Adaptively routing transactions to servers

• US 9,262,490 B2
• Filed: 05/07/2010
• Issued: 02/16/2016
• Est. Priority Date: 08/12/2004
• Status: Active Grant

• Current Assignee: Oracle International Corporation (Oracle Corporation)
• Original Assignee: Oracle International Corporation (Oracle Corporation)
• Inventors: Colrain, Carol Lyndall, Zoll, Michael Ullrich, Begun, Vladimir, Irudayaraj, Rajkumar Arockia, Bruns, Logan O'Sullivan
• Primary Examiner(s): CHANNAVAJJALA, SRIRAMA T
• Application Number:
  US12/776,341
• Publication Number:
  US 20110276579A1
• Time in Patent Office: 2,111 Days
• Field of Search: 707/720, 707/737, 707/756, 707/770, 707/E17.044, 707/966, 707/971, 707/972, 707/999.002, 711161-163, 711118-123, 711/126
• US Class Current: 1/1
• CPC Class Codes: G06F 16/2379 Updates performed during online database operations; commit processing
• CPC Class Codes: G06F 16/25 Integrating or interfacing systems involving database management systems
• CPC Class Codes: G06F 16/284 Relational databases
• CPC Class Codes: G06F 16/84 Mapping; Conversion
• CPC Class Codes: G06F 16/86 Mapping to a database
• CPC Class Codes: G06F 2212/7201 Logical to physical mapping or translation of blocks or pages
• CPC Class Codes: G06F 9/5011 the resources being hardware resources other than CPUs, Servers and Terminals
• CPC Class Codes: G06F 9/5027 the resource being a machine, e.g. CPUs, Servers, Terminals
• CPC Class Codes: H04L 67/60 Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources admission control or resource allocation H04L47/70

Abstract

A method and apparatus is provided for dynamically generating data to dataset mappings and dataset to logical instance mappings and distributing these mappings to subscribers for the purpose of routing work. An advisory engine updates the mappings of data to data sets based at least in part on changing demands and patterns of access to data, and updates the mappings of datasets to logical instances based at least in part on anticipated capacity and change in conditions of the database servers. The subscribers receive and store the mapping information that maps the data to dataset identifiers, and dataset identifiers to logical database instances. The subscribers may also receive information that describes the current availability conditions of database instances, which are used to override the mapping of logical database instances to physical database instances. Subscribers may use the data to dataset mapping to map data identified from in an incoming database request to a dataset. Subscribers may use the dataset to logical instance mapping to map a dataset to a possible database server, and the logical instance to physical instance mapping to map a possible database server to an available database server. In this manner, the subscribers may route data efficiently even when access patterns and server conditions change.

Claims

1. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause:

causing mappings to be stored for use by one or more mid-tier servers that provide connections to database servers, the mappings comprising;

a data-to-dataset mapping that maps first requests associated with data having one or more first key values to one or more first datasets and other requests associated with other data having one or more other key values to one or more other datasets,

a first dataset-to-candidate-physical-database-server mapping that maps the one or more first datasets to a first candidate physical database server and the one or more other datasets to one or more other candidate physical database servers, and

a candidate-physical-database-server-to-physical-database-server mapping that maps the first candidate physical database server to a first physical database server and the one or more other candidate physical database servers to one or more other physical database servers;

wherein when, based on a change in one or more conditions of the first candidate physical database server after causing the mappings to be stored for use, the first candidate physical database server does not satisfy stored criteria, one or more of the first requests, which are mapped to the first candidate physical database server, are caused to be routed to a first other physical database server, and

when a second candidate physical database server does not satisfy the stored criteria, one or more second requests, which are mapped to the second candidate physical database server, are caused to be routed to a second other physical database server;

wherein the second candidate physical database server is different from the first candidate physical database server;

receiving information that indicates a change of at least one of a condition or a characteristic of one or more physical database servers;

using the information to automatically generate a second dataset-to-candidate-physical-database-server mapping that is different from the first dataset-to-candidate-physical-database-server mapping, wherein the second dataset-to-candidate-physical-database-server mapping maps at least one of the first one or more datasets to the second candidate physical database server;

causing the one or more mid-tier servers to use the second dataset-to-candidate-physical-database-server mapping instead of the first dataset-to-candidate-physical-database-server mapping.

