Method to efficiently track I/O access history using efficient memory data structures

• US 9,798,754 B1
• Filed: 06/12/2014
• Issued: 10/24/2017
• Est. Priority Date: 06/12/2014
• Status: Active Grant

• Current Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.), Telefonaktiebolaget LM Ericsson
• Original Assignee: Emc IP Holding Company LLC (Dell Technologies Inc.)
• Inventors: Shilane, Philip, Wallace, Grant
• Primary Examiner(s): Perveen, Rehana
• Assistant Examiner(s): Skhoun, Hicham
• Application Number:
  US14/303,427
• Time in Patent Office: 1,230 Days
• Field of Search: 707745
• CPC Class Codes: G06F 11/3037 where the computing system component is a memory, e.g. virtual memory, cache accessing, addressing or allocating within memory systems or architectures G06F12/00; checking stores for correct operation G11C29/00
• CPC Class Codes: G06F 11/3476 Data logging G06F11/14, G06F11/2205 take precedence
• CPC Class Codes: G06F 11/3485 for I/O devices
• CPC Class Codes: G06F 12/0862 with prefetch
• CPC Class Codes: G06F 12/0888 using selective caching, e.g. bypass
• CPC Class Codes: G06F 12/127 using additional replacement algorithms
• CPC Class Codes: G06F 16/2237 Vectors, bitmaps or matrices
• CPC Class Codes: G06F 2201/885 Monitoring specific for caches

Abstract

An embodiment is described in which a memory device stores a record of I/O accesses to data blocks. And each access record indicates which data block was accessed and during which time period the access occurred. A memory-efficient data structure (MEDS) may be generated and stored in a cache or storage device and the access data moved from the memory device into the MEDS. The MEDS represents blocks that were accessed during a particular time period. When a second data block is accessed, a query function is applied to the second block'"'"'s identifier to return a value based on data stored in the MEDS. The return value from the query function indicates whether the second data block was accessed during the particular time period associated with the MEDS. A storage management action is performed based on whether the second data block was accessed during the particular time period.

Claims

1. A computer-implemented method for determining access patterns of data stored in a data processing system, the method comprising:

responsive to receiving a request to access a first data block, recording in a first memory device an access indicator that indicates a first identifier, of the first data block, in association with a time period at which the first data block was accessed, wherein the access indicator is a bloom filter, and a plurality of functions of an identifier of a particular data block are used to identify a corresponding plurality of bit positions in an access bitmap, a value of each bit at the plurality of bit positions together indicating access of the particular data block;

generating a memory-efficient data structure (MEDS) based on a plurality of access indicators recorded in the first memory device, representing an access pattern corresponding to the time period during which the first data block was accessed;

compressing access information from the plurality of access indicators, wherein the MEDS occupies less storage space on a second memory device as compared to an amount of storage space that the plurality of access indicators occupy on the first memory device;

subsequently in response to a request for accessing a second data block, uncompressing the access information stored in the MEDS before applying a query function to a second identifier of the second data block, wherein the query function returns a value based on data stored in the MEDS to indicate whether the second data block is associated with an access pattern represented by the MEDS; and

Performing a storage management action based on the value returned by the query function, the storage management action including pre-fetching a data block, caching a data block, or evicting a data block, depending on the value returned by the query function.

2. The method of claim 1, wherein generating the MEDS representing a particular access time period is performed offline.

3. The method of claim 1, wherein the access indicator is the access bitmap.

4. The method of claim 3, wherein:

each bit in the access bitmap indicates access of the particular data block, the particular data block occupying a particular size in the first memory device.

5. The method of claim 1, wherein:

each access indicator indicates access of a data block of a first particular size; and

the MEDS includes a plurality of data elements, each data element having a value that indicates access of a data block of a second particular size; and

the second particular size is larger than the first particular size.

6. The method of claim 1, further comprising:

selecting one or more sets of data blocks, wherein each set of data blocks comprises a plurality of data blocks, wherein generating the memory-efficient data structure (MEDS) is performed based on access indicators indicating access of the selected one or more sets of data blocks, and wherein the MEDS does not include an indication of access for at least one data block that has an access indicator stored in the first memory device.

7. The method of claim 1, wherein the MEDS is stored to a second memory device different than the first memory device comprising non-volatile storage and can be queried at a future time.

8. The method of claim 7, wherein the non-volatile storage is a database.

9. The method of claim 1, wherein the first memory device is a volatile random access memory device.

10. A non-transitory computer-readable storage medium storing instructions for determining access patterns of data stored in a data processing system, the instructions that when executed by one or more processors, cause the one or more processors to:

record, responsive to receipt of a request to access a first data block, in a first memory device an access indicator that indicates a first identifier, of the first data block, in association with a time period at which the first data block was accessed, wherein the access indicator is a bloom filter, and a plurality of functions of an identifier of a particular data block are used to identify a corresponding plurality of bit positions in an access bitmap, a value of each bit at the plurality of bit positions together indicating access of the particular data block;

generate a memory-efficient data structure (MEDS) based on a plurality of access indicators recorded in the first memory device, the MEDS representing an access pattern corresponding to the time period during which the first data block was accessed;

compress access information from the plurality of access indicators, wherein the MEDS occupies less storage space on a second memory device as compared to an amount of storage space that the plurality of access indicators occupy on the first memory device;

uncompress the access information stored in the MEDS before apply a query function, in response to a request for access to a second data block, to a second identifier of the second data block, wherein the query function returns based on data stored in the MEDS, an indication of whether the second data block was accessed during the time period associated with the MEDS; and

Perform a storage management action based on the indication returned by the query function, the storage management action including pre-fetching a data block, caching a data block, or evicting a data block, depending on the value returned by the query function.

11. The non-transitory computer-readable storage medium of claim 10, wherein the MEDS representing a particular access time period is generated offline.

12. The non-transitory computer-readable storage medium of claim 10, wherein the access indicator is the access bitmap.

13. The non-transitory computer-readable storage medium of claim 12, wherein:

each bit in the access bitmap indicates access of the particular data block, the particular data block occupies a particular size in the first memory device.

14. The non-transitory computer-readable storage medium of claim 10, wherein:

each access indicator indicates access of a data block of a first particular size; and

the MEDS includes a plurality of data elements, each data element having a value that indicates access of a data block of a second particular size; and

the second particular size is larger than the first particular size.

15. The non-transitory computer-readable storage medium of claim 10, the instructions further cause the one or more processors to:

select one or more sets of data blocks, wherein each set of data blocks comprises a plurality of data blocks, wherein generation of the memory-efficient data structure (MEDS) is performed based on access indicators that indicate access of the selected one or more sets of data blocks, and wherein the MEDS does not include an indication of access for at least one data block that has an access indicator stored in the first memory device.

16. The non-transitory computer-readable storage medium of claim 10, wherein the MEDS is stored to a second memory device different than the first memory device comprising non-volatile storage and can be queried at a future time.