Automatic assignment of services to servers in a multi-server system

• US 20050246441A1
• Filed: 07/26/2004
• Published: 11/03/2005
• Est. Priority Date: 04/16/2004
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Young, Gary D., Chandrasekaran, Sashikanth
• Granted Patent: US 8,560,690 B2
• Time in Patent Office: Days
• US Class Current: 709/226
• CPC Class Codes: H04L 67/1001 for accessing one among a plurality of replicated servers
• CPC Class Codes: H04L 67/1008 based on parameters of servers, e.g. available memory or workload monitoring of computer activity G06F11/30
• CPC Class Codes: H04L 67/63 Routing a service request depending on the request content or context

Abstract

Groups of services are created based on one or both of the following: (a) what is accessed by each service, and (b) a likelihood of conflicts between accesses (e.g., a database object read operation or a database object write operation) performed by each service and accesses performed by each other service. Groups of services, which may be the groups created or subsets of those groups, are assigned to run on particular servers based on the processing power requirements of each group and the processing power available to each server. In the grouping process, accesses can be biased by giving more weight to object write operations than to object read operations, and to access operations on an index or dictionary than to access operations on a table.

Claims

1. A method for automatically assigning a plurality of services to a plurality of servers, the method comprising the computer-implemented steps of:

creating groups of one or more services by

grouping services of said plurality of services based on at least one of

(a) what is accessed by each service, and

(b) likelihood of conflicts between accesses performed by each service and accesses performed by each other service; and

assigning said groups to servers of said plurality of servers based on processing power requirements of each group and processing power available to each server of said plurality of servers.

2. The method of claim 1, wherein the step of creating groups comprises creating

groups based on what is accessed by each service, based on which objects in a database are accessed by each service.

3. The method of claim 1, wherein the step of creating groups comprises creating

groups based on likelihood of conflicts between accesses performed by each service and accesses performed by each other service, based on conflicts between accesses of objects in a database.

4. The method of claim 1,

wherein the step of creating groups comprises creating initial groups of one or more services;

and

wherein the step of assigning said groups comprises

if an initial group cannot be assigned to a particular server of said plurality of servers because said processing power requirements of said initial group exceeds said processing power available to said particular server, then subdividing said initial group into sub-groups of one or more services; and

assigning a first sub-group to said particular server and a second sub-group to another server of said plurality of servers that is not said particular server.

5. The method of claim 4,

wherein the step of creating groups comprises creating groups based only on (a) what is accessed by each service;

and

wherein the step of subdividing said initial group comprises

creating said first sub-group by removing, from said initial group, a first service that has the least similarity in what is accessed by said first service in comparison with what is accessed by services, of said initial group, other than said first service; and

creating said second sub-group comprising said first service.

6. The method of claim 4,

wherein the step of creating groups comprises creating groups based only on (b) likelihood of conflicts between accesses performed by each service and accesses performed by each other service;

and

wherein the step of subdividing said initial group comprises

creating said first sub-group by removing, from said initial group, a first service that has the least cost of conflicts between accesses performed by said first service and accesses performed by services, of said initial group, other than said first service; and

creating said second sub-group comprising said first service.

7. The method of claim 1,

wherein the step of assigning said groups comprises

if no services, from a group of said groups, can be assigned to any single server of said plurality of servers because said processing power requirement of each service from said group exceeds said processing power available to each server, then assigning a first service from said initial group to all servers of said plurality of servers;

wherein said first service is a service having a processing power requirement that equals or exceeds said processing power requirement of each service, of said group, other than said first service.

8. The method of claim 1,

wherein the step of assigning said groups comprises

attempting to assign each of said groups to a respective single server of said plurality of servers;

and

if a group of said groups cannot be assigned to any single server of said plurality of servers because said processing power requirement of said group exceeds said processing power available to each server, then assigning said group to two or more servers of said plurality of servers,

wherein said processing power available to said two or more servers equals or exceeds said processing power requirement of said group.

9. The method of claim 1, further comprising the computer-implemented step of:

computing said processing power requirements of each service in each group by summing the processing power used by all sessions in said service.

10. The method of claim 1, wherein the step of creating said groups comprises creating

groups based only on (a) what is accessed by each service.

11. The method of claim 10, wherein the step of creating said groups comprises:

representing database object access statistics for said plurality of services as respective strings; and

comparing strings to determine similarity between object access patterns of services that are associated with each of said strings being compared.

12. The method of claim 1, wherein the step of creating said groups comprises creating groups based only on (b) likelihood of conflicts between accesses performed by each service and accesses performed by each other service.

13. The method of claim 1, wherein the step of creating said groups comprises creating groups based on (a) what is accessed by each service and (b) likelihood of conflicts between accesses performed by each service and accesses performed by each other service.

14. The method of claim 1, wherein the step of creating said groups comprises

creating groups based only on (a) what is accessed by each service, by grouping said services based on said services within a group having similar access patterns regarding what database objects are accessed by each of said services.

15. The method of claim 14, further comprising the computer-implemented step of:

generating an access pattern for each service based on past accesses to database objects;

generating weighted access patterns by weighting, within respective access patterns, (a) write operations on an index or a data dictionary higher than read operations on said index or said data dictionary, (b) write operations on an index or a data dictionary higher than write operations on a table, (c) write operations on a table higher than read operations on said table, and (d) read operations on an index or a data dictionary higher than read operations on a table; and

determining similar access patterns by comparing respective weighted access patterns for services.

16. The method of claim 1, wherein the step of creating said groups comprises

creating groups based only on (b) likelihood of conflicts between accesses performed by each service and accesses performed by each other service, by grouping said services based on said services within a group having a high likelihood of conflicts between accesses performed by services within said group and accesses performed by other services within said group, relative to a likelihood of conflicts between accesses performed by services within said group and accesses performed by services not within said group.

17. The method of claim 16, wherein the step of creating groups comprises:

generating weighted probabilities of conflicts between accesses performed by each service and accesses performed by each other service, based on weighting (a) write conflicts on an index or a data dictionary higher than read conflicts on said index or said data dictionary, (b) write conflicts on an index or a data dictionary higher than write conflicts on a table, (c) write conflicts on a table higher than read conflicts on said table, and (d) read conflicts on an index or a data dictionary higher than read conflicts on a table; and

computing said likelihoods of conflicts using said weighted probabilities of conflicts.

18. The method of claim 1, wherein said plurality of servers are configured as a database cluster in which each server of said plurality of servers manages data stored in a same database.

19. The method of claim 1, wherein the step of assigning comprises pre-computing an assignment of said groups to servers based on termination of a particular node of a plurality of nodes on which said plurality of servers execute.

20. The method of claim 19, wherein the step of assigning comprises in response to termination of said particular node, assigning said groups according to said assignment.

21. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 1.

22. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 2.

23. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 3.

24. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 4.

25. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 5.

26. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 6.

27. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 7.

28. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 8.

29. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 9.

30. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 10.

31. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 11.

32. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 12.

33. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 13.

34. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 14.

35. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 15.

36. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 16.

37. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 17.

38. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 18.

39. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 19.

40. A computer-readable medium carrying one or more sequences of instructions which, when executed by one or more processors, causes the one or more processors to perform the method recited in claim 20.

41. A computer system that can automatically assign a plurality of services to a plurality of servers, the system comprising:

means for creating groups of one or more services by

grouping services of said plurality of services based on at least one of

(a) what is accessed by each service, and

(b) likelihood of conflicts between accesses performed by each service and accesses performed by each other service; and

means for assigning said groups to servers of said plurality of servers based on processing power requirements of each group and processing power available to each server of said plurality of servers.