Distributed redundant database

• US 5,890,156 A
• Filed: 05/02/1996
• Issued: 03/30/1999
• Est. Priority Date: 05/02/1996
• Status: Expired due to Term

• Current Assignee: Sound View Innovations, LLC (Sound View Innovation Holdings, LLC)
• Original Assignee: Alcatel-Lucent USA, Inc. (Nokia Corporation)
• Inventors: Vatteroni, James R., Robertson, Walter C., Kasam, Rashmi K., Rekieta, Thomas W., Chen, Chyuan-Meei
• Primary Examiner(s): Amsbury, Wayne
• Assistant Examiner(s): LEWIS, CHERYL RENEA
• Application Number:
  US08/640,544
• Time in Patent Office: 1,062 Days
• Field of Search: 395/608, 395/620, 395/200, 395/712, 707/10, 707/8, 707/9, 707/200
• US Class Current: 707/610
• CPC Class Codes: G06F 16/27 Replication, distribution or synchronisation of data between databases or within a distributed database system; Distributed database system architectures therefor
• CPC Class Codes: G06F 9/50 Allocation of resources, e.g. of the central processing unit [CPU]
• CPC Class Codes: Y10S 707/953 Organization of data
• CPC Class Codes: Y10S 707/959 Network
• CPC Class Codes: Y10S 707/99938 Concurrency, e.g. lock management in shared database
• CPC Class Codes: Y10S 707/99939 Privileged access

Abstract

A distributed redundant database architecture (10) primarily for storing a plurality of wireless service subscriber files (84, 86) is provided. The database architecture (10) include a first set of processors (44 46, 48) each having a shared memory (72, 74, 76), where the processors (44, 46, 48) are divided into processor groups (38, 40, 42). A memory storage device (80, 82) is coupled to each processor group (38, 40, 42) in a multi-initiator mode so that each memory storage device (80, 82) is accessible by the processors (44, 46, 48) in its respective processor group (38, 40, 42). The plurality of subscriber files (84, 86) are assigned to the processor groups (38, 40, 42) and further assigned to the processors (44, 46, 48) in each processor group (38, 40, 42) to achieve a substantially even distribution. The plurality of subscriber files are stored in the memory storage devices (80, 82) of respective assigned processor groups (38, 40, 42), and are mapped into the respective shared memories (72, 74, 76) of the processors. A database manager (52, 54, 66) associated with the processors and processor groups is provided for managing the distribution and redistribution of the subscriber files to the processors. A second set of similarly configured processors residing at a remote site (14) maintains a second copy of the subscriber files (84, 86) to provide database backup.

Claims

1. A distributed redundant database architecture for storing a plurality of files, comprising:

a first set of processors each having a shared memory, said processors being divided into processor groups, each processor of a particular processor group having access to the shared memory of each of the other processors within the particular processor group;

a memory storage device coupled to each processor group, each memory storage device being simultaneously accessible by processors in its respective processor group;

said plurality of files each having a plurality of records, and being assigned to said processor groups and further to said processors in each processor group to achieve a substantially even distribution;

said plurality of files being stored in said memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said plurality files to said processors.

2. The distributed redundant database architecture, as set forth in claim 1, further comprising:

a second set of processors being physically remote from said first set of processors, and being coupled therewith, said second set of processors also being grouped into processor groups, each processor group being coupled to a memory storage device;

said plurality of files being assigned to said processor groups and further to said processors in each processor group to achieve a substantially even distribution in a similar manner as said first set of processors;

said plurality of files being stored in said memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors in a similar manner as said first set of processors; and

a second database manager associated with said processor groups for managing the distribution of said plurality of files to said second set of processors.

3. The distributed redundant database architecture, as set forth in claim 2, wherein said first and second sets of processors each actively process database queries and updates associated with a different half of records stored in said plurality of files, said memory storage devices of both sets of processors each storing a copy of said plurality of files.

4. The distributed redundant database architecture, as set forth in claim 3, further comprising first and second database synchronization processes associated with said first and second sets of processors respectively for transmitting updated data in respective half of records stored in said plurality of files to one another.

