Transaction processing system and method with reduced locking

• US 5,247,672 A
• Filed: 02/15/1990
• Issued: 09/21/1993
• Est. Priority Date: 02/15/1990
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Mohan, Chandrasekaran
• Primary Examiner(s): Shaw, Gareth D.
• Assistant Examiner(s): Katbab, A.
• Application Number:
  US07/480,700
• Time in Patent Office: 1,314 Days
• Field of Search: 364/DIG. 1, 364/DIG. 2, 395/650
• US Class Current: 711/152
• CPC Class Codes: G06F 2201/825 the problem or solution involving locking
• CPC Class Codes: G06F 2201/87 Monitoring of transactions
• CPC Class Codes: G06F 9/466 Transaction processing
• CPC Class Codes: G06F 9/526 Mutual exclusion algorithms
• CPC Class Codes: Y10S 707/99938 Concurrency, e.g. lock management in shared database

Abstract

Apparatus and method for reading data pages in a transaction processing system without locking the pages are disclosed. The system maintains a Global_-- Committed_-- LSN identifying the oldest uncommitted transaction accessing any of the data, and Object_-- Committed_-- LSNs identifying the oldest uncommitted transactions accessing particular files, tables and indexes. Each data page includes a Page_-- LSN identifying the last transaction to have updated the page. To read a page, a transaction first latches the pages, and compares the page'"'"'s Page_-- LSN with the Global_-- Committed_-- LSN, or with the page'"'"'s respective Object_-- Committed_-- LSN. If the Page_-- LSN is older than the Committed_-- LSN with which it was compared, then the transaction reads the page without locking it, since there can be no uncommitted transaction in process which might have updated the page'"'"'s data. However if the Page_-- LSN is younger than the Committed_-- LSN, the page is locked before being read.

Claims

1. In a computerized data processing system having a plurality of data units, each data unit being identified by a respective first identifier which reflects an age not older than the age of the youngest modification of the data unit, a method of accessing a selected one of the data units by a selected one of a plurality of processes comprising the steps performed by the data processing system of:

(a) maintaining a second identifier which reflects an age at least as old as the then-oldest in-process modification of a first plurality of the plurality of data units;

(b) comparing the second identifier with the selected data unit'"'"'s respective first identifier;

(c) obtaining access to the data unit for the selected process compatibly with other process'"'"' accesses to the data unit, said step of obtaining access being performed using fewer computer instructions if the selected data unit'"'"'s respective first identifier is older than the second identifier than if the selected data unit'"'"'s respective first identifier is not older than the second identifier; and

(d) reading the data unit by the selected process.

2. The method of claim 1, wherein the step of maintaining a second identifier includes identifying the then-oldest in-process modification taking place on any of the plurality of data units.

3. The method of claim 1 wherein the step of maintaining a second identifier includes identifying the then-oldest in-process modification taking place on any of a subset of the plurality of data units, the subset including the selected data unit and including one of (i) all of the plurality of data units which are files, (ii) all of the plurality of data units which are tables, and (iii) all of the plurality of data units which are indexes.

4. The method of claim 1, wherein the step (c) of obtaining access comprises latching the data unit if the selected data unit'"'"'s respective first identifier is older than the second identifier, and locking the data unit if the selected data unit'"'"'s respective first identifier is not older than the second identifier, wherein said step of latching includes fewer computer instructions than said step of locking.

5. The method of claim 1, wherein the data units comprise data pages.

6. The method of claim 1, wherein the data units comprise data records, multiple such records being included in a single data page.

7. The method of claim 1, wherein a process for modifying the selected data units includes making a plurality of modifications, and the process responsive to the step of obtaining access, either performs all of the plurality of modifications or performs no modifications.

8. The method of claim 7, wherein the second identifier is updated in response to every modifying process.

9. A data processing system comprising:

a plurality of data units, each data unit being identified by a first respective identifier means which reflects an age not older than the age of the youngest modification of the data unit;

means for reading a selected one of the plurality of data units to which access has been obtained;

means for maintaining a second identifier which reflects an age at least as old as the then-oldest in-process modification of a first plurality of the plurality of data units;

means for comparing the second identifier with the selected data unit'"'"'s respective first identifier means; and

means for obtaining access to the selected data unit for the selected process compatibly with other process'"'"' accesses to the selected data unit, said access obtaining means comprising fewer computer instruct if the selected data unit'"'"'s respective first identifier is older than the second identifier than if the selected data unit'"'"'s respective first identifier is not older than the second identifier.

10. The system of claim 9, wherein the access obtaining means comprises computer instructions for latching the selected data unit if the selected data unit'"'"'s identifier is older than the second identifier, and computer instructions for locking the selected data unit if the selected data unit'"'"'s identifier is not older than the second identifier, said latching computer instructions being fewer than said locking computer instructions.

15. A system as recited in claim 9, wherein the means for maintaining a second identifier includes means for maintaining a second identifier which reflects an age at least as old as the then-oldest in-process modification of any of the plurality of data units.

16. A system as recited in claim 9, wherein the means for maintaining a second identifier includes means for maintaining a second identifier which reflects an age at least as old as the then-oldest in-process modification of the first plurality of data units, the first plurality including one of (i) all of the plurality of data units which are files, (ii) all of the plurality of data units which are tables, and (iii) all of the plurality of data units which are indexes.

11. A method for accessing a selected one of a plurality of data units, maintained on a data processing system, by a selected one of a plurality of processes, the method comprising the steps of:

maintaining, for each respective one of the plurality of data units, a first identifier which identifies a most recent time at which the respective data unit was modified;

maintaining a second identifier which identifies a time of commencement of an oldest in-progress modification of any one of a first plurality of the plurality of data units, the first plurality of data units including the selected one of the plurality of data units;

comparing the second identifier with the first identifier of the selected one of the plurality of data units to determine whether the most recent modification of the selected data unit occurred prior to the commencement of the oldest in-progress modification of any one of the first plurality of data units;

accessing the selected one of the plurality of data units according to a first set of computer instructions if the step of comparing determines that the most recent modification of the selected data unit occurred prior to the commencement of the oldest in-progress modification of any one of the first plurality of data units; and

accessing the selected one of the plurality of data units according to a second set of computer instructions otherwise.

12. The method of claim 11, wherein:

other ones of the plurality of processes may concurrently require access to the selected one of the plurality of data units;

any one of the plurality of processes which accesses one of the plurality of data units does so according to one of a plurality of modes, certain ones of the modes being mutually compatible for concurrent access, certain other ones of the modes being mutually incompatible for concurrent access and requiring consecutive access; and

the steps of accessing each include (a) accessing concurrently with at least one other process if each of the at least one other process and the selected process access according to mutually compatible modes, and (b) accessing after at least one other process, currently accessing, is completed, if the selected process accesses according to a mode mutually incompatible with a mode according to which at least one other process is currently accessing the data unit.

13. A method as recited in claim 11 wherein the step of maintaining includes maintaining a second identifier which identifies a time of commencement of an oldest in-process modification of any of the plurality of data units.

14. A method as recited in claim 11 wherein the step of maintaining includes maintaining a second identifier which identifies a time of commencement of an oldest in-process modification of the first plurality of data units, the first plurality including one of (i) all of the plurality of data units which are files, (ii) all of the plurality of data units which are tables, and (iii) all of the plurality of data units which are indexes.