Extensible file system which layers a new file system with an old file system to provide coherent file data

• US 5,561,799 A
• Filed: 01/11/1996
• Issued: 10/01/1996
• Est. Priority Date: 06/17/1993
• Status: Expired due to Term

• Current Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Khalidi, Yousef A., Nelson, Michael N.
• Primary Examiner(s): AMSBURY, WAYNE P
• Application Number:
  US08/584,878
• Time in Patent Office: 264 Days
• Field of Search: 395/600, 395/650
• US Class Current: 707/809
• CPC Class Codes: G06F 16/10 File systems; File servers
• CPC Class Codes: Y10S 707/966 Distributed

Abstract

In this disclosure an architecture for extensible file systems is described. Also disclosed is an implementation of the architecture. The architecture enables the extension of file system functionality by stacking (or composing) new file systems (layers) on top of existing file systems. The implementor of a new layer has the option of keeping the files exported by the new layer coherent with files of the underlying layer, as well as the option of sharing the same cached memory with the files of the underlying layer. A flexible framework is also disclosed for arranging the file systems'"'"' name spaces. Composing of new layers on top of existing ones can be done statically (at compile/configuration time) or dynamically (at boot/run time). In addition, the file system layers can reside in the same address space or in different address spaces, and be implemented on a local computer node or on a remote computer node.

Claims

1. In a computer system having a naming service, a computer implemented method for associating a new file system with an existing file system, said method comprising the steps of:

determining the location of said existing file system located in a first layer, by interrogating said naming service;

connecting a new file system to said naming service, said new file system located in a second layer, said first layer and said second layer being program code sections which may reside in different computers and such that files in said existing file system can be accessed through said new file system; and

connecting said new file system to said existing file system by a data movement facility wherein said data movement facility provides coherent file data from said existing file system by means of a pager program located in said first layer, said pager program being connected to a cache program located in said second layer, said cache program supplying said coherent file data from said existing file system to said new file system.

2. The method described in claim 1 wherein said files in said existing file system remain directly accessible as well as accessible through said new file system.

3. The method described in claim 1 wherein said data movement facility provides coherent file attribute data from said existing file system by means of an attribute pager program located in said first layer, said attribute pager program being connected to an attribute cache program located in said second layer, said attribute cache program supplying said coherent file attribute data from said existing file system to said new file system.

4. The method described in claim 1 wherein said new file system can be implemented in a kernel of an operating system in said computer.

5. The method described in claim 1 wherein said new file system can be implemented in a user process in said computer.

6. The method described in claim 1 wherein said computer in which said new file system can be implemented, is a remote computer which does not contain said existing file system.

7. The method described in claim 1 wherein said new file system can be connected to said existing file system dynamically at system run time or statically at compile time.

8. The method described in claim 1 further comprising additional steps of interposing an additional file system in a third layer on top of said second layer containing said new file system and connecting said additional file system to said naming service and to said new file system.

9. The method described in claim 8 wherein said second layer containing said new file system which can cache coherent file data from said existing file system, can also act as a pager for supplying file data from said new file system to said third layer, said third layer also having the ability to cache said file data from said new file system for use by said additional file system.

10. The method described in claim 9 wherein said data movement facility coordinates said caching of said file data in said second layer and said third layer and provides the ability to maintain said cached file data coherent.

11. The method described in claim 10 wherein said coherent cached file data may be either file data or file attribute data.

12. The method described in claim 1 wherein said computer system is a distributed computer system with a plurality of computers.

13. The method described in claim 12 further comprising additional steps of implementing a plurality of additional file systems, each of which said additional file systems being implemented on a computer anywhere in said distributed computer system, and implementing each of said additional file systems as layers connected to each other in any order desired by an implementor wherein said connections are by means of naming services and data movement facilities capable of moving data from one layer to a next layer, and wherein a given layer may stack on more than one underlying layer.

14. In a computer system, having an existing file system located in a first layer connected to a naming service, an extensible file system comprising:

an identification device for identifying said existing file system by accessing said naming service to locate said existing file system;

a new file system coupled to said computer system said new file system located in a second layer, said first layer and said second layer being program code mechanisms which may reside in different computers;

a connecting mechanism coupled to said computer system for connecting said new file system to said naming service such that files in said existing file system can be accessed by reference to said new file system; and

a data movement facility coupled to said computer system wherein said data movement facility provides coherent file data from said existing file system by means of a pager program located in said first layer, said pager program being connected to a cache program located in said second layer, said cache program supplying said coherent file data from said existing file system to said new file system.

15. In a distributed computer system having one or more computers, each computer having an operating system and an existing file system located in a first layer and a naming service which can map names to files, an extensible file system comprising:

an identification device in a computer for identifying a location of said existing file system by accessing said naming service on said computer;

a file system creation mechanism coupled to said computer for creating a new file system by creating an instance of said new file system which can be connected to said existing file system, wherein said new file system is located in a second layer, said first layer and said second layers being program code mechanisms which reside in different computers;

a connecting mechanism coupled to said computer for connecting said new file system to said naming service whereby a client can access files in said existing file system, file accesses being routed through said new file system without said client having any knowledge of said new file system; and

a data movement facility coupled to said computer wherein said data movement facility provides coherent file data from said existing file system by means of a pager program located in said first layer, said pager program being connected to a cache program located in said second layer, said cache program supplying said coherent file data from said existing file system to said new file system; and

wherein said data movement facility caches file attribute data in said second layer and maintains said cached file attribute data coherent with said files in said existing file system.

16. The extensible file system described in claim 15 wherein a client can access said new file system without having any knowledge of said existing file system.

17. The extensible file system described in claim 15 wherein said files in said existing file system remain directly accessible even when said new file system is interposed on top of said existing file system, with coherency of data in said files maintained when accessed by different file system layers.

18. The extensible file system described in claim 15 wherein the implementor of said new file system decides whether it is necessary to maintain files in said new file system coherent with said files in said existing file system, where a particular coherency procedure to be used is chosen by said existing file system in its capacity as a pager.

19. The extensible file system described in claim 15 wherein said new file system can be implemented in a kernel of an operating system in said computer.

20. The extensible file system described in claim 15 wherein said new file system can be implemented in a user process in said computer.

21. The extensible file system described in claim 15 wherein said computer in which said new file system can be implemented, is a remote computer which does not contain said existing file system.

22. The extensible file system described in claim 15 wherein said new file system can be connected to said existing file system dynamically at system run time or statically at compile time as the implementor of said new file system desires.

23. The extensible file system described in claim 15 further comprising an additional file system in a third layer which can be interposed on top of said second layer containing said new file system and connected to said naming service and to said new file system.

24. The extensible file system described in claim 23 wherein said second layer containing said new file system which can cache coherent file data from said existing file system, can also act as a pager for supplying file data from said new file system to said third layer, said third layer also having the ability to cache said file data from said new file system for use by said additional file system.

25. The extensible file system described in claim 24 wherein said data movement facility coordinates said caching of said file data in said second layer and said third layer and provides the ability to maintain said cached file data coherent.

26. The extensible file system described in claim 25 wherein said coherent cached file data may be either file data or file attribute data.

27. The extensible file system described in claim 15 further comprising a plurality of additional file systems, each of which said additional file systems can be implemented on a computer anywhere in said distributed computer system, and each of which said additional file systems can be implemented as layers connected to each other in any order desired by an implementor wherein said connections are by means of naming facilities and data movement facilities capable of moving data from one layer to a next layer.