Automatically identifying an optimal set of attributes to facilitate generating best practices for configuring a networked system

• US 7,779,122 B2
• Filed: 05/28/2008
• Issued: 08/17/2010
• Est. Priority Date: 01/21/2008
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Routray, Ramani Ranjan, Sarkar, Prasenjit, Tan, Chung-hao, Butler, Eric Kevin, Yang, Kiyoung
• Primary Examiner(s): Dinh; Khanh Q
• Application Number:
  US12/127,916
• Publication Number:
  US 20090187518A1
• Time in Patent Office: 811 Days
• Field of Search: 709/224, 709/226, 709/228, 709/231, 455/456.1, 382/240
• US Class Current: 709/224
• CPC Class Codes: H04L 41/08 Configuration management of networks or network elements address allocation H04L61/50

Abstract

A method and system for automatically identifying an optimal set of attributes of entities included in a networked system. Entity types are ranked based on information gain. A first classification accuracy relative to a first entity type is determined. The first entity type is the top-ranked entity type or a first aggregate entity type. A second entity type is selected base on the ranking. A database join of a first set of attributes associated with the first entity type and a second set of attributes associated with the second entity type is performed. A second classification accuracy relative to a second aggregate entity type generated by the join is determined. In response to determining that the second classification accuracy is not greater than the first classification accuracy, an optimal set of attributes contributing to a problem in the networked system is identified as the first set of attributes.

Claims

1. A computer-implemented method of automatically identifying an optimal set of attributes of entities to facilitate generating best practices for configuring a networked system, comprising:

ranking, by a computing system and based on a plurality of information gain values, a plurality of entity types of a plurality of entities included in said networked system;

determining, by said computing system and subsequent to said ranking, a first classification accuracy relative to a first entity type, wherein said first entity type is a highest ranked entity type of said plurality of entity types based on said ranking or is a first aggregate entity type associated with two or more entity types of said plurality of entity types;

selecting, by said computing system and subsequent to said determining said first measurement, a second entity type of said plurality of entity types, wherein said selecting is based on said ranking;

performing, by said computing system, a database join operation on a first set of one or more attributes of one or more entities of said first entity type and a second set of one or more attributes of one or more entities of said second entity type, wherein a result of said performing is a second aggregate entity type;

determining, by said computing system, a second classification accuracy relative to said second aggregate entity type;

determining, by said computing system, that said second classification accuracy is less than or equal to said first classification accuracy;

identifying, by said computing system and in response to said determining that said second measurement is less than or equal to said first measurement, an optimal set of one or more attributes as said first set of one or more attributes, wherein said optimal set contributes to a problem associated with said networked system; and

storing said optimal set in a data repository coupled to said computing system.

2. The method of claim 1, further comprising generating a set of best practices for configuring said networked system to avoid said problem, wherein a configuration of said networked system that follows said set of best practices excludes any entity having said optimal set of one or more attributes.

3. The method of claim 1, wherein said ranking said plurality of entity types comprises determining an information entropy value

$H$

( T ) ⁢

⁢

as ⁢

⁢

∑

j = 1 n ⁢

⁢

p ⁡

( j ) ⁢

log 2 ⁡

( 1 p ⁡

( j ) ) ,

wherein T is an entity type of said plurality of entity types, wherein p(j) is a probability of an occurrence of an outcome j of T, and wherein n is a total number of possible outcomes of T.

4. The method of claim 3, wherein said ranking said plurality of entity types further comprises:

identifying a set of entities that report said problem, wherein said set of entities is included in said plurality of entities;

determining, subsequent to said identifying said set of entities, P true data paths that are internal to said set of entities, wherein P_i true data paths of said P true data paths are included in an i-th partition of k partitions of a configuration of said networked system, and wherein k>

1;

determining, subsequent to said identifying said set of entities, N false data paths that are external to said set of entities, wherein N_i false data paths of said N false data paths are included in said i-th partition of said configuration; and

determining an entropy value Entropy(P_i,N_i) as

$-\left(\frac{P_i}{P_i+N_i}\times <br><br>{\log }_2<br><br>\frac{P_i}{P_i+N_i}+\frac{N_i}{P_i+N_i}\times <br><br>{\log }_2<br><br>\frac{N_i}{P_i+N_i}\right).$

5. The method of claim 4, wherein said ranking said plurality of entity types further comprises:

determining an information gain value Gain(T) of said plurality of information gain values as

$H<br><br>\left(T\right)-\underset{i=1}{\overset{k}{\sum <br><br>}}<br><br><br><br>\frac{P_i+N_i}{P+N}\times <br><br>\mathrm{Entropy}<br><br><br><br>\left(P_i,N_i\right),$

wherein Gain(T) indicates an information gain of said entity type T; and

ordering said entity type T relative to one or more other entity types of said plurality of entity types based on Gain(T).