5. The distributed redundant database architecture, as set forth in claim 4, wherein said updated data being transmitted between said first and second sets of processors are transient data.

6. The distributed redundant database architecture, as set forth in claim 4, further comprising:

a route table recording the file assignment of said plurality of files to processors in said first and second sets of processors; and

first and second database sync managers exchanging said route tables with one another.

7. The distributed redundant database architecture, as set forth in claim 1, further comprising a database reconfiguration process executed by said database manager for redistributing said files to said processors as one or more of said processors become inactive and as inactive processors become active.

8. The distributed redundant database architecture, as set forth in claim 1, further comprising a load balancing process executed by said database manager for moving one or more selected files from one processor to another to evenly distribute the amount of processing required to service queries and updates to said plurality of files.

9. The distributed redundant database architecture, as set forth in claim 1, further comprising a memory synchronization process executed by said database manager for copying said plurality of files stored in said shared memories of said processors to its respective memory storage devices on a periodic basis.

10. The distributed redundant database architecture, as set forth in claim 1, wherein said memory storage device includes a redundant memory storage device storing a second copy of said files, both said memory storage devices being updated simultaneously.

11. The distributed redundant database architecture, as set forth in claim 1, wherein each file includes a plurality of records, each record containing information related to a subscriber of a wireless service having a personal subscriber number, each subscriber record being contained in a file having a file index equal to the result of a modulo operation on said personal subscriber number.

12. A distributed redundant database architecture for storing a plurality of files, comprising:

a first set of processors each having a shared memory, said processors being divided into processor groups;

a memory storage device coupled to each processor group, each memory storage device being simultaneously accessible by processors in its respective processor group;

said plurality of files each having a plurality of records, and being assigned to said processor groups and further to said processors in each processor group to achieve a substantially even distribution;

said plurality of files being stored in said memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said plurality files to said processors, wherein each file is associated with a Red-Black Tree data structure having a plurality of nodes corresponding to a plurality of records of the file, said nodes of said Red-Black Tree being organized for speedy searches to facilitate quick location and access of corresponding records in said file.

13. A distributed redundant database architecture for storing a plurality of files, comprising:

a first set of processors each having a shared memory, said processors being divided into processor groups;

a memory storage device coupled to each processor group, each memory storage device being simultaneously accessible by processors in its respective processor group;

said plurality of files each having a plurality of records, and being assigned to said processor groups and further to said processors in each processor group to achieve a substantially even distribution;

said plurality of files being stored in said memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said plurality files to said processors, wherein said memory storage devices are multi-port disks.

14. A distributed redundant database architecture for storing a plurality of files, comprising:

a first set of processors each having a shared memory, said processors being divided into processor groups;

a memory storage device coupled to each processor group, each memory storage device being simultaneously accessible by processors in its respective processor group;

said plurality of files each having a plurality of records, and being assigned to said processor groups and further to said processors in each processor group to achieve a substantially even distribution;

said plurality of files being stored in said memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said plurality files to said processors, wherein said memory storage devices are coupled in a multi-initiator mode with said processor group.

15. A distributed redundant database architecture for storing a plurality of files, comprising:

a first set of processors each having a shared memory, said processors being divided into processor groups;

a memory storage device coupled to each processor group, each memory storage device being simultaneously accessible by processors in its respective processor group;

said plurality of files each having a plurality of records, and being assigned to said processor groups and further to said processors in each processor group to achieve a substantially even distribution;

said plurality of files being stored in said memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said plurality files to said processors, wherein said memory storage device is partitioned into X filesystems, said files being generally evenly distributed among said filesystems.

16. The distributed redundant database architecture, as set forth in claim 15, wherein each file includes a plurality of records, each record containing information related to a subscriber of a wireless service having a personal subscriber number, said file having a file index being equal to the result of a modulo Y operation on said personal subscriber number, where Y is the total number of subscriber files.

