Systems and methods for proactive caching utilizing OLAP variants

• US 20040193576A1
• Filed: 03/28/2003
• Published: 09/30/2004
• Est. Priority Date: 03/28/2003
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Corporation
• Inventors: Petculescu, Cristian, Conlon, Thomas P., Netz, Amir
• Granted Patent: US 7,269,581 B2
• Time in Patent Office: Days
• US Class Current: 1/1
• CPC Class Codes: G06F 16/283 Multi-dimensional databases or data warehouses, e.g. MOLAP or ROLAP
• CPC Class Codes: Y10S 707/954 Relational
• CPC Class Codes: Y10S 707/957 Multidimensional
• CPC Class Codes: Y10S 707/99931 Database or file accessing
• CPC Class Codes: Y10S 707/99933 Query processing, i.e. searching

Abstract

The present invention leverages MOLAP performance for ROLAP objects (dimensions, partitions and aggregations) by building, in a background process, a MOLAP equivalent of that object. When the background processing completes, queries are switched from ROLAP queries to MOLAP queries. When changes occur to relevant relational objects (such as tables that define content of OLAP objects), an OLAP object is switched back to a ROLAP mode, and all relevant caches are dropped while, as a background process, a new MOLAP equivalent is created.

Claims

1. A system for caching information, comprising:

a multidimensional objects interface that allows for accessing at least one multidimensional object providing dynamic multidimensional analysis data derived from a database;

a cache interface that allows for accessing and controlling at least one cache providing dynamic multidimensional analysis data derived from the multidimensional object; and

at least one analysis component coupled to the multidimensional object and the cache for proactively controlling access to the multidimensional object and the cache.

2. The system of claim 1, the database comprising a relational database.

3. The system of claim 1, the multidimensional object comprising a ROLAP object.

4. The system of claim 1, the cache comprising a MOLAP cache.

5. The system of claim 1, the multidimensional object comprising real-time access analysis data and the cache comprising quick access analysis data.

6. The system of claim 1, the analysis component, comprising a Unified Dimensional Model (UDM).

7. The system of claim 1, the analysis component, comprising:

at least one input for determining when to proactively control the cache; and

a query interpreter for receiving and qualifying queries based upon the input.

8. The system of claim 7, the input comprising at least one selected from the group consisting of a user input and a system input for proactively controlling the cache.

9. The system of claim 8, the system input comprising a database update notification.

10. The system of claim 8, the user input comprising at least one selected from the group consisting of a quiet time delay, a quiet time delay override, a forced refresh time, and a user initiated partial cache rebuild.

11. The system of claim 8, the system input comprising at least one selected from the group consisting of a last database update tracker, tables affecting OLAP objects, and a dependent OLAP objects data source tracker.

12. The system of claim 1, the analysis component having proactive caching operational modes comprising at least one selected from the group consisting of a ROLAP mode, a MOLAP mode, and a MOLAP/ROLAP mode.

13. The system of claim 12, the ROLAP mode operatively triggered by at least one selected from the group consisting of quiet time delay met, quiet time delay override met, forced refresh time met, and database updated.

14. The system of claim 12, the MOLAP mode operatively triggered by at least one selected from the group consisting of:

equivalent query data available from both a MOLAP cache and a ROLAP object; and

a user demand.

15. The system of claim 12, the MOLAP/ROLAP mode operatively triggered by at least one selected from the group consisting of both real-time and low latency queries required and partial rebuilding of a MOLAP cache required.

16. The system of claim 1, the analysis component having capabilities to process multiple queries each requiring different proactive cache handling.

17. The system of claim 1, the analysis component having capabilities to process a parsed query, each part requiring different proactive cache handling.

18. The system of claim 1, the analysis component comprising an adaptive tuning component for optimizing the analysis component.

19. The system of claim 18, the adaptive tuning component, comprising:

a query interpreter for handling queries and instigating query sessions with the multidimensional object and the cache;

a result comparator operatively connected to the query interpreter for determining differences between the query sessions of the multidimensional object and the cache; and

a performance optimizer operatively connected to the result comparator for optimizing the query interpreter based on the differences determined by the result comparator.

64. A device employing the system of claim 1 comprising at least one from a group consisting of a computer, a server, and a handheld electronic device.

20. A method of database serving, comprising:

coupling operatively to at least one multidimensional object derived from at least one database and to at least one cache comprising dynamic multidimensional analysis data derived from the multidimensional objects; and

controlling proactively access to the multidimensional object and the cache in response to a query.

21. The method of claim 20, the multidimensional object comprising a ROLAP object.

22. The method of claim 20, the database comprising a relational database.

23. The method of claim 20, the cache comprising a MOLAP cache.

24. The method of claim 20, further comprising:

processing the multidimensional objects to produce the cache.

25. The method of claim 24, the processing comprising a background transaction.

26. The method of claim 24, further comprising:

monitoring the database for relevant changes; and

reprocessing the multidimensional objects to produce an updated cache when relevant changes occur.

27. The method of claim 24, further comprising:

canceling the processing of the multidimensional objects when at least one selected from the group consisting of an automatic condition detected and a user initiated action detected occurs.

28. The method of claim 27, the automatic condition detected comprising a database change.

29. The method of claim 27, the user initiated action detected comprising at least one selected from the group consisting of committing processing of a multidimensional object and starting another transaction on a multidimensional object.

30. A system for caching information, comprising:

first means for providing dynamic multidimensional analysis data derived from a database;

second means for providing dynamic multidimensional analysis data derived from the first means; and

third means for coupling the first means and the second means for proactively controlling access to the first and second means to provide a desired query response regarding the database.

