System and method for managing a hierarchical storage system through improved data migration

• US 5,644,766 A
• Filed: 08/23/1996
• Issued: 07/01/1997
• Est. Priority Date: 03/22/1994
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Coy, Henry Robert, Cabrera, Luis Felipe, Rees, Robert M.
• Primary Examiner(s): Black, Thomas G.
• Assistant Examiner(s): ROBINSON, GRETA LEE
• Application Number:
  US08/697,397
• Time in Patent Office: 312 Days
• Field of Search: 395/600, 395/441, 395/488, 395/497.02, 395/848, 395/444, 364/228.1, 364/222.81, 364/246.3, 364/243, 364/282.1, 364/246, 364/283.2
• US Class Current: 1/1
• CPC Class Codes: G06F 16/10 File systems; File servers
• CPC Class Codes: Y10S 707/99953 Recoverability
• CPC Class Codes: Y10S 707/99955 Archiving or backup

Abstract

A system and method are provided for preserving spacial and temporal locality of sets of related objects when moving the sets within a storage hierarchy via a common server. The appropriate meta data is gathered to track the spacial and temporal locality of the sets of objects being moved within the storage hierarchy and the algorithm uses the meta data to preserve the spacial and temporal locality when moving the objects. A collection of logically clustered data objects is identified. The logical cluster is then moved down through the storage hierarchy together to be stored in less costly storage devices. The logical cluster of data objects is then retrievable more efficiently as a whole when requested.

Claims

1. A computerized data storage management system for managing data stored in a storage hierarchy having memory, central processing unit (CPU), and a plurality of levels of media instances wherein data is stored more economically in a lower level media instance, said system comprising:

means for identifying a collection of related data objects in a given level of the storage hierarchy;

means for identifying at least one media instance in a lower level for storing said collection so that a maximum number of data objects in the collection can be stored together in the same media instance; and

storage means for storing said collection clustered together in a minimum number of media instances.

2. The system as in claim 1, further comprising means for moving data objects between levels of the storage hierarchy together as a related collection.

3. The system as in claim 1, further comprising means for retrieving the collection from the media instance as a whole, thereby minimizing retrieval cost.

4. The system as in claim 1, further comprising means for retaining a copy of data objects moved down the hierarchy from a given level until storage space in the level is needed.

5. The system as in claim 1, further comprising means for identifying said collections based on a user ID and a file system.

6. The system of claim 1, further comprising a server in communication with a plurality of clients wherein the management system is controlled from the server and data objects are transmitted from the clients to the server for storage in the storage hierarchy.

7. The system of claim 6, further comprising means for the server to determine when data should be transferred from a client to the storage hierarchy.

8. A computerized distributed hierarchical data storage management system comprising:

a server computer system having a central processing unit and memory;

a plurality of client computer systems each having a central processing unit and a non-volatile storage device, in communication with the server;

a hierarchical data storage system in communication with the server and having a plurality of levels of media instances wherein data is stored more economically on a lower level;

means for identifying a logical cluster of data objects in a first level;

means for identifying at least one media instance in a lower level for storing said logical cluster by first attempting to identify at least one media instance already containing other data objects from said logical cluster, second, attempting to identify at least one empty media instance and third, identifying a media instance with the largest available space; and

storage means for storing said logical cluster in a minimum number of identified media instances.

9. The system as in claim 8, further comprising means for moving data through each level together as a logical cluster until all of the data in the logical cluster is stored in the lowest level.

10. The system of claim 8, further comprising means for removing copies of data from the first level when storage space in the first level is needed.

11. The system of claim 8, further comprising means for identifying logical clusters based on an owner of the cluster and a file system.

12. The system of claim 8, further comprising means for the server to poll each of the clients to determine when data objects stored at the client need to be transferred to the storage hierarchy.

13. In a computerized data storage system comprising a central processing unit (CPU) and main memory in communication with a plurality of media instances arranged in a hierarchy, a method for moving data objects between levels of the hierarchy comprising the steps performed by the CPU of:

(A) identifying logically clustered collections of objects within a first level of the hierarchy;

(B) identifying a minimum number of media instances within a next lower level of a storage hierarchy for storing the collection so that a maximum number of objects in the collection can be stored together in the same media instance; and

(C) moving the collection to the media instances within the next lower level of the storage hierarchy, storing objects in a logically clustered collection together within a level of a hierarchy.

14. The method of claim 13, further comprising repeating steps (A) through (C) until the objects in the collection are on the most economical level of the storage hierarchy.

15. The method of claim 13 wherein the collections are user defined named collections.

16. The method of claim 13, wherein media instances are identified by first identifying a media instance containing objects belonging to the same collection and identifying media instances that are empty.

17. The method as in claim 13, further comprising the step of:

upon request to retrieve a logical cluster, transmitting to the requester the collection from the media instance as a whole thereby minimizing retrieval cost.

18. An article of manufacture for use in a computerized data storage system in communication with a plurality of media instances of different types arranged in a hierarchy, for moving data objects between levels of the hierarchy, the storage system having means to write the data objects to the media instances, and run time environment means,

said article of manufacture comprising a computer-readable storage medium having a computer program code embodied in said medium which causes the computer to:

identify logically clustered collections of objects within a first level of the hierarchy;

identify a minimum number of media instances within a next lower level of a storage hierarchy for storing the collection by first identifying the media instances within the next lower level already containing objects belonging to the collection and media instances that are empty; and

move the collection to the media instances within the next lower level of the storage hierarchy, storing objects within a logically clustered collection together within a level of a hierarchy.

19. The article of manufacture of claim 18, wherein the computer program code causes the storage system to continue moving objects in the hierarchy until the objects in the collection are on the most economical level of the storage hierarchy.

20. The article of manufacture of claim 18, wherein the collections are user defined named collections.