Multi-tier cache and method for implementing such a system

• US 5,787,466 A
• Filed: 05/01/1996
• Issued: 07/28/1998
• Est. Priority Date: 05/01/1996
• Status: Expired due to Term

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Berliner, Brian
• Primary Examiner(s): Swann, Tod R.
• Assistant Examiner(s): KING, CONLEY BURRELL
• Application Number:
  US08/641,653
• Time in Patent Office: 818 Days
• Field of Search: 395/440, 395/449, 395/460, 395/461, 395/463, 395/487, 395/486, 711/113, 711/122, 711/133, 711/134, 711/136, 711/159, 711/160, 711/117
• US Class Current: 711/117
• CPC Class Codes: G06F 12/0868 Data transfer between cache memory and other subsystems, e.g. storage devices or host systems
• CPC Class Codes: G06F 12/0897 with two or more cache hierarchy levels with multilevel cache hierarchies G06F12/0811
• CPC Class Codes: G06F 2212/221 Static RAM
• CPC Class Codes: G06F 2212/224 Disk storage

Abstract

A multi-tier cache system and a method for implementing the multi-tier cache system is disclosed. The multi-tier cache system has a small cache in random access memory (RAM) that is managed in a Least Recent Used (LRU) fashion. The RAM cache is a subset of a much larger non-volatile cache on rotating magnetic media (e.g., a hard disk drive). The non-volatile cache is, in turn a subset of a local CD-ROM or of a CD-ROM or mass storage device controlled by a server system. In a preferred embodiment of the invention, a heuristic technique is employed to establish a RAM cache of optimum size within the system memory. Also in a preferred embodiment, the RAM cache is made up of multiple identically-sized sub-blocks. A small amount of RAM is utilized to maintain a table which implements a Least Recently Used (LRU) RAM cache purging scheme. A hashing mechanism is employed to search for the "bucket" within the RAM cache in which the requested data may be located. If the requested data is in the RAM cache, the request is satisfied with that data. If the requested data is not in the RAM cache, the least recently used sub-block is purged from the cache if the cache is full, and the RAM cache is updated from the non-volatile cache whenever possible, and from the cached storage device when the non-volatile cache does not contain the requested data.

Claims

1. A method for implementing a multi-tier caching system, said multi-tier caching system being implemented on a computer system having a random access memory (RAM), a portion of which is set aside and initialized as a RAM cache for storing first-tier data within a plurality of RAM data blocks, and at least one local non-volatile local mass storage device on which data can be alterably stored and on which a portion thereof functions as a non-volatile cache for storing second-tier data within non-volatile data blocks, said computer system having a data path to a mass storage device with an access speed slower than that of the local mass storage device, said slower mass storage device storing a full set of file data, said method comprising the steps of:

whenever a received file input/output request involves data resident on said slower mass storage device and the requested data is not available as first-tier data nor available as second-tier data, determining whether or not all RAM data blocks are already filled with first-tier data;

purging first-tier data from a least-recently-used RAM data block if all RAM data blocks are already filled;

retrieving a quantity of third-tier data which includes the requested data;

writing all retrieved third-tier data to a non-volatile data block, as well as to an empty RAM data block; and

resolving the file input/output request from the retrieved third-tier data written to the empty RAM data block.

2. The method of claim 1, which further comprises the step of determining, upon receipt of a file input/output request, whether or not the request involves third-tier data resident on the slower mass storage device.

3. The method of claim 1, wherein said method further comprises the step of determining, for a file input/output request involving data resident on said slower mass storage device, whether or not first tier data which satisfies the request is already resident in one of said RAM data blocks.

4. The method of claim 3, which further comprises the step of resolving the request from such first-tier data if such first-tier data which satisfies the request is already resident in one of said RAM data blocks.

5. The method of claim 1, which, if a received input/output request involves data resident on said slower mass storage device, and if the requested data is not available as first-tier data, but is available as second-tier data, further comprises the steps of:

determining whether or not all RAM data blocks are already filled with first-tier data;

purging a least-recently-used RAM data block if all RAM data blocks are already filled;

retrieving a quantity of second-tier data which includes the requested data;

writing retrieved second-tier data to an empty RAM data block; and

resolving the file input/output request from the retrieved second-tier data written to the empty RAM data block.

