Distributed token manager with transactional properties

• US 7,281,050 B2
• Filed: 04/08/2003
• Issued: 10/09/2007
• Est. Priority Date: 04/08/2003
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Dinker, Darpan, Gopinath, Pramod, Nadipalli, Suveen R., Tonse, Sudhir
• Primary Examiner(s): Hu; Jinsong
• Application Number:
  US10/408,885
• Publication Number:
  US 20040215772A1
• Time in Patent Office: 1,645 Days
• Field of Search: 709/229, 707/8, 707/203, 707/205
• US Class Current: 709/229
• CPC Class Codes: H04L 63/10 for controlling access to devices or network resources
• CPC Class Codes: H04L 67/1095 Replication or mirroring of data, e.g. scheduling or transport for data synchronisation between network nodes
• CPC Class Codes: H04L 67/1097 for distributed storage of data in networks, e.g. transport arrangements for network file system [NFS], storage area networks [SAN] or network attached storage [NAS]
• CPC Class Codes: H04L 69/329 in the application layer [OSI layer 7]
• CPC Class Codes: H04L 9/40 Network security protocols
• CPC Class Codes: Y10S 707/99938 Concurrency, e.g. lock management in shared database
• CPC Class Codes: Y10S 707/99953 Recoverability

Abstract

A system and method for controlling access to data in a distributed computer system. Distributed Token Manager (DTM) is a system-level service that coordinates read/write access of data objects (tokens) in a multi-process and multi-threaded environment. The DTM may support a transactional model such that write operations to a data object performed by a client process or thread can be either committed or rolled back.

Claims

1. A method for coordinating access to a data object utilized in a distributed software application, the method comprising:

a server computer granting a lock for writing to the data object to a first client computer;

the first client computer performing one or more write operations to the data object according to a transactional model, wherein said performing the one or more write operations to the data object according to the transactional model enables the one or more write operations to be either committed or rolled back;

the server computer re-acquiring the lock for writing to the data object from the first client computer; and

the server computer performing one of;

committing the one or more write operations performed by the first client computer;

or

rolling back the one or more write operations performed by the first client computer in response to a failure;

the server computer communicating state information for the data object to a backup server computer to maintain a backup of the state information on the backup server computer, wherein the state information comprises information regarding the lock for writing.

2. The method of claim 1,

wherein said first client computer performing one or more write operations to the data object results in one or more changes to the data object;

wherein said committing the one or more write operations performed by the first client computer comprises making the one or more changes to the data object permanent.

3. The method of claim 1,

wherein said first client computer performing one or more write operations to the data object results in one or more changes to the data object;

wherein said rolling back the one or more write operations performed by the first client computer comprises undoing the one or more changes to the data object.

4. The method of claim 1, further comprising:

the server computer receiving a request from the first client computer to commit the one or more write operations;

wherein said sewer computer committing the one or more write operations is performed in response to the request.

5. The method of claim 4, wherein said server computer re-acquiring the lock for writing to the data object from the first client computer is performed in response to the request to commit the one or more write operations.

6. The method of claim 1, further comprising:

the server computer receiving a request from the first client computer to roll back the one or more write operations;

wherein said server computer rolling back the one or more write operations is performed in response to the request.

7. The method of claim 6, wherein said server computer re-acquiring the lock for writing to the data object from the first client computer is performed in response to the request to roll back the one or more write operations.

8. The method of claim 1, further comprising:

the server computer determining that an error condition occurred;

wherein said server computer rolling back the one or more write operations is performed in response to the error condition.

9. The method of claim 8, wherein said server computer re-acquiring the lock for writing to the data object from the first client computer is performed in response to said server computer determining the error condition.

10. The method of claim 8, wherein said server computer determining that an error condition occurred comprises the server computer determining that a timeout occurred.

11. The method of claim 8, wherein said server computer determining that an error condition occurred comprises the server computer determining that a failure of the first client computer occurred.

12. The method of claim 1, further comprising:

the server computer storing the state information for the data object;

wherein said server computer rolling back the one or more write operations performed by the first client computer comprises the server computer using the state information to restore a previous state of the data object.

13. The method of claim 1, wherein said first client computer performing one or more write operations to the data object comprises the first client computer communicating with the server computer to perform the one or more write operations to the data object.

14. The method of claim 13, wherein the server computer has exclusive write access to the data object.

15. The method of claim 13,

wherein the server computer logs information regarding the one or more write operations to the data object;

wherein the information regarding the one or more write operations to the data object is usable to roll back the one or more write operations.

