Techniques for automated discovery of storage devices and their performance characteristics

• US 8,868,797 B1
• Filed: 03/30/2012
• Issued: 10/21/2014
• Est. Priority Date: 03/30/2012
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Original Assignee: EMC Corporation (Dell Technologies Inc.)
• Inventors: Kirac, Ahmet, Sahin, Adnan, Naamad, Amnon
• Primary Examiner(s): Tsai, Henry
• Assistant Examiner(s): BORROMEO, JUANITO C
• Application Number:
  US13/435,462
• Time in Patent Office: 935 Days
• Field of Search: None
• US Class Current: 710/15
• CPC Class Codes: G06F 3/061 Improving I/O performance
• CPC Class Codes: G06F 3/0631 by allocating resources to storage systems
• CPC Class Codes: G06F 3/0685 Hybrid storage combining heterogeneous device types, e.g. hierarchical storage, hybrid arrays

Abstract

Described are techniques for discovering physical device dependencies. A set of baseline response times for logical devices of a data storage system are received. A first of the logical devices is selected and a remaining set of logical devices includes each of the logical devices other than the first logical device. Data operations are issued to the first logical device at a first rate while also issuing data operations to the remaining logical devices. A set of observed response times for the remaining logical devices is received. For each remaining logical device, it is determined whether the logical device has a corresponding observed response time that is significantly more the logical device'"'"'s baseline response time. A first logical device set is determined having physical device dependencies that includes the first logical device set and each remaining logical device having an observed response time significantly greater than it'"'"'s baseline response time.

Claims

1. A method for discovering physical device dependencies between logical devices comprising:

receiving, for a plurality of logical devices of a data storage system, a set of baseline response times obtained when issuing data operations to the plurality of logical devices, each of said plurality of logical devices having storage provisioned on one or more physical devices;

selecting a first of the plurality of logical devices, wherein a remaining set of logical devices includes each of the plurality of logical devices other than the first logical device;

issuing data operations to the first logical device at a first rate while also issuing data operations to each logical device of the remaining set, wherein said first rate is sufficient to result in a first observed response time for the first logical device that is significantly more than a corresponding one of the baseline response times for the first logical device, and wherein said first rate is sufficient to result in an observed response time for each logical device of the remaining set having a portion of data stored on a same physical device as a portion of data of the first logical device whereby the observed response time for said each logical device of the remaining set is significantly more than a corresponding one of the baseline response times for said each logical device;

receiving a set of observed response times for the remaining set of logical devices;

determining, for each logical device of the remaining set of logical devices, whether said each logical device has a corresponding one of the observed response times of the set that is significantly more than a corresponding one of the baseline response times for said each logical device; and

determining a first logical device set of one or more logical devices having physical device dependencies, said first logical device set including the first logical device and each logical device of the remaining set determined to have a corresponding observed response time of the set of observed response times that is significantly greater than a corresponding one of the baseline response times for said each logical device.

2. The method of claim 1, wherein each logical device of the first logical device set has at least a portion of data stored on a physical device of a same physical device group of one or more physical devices.

3. The method of claim 2, wherein each logical device of the first logical device set other than the first logical device has at least some data stored on a same physical device as at least some data of the first logical device.

4. The method of claim 1, wherein said set of observed response times are obtained using data collected while performing said issuing data operations to the first logical device and while also performing said issuing data operations to each logical device in the remaining set.

5. The method of claim 1, wherein said issuing data operations to each logical device of the remaining set issues data operations to said each logical device of the remaining set at about a same rate as when a corresponding one of the baseline response times is obtained for said each logical device.

6. The method of claim 1, wherein the set of observed response times is an adjusted set of observed response times determined by adjusting an original set of observed response times using the set of baseline response times.

7. The method of claim 1, wherein said first rate is higher than a rate at which data operations are issued to the first logical device in obtaining the set of baseline response times.

8. The method of claim 1, wherein said set of baseline response times are obtained by maintaining a same average queue depth for each of the plurality of logical devices.

9. The method of claim 8, wherein said average queue depth for said each logical device represents an average number of I/O operations outstanding for said each logical device.

10. The method of claim 9, wherein said average queue depth maintained for each of the plurality of logical devices in obtaining the set of baseline response times is a first average queue depth and said issuing data operations to each logical device of the remaining set is at about a same rate as when a corresponding one of the baseline response times is obtained for said each logical device.

11. The method of claim 10, wherein said first rate is determined in accordance with maintaining a second average queue depth for the first logical device, the second average queue depth being larger than said first average queue depth.

12. The method of claim 11, wherein said first average queue depth is equal to or less than 1 and said second average queue depth is 32.