31. A method of caching data, comprising:

providing at least one multidimensional object providing dynamic multidimensional analysis data derived from a database;

constructing at least one cache providing dynamic multidimensional analysis data derived from at least one multidimensional object;

switching an operational mode of an analysis component to access the cache only for query analysis; and

analyzing queries utilizing the cache.

32. The method of claim 31, further comprising:

determining if any changes to relevant multidimensional objects have occurred; and

switching the analysis component operational mode to access the multidimensional objects when relevant changes have occurred.

33. The method of claim 32, further comprising:

removing all cache related to the relevant changes; and

rebuilding the cache based on changed multidimensional objects.

34. The method of claim 31, further comprising:

determining if a query requires at least one selected from the group of low latency data and real-time data.

35. The method of claim 34, further comprising:

switching the analysis component to a MOLAP/ROLAP operational mode when both low latency and real-time data are required.

36. The method of claim 31, the multidimensional objects comprising ROLAP objects.

37. The method of claim 31, the database comprising a relational database.

38. The method of claim 31, the cache comprising a MOLAP cache.

39. The method of claim 31, the analysis component comprising a UDM.

40. The method of claim 31, the analysis component having operational modes comprising at least one selected from the group consisting of a MOLAP mode, a ROLAP mode, and a MOLAP/ROLAP mode.

41. A method of proactive caching, comprising:

providing an input related to determining an operational mode;

determining the operational mode based, at least in part, upon the input;

providing at least one multidimensional object providing dynamic multidimensional analysis data derived from a database;

building at least one cache providing dynamic multidimensional analysis data derived from at least one multidimensional object;

switching an analysis component having operational modes to the determined operational mode; and

processing queries via the analysis component utilizing the determined operational mode.

42. The method of claim 41, further comprising:

determining if any changes to relevant multidimensional objects have occurred; and

switching the analysis component operational mode to access the multidimensional objects when relevant changes have occurred.

43. The method of claim 42, further comprising:

removing all cache related to the relevant changes; and

rebuilding the cache based on the changed multidimensional objects.

44. The method of claim 41, further comprising:

determining relevant changes during the building of the cache; and

canceling the building of the cache when relevant changes occur.

45. The method of claim 41, the input comprising at least one selected from the group consisting of at least one user input and at least one system input.

46. The method of claim 45, the user input comprising at least one selected from the group consisting of a low latency query request and a real-time query request.

47. A method of proactive caching, comprising:

providing at least one multidimensional object providing dynamic multidimensional analysis data derived from a database;

building at least one cache providing dynamic multidimensional analysis data derived from at least one multidimensional object;

providing an input for designating data;

determining if any changes to relevant multidimensional objects have occurred;

switching the analysis component operational mode to access the multidimensional objects when relevant changes have occurred; and

rebuilding the cache based on the changed relevant multidimensional objects.

48. The method of claim 47, the input comprising data desired to be updated.

49. The method of claim 48, changes to relevant multidimensional objects including the data desired to be updated.

50. The method of claim 47, the input comprising information for linking data to relevant multidimensional objects.

51. The method of claim 50, further comprising:

switching the analysis component to an operational mode to access the cache only after rebuilding the cache; and

processing queries via the analysis component utilizing the cache.

52. A method of proactive caching, comprising:

providing at least one multidimensional object providing dynamic multidimensional analysis data derived from a database;

building at least one cache providing dynamic multidimensional analysis data derived from at least one multidimensional object;

providing an input for determining a cache rebuild parameter;

determining if the cache rebuild parameter has been satisfied;

switching the analysis component operational mode to access the multidimensional objects when the cache rebuild parameter has been satisfied; and

rebuilding the cache based on relevant multidimensional objects.

53. A method of claim 52, the cache rebuild parameter comprising at least one selected from the group consisting of a forced refresh rate, a quiet time delay, and a quiet time delay override.

54. A method of optimizing proactive caching, comprising:

inputting a query into a query interpreter;

loading the query into at least two sessions, each session accessing different database derived data sources;

comparing resulting data from the sessions to determine differences via a result comparator; and

analyzing the differences of the sessions to adaptively optimize the proactive caching by the query interpreter.

55. A data packet transmitted between two or more computer components that facilitates data analysis, the data packet comprising dynamic multidimensional analysis data, based, in part, on a proactive caching structure.

56. The data packet of claim 55, the dynamic multidimensional analysis data comprising at least one selected from the group consisting of a multidimensional object data derived from a database and a dynamic multidimensional analysis data derived from at least one multidimensional object.

57. The data packet of claim 56, the dynamic multidimensional analysis data derived from at least one multidimensional object comprising data from a MOLAP cache.

58. The data packet of claim 56, the multidimensional object data comprising data from an OLAP object.

59. The data packet of claim 58, the OLAP object comprising a ROLAP object.

60. The data packet of claim 56, the database comprising a relational database.

61. A computer readable medium storing computer executable components of a system for facilitating data analysis, comprising a proactive caching system that provides information associated with a data set, based, at least in part, upon at least one OLAP variant cache.

62. The medium of claim 61, the OLAP variant cache comprising MOLAP cache.

63. The medium of claim 61, the proactive caching system comprising:

at least one analysis component capable of being proactively coupled to a multidimensional object and to a cache comprising dynamic multidimensional analysis data derived from the multidimensional object, proactively controlling access to the multidimensional object and the cache.