16. The method of claim 1, further comprising:

the server computer receiving a request from the first client computer to acquire the lock for writing to the data object;

wherein said sewer computer granting the lock to the first client computer comprises the sewer computer granting the lock to the first client computer in response to the request.

17. The method of claim 1, further comprising:

the server computer receiving a request from a second client computer for a lock for writing to the data object; and

the server computer queuing the request from the second client computer until after said server computer re-acquiring the lock for writing to the data object from the first client computer.

18. The method of claim 1,

wherein said server computer granting the lock for writing to the data object to the first client computer includes storing information indicating that the first client computer holds the lock for writing to the data object;

wherein said server computer re-acquiring the lock for writing to the data object from the first client computer includes updating the information to indicate that the first client computer no longer holds the lock for writing to the data object.

19. The system of claim 1, wherein the data object comprises one of:

HTTP session data;

IIOP session data;

a component having callable methods.

20. The method of claim 1, wherein the data object is located on the first client computer.

21. The method of claim 1, wherein the data object is located on the server computer.

22. The method of claim 1, wherein the data object is located on a computer coupled to the first client computer.

23. A method for coordinating access to a data object utilized in a distributed software application, the method comprising:

granting a lock for writing to the data object to a first client computer;

storing lock information indicating that the first client computer holds the lock for writing to the data object;

determining that a failure occurred on the first client computer;

updating the lock information to indicate that the first client computer no longer holds the lock for writing to the data object in response to said determining that the failure occurred on the first client computer; and

queuing a request by a second client computer to acquire the lock for writing to the data object after said granting the lock for writing to the data object to the first client computer;

wherein the first client computer performs one or more write operations on the data object;

wherein the method further comprises rolling back the one or more write operations in response to determining that the failure occurred on the first client computer.

24. The method of claim 23, further comprising:

granting the lock for writing to the data object to the second client computer after said updating the lock information to indicate that the first client computer no longer holds the lock for writing to the data object; and

storing lock information indicating that the second client computer holds the lock for writing to the data object.

25. A server computer operable to process requests sent by a plurality of client computers, the server computer comprising:

a first memory storing program instructions;

a first processor coupled to the first memory;

wherein the first processor is operable to execute program instructions stored in the first memory to implement a manager service;

wherein the manager service is operable to;

grant a lock for writing to a data object to a first client computer, wherein the data object is located on the first client computer;

perform one or more write operations to the data object in response to one or more requests from the first client computer, wherein the one or more write operations to the data object are performed according to a transactional model, wherein said performing the one or more write operations to the data object according to the transactional model enables the one or more write operations to be either committed or rolled back;

re-acquire the lock for writing to the data object from the first client computer; and

perform one of;

commit the one or more write operations to the data object;

or

roll back the one or more write operations to the data object in response to a failure.

26. The server computer of claim 25,

wherein said performing the one or more write operations to the data object results in one or more changes to the data object;

wherein said committing the one or more write operations comprises making the one or more changes to the data object permanent.

27. The server computer of claim 25,

wherein said performing the one or more write operations to the data object results in one or more changes to the data object;

wherein said rolling back the one or more write operations comprises undoing the one or more changes to the data object.

28. A method for coordinating access to HTTP session data utilized in a distributed software application, the method comprising:

a server computer granting a lock for writing to the HTTP session data to a first client computer;

the first client computer performing one or more write operations to the HTTP session data according to a transactional model, wherein said performing the one or more write operations to the HTTP session data according to the transactional model enables the one or more write operations to be either committed or rolled back;

the server computer re-acquiring the lock for writing to the HTTP session data from the first client computer; and

the server computer performing one of;

committing the one or more write operations performed by the first client computer;

or

rolling back the one or more write operations performed by the first client computer in response to a failure.

29. A distributed system operable to coordinate access to shared data, the system comprising:

a plurality of client computers;

wherein the client computers execute client processes operable to send requests to a manager service to acquire access rights for accessing data objects stored on a first computer;

wherein the manager service is operable to respond to the requests sent by the client processes to grant access rights for the data objects;

wherein the manager service is operable to support a transactional model for writing to the data objects such that for each client process that acquires access rights for writing to a data object, the manager service is operable to either commit or roll back write operations which the client process performs on the data object;

wherein the manager service is further operable to send a reclaim request to a client process to re-acquire access rights granted to the client; and

wherein the manager service is further operable to perform one of;

committing the one or more write operations performed by the first client computer;

or

rolling back the one or more write operations performed by the first client computer in response to a failure.