Relational data base memory utilization analyzer

• US 5,237,681 A
• Filed: 05/24/1991
• Issued: 08/17/1993
• Est. Priority Date: 05/24/1991
• Status: Expired due to Fees

• Current Assignee: Telcordia Technologies Incorporated (Telefonaktiebolaget LM Ericsson)
• Original Assignee: Bell Communications Research, Inc. (Telefonaktiebolaget LM Ericsson)
• Inventors: Kagan, Jeremy S., Sanders, Leo S., Lutin, John N.
• Primary Examiner(s): Lee, Thomas C.
• Assistant Examiner(s): Loomis, John C.
• Application Number:
  US07/705,187
• Time in Patent Office: 816 Days
• Field of Search: 364/DIG. 1, 364/DIG.2, 395/575, 395/600, 395/650, 395/700, 371/19, 371/21.1.21.6
• US Class Current: 1/1
• CPC Class Codes: G06F 16/284 Relational databases
• CPC Class Codes: Y10S 707/99945 Object-oriented database structure processing
• CPC Class Codes: Y10S 707/99948 Application of database or data structure, e.g. distributed, multimedia, or image

Abstract

Methods and apparatus are set forth for identifying the actual population of data within computer memory utilized to support a relational data base, where memory is defined herein as being "populated" if it is both allocated and actually filled with data. By identifying data population (thereby identifying unpopulated areas of memory versus simply making a determination of memory allocation without regard to how it is being used), utilities can be designed to recover memory resources, data base management techniques can be revised to more conservatively allocate memory, etc. Accordingly, memory resources can be more efficiently used and in certain instances the cost of adding memory to enhance a system can be delayed or eliminated.

Claims

1. A computer implemented method for performing relational data base memory utilization analysis for a data base having allocated memory such that memory actually utilized is determined, comprising the computer implemented steps of:

(a) selecting a set of data base files to be analyzed;

(b) specifying a set of relations, said set of relations being a selected subset of all of the relations located within said set of data base files, for which memory utilization information is to be determined, said relations including tuples for storing data, each tuple having a size and a data activity state from the group of states including active and inactive;

(c) identifying tuples of each said relation in said set of relations in the active state and thus actually storing data; and

(d) determining the memory actually utilized for each said relation based upon said identified tuples and the size of said identified tuples such that the memory actually utilized in said specified set of relations within said selected set of data base files is determined.

2. A method as set forth in claim 1 wherein a plurality of said relations have a data page structure, said structure including a plurality of data pages and each said data page having a status bit map, said status bit map indicating said data activity state of each said tuple of said data page and wherein said step of identifying further comprises the step of detecting the contents of said status bit map of each said data page to determine whether the tuples of each said data page are in the active state and thus actually storing data.

3. A method as set forth in claim 2 wherein said step of identifying comprises the step of analyzing the data contained in each tuple on each page of each said relation that does not have a data page structure that includes a status bit map.

4. A computer implemented method for performing relational data base memory utilization analysis for a data base having allocated memory such that memory actually utilized is determined, comprising the computer implemented steps of:

(a) selecting a set of data base files to be analyzed;

(b) specifying a set of relations, said set of relations being a selected subset of all of the relations located within said set of data base files, for which memory utilization information is to be determined, each said relation including a plurality of tuples for storing data and each said tuple having a size and a data activity state from the group of states including active and inactive;

(c) determining which of said relations in said set of relations has a data page structure having a plurality of data pages, each said data page including a status bit map indicating said data activity state of each tuple;

(d) detecting the contents of said status bit map on each page of each relation having said data page structure that includes said status bit map and identifying tuples in said active state and thus actually storing data;

(e) analyzing the data contained in each tuple on each page of each relation that does not have a data page structure that includes a status bit map and identifying tuples in said active state and thus actually storing data; and

(f) determining the memory actually utilized for each said relation based upon said identified tuples and the size of said identified tuples such that the memory actually utilized in said specified set of relations within said selected set of data base files is determined.