6. The method of claim 1, wherein said portion of the RAM is set aside and initialized as a RAM cache at operating system boot-time.

7. The method of claim 1, wherein said RAM data blocks and said non-volatile data blocks are of equal size.

8. The method of claim 1, wherein said slower mass storage device is a CD-ROM drive controlled by said computer system.

9. The method of claim 1, wherein said slower mass-storage device is a CD-ROM drive controlled by a server data processing system to which the computer system is coupled via a network data link.

10. The method of claim 1, wherein said slower mass-storage device is a hard disk drive controlled by a server data processing system to which the computer system is coupled via a network data link.

11. A method for implementing a multi-tier caching system, said multi-tier caching system being implemented on a computer system having a random access memory (RAM), a portion of which is set aside and initialized as a RAM cache for storing first-tier data within a plurality of RAM data blocks, and at least one local non-volatile local mass storage device on which data can be alterably stored and on which a portion thereof functions as a non-volatile cache for storing second-tier data within non-volatile data blocks, said computer system having a data path to a mass storage device with an access speed slower than that of the local mass storage device, said slower mass storage device storing a full set of file data, said method comprising the steps of:

whenever a received file input/output request involves data resident on said slower mass storage device, and the requested data is not available as first-tier data nor available as second-tier data, providing for determining whether or not all RAM data blocks are already filled with first-tier data;

providing for purging first-tier data from a least-recently-used RAM data block if all RAM data blocks are already filled;

providing for retrieving a quantity of third-tier data which includes the requested data;

providing for writing all retrieved third-tier data to a non-volatile data block, as well as to an empty RAM data block; and

providing for resolving the file input/output request from the retrieved third-tier data written to the empty RAM data block.

12. The method of claim 11, which further comprises the step of providing for determining, upon receipt of a file input/output request, whether or not the request involves third-tier data resident on the slower mass storage device.

13. The method of claim 11, wherein said method further comprises the step of providing for determining, for a file input/output request involving data resident on said slower mass storage device, whether or not first tier data which satisfies the request is already resident in one of said RAM data blocks.

14. The method of claim 13, which further comprises the step of providing for resolving the request from such first-tier data if such first-tier data which satisfies the request is already resident in one of said RAM data blocks.

15. The method of claim 11, which, if a received input/output request involves data resident on said slower mass storage device, and if the requested data is not available as first-tier data, but is available as second-tier data, further comprises the steps of:

providing for determining whether or not all RAM data blocks are already filled with first-tier data;

providing for purging a least-recently-used RAM data block if all RAM data blocks are already filled;

providing for retrieving a quantity of second-tier data which includes the requested data;

providing for writing all retrieved second-tier data to an empty RAM data block; and

providing for resolving the file input/output request from the retrieved second-tier data written to the empty RAM data block.

16. A computer program product comprising a computer usable medium having computer readable code embodied therein for implementing a multi-tier caching system on a computer system, said computer system having a random access memory (RAM), a portion of said RAM being is set aside and initialized as a RAM cache for storing first-tier data within a plurality of equal sized RAM data blocks, and at least one local non-volatile local mass storage device on which data can be alterably stored, a portion of said local mass storage device functioning as a non-volatile cache for storing second-tier data in non-volatile data blocks, said computer program product further comprising:

computer readable program code devices configured to cause a computer, whenever a received input/output request involves data resident on a mass storage device identified for caching, and if the requested data is not available as first-tier data nor available as second-tier data, to effect determining whether or not all RAM data blocks are already filled with first-tier data, to effect purging first-tier data from a least-recently-used RAM data block if all RAM data blocks are already filled, to effect retrieving a quantity of third-tier data which includes the requested data, to effect writing retrieved third-tier data to a non-volatile data block, as well as to an empty RAM data block, and to effect resolving the file input/output request from the retrieved third-tier data written to the empty RAM data block.

