System and method of proactively and reactively diagnosing a data communication network