5. A method as set forth in claim 4 wherein said step of analyzing the data contained in each tuple on each page further comprises the steps of:

(b) determining the number of tuples on each data page for each relation that does not have a data page structure that includes a status bit map;

(b) utilizing the tuple size and the determined number of tuples on a given data page to locate each tuple on said given data page; and

c) analyzing the content of each tuple on said given data page.

6. A method as set forth in claim 5 wherein said tuples are sequentially ordered on said data page and said step of analyzing the data contained in each tuple on each page further comprises the steps of:

(a) locating the first tuple on a given datapage; and

(b) utilizing the tuple size and the determined number of tuples located on a given data page to sequentially analyze the contents of each tuple on said given data beginning with the first tuple located thereon.

7. A method as set forth in claim 5 wherein said step of analyzing the data contained in each tuple on each page further comprises the step of determining if a preselected bit pattern, indicating that a given tuple is unused, is present in the tuplex being examined on said given data page.

8. A method as set forth in claim 4 further comprising the steps of:

(a) transforming said activity state of each tuple for each relation in said set of relations and for each file in said set of data base files, into a set of signals indicating memory population; and

(b) outputting said set of signals indicating memory population to thereby provide an analysis of data base memory utilization.

9. A method as set forth in claim 8 wherein said relational data base memory includes a set of Office Dependent Data (ODD) files created by a telephone switching system.

10. A method as set forth in claim 9 wherein each relation in each of said ODD files is represented by an entry in an access dictionary having access information that it utilized to access data pages of each relation in said set of relations in order to determine said data activity state of each tuple on each data page of each relation.

11. A computer implemented method for identifying the actual population of data within computer system memory utilized to support a set of relational data bases, comprising the computer implemented steps of:

(a) creating a list of files, wherein the set of files on said list constitute a preselected subset of said set of data bases and further wherein each file on said list contains a preselected set of relations, said preselected set of relations being a subset of all relations in said preselected subset of said set of data bases, to be analyzed, each relation including tuples for storing data, each tuple having a size and a data activity state from the group of states including active and inactive wherein said activity state indicates whether the tuple is populated with data;

(b) creating a list of said preselected set of relations from each of said files of said list of files;

(c) loading said list of files and said list of said preselected set of relations into a digital computer that is offline with respect to said computer system;

(d) determining in said digital computer said activity state of each tuple in each relation on said list of relations, said digital computer having access to said set of files; and

(e) compiling information based upon said activity state of each tuple and its size to provide an indication of actual memory usage.

12. A method as set forth in claim 11 wherein the step of determining further comprises the steps of:

(a) determining which of the relations in said preselected set of relations has a data page structure that includes a status bit map for keeping track of tuples in a given relation in the active state and tuples in the given relation in the inactive state;

(b) identifying the number of tuples in said active state for each relation in said set of relations that has a data page structure that includes a status bit map by examining the contents of the status bit map on each page of each such relation; and

(c) identifying the number of tuples in said active state for each relation in said set of relations that does not have a data page structure that includes a status bit map by analyzing the contents of each tuple on each data page in each such relation.

13. A method as set forth in claim 12 wherein said set of relational data bases includes Office Dependent Data (ODD) files created by a telephone switching system that includes an administrative module, a communications module and a set of switch modules in which binary data files, corresponding to said ODD files, are maintained and updated in real time.

14. A method as set forth in claim 13 wherein each relation in each of said ODD files is represented by an entry in an access dictionary that is utilized to access the data pages of each relation in said set of relations in order to determine said data activity status of each tuple on each data page each relation.

15. A method as set forth in claim 14 wherein said step of compiling further comprises the step of tabulating information on said activity sate of each tuple and storing said tabulated information in a statistical file, said file reflecting, for each relation analyzed, a total number of tuples in a given relation compared with the number of tuples in said active state.

16. A method as set forth in claim 14 wherein each said data page contains a maximum number of tuples and said step of compiling further comprises the step of

tabulating the number of tuples in said active state for each data page of each relation,

computing, based upon said number of tuplex in an active state and said maximum number of tuplex per data page a percentage fill for each data page of each relation, and

storing said computed percentage fill and said number of tuples in said active state in a verbose file.