6. The method of claim 1, wherein said determining said first classification accuracy includes applying a decision tree algorithm on said first entity type, and wherein said determining said second classification accuracy includes applying said decision tree algorithm on said second aggregate entity type.

7. The method of claim 1, wherein said first entity type is said first aggregate entity type and wherein said method further comprises performing, by said computing system, prior to said determining said first measurement and in an order determined by said ranking, said database join operation recursively on n sets of one or more attributes associated with n entity types of said plurality of entity types in a one-to-one correspondence, wherein a result of said performing said database join operation recursively on said n sets of one or more attributes is said first aggregate entity type.

8. The method of claim 7, wherein said selecting said second entity type includes determining that a rank of said second entity type is lower than n ranks associated with said n entity types, and wherein said rank and said n ranks are determined by said ranking.

9. The method of claim 1, wherein said first entity type is said highest ranked entity type, and wherein said selecting said second entity type includes determining that a rank of said second entity type is a next highest ranked entity type of said plurality of entity types based on said ranking.

10. The method of claim 1, wherein said networked system includes a storage area network.

11. A computing system comprising a processor and a computer-readable memory unit coupled to said processor, said memory unit containing instructions that when executed by said processor implement a method of automatically identifying an optimal set of attributes of entities to facilitate generating best practices for configuring a networked system, wherein said method comprises:

ranking, based on a plurality of information gain values, a plurality of entity types of a plurality of entities included in said networked system;

determining, subsequent to said ranking, a first classification accuracy relative to a first entity type, wherein said first entity type is a highest ranked entity type of said plurality of entity types based on said ranking or is a first aggregate entity type associated with two or more entity types of said plurality of entity types;

selecting, subsequent to said determining said first measurement, a second entity type of said plurality of entity types, wherein said selecting is based on said ranking;

performing a database join operation on a first set of one or more attributes of one or more entities of said first entity type and a second set of one or more attributes of one or more entities of said second entity type, wherein a result of said performing is a second aggregate entity type;

determining a second classification accuracy relative to said second aggregate entity type;

determining that said second classification accuracy is less than or equal to said first classification accuracy;

identifying, in response to said determining that said second measurement is less than or equal to said first measurement, an optimal set of one or more attributes as said first set of one or more attributes, wherein said optimal set contributes to a problem associated with said networked system; and

storing said optimal set in a data repository coupled to said computing system.

12. The system of claim 11, wherein said method further comprises generating a set of best practices for configuring said networked system to avoid said problem, wherein a configuration of said networked system that follows said set of best practices excludes any entity having said optimal set of one or more attributes.

13. The system of claim 11, wherein said ranking said plurality of entity types comprises determining an information entropy value

$H$

( T ) ⁢

⁢

as ⁢

⁢

∑

j = 1 n ⁢

⁢

p ⁡

( j ) ⁢

log 2 ⁡

( 1 p ⁡

( j ) ) ,

wherein T is an entity type of said plurality of entity types, wherein p(j) is a probability of an occurrence of an outcome j of T, and wherein n is a total number of possible outcomes of T.

14. The system of claim 13, wherein said ranking said plurality of entity types further comprises:

identifying a set of entities that report said problem, wherein said set of entities is included in said plurality of entities;

determining, subsequent to said identifying said set of entities, P true data paths that are internal to said set of entities, wherein P_i true data paths of said P true data paths are included in an i-th partition of k partitions of a configuration of said networked system, and wherein k>

1;

determining, subsequent to said identifying said set of entities, N false data paths that are external to said set of entities, wherein N_i false data paths of said N false data paths are included in said i-th partition of said configuration; and

determining an entropy value Entropy(P_i,N_i) as

$-\left(\frac{P_i}{P_i+N_i}\times <br><br>{\log }_2<br><br>\frac{P_i}{P_i+N_i}+\frac{N_i}{P_i+N_i}\times <br><br>{\log }_2<br><br>\frac{N_i}{P_i+N_i}\right).$

