Update manager for database system

• US 8,126,919 B2
• Filed: 05/18/2010
• Issued: 02/28/2012
• Est. Priority Date: 04/11/2006
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP AG (SAP SE)
• Inventors: Eberlein, Peter
• Primary Examiner(s): Filipczyk, Marc
• Application Number:
  US12/782,401
• Publication Number:
  US 20100228788A1
• Time in Patent Office: 651 Days
• Field of Search: 707/783, 707/955, 707/999.006, 707/999.101, 707/999.201, 707/999.009, 709/201, 709/203, 713/168, 713/178, 713/185
• US Class Current: 707/783
• CPC Class Codes: G06F 16/2315 Optimistic concurrency control
• CPC Class Codes: Y10S 707/955 Object-oriented
• CPC Class Codes: Y10S 707/99936 Pattern matching access

Abstract

A database management protocol exchanges update tokens between a client and a server on which the database resides. When a client requests data to be read from a database, an update manager either reads an update token stored therein or generates one dynamically. The update token represents a current state of the data object being read. Sometime thereafter, if the client requests new data to be stored in the object, the client may furnish the update token back to the update manager. The update manager compares the client'"'"'s update token to a local update token representing a current state of the database and, if they match, determines that the state has not changed. If they do not match, an error results.

Claims

1. A data management method, comprising:

in response to a read request received from an application;

reading, from a database management system of a group of databases, a data object referenced by the read request;

applying a hash function to select data fields of the data object, generating therefrom a first update token, the hash function taking as inputs contents of data field that for which data inconsistencies cannot be tolerated and omitting other data fields, and

sending the requested data object and the first update token to a sender of the read request; and

in response to a write request received from the application subsequent to the read request, the write request addressed to the data object and including the first update token;

reading, from the database management system, a current version of the data object;

applying the hash function to the select data fields of the current version of data object, generating therefrom a second update token, and

comparing the first update token received in the write request to the second update token generated from the current version of the data object;

if the first and second update tokens match, executing the write request at the database management system;

wherein the group of databases comprise a first database that provides native support for update control that prevents inconsistencies introduced by multiple attempts to update same data simultaneously and a second database that does not provide native support for update control that prevents inconsistencies introduced by multiple attempts to update same data simultaneously.

2. The method of claim 1, wherein the database management system natively supports update control.

3. The method of claim 2, wherein the database management system natively supports update control by a timestamp stored in the database management system in association with stored data objects.

4. The method of claim 2, wherein the database management system natively supports update control by a count value representing a number of times a data object has been updated in the database management system, the count value stored in the database management system in association with the data object.

5. The method of claim 1, further comprising, if a match does not occur, transmitting the current version of the data object and the second update token to the application in response to the write request.

6. The method of claim 1, further comprising, after the write request is executed, generating a third update token by applying the hash function to the selected data fields of a version of the data object generated after the write request is executed.

7. A data management method in a multi-application computer system, comprising:

in response to a write request received from a first application, the write request addressed to a data object stored in a database of a group of databases and including an update token received by the first application during a prior interaction with the data object, determining whether the data object has been updated by another application during a time period between generation of the first update token and a time of the write request;

reading, from the database, a current version of the data object;

applying a hash function to select data fields of the data object, generating therefrom a second update token, the hash function taking as inputs fields of the data object for which data inconsistencies cannot be tolerated, the hash function omitting other data fields of the data object;

comparing the first update token received in the write request to the second update token generated from the hash function; and

if the first and second update tokens match, executing the write request at the database management system;

wherein the group of databases comprise a first database that provides native support for update control that prevents inconsistencies introduced by multiple attempts to update same data simultaneously and a second database that does not provide native support for update control that prevents inconsistencies introduced by multiple attempts to update same data simultaneously.

8. The data management method of claim 7, further comprising, if the first and second update tokens do not match, transmitting the current version of the data object and the second update token to the first application in response to the write request.

9. The data management method of claim 7, wherein the database natively supports update control.

10. The data management method of claim 9, wherein the database natively supports update control by a timestamp stored in the database in association with stored data objects.

11. The data management method of claim 9, wherein the database natively supports update control by a count value representing a number of times a data object has been updated in the database, the count value stored in the database in association with the data object.

12. An apparatus for data management, the apparatus comprising:

a memory device having executable instructions stored therein; and

a processor in operable communication with the memory device and operative to;

in response to receiving a read request to read data;

forward the read request to a database of a group of databases;

receive a requested data object referenced by the read request;

select data fields from the requested data object for which data inconsistencies cannot be tolerated;

apply a hash function to the selected data fields to generate a first update token and omitting other data fields not selected, wherein the other data fields not selected are data fields for which data inconsistencies can be tolerated, and

send the requested data object and the first update token to a sender of the read request; and

in response to receiving a write request to write data to a destination data object;

compare a second update token contained in the write request to a local update token associated with the destination data object;

if the second update token and the local update token match, store data received in the write request to the database;

wherein the second and local update tokens will match even if updates to the destination data object have occurred during a time between generation of the second and local update tokens, provided the updates to the destination data object are directed to data fields thereof that are tolerant of data inconsistencies;

wherein the group of databases comprise a first database that provides native support for update control that prevents inconsistencies introduced by multiple attempts to update same data simultaneously and a second database that does not provide native support for update control that prevents inconsistencies introduced by multiple attempts to update same data simultaneously.

13. The apparatus of claim 12, wherein the database natively supports update control.

14. The apparatus of claim 13, wherein the database natively supports update control by a timestamp stored in the database in association with stored data objects.

15. The apparatus of claim 13, wherein the database natively supports update control by a count value representing a number of times a data object has been updated in the database, the count value stored in the database in association with the data object.

16. The apparatus of claim 12, wherein the processor is further operative to:

if a match does not occur, transmit the destination data object and the local update token to a sender of the write request.

17. The apparatus of claim 12, wherein the processor is further operative to:

after the storing the data received in the write request, update the local update token by applying the hash function to selected data fields of the destination data object to generate an updated local token and omitting other data fields not selected,

wherein the selected data fields are data fields for which data inconsistencies cannot be tolerated and the other data fields not selected are data fields for which data inconsistencies can be tolerated.