2. The one or more storage media as recited in claim 1, wherein the first dataset-to-candidate-physical-database-server mapping was generated based at least in part on a past response time of the one or more physical database servers, wherein the information comprises an updated response time, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the updated response time.

3. The one or more storage media as recited in claim 1, wherein the first dataset-to-candidate-physical-database-server mapping was generated based at least in part on at least one of a past load or past availability of the one or more physical database servers, wherein the information comprises an updated load or an updated availability, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the at least one of the updated load or the updated availability.

4. The one or more storage media as recited in claim 1, wherein the first dataset-to-candidate physical-database-server mapping was generated based at least in part on at least one of a past cache miss rate or a past cache hit rate of the one or more physical database servers, wherein the information comprises an updated cache miss rate or an updated cache hit rate, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the at least one of the updated cache miss rate or the updated cache hit rate.

5. The one or more storage media as recited in claim 1, wherein the first dataset-to-candidate-physical-database-server mapping, when generated, was not based on a particular characteristic of the one or more physical database servers, wherein the information comprises the particular characteristic, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the particular characteristic.

6. The one or more storage media as recited in claim 1, wherein the instructions, when executed by the one or more computing devices, further cause detecting that the at least one of the condition or the characteristic satisfies a threshold, and wherein the performance of using the information to automatically generate the second dataset-to-candidate-physical-database-server mapping is in response to the performance of detecting.

7. The one or more storage media as recited in claim 1, wherein the performance of using the information to automatically generate the second dataset-to-candidate-physical-database-server mapping that is different from the first dataset-to-candidate-physical-database-server mapping comprises performance of restructuring the first dataset-to-candidate-physical-database-server mapping, based at least in part on the information, to balance load across a plurality of physical database servers that include the one or more physical database servers.

8. The one or more storage media as recited in claim 1, wherein the information is first information, wherein the data-to-dataset mapping comprises a first data-to-dataset mapping, and wherein the instructions, when executed by the one or more computing devices, further cause:

receiving second information that indicates at least one of a change to a pattern of data access or a demand for data; and

using the second information to automatically generate a second data-to-dataset mapping that is different from the first data-to-dataset mapping.

9. One or more non-transitory storage media storing instructions which, when executed by one or more computing devices, cause:

storing a first mapping of data to respective datasets, a second mapping of datasets to respective candidate physical database servers, and a third mapping of candidate physical database servers to respective physical database servers,

wherein the first mapping maps requests associated with data having one or more key values to a one or more particular datasets and other requests associated with other data having one or more other key values to one or more other datasets,

wherein the second mapping maps the one or more particular datasets to a particular candidate physical database server and the one or more other datasets to one or more other candidate physical database servers, and

wherein the third mapping maps the particular candidate physical database server to a particular physical database server and the one or more other candidate physical database servers to one or more other physical database servers;

at a mid-tier server that routes requests to database servers, receiving a particular request to access a database;

in response to receiving the particular request, routing the particular request to a physical database server by;

mapping the particular request to the particular dataset using the first mapping;

mapping the particular dataset to the particular candidate physical database server using the second mapping;

accessing stored information that indicates the particular candidate physical database server does not satisfy stored criteria;

mapping the particular candidate physical database server to the other physical database server using the third mapping; and

wherein when, based on a change in one or more conditions of the particular candidate physical database server after the third mapping is stored, the particular candidate physical database server does not satisfy the stored criteria, the mid-tier server routes the particular request from the particular candidate physical database server to the other physical database server.

10. The one or more storage media as recited in claim 9, wherein the instructions, when executed by the one or more computing devices, further cause:

receiving an update to the first mapping that accounts for at least one of a change to a pattern of data access or a demand for data; and

wherein the performance of routing the particular request to the physical database server reflects the update and maps the particular request to a different physical database server than the particular request would be mapped without the update.

11. The one or more storage media as recited in claim 9, wherein the instructions, when executed by the one or more computing devices, further cause:

receiving an update to the second mapping that accounts for at least one of a change of a condition or a characteristic of one or more physical database servers; and

wherein the performance of routing the particular request to the physical database server reflects the update and maps the particular request to a different physical database server than the particular request would be mapped without the update.