15. The system of claim 14, wherein said ranking said plurality of entity types further comprises:

determining an information gain value Gain(T) of said plurality of information gain values as

$H<br><br>\left(T\right)-\underset{i=1}{\overset{k}{\sum <br><br>}}<br><br><br><br>\frac{P_i+N_i}{P+N}\times <br><br>\mathrm{Entropy}<br><br><br><br>\left(P_i,N_i\right),$

wherein Gain(T) indicates an information gain of said entity type T; and

ordering said entity type T relative to one or more other entity types of said plurality of entity types based on Gain(T).

16. A computer program product, comprising a computer non-transitory storage medium having a computer readable program code embodied therein, said computer readable program code containing instructions that when executed by a processor of a computing system implement a method of automatically identifying an optimal set of attributes of entities to facilitate generating best practices for configuring a networked system, said method comprising:

ranking, based on a plurality of information gain values, a plurality of entity types of a plurality of entities included in said networked system;

determining, subsequent to said ranking, a first classification accuracy relative to a first entity type, wherein said first entity type is a highest ranked entity type of said plurality of entity types based on said ranking or is a first aggregate entity type associated with two or more entity types of said plurality of entity types;

selecting, subsequent to said determining said first measurement, a second entity type of said plurality of entity types, wherein said selecting is based on said ranking;

performing a database join operation on a first set of one or more attributes of one or more entities of said first entity type and a second set of one or more attributes of one or more entities of said second entity type, wherein a result of said performing is a second aggregate entity type;

determining a second classification accuracy relative to said second aggregate entity type;

determining that said second classification accuracy is less than or equal to said first classification accuracy;

identifying, in response to said determining that said second measurement is less than or equal to said first measurement, an optimal set of one or more attributes as said first set of one or more attributes, wherein said optimal set contributes to a problem associated with said networked system; and

storing said optimal set in a data repository coupled to said computing system.

17. The program product of claim 16, wherein said method further comprises generating a set of best practices for configuring said networked system to avoid said problem, wherein a configuration of said networked system that follows said set of best practices excludes any entity having said optimal set of one or more attributes.

18. The program product of claim 16, wherein said ranking said plurality of entity types comprises determining an information entropy value

$H$

( T ) ⁢

⁢

as ⁢

⁢

∑

j = 1 n ⁢

⁢

p ⁡

( j ) ⁢

log 2 ⁡

( 1 p ⁡

( j ) ) ,

wherein T is an entity type of said plurality of entity types, wherein p(j) is a probability of an occurrence of an outcome j of T, and wherein n is a total number of possible outcomes of T.

19. The program product of claim 18, wherein said ranking said plurality of entity types further comprises:

identifying a set of entities that report said problem, wherein said set of entities is included in said plurality of entities;

determining, subsequent to said identifying said set of entities, P true data paths that are internal to said set of entities, wherein P_i true data paths of said P true data paths are included in an i-th partition of k partitions of a configuration of said networked system, and wherein k>

1;

determining, subsequent to said identifying said set of entities, N false data paths that are external to said set of entities, wherein N_i false data paths of said N false data paths are included in said i-th partition of said configuration; and

determining an entropy value Entropy(P_i,N_i) as

$-\left(\frac{P_i}{P_i+N_i}\times <br><br>{\log }_2<br><br>\frac{P_i}{P_i+N_i}+\frac{N_i}{P_i+N_i}\times <br><br>{\log }_2<br><br>\frac{N_i}{P_i+N_i}\right).$

20. The program product of claim 19, wherein said ranking said plurality of entity types further comprises:

determining an information gain value Gain(T) of said plurality of information gain values as

$H<br><br>\left(T\right)-\underset{i=1}{\overset{k}{\sum <br><br>}}<br><br><br><br>\frac{P_i+N_i}{P+N}\times <br><br>\mathrm{Entropy}<br><br><br><br>\left(P_i,N_i\right),$

wherein Gain(T) indicates an information gain of said entity type T; and

ordering said entity type T relative to one or more other entity types of said plurality of entity types based on Gain(T).