whether the logical device has any associated service level objective denoting a performance goal to be maintained with respect to I/O operations for the logical device; and
responsive to determining that the logical device does not have any associated service level objective, using a default ranking of the plurality of storage tiers to select the storage tier as a highest ranked one of the plurality of storage tiers having available capacity, wherein said plurality of tiers are configured into RAID groups having write performance penalties, wherein said default ranking of the plurality of storage tiers is ranked in accordance with one or more primary criteria including media type characteristics and one or more secondary criteria including the write performance penalties associated with different RAID types and configurations, wherein each of the different RAID types with an associated configuration has an associated one of the write performance penalties denoting a number of additional back-end I/O operations incurred when issuing a single write operation that stores data in a RAID group configured with said each RAID type and associated configuration;
allocating first physical storage from the storage tier;
mapping a logical address subrange of the logical device to the first physical storage; and
storing the write data in the first physical storage at a physical storage location mapped to the target logical address.
13. A non-transitory computer readable medium comprising code stored thereon that, when executed, performs a method that stores data comprising:
receiving a write operation directed to a target logical address of a logical device where physical storage has not been allocated for the target logical address, said logical device having a service level objective denoting a performance goal to be maintained with respect to I/O operations for the logical device;
selecting a storage tier from a plurality of storage tiers from which to allocate storage for storing write data of the write operation, wherein said selecting is performed in accordance with determining a match between an expected level of performance of said storage tier of physical storage devices having a same physical device type and said performance goal of the logical device as denoted by the service level objective, wherein said selecting includes:
determining whether the logical device has any associated service level objective denoting a performance goal to be maintained with respect to I/O operations for the logical device; and
responsive to determining that the logical device does not have any associated service level objective, using a default ranking of the plurality of storage tiers to select the storage tier as a highest ranked one of the plurality of storage tiers having available capacity, wherein said plurality of tiers are configured into RAID groups having write performance penalties, wherein said default ranking of the plurality of storage tiers is ranked in accordance with one or more primary criteria including media type characteristics and one or more secondary criteria including the write performance penalties associated with different RAID types and configurations, wherein each of the different RAID types with an associated configuration has an associated one of the write performance penalties denoting a number of additional back-end I/O operations incurred when issuing a single write operation that stores data in a RAID group configured with said each RAID type and associated configuration;
allocating first physical storage from the storage tier;
mapping a logical address subrange of the logical device to the first physical storage; and
storing the write data in the first physical storage at a physical storage location mapped to the target logical address.
14. The non-transitory computer readable medium of claim 13 , wherein each of the plurality of storage tiers has characteristics denoting a media type of physical storage devices of said each tier and denoting a RAID type and configuration of physical storage devices.
15. The non-transitory computer readable medium of claim 14 , wherein the media type associated with said each storage tier denotes an expected level of performance with respect to I/O operations directed to said each tier.
16. The non-transitory computer readable medium of claim 13 , wherein the performance goal of the service level objective and the expected level of performance of said storage tier are response times.
17. The non-transitory computer readable medium of claim 14 , wherein said plurality of storage tiers includes a first storage tier having first characteristics identifying a media type with an expected level of performance matching the performance goal of the service level objective and identifying a first RAID type and configuration, and a second storage tier having second characteristics identifying a media type with an expected level of performance matching the performance goal of the service level objective and identifying a second RAID type and configuration, and wherein said selecting further comprises:
determining a first write penalty associated with the first RAID type and configuration;
determining a second write penalty associated with the second RAID type and configuration; and
selecting the first storage tier rather than the second storage tier from which to allocate storage for storing the write data if the first write penalty is less than said second write penalty, otherwise selecting the second storage tier rather than the first storage tier.
18. The non-transitory computer readable medium of claim 13 , wherein the storage tier is characterized by a first media type denoting that the storage tier includes physical devices of the first media type and wherein the storage tier is characterized by a first RAID type and configuration denoting that physical devices of the storage tier are configured into one or more RAID groups of the first RAID type and configuration, and the method further comprising:
collecting I/O workload information for a data portion including the target logical address and the logical address subrange of the logical device;
determining, based on the I/O workload information, a preferred media type and a preferred RAID type and configuration for the data portion;
determining, for first data of the data portion stored in the storage tier, whether said first RAID type and configuration matches the preferred RAID type and configuration, and whether said first media type matches said preferred media type;
responsive to determining that any of said first RAID type and configuration does not match the preferred RAID type and configuration, and said first media type does not match said preferred media type, selecting the first data of the data portion to be moved from the storage tier to a target storage tier.
19. The non-transitory computer readable medium of claim 18 , wherein the target storage tier has said preferred RAID type and configuration and said preferred media type.
20. The non-transitory computer readable medium of claim 18 , wherein the I/O workload information includes a write probability for said data portion and an I/O density for the data portion.