17. The computer program product of claim 16, which further comprises computer readable program code devices configured to cause the computer system to effect determining, upon receipt of a file input/output request from the operating system, whether or not the request involves third-tier data resident on a mass storage device, other than the local mass storage device, which is identified for caching of resident file data.

18. The computer program product of claim 16, which further comprises computer readable program code devices configured to cause a computer to effect determining, for a file input/output request from an operating system controlling the computer system'"'"'s operation, a request involving data resident on a mass storage device identified for caching, whether or not first tier data which satisfies the request is already resident in one of said sub-blocks of the RAM cache, and, if first-tier data which satisfies the request is already resident in one of said sub-blocks, to effect resolving the request from such first-tier data.

19. The computer program product of claim 16, which further comprises computer readable program code devices configured to cause a computer, if a received input/output request involves data resident on a mass storage device identified for caching, and if the requested data is not available as first-tier data, but is available as second-tier data, to effect determining whether or not all sub-blocks are already filled with first-tier data, to effect purging a least-recently-used sub-block if all sub-blocks are already filled, to effect retrieving a quantity of second-tier data which includes the requested data, to effect writing retrieved second-tier data to an empty sub-block, and to effect resolving the file input/output request from the retrieved second-tier data written to the empty sub-block.

20. A multi-tier caching system implemented in conjunction with a computer system having a random access memory (RAM), and at least one local non-volatile local mass storage device on which data can be alterably stored, said multi-tier caching system comprising:

a mass storage device having a slower access time than that of said local mass storage device, said slower mass storage device having been designated for caching, and said slower mass storage device providing storage for third-tier data;

a non-volatile cache established on a portion of said local mass storage device, said non-volatile cache providing storage for non-volatile blocks of second-tier data, said non-volatile cache having a total storage capacity less than that of said slower mass storage device, said second-tier data being a sub-set of said third-tier data always comprising at least the most recently used third-tier data; and

a RAM cache established in a portion of the RAM, said RAM cache providing storage for RAM blocks of first-tier data, and said RAM cache having a total storage capacity less than that of said non-volatile cache, said first-tier data being a subset of said second-tier data comprising all of the most recently used third-tier data, said RAM cache coupled to supply all requests for data stored on said mass storage device having a slower access time.

21. The multi-tier caching system of claim 20, wherein an unused non-volatile data block is randomly selected to be purged of data during non-volatile cache updates when an empty non-volatile data block is unavailable.

22. The multi-tier caching system of claim 20, wherein a least recently used RAM data block is purged of data during RAM cache updates when no RAM data block is empty.

23. The multi-tier caching system of claim 20, wherein said slower mass storage device is a CD-ROM drive controlled by said computer system.

24. The multi-tier caching system of claim 20, wherein said slower mass storage device is a CD-ROM drive controlled by a server data processing system to which the computer system is coupled via a network data link.

25. The multi-tier caching system of claim 20, wherein said slower mass storage device is a hard disk drive controlled by a server data processing system to which the computer system is coupled via a network data link.

26. The multi-tier caching system of claim 20, wherein if an incoming file I/O request involves data resident on both said slower mass storage device and within said non-volatile cache but not within said RAM cache, said RAM cache is updated from the non-volatile cache.

27. The multi-tier caching system of claim 20, wherein if an incoming file I/O request involves data resident on said slower mass storage device, but not resident within said non-volatile cache or within said RAM cache, said RAM cache and said non-volatile volatile cache are both updated from said slower mass storage device.

28. The multi-tier caching system of claim 20, wherein the non-volatile data blocks within said non-volatile memory are the same size as the RAM data blocks within said RAM cache.

29. The multi-tier caching system of claim 20, wherein a portion of said RAM is utilized for maintaining a table which implements garbage collection involving the least recently used sub-block within the RAM cache during RAM cache updates when no RAM data block is empty.

30. The multi-tier caching system of claim 29, wherein said RAM cache and said table are established in RAM upon the receipt of a file I/O request at system boot-time initialization.