System and Method for Maintaining Cache Coherency

• US 20150161048A1
• Filed: 12/09/2013
• Published: 06/11/2015
• Est. Priority Date: 12/09/2013
• Status: Active Grant

• Current Assignee: PernixData, Inc. (Nutanix Incorporated)
• Original Assignee: PernixData, Inc. (Nutanix Incorporated)
• Inventors: Vaghani, Satyam B., Patil, Kaustubh Sambhaji, Vilayannur, Murali Natarajan, Ostrowski, Michal, Jung, Woon Ho, Kumar, Poojan
• Granted Patent: US 9,489,389 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 12/0871 Allocation or management of cache space
• CPC Class Codes: G06F 16/172 Caching, prefetching or hoarding of files
• CPC Class Codes: G06F 2212/1016 Performance improvement
• CPC Class Codes: G06F 2212/465 Structured object, e.g. database record
• CPC Class Codes: G06F 2212/466 Metadata, control data

Abstract

A cache manager maintains coherency of cached objects used by applications in a computing system. The cache manager maintains a cache memory of objects as they are accessed by an application. The cache manager tracks cached objects on a per-application basis using a cache database, storing for each object metadata comprising the generation number of the application using the cached object, and the file generation number of the file from which the object was read. The cache manager determines whether cached objects are valid and can be used by the application, or are stale and must be reloaded, by using these generation numbers when an application attempts to use a file or object. In this fashion, cache coherency is maintained even in a distributed environment without the centralized cache management bottlenecks of prior approaches and without the need to eagerly purge the cache of stale versions of objects.

Claims

1. In a computing system having a cache memory and a storage device, a cache coherency method comprising:

intercepting, by a cache manager running on the computing system, a read request from an application running on the computing system;

determining, by the cache manager from the intercepted read request, an identifier for a required object and a required generation number; and

determining by the cache manager;

if the required object is not stored in the cache memory, then storing in the cache memory a copy of the required object retrieved from the storage device, adding metadata regarding the stored object to the cache database, the stored object metadata comprising an object generation number equal to the application generation number, adding a reference to the stored object to a list of references to object stored in the cache memory that are being used by the application, and returning the copy of the object to the application,

else

if the required object is stored in the cache memory, and the stored object is of a generation number that is not less than the required generation number, then adding a reference to the stored object to the list of references to objects stored in the cache memory that are being used by the application, and returning the stored object from the cache memory to the application,

else

if the required object is stored in the cache memory, and the stored object is of a generation number that is less than the required generation number, then storing in the cache memory a copy of the required object retrieved from the storage device, adding metadata regarding the stored object to the cache database, the stored object metadata comprising an object generation number equal to the application generation number, adding a reference to the stored object to a list of references to objects stored in the cache memory that are being used by the application, and returning the copy of the object to the application.

2. The method of claim 1 wherein the required generation number is the application generation number.

3. The method of claim 1 where the storage device is coupled to the computing system over a computer network.

4. The method of claim 1 where the cache memory is distributed over a plurality of computing systems in a network.

5. The method of claim 4 where the cache memory distributed over the plurality of computing systems in a network is maintained as one cache.

6. The method of claim 4 where the cache manager is distributed over a plurality of computing systems in a network.

7. The method of claim 1 where the application is a virtual machine.

8. The method of claim 1 wherein adding metadata regarding the stored object to the cache database further comprises adding file system metadata about a file containing the stored object, the file system metadata comprising a file identifier and a file generation number, and the method further comprising, prior to the step of intercepting the read request:

intercepting, by the cache manager running on the computing system, a file open request for a file object on the storage device from an application running on the computing system;

obtaining, by the cache manager, a file open response containing a file identifier and a file generation number for the requested file object from the storage device, and

determining by the cache manager;

if the list of references to objects stored in the cache memory that are being used by the application does not contain a reference to a cached object from the file object, then returning the file open response to the application,

else

if the file generation number contained in the file open response matches the file generation number stored as file system metadata for the cached object, then returning the file open response to the application,

else

incrementing the application generation number and returning the file open response to the application.

9. The method of claim 1 wherein adding metadata regarding the stored object to the cache database further comprises adding storage system metadata about a data object containing the stored object, the storage system metadata comprising a data object identifier and a data object generation number, and the method further comprising, prior to the step of intercepting the read request:

intercepting, by the cache manager running on the computing system, a data object access request for a data object on the storage device from an application running on the computing system;

obtaining, by the cache manager, a data object access response containing a data object identifier and a data object generation number for the requested data object from the storage device, and

determining by the cache manager;

if the list of references to objects stored in the cache memory that are being used by the application does not contain a reference to a cached object from the data object, then returning the data object access response to the application,

else

if the data object generation number contained in the data object access response matches the cached data object generation number stored as data object metadata for the cached object, then returning the data object access response to the application,

else

incrementing the application generation number and returning the data object access response to the application.