17. A distributed redundant database architecture for a wireless communications network, comprising:

first and second service control points, each service control point servicing one half of wireless service subscribers, each service control point including;

a first set of processors each having a shared memory, said processors being divided into processor groups, each processor of a particular processor group having access to the shared memory of each of the other processors within the particular processor group, said processors of said processor groups being assigned subscriber files having records containing subscriber information, said files being substantially evenly distributed among said processors in each processor group;

a pair of mirrored memory storage devices coupled to each processor group, each pair of mirrored memory storage device being accessible by its respective processor group;

said subscriber files being stored in said pairs of mirrored memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said files to said processors.

18. The distributed redundant database architecture, as set forth in claim 17, further comprising:

a route table recording the file assignment of said subscriber files to processors in said first and second service control points; and

first and second database sync processes in respective service control points exchanging said route tables with one another.

19. The distributed redundant database architecture, as set forth in claim 17, further comprising a database reconfiguration process executed by said database manager for redistributing said subscriber files to said processors as one or more of said processors become inactive or one or more inactive processors become active.

20. The distributed redundant database architecture, as set forth in claim 17, further comprising a load balancing process executed by said database manager for moving one or more selected subscriber files from one processor to another to evenly distribute the amount of processing required to service queries and updates to said plurality of subscriber files.

21. The distributed redundant database architecture, as set forth in claim 17, further comprising a memory synchronization process executed by said database manager for copying said plurality of subscriber files stored in said shared memories of said processors to its respective pairs mirrored memory storage devices on a periodic basis.

22. The distributed redundant database architecture, as set forth in claim 17, wherein each subscriber file includes a plurality of records, each record containing information related to a wireless service subscriber having a personal subscriber number, said record of said subscriber being stored in a subscriber file having a file index equal to the result of a modulo operation on said personal subscriber number.

23. A distributed redundant database architecture for a wireless communications network, comprising:

first and second service control points, each service control point servicing one half of wireless service subscribers, each service control point including;

a first set of processors each having a shared memory, said processors being divided into processor groups, said processors of said processor groups being assigned subscriber files having records containing subscriber information, said files being substantially evenly distributed among said processors in each processor group;

a pair of mirrored memory storage devices coupled to each processor group, each pair of mirrored memory storage device being accessible by its respective processor group;

said subscriber files being stored in said pairs of mirrored memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said files to said processors; and

first and second database synchronization processes associated with said first and second service control points respectively for transmitting updated transient data, such as subscriber location, in said subscriber files to one another.

24. A distributed redundant database architecture for a wireless communications network, comprising:

first and second service control points, each service control point servicing one half of wireless service subscribers, each service control point including;

a first set of processors each having a shared memory, said processors being divided into processor groups, said processors of said processor groups being assigned subscriber files having records containing subscriber information, said files being substantially evenly distributed among said processors in each processor group;

a pair of mirrored memory storage devices coupled to each processor group, each pair of mirrored memory storage device being accessible by its respective processor group;

said subscriber files being stored in said pairs of mirrored memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said files to said processors, wherein each subscriber file is associated with a Red-Black Tree data structure having a plurality of nodes corresponding to a plurality of records of the subscriber file, said nodes of said Red-Black Tree being organized for speedy searches to facilitate quick location and access of corresponding records in said subscriber file.

25. A distributed redundant database architecture for a wireless communications network, comprising:

first and second service control points, each service control point servicing one half of wireless service subscribers, each service control point including;

a first set of processors each having a shared memory, said processors being divided into processor groups, said processors of said processor groups being assigned subscriber files having records containing subscriber information, said files being substantially evenly distributed among said processors in each processor group;

a pair of mirrored memory storage devices coupled to each processor group, each pair of mirrored memory storage device being accessible by its respective processor group;

said subscriber files being stored in said pairs of mirrored memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said files to said processors, wherein said mirrored memory storage devices are multi-port disks.