12. The one or more storage media as recited in claim 9, wherein the performance of mapping the particular candidate physical database server to the other physical database server is based at least in part on determining that the particular candidate physical database server does not satisfy the stored criteria, wherein the other physical database server is re-mapped to a particular other physical database server, and the re-mapping is repeated until the particular other physical database server satisfies the stored criteria.

13. The one or more storage media as recited in claim 9, wherein the performance of mapping the particular dataset to the particular candidate physical database server using the second mapping maps the particular dataset to a bit in a bitmap that represents one or more conditions of the physical database servers, wherein the bit is associated with the particular candidate physical database server;

and wherein the performance of mapping the particular candidate physical database server to the other physical database server using the third mapping re-maps the particular candidate physical database server to the other physical database server if the bit indicates that the particular candidate physical database server does not satisfy the stored criteria.

14. The one or more storage media as recited in claim 9, wherein the instructions, when executed by the one or more computing devices, further cause providing an interface to specify a data key at connection checkout.

15. The one or more storage media as recited in claim 9, wherein the performance of mapping the particular request to the particular dataset using the first mapping scans the particular request for data keys and uses the data keys in a rank order specified by the first mapping.

16. The one or more storage media as recited in claim 9, wherein the performance of mapping the particular request to the particular dataset using the first mapping uses a property of the particular request as a data key to identify the particular dataset, wherein the property comprises at least one of a network domain, a client identifier, or a username.

17. A method comprising:

causing mappings to be stored for use by one or more mid-tier servers that provide connections to database servers, the mappings comprising;

a data-to-dataset mapping that maps first requests associated with data having one or more first key values to a one or more first datasets and other requests associated with other data having one or more other key values to one or more other datasets,

a first dataset-to-candidate-physical-database-server mapping that maps a first the one or more first datasets to a first candidate physical database server and the one or more other datasets to one or more other candidate physical database servers, and

a candidate-physical-database-server-to-physical-database-server mapping that maps the first candidate physical database server to a first physical database server and the one or more other candidate physical database servers to one or more other physical database servers;

wherein when, based on a change in one or more conditions of the first candidate physical database server after causing the mappings to be stored for use, the first candidate physical database server does not satisfy stored criteria, one or more of the first requests, which are mapped to the first candidate physical database server, are caused to be routed to a first other physical database server, and

when a second candidate physical database server does not satisfy the stored criteria, one or more second requests, which are mapped to the second candidate physical database server, are caused to be routed to a second other physical database server;

wherein the second candidate physical database server is different from the first candidate physical database server;

receiving information that indicates a change of at least one of a condition or a characteristic of one or more physical database servers;

using the information to automatically generate a second dataset-to-candidate-physical-database-server mapping that is different from the first dataset-to-candidate-physical-database-server mapping, wherein the second dataset-to-candidate-physical-database-server mapping maps at least one of the first one or more datasets to the second candidate physical database server;

causing the one or more mid-tier servers to use the second dataset-to-candidate-physical-database-server mapping instead of the first dataset-to-candidate-physical-database-server mapping;

wherein the method is performed by one or more computing devices.

18. The method as recited in claim 17, wherein the first dataset-to-candidate-physical-database-server mapping was generated based at least in part on a past response time of the one or more physical database servers, wherein the information comprises an updated response time, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the updated response time.

19. The method as recited in claim 17, wherein the first dataset-to-candidate-physical-database-server mapping was generated based at least in part on at least one of a past load or past availability of the one or more physical database servers, wherein the information comprises an updated load or an updated availability, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the at least one of the updated load or the updated availability.

20. The method as recited in claim 17, wherein the first dataset-to-candidate physical-database-server mapping was generated based at least in part on at least one of a past cache miss rate or a past cache hit rate of the one or more physical database servers, wherein the information comprises an updated cache miss rate or an updated cache hit rate, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the at least one of the updated cache miss rate or the updated cache hit rate.

21. The method as recited in claim 17, wherein the first dataset-to-candidate-physical-database-server mapping, when generated, was not based on a particular characteristic of the one or more physical database servers, wherein the information comprises the particular characteristic, and wherein the second dataset-to-candidate-physical-database-server mapping is based at least in part on the particular characteristic.