10. A system for maintaining cache coherency in a computing system comprising:

a cache memory;

a cache database;

a storage device; and

a cache manager running on the computing system, the cache manager configured to;

intercept a read request from an application running on the computing system;

determine from the intercepted read request an identifier for a required object and a required generation number; and

determine;

if the required object is not stored in the cache memory, then store in the cache memory a copy of the required object retrieved from the storage device, add metadata regarding the stored object to the cache database, the stored object metadata comprising an object generation number equal to the application generation number, add a reference to the stored object to a list of references to objects stored in the cache memory that are being used by the application, and return the copy of the object to the application,

else

if the required object is stored in the cache memory, and the stored object is of a generation number that is not less than the required generation number, then add a reference to the stored object to the list of references to objects stored in the cache memory that are being used by the application, and return the stored object from the cache memory to the application,

else

if the required object is stored in the cache memory, and the stored object is of a generation number that is less than the required generation number, then store in the cache memory a copy of the required object retrieved from the storage device, add metadata regarding the stored object to the cache database, the stored object metadata comprising an object generation number equal to the application generation number, add a reference to the stored object to the list of references to objects stored in the cache memory that are being used by the application, and return the copy of the object to the application.

11. The system of claim 10 where the cache manager is configured to set the required generation number to the application generation number.

12. The system of claim 10 where the storage device is coupled to the computing system via a computer network.

13. The system of claim 10 wherein the storage device is located in the computing system.

14. The system of claim 10 where the application is a virtual machine.

15. The system of claim 10 wherein the cache manager is configured to add metadata regarding the stored object to the cache database by being further configured to add file system metadata regarding the stored object to the cache database, the file system metadata comprising a file identifier and a file generation number, and wherein the cache manager is further configured to:

before intercepting the read request,

intercept a file open request for a file object on the storage device from the application running on the computing system;

obtain a file open response containing a file identifier and a file generation number for the requested file object from the storage device, and

determine;

if the list of references to objects stored in the cache memory that are being used by the application does not contain a reference to a cached object from the file object, then return the file open response to the application,

else

if the file generation number contained in the file open response matches the file generation number stored as file system metadata for the cached object, then return the file open response to the application,

else

increment the application generation number and return the file open response to the application.

16. The system of claim 10 wherein the cache manager is configured to add metadata regarding the stored object to the cache database by being further configured to add storage system metadata regarding the stored object to the cache database, the storage system metadata comprising a data object identifier and a data object generation number;

and wherein where the cache manager is further configured to;

before intercepting the read request,

intercept a data object access request for a data object on the storage device from the application running on the computing system;

obtain a data object access response containing a data object identifier and a data object generation number for the requested data object from the storage device, and

determine;

if the list of references to objects stored in the cache memory that are being used by the application does not contain a reference to a cached object from the data object, then return the data object access response to the application,

else

if the data object generation number contained in the data object open response matches the data object generation number stored as storage system metadata for the cached object, then return the data object access response to the application,

else

increment the application generation number and return the data object access response to the application.

17. A non-transitory computer readable medium having stored thereupon computing instructions comprising:

a code segment to intercept a read request from an application running on a computing system;

a code segment to determine from the intercepted read request, an identifier for a required object and a required generation number; and

a code segment to determine;

if the required object is not stored in a cache memory, then store in the cache memory a copy of the required object retrieved from a storage device, add metadata regarding the stored object to the cache database, the stored object metadata comprising an object generation number equal to the application generation number, add a reference to the object to a list of references to objects stored in the cache memory that are being used by the application, and return the copy of the object to the application,

else

if the required object is stored in the cache memory, and the stored object is of a generation number that is not less than the required generation, then add a reference to the stored object to the list of references to objects stored in the cache memory that are being used by the application, and return the stored object from the cache memory to the application,

else

if the required object is stored in the cache memory, and the stored object is of a generation number that is less than the required generation number, store in the cache memory a copy of the required object retrieved from a storage device, add metadata regarding the stored object to the cache database, the stored object metadata comprising an object generation number equal to the application generation number, add a reference to the object to the list of references to objects stored in the cache memory that are being used by the application, and return the copy of the object to the application.

18. The non-transitory computer readable medium of claim 17 where the code segment to determine from the intercepted read request a required generation number comprises a code segment to set the required generation number to the application generation number.

19. The non-transitory computer readable medium of claim 17 where the application is a virtual machine.

20. The non-transitory computer readable medium of claim 17 wherein the code segment to add metadata regarding the stored object to the cache database further includes a code segment to:

add file system metadata about a file containing the stored object, the file system metadata comprising a file identifier and a file generation number;

and the non-transitory computer readable medium further comprises a code segment to;

before intercepting the read request,

intercept a file open request for a file object on the storage device from the application running on the computing system;

obtain a file open response containing a file identifier and a file generation number for the requested file object from the storage device, and

determine;

if the list of references to objects stored in the cache memory that are being used by the application does not contain a reference to a cached object from the file object, then return the file open response to the application,

else

if the file generation number contained in the file open response matches the file generation number stored as file system metadata for the cached object, then return the file open response to the application,

else

increment the application generation number and return the file open response to the application.

21. The non-transitory computer readable medium of claim 17 wherein the code segment to add metadata regarding the stored object to the cache database further includes a code segment to add storage system metadata about a data object containing the stored object, the storage system metadata comprising a data object identifier and a data object generation number;

and the non-transitory computer readable medium further includes a code segment to;

before intercepting the read request,

intercept a data object access request for a data object on the storage device from the application running on the computing system;

obtain a data object access response containing a data object identifier and a data object generation number for the requested data object from the storage device, and

determine;

if the list of references to objects stored in the cache memory that are being used by the application does not contain a reference to a cached object from the data object, then return the data object access response to the application,

else

if the data object generation number contained in the data object open response matches the data object generation number stored as storage system metadata for the cached object, then return the data object access response to the application,

else

increment the application generation number and return the data object access response to the application.