26. A distributed redundant database architecture for a wireless communications network, comprising:

first and second service control points, each service control point servicing one half of wireless service subscribers, each service control point including;

a first set of processors each having a shared memory, said processors being divided into processor groups, said processors of said processor groups being assigned subscriber files having records containing subscriber information, said files being substantially evenly distributed among said processors in each processor group;

a pair of mirrored memory storage devices coupled to each processor group, each pair of mirrored memory storage device being accessible by its respective processor group;

said subscriber files being stored in said pairs of mirrored memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said files to said processors, wherein said mirrored memory storage devices are coupled to said processors in each processor group in a multi-initiator mode.

27. A distributed redundant database architecture for a wireless communications network, comprising:

first and second service control points, each service control point servicing one half of wireless service subscribers, each service control point including;

a first set of processors each having a shared memory, said processors being divided into processor groups, said processors of said processor groups being assigned subscriber files having records containing subscriber information, said files being substantially evenly distributed among said processors in each processor group;

a pair of mirrored memory storage devices coupled to each processor group, each pair of mirrored memory storage device being accessible by its respective processor group;

said subscriber files being stored in said pairs of mirrored memory storage devices of respective assigned processor groups, and further being mapped into said respective shared memories of said processors; and

a database manager associated with said processor groups for managing the distribution of said files to said processors, wherein said mirrored memory storage devices are each partitioned into X filesystems, said files being generally evenly distributed among said filesystems.

28. The distributed redundant database architecture, as set forth in claim 27, further comprising a database reconfiguration process for redistributing said filesystems among said processors in a processor group in the event one or more processors in said processor group becomes inactive or one or more inactive processors become active.

29. The distributed redundant database architecture, as set forth in claim 27, wherein each file includes a plurality of records, each record containing information related to a subscriber of a wireless service having a personal subscriber number, said file having a file index being equal to the result of a modulo Y operation on said personal subscriber number, where Y is the total number of subscriber files.

30. A method for organizing data storage and access in a database, comprising the steps of:

organizing data in said database into a plurality of files, each file having a file index and contain a plurality of records;

providing a first set of processors and dividing said processors into processor groups, each processor group coupled to a memory storage device, each processor within a particular processor group having a shared memory that is accessible by every other processor within the particular processor group;

assigning said plurality of files to said processor groups and further to said processors in each processor group in an evenly distributed manner; and

storing said assigned files in said memory storage device of each processor group according to said assignment, and mapping said files assigned to each processor from said memory storage device to the shared memory of each processor.

31. The method, as set forth in claim 30, further comprising the steps of:

receiving database queries with reference to specific files;

consulting a route table for file assignment to processors; and

routing said received database queries to processors assigned to said specific files.

32. The method, as set forth in claim 30, further comprising the steps of:

receiving data updates with reference to specific files;

consulting a route table for file assignment to processors; and

updating data stored in said shared memory of said processors assigned to said specific files.

33. The method, as set forth in claim 32, further comprising the steps of updating data stored in said memory storage device with data stored in said shared memories of each processor in said respective processor group.

34. The method, as set forth in claim 30, further comprising the steps of:

receiving data updates with reference to specific files;

consulting a route table for file assignment to processors; and

updating data stored in said shared memory of said processors assigned to said specific files and said respective memory storage device of the processor group.

35. The method, as set forth in claim 30, further comprising the step of auditing said processors to determine file assignments thereto.

36. The method, as set forth in claim 30, further comprising the steps of:

detecting one or more processors going inactive; and

reconfiguring the distribution of files to remaining active processors.

37. The method, as set forth in claim 30, further comprising the steps of:

monitoring data access and update traffic being processed by each processor; and

moving one or more files among said processors to balance the traffic load on the processors.

38. The method, as set forth in claim 30, further comprising the steps of:

providing a backup memory storage device as a mirror device for each memory storage device; and

performing all data updates to both memory storage devices.

39. The method, as set forth in claim 30, further comprising the steps of:

providing a second set of processors and dividing said processors into processor groups, each processor group coupled to a memory storage device;

assigning said plurality of files to said processor groups and further to said processors in each processor group in an evenly distributed manner as in said first set of processors; and

storing said assigned files in said memory storage device of each processor group according to said assignment, and mapping said files assigned to each processor from said memory storage device to a shared memory of each processor.