13. The method of claim 1, wherein each logical device of the first logical device set has a corresponding observed response time that is greater than a corresponding one of the baseline response times for said each logical device by a threshold amount determined in accordance with a threshold factor and the corresponding one baseline response time for said each logical device.

14. The method of claim 13, wherein the threshold factor is any of one and a half or twice the corresponding one baseline response time for said each logical device thereby resulting in a varying threshold amount for each of the plurality of logical devices, wherein each logical device of the first logical device set has a corresponding observed response time that is greater than a corresponding one of the baseline response times for said each logical device by an amount, said amount being at least the threshold factor multiplied by the corresponding one baseline response time for said each logical device.

15. The method of claim 1, wherein logical devices of the first logical device set have data stored on a first physical device group of one or more physical devices and the method further comprising:

determining a set of one or more aggregated response times for said first physical device group using response times for the first logical device set; and

classifying said one or more physical devices of the first physical device group as being physical devices of a first of a plurality of storage device tiers, said classifying using information including the set of one or more aggregated response times in comparison to expected response times for physical devices of each of the plurality of storage tiers.

16. The method of claim 1, wherein the method is executed on any of a host to determine information about a data storage system, a first data storage system to determine information about the plurality of logical devices included in a second data storage system, and an appliance presenting to a client a set of virtualized storage devices having storage provisioned across a plurality of physical devices on one or more data storage systems.

17. The method of claim 1, wherein a current set of logical devices is determined by removing the first logical device set from the plurality of logical devices, and the method includes performing said selecting, said issuing data operations, said receiving, said first determining and said second determining with respect to the current set of logical devices rather than the plurality of logical devices.

18. A non-transitory computer readable medium comprising code stored thereon for discovering physical device dependencies between logical devices, the non-transitory computer readable medium comprising code stored thereon that, when executed by a processor, performs a method comprising:

receiving, for a plurality of logical devices of a data storage system, a set of baseline response times obtained when issuing data operations to the plurality of logical devices, each of said plurality of logical devices having storage provisioned on one or more physical devices;

selecting a first of the plurality of logical devices, wherein a remaining set of logical devices includes each of the plurality of logical devices other than the first logical device;

issuing data operations to the first logical device at a first rate while also issuing data operations to each logical device of the remaining set, wherein said first rate is sufficient to result in a first observed response time for the first logical device that is significantly more than a corresponding one of the baseline response times for the first logical device, and wherein said first rate is sufficient to result in an observed response time for each logical device of the remaining set having a portion of data stored on a same physical device as a portion of data of the first logical device whereby the observed response time for said each logical device of the remaining set is significantly more than a corresponding one of the baseline response times for said each logical device;

receiving a set of observed response times for the remaining set of logical devices;

determining, for each logical device of the remaining set of logical devices, whether said each logical device has a corresponding one of the observed response times of the set that is significantly more than a corresponding one of the baseline response times for said each logical device; and

determining a first logical device set of one or more logical devices having physical device dependencies, said first logical device set including the first logical device and each logical device of the remaining set determined to have a corresponding observed response time of the set of observed response times that is significantly greater than a corresponding one of the baseline response times for said each logical device.

19. The non-transitory computer readable medium of claim 18, wherein each logical device of the first logical device set has at least a portion of data stored on a physical device of a same physical device group of one or more physical devices.

20. A system comprising:

a set of one or more data storage systems; and

a first component in communication with the one or more data storage systems, the first component including a computer readable medium comprising code stored thereon that, when executed by a processor, performs a method comprising;

receiving, for a plurality of logical devices of the one or more data storage systems, a set of baseline response times obtained when issuing data operations to the plurality of logical devices, each of said plurality of logical devices having storage provisioned on one or more physical devices;

selecting a first of the plurality of logical devices, wherein a remaining set of logical devices includes each of the plurality of logical devices other than the first logical device;

issuing data operations to the first logical device at a first rate while also issuing data operations to each logical device of the remaining set, wherein said first rate is sufficient to result in a first observed response time for the first logical device that is significantly more than a corresponding one of the baseline response times for the first logical device, and wherein said first rate is sufficient to result in an observed response time for each logical device of the remaining set having a portion of data stored on a same physical device as a portion of data of the first logical device whereby the observed response time for said each logical device of the remaining set is significantly more than a corresponding one of the baseline response times for said each logical device;

receiving a set of observed response times for the remaining set of logical devices;

determining, for each logical device of the remaining set of logical devices, whether said each logical device has a corresponding one of the observed response times of the set that is significantly more than a corresponding one of the baseline response times for said each logical device; and

determining a first logical device set of one or more logical devices having physical device dependencies, said first logical device set including the first logical device and each logical device of the remaining set determined to have a corresponding observed response time of the set of observed response times that is significantly greater than a corresponding one of the baseline response times for said each logical device.