17. A method as set forth in claim 14 wherein said step of compiling further comprises the step of tabulating said information regarding said activity state of each tuple, identifying relations having data pages which contain no data, and storing information related to said identified relations in an empty file to provide information regarding all data pages within a given switch module that contain no data.

18. A method as set forth in claim 14 wherein each data page of each relation has an identification number and wherein said step of compiling further comprises the steps of

comparing said identification number of each said data page of each said relation to an offset which indicates a range of files to be analyzed and,

if said offset is greater than said identification number, storing said identification number in a range check file to provide an error indication whenever the page identifier for a given page is outside the range of the switch module within which the page is properly located.

19. A computer implemented method for performing memory population analysis on a relational data base, wherein the contents of said data base include a first set of relations, each relation including tuples for storing data, each said tuple having a size and a data activity sate from the ground of sates including active and inactive and each relation having an associated status bit map for indicating the data activity state of each tuple within a given relation, and a second set of relations each created utilizing a storage and access method that does not maintain an associated bit map, comprising the computer implemented steps of:

(a) inputting, to a digital computer having access to the contents of said data base, a first set of signals specifying the relations to be analyzed, said relations to be analyzed being a selected subset of all relations in said data base,

(b) inputting, to said digital computer, a second set of signals specifying the storage and access method associated with each relation of said relations to be analyzed;

(c) determining, from said second set of signals, whether a given relation being analyzed has an associated status bit map;

(d) accessing the status bit map for each relation being analyzed and belonging to said first set of relations;

(e) identifying tuples in the inactive data activity state in each relation being analyzed and belonging to said first set of relations, directly from the status bit map associated with each such relation; and

(f) compiling information based on said identified tuples in the inactive data activity sate and the size of each said identified tuple.

20. A method as set forth in claim 19 further comprising the step of, whenever said second set of relations in non-empty, counting tuples within each such relation by analyzing the contents of each tuple to determine the data activity state of each tuple in each such relation.

21. A computer implemented method for analyzing a set of relational data bases that are stored in the memory of a set of switch modules includes in a telephone switching system, wherein said set of relational data bases comprise static protected Office Dependent (ODD) that is classified as non-redundant if required for call processing based on special features and service required by the specific lines and trunks which are assigned to a particular switch module, and redundant if stored in the memory of every switch module without regard to the nature of the lines and trunks connected, comprising the computer implemented steps of:

(a) entering a list of ODD files to be analyzed into a digital computer;

(b) initializing a hard coded table, stored within said computer, to include a list of relations containing redundant data within said ODD files;

(c) initializing a hard coded table, stored within said computer, to include a list of relations containing non-redundant data within said ODD files;

(d) inputting to said digital computer, from media containing the ODD files in binary form, an ODD file to be analyzed;

(e) initializing the ODD file input to said digital computer;

(f) reading an access dictionary to determine head table, intermediate table and data page location information for each relation listed in the redundant and non-redundant hard coded tables, said relations listed in the redundant and non-redundant hard coded tables being a subset of all relations in said ODD files, each said relation including tuples for storing data, each said tuple having a size and a data activity state from the group of states including active and inactive wherein if said tuple is in an active state, then the tuple is populated and stores data;

(g) processing each relation listed in the redundant and non-redundant hard coded tables for the ODD file input in step (d) to determine said data activity sate of each tuple in each listed relation;

(h) outputting memory population information for each processed relation based on the activity state and size of each tuple; and

(i) if further ODD files on said list of ODD files remain to be processed, returning to step (d) and specifying the next ODD file on said list of ODD files as the ODD file to be analyzed if any such files remain and terminating the relational data base analysis if not further ODD files remain to be processed.

22. A method as set forth in claim 21 wherein said memory population information output is provided in terms of pages, tuples and bytes.