40. The method, as set forth in claim 30, further comprising the steps of synchronizing the data stored in shared memories of said first and second sets of processors.

41. The method, as set forth in claim 40, wherein the data synchronizing step further comprises the steps of exchanging data updates to said files.

42. The method, as set forth in claim 30, further comprising the steps of:

detecting one or more processors going inactive;

reconfiguring the distribution of files to remaining active processors;

monitoring data access and update traffic being processed by each processor;

moving one or more files among said processors to balance the traffic load on the processors; and

coordinating said reconfiguration and file moving steps.

43. The method, as set forth in claim 30, further comprising the steps of:

detecting one or more inactive processors becoming active; and

reconfiguring the distribution of files to said newly active processors.

44. A method for organizing data storage and access in a database, comprising the steps of:

organizing data in said database into a plurality of files, each file having a file index and contain a plurality of records;

providing a first set of processors and dividing said processors into processor groups, each processor group coupled to a memory storage device;

assigning said plurality of files to said processor groups and further to said processors in each processor group in an evenly distributed manner; and

storing said assigned files in said memory storage device of each processor group according to said assignment, and mapping said files assigned to each processor from said memory storage device to a shared memory of each processor;

receiving data updates with reference to specific files;

consulting a route table for file assignment to processors; and

updating data stored in said shared memory of said processors assigned to said specific files;

updating data stored in said memory storage device with data stored in said shared memories of each processor in said respective processor group, wherein said data updating step includes the steps of;

initializing a synchronization timer;

copying at least a portion of a file at the expiration of said synchronization timer to said memory storage device; and

resetting said synchronization timer and repeating said copying and timer resetting steps to copy all files.

45. A method for organizing data storage and access in a database, comprising the steps of:

organizing data in said database into a plurality of files, each file having a file index and contain a plurality of records;

providing a first set of processors and dividing said processors into processor groups, each processor group coupled to a memory storage device;

assigning said plurality of files to said processor groups and further to said processors in each processor group in an evenly distributed manner;

storing said assigned files in said memory storage device of each processor group according to said assignment, and mapping said files assigned to each processor from said memory storage device to a shared memory of each processor;

synchronizing the data stored in shared memories of said first and second sets of processors;

marking each record containing modified data as dirty as the data is being modified, said files having a plurality of records;

transferring a copy of all dirty record of all the files to a buffer; and

transmitting the buffer to said second set of processors.

46. The method, as set forth in claim 45, wherein said transferring step includes the step of transferring records of files serviced by each processor individually, and said transmitting step includes the step of transmitting said buffer to a corresponding processor in said second set of processors.

47. A method for organizing data storage and access in a database, comprising the steps of:

organizing data in said database into a plurality of files, each file having a file index and contain a plurality of records;

providing a first set of processors and dividing said processors into processor groups, each processor group coupled to a memory storage device;

assigning said plurality of files to said processor groups and further to said processors in each processor group in an evenly distributed manner; and

storing said assigned files in said memory storage device of each processor group according to said assignment, and mapping said files assigned to each processor from said memory storage device to a shared memory of each processor;

associating a file index with each file;

storing information on a wireless service subscriber in a record in one of said files;

associating a personal subscriber number with said wireless service subscriber; and

performing a modulo operation on said personal subscriber number to compute for said file index thereof.

48. A method for organizing data storage and access in a database, comprising the steps of:

organizing data in said database into a plurality of files, each file having a file index and contain a plurality of records;

providing a first set of processors and dividing said processors into processor groups, each processor group coupled to a memory storage device;

assigning said plurality of files to said processor groups and further to said processors in each processor group in an evenly distributed manner; and

storing said assigned files in said memory storage device of each processor group according to said assignment, and mapping said files assigned to each processor from said memory storage device to a shared memory of each processor;

associating each record of a file with a node in a Red-Black Tree data structure; and

searching among nodes of said Red-Black Tree data structure to locate a specific record.

49. The method, as set forth in claim 48, wherein said associating step comprises the step of defining a pointer in each node for pointing at respective records.