22. The method as recited in claim 17, further comprising detecting that the at least one of the condition or the characteristic satisfies a threshold, and wherein the performance of using the information to automatically generate the second dataset-to-candidate-physical-database-server mapping is in response to the performance of detecting.

23. The method as recited in claim 17, wherein the performance of using the information to automatically generate the second dataset-to-candidate-physical-database-server mapping that is different from the first dataset-to-candidate-physical-database-server mapping comprises performance of restructuring the first dataset-to-candidate-physical-database-server mapping, based at least in part on the information, to balance load across a plurality of physical database servers that include the one or more physical database servers.

24. The method as recited in claim 17, wherein the information is first information, wherein the data-to-dataset mapping comprises a first data-to-dataset mapping, the method further comprising:

receiving second information that indicates at least one of a change to a pattern of data access or a demand for data; and

using the second information to automatically generate a second data-to-dataset mapping that is different from the first data-to-dataset mapping.

25. A method comprising:

storing a first mapping of data to respective datasets, a second mapping of datasets to respective candidate physical database servers, and a third mapping of candidate physical database servers to respective physical database servers,

wherein the first mapping maps requests associated with data having one or more key values to a one or more particular datasets and other requests associated with other data having one or more other key values to one or more other datasets,

wherein the second mapping maps the one or more particular datasets to a particular candidate physical database server and the one or more other datasets to one or more other candidate physical database servers, and

wherein the third mapping maps the particular candidate physical database server to a particular physical database server and the one or more other candidate physical database servers to one or more other physical database servers;

at a mid-tier server that routes requests to database servers, receiving a particular request to access a database;

in response to receiving the particular request, routing the particular request to a physical database server by;

mapping the particular request to the particular dataset using the first mapping;

mapping the particular dataset to the particular candidate physical database server using the second mapping;

accessing stored information that indicates the particular candidate physical database server does not satisfy stored criteria;

mapping the particular candidate physical database server to the other physical database server using the third mapping; and

wherein when, based on a change in one or more conditions of the particular candidate physical database server after the third mapping is stored, the particular candidate physical database server does not satisfy the stored criteria, the mid-tier server routes the particular request from the particular candidate physical database server to the other physical database server;

wherein the method is performed by one or more computing devices.

26. The method as recited in claim 25, further comprising:

receiving an update to the first mapping that accounts for at least one of a change to a pattern of data access or a demand for data; and

wherein the performance of routing the particular request to the physical database server reflects the update and maps the particular request to a different physical database server than the particular request would be mapped without the update.

27. The method as recited in claim 25, further comprising:

receiving an update to the second mapping that accounts for at least one of a change of a condition or a characteristic of one or more physical database servers; and

wherein the performance of routing the particular request to the physical database server reflects the update and maps the particular request to a different physical database server than the particular request would be mapped without the update.

28. The method as recited in claim 25, wherein the performance of mapping the particular candidate physical database server to the other physical database server is based at least in part on determining that the particular candidate physical database server does not satisfy the stored criteria, wherein the other physical database server is re-mapped to a particular other physical database server, and the re-mapping is repeated until the particular other physical database server satisfies the stored criteria.

29. The method as recited in claim 25, wherein the performance of mapping the particular dataset to the particular candidate physical database server using the second mapping maps the particular dataset to a bit in a bitmap that represents one or more conditions of the physical database servers, wherein the bit is associated with the particular candidate physical database server;

and wherein the performance of mapping the particular candidate physical database server to the other physical database server using the third mapping re-maps the particular candidate physical database server to the other physical database server if the bit indicates that the particular candidate physical database server does not satisfy the stored criteria.

30. The method as recited in claim 25, further comprising providing an interface to specify a data key at connection checkout.

31. The method as recited in claim 25, wherein the performance of mapping the particular request to the particular dataset using the first mapping scans the particular request for data keys and uses the data keys in a rank order specified by the first mapping.

32. The method as recited in claim 25, wherein the performance of mapping the particular request to the particular dataset using the first mapping uses a property of the particular request as a data key to identify the particular dataset, wherein the property comprises at least one of a network domain, a client identifier, or a username.