23. A relational data base memory analyzer, for identifying the actual population of data within a computer system memory utilized to support a set of relational data bases, wherein said data base analyzer is provided with a list of files, an a list of preselected relations, said list of preselected relations being a subset of all relations in said files, within each file, to be analyzed, comprising:

(a) a digital computer which is associated with but offline with respect to said computer system;

(b) means for loading said list of files and said list of said preselected relations into said digital computer, each relation including tuples for storing data, each tuple having a size and a data activity state from the group of states including active and inactive wherein said activity state indicates whether the tuple is populated with data;

(c) program means for determining said data activity state of each tuple in each relation on said list of relations, said means having access to the binary version of the set of files on said list of files;

(d) means for compiling data activity state for each tuple and

(e) means for outputting said compiled data activity states to the computer system to provide, based upon said compiled activity states and the size of each tuple, and indication of actual memory usage.

24. Apparatus for performing memory population analysis on a relation data base, wherein each relation of the relational data base includes tuples for storing data, each tuple having a size and a data activity state from the group of states including active and inactive, said activity state indicating whether the tuple is populated with data and wherein the contents of said data base include a first set of relations each having an associated status bit map for indicating the data activity state of each tuple in the relation within a given relation, and a second set of relations each created utilizing a storage and access method that does not maintain an associated bit map, comprising:

(a) a digital computer having access to the contents of said data base;

(b) means for inputting a first set of signals to said digital computer, wherein said first set of signals specify the relations to be analyzed, said relations to be analyzed being a selected subset of all relations in said data base,

(c) means for inputting a second set of signals to said digital computer, wherein said second set of signal specify the storage and access method associated with each relation of said relations to be analyzed;

(d) tuple status determination means, located within said digital computer, for (1) determining, from said second set of signals, whether a given relation being analyzed has an associated status bit map;

(2) accessing the status bit map for each relation being analyzed and belonging to said first set of relations;

(3) identifying tuples in the inactive data activity sate known as unused tuples in each relation being analyzed and belonging to said first set of relations, directly from the status bit map associated with each such relation; and

(4) compiling information based upon the identified tuples and the size of each identified tuple, said information being known as memory status information; and

(e) means for outputting said memory status information from said digital computer.

25. Apparatus as set forth in claim 24 wherein said tuple status determination means is further operative to determine unused tuples within each relation in said second set of relations by analyzing the contents of each tuple located therein.

26. Apparatus as set forth in claim 25 wherein said tuple status determination means is a computer implementing computer program means.

27. Apparatus for analyzing a set of relational data bases that are stored in the memory of a set of switch modules included in a telephone switching system, wherein said set of relational data bases comprise static protected Office Dependent Data (ODD) that is classified as non-redundant if required for call processing based on special features and services required by the specific lines and trunks which are assigned to a particular switch module, and redundant if stored in the memory of every switch module without regard to the nature of the lines and trunks connected, comprising:

(a) a programmable digital computer;

(b) means for entering a list of ODD files to be analyzed into said digital computer;

(c) means for initializing a hard coded table, stored within said computer, to include a list of relations containing redundant data within said ODD files;

(d) means for initializing a hard coded table, stored within said computer, to include a list of relations containing non-redundant data within said ODD files;

(e) means for inputting to said digital computer, from media containing the ODD files in binary form, an ODD file to be analyzed;

(f) means for (1) initializing the ODD file input to said digital computer;

(2) reading an access dictionary to determine head table, intermediate table and data page location information for each relation listed in the redundant and non-redundant hard coded tables, said relations listed in the redundant and non-redundant hard coded tables being a subset of all relations in said ODD files, each relation including tuples for storing data, each tuple having a size and a data activity state from the group of states including active and inactive wherein said activity state indicates whether the tuple is populated with data;

(3) processing each relation listed in the redundant and non-redundant hard coded tables for the ODD file, by said means for inputting, to determine the data activity state of each tuple in each listed relation;

(4) outputting information based upon tuples in the inactive data activity state and the size of each tuple in the inactive data state known as memory population information for each processed relation; and

(5) determining if further ODD files on said list of ODD files remain to be processed, returning to step (d) and specifying the next ODD file on said list of ODD files as the ODD file to be analyzed if any such files remain and terminating the relational data base analysis if not further ODD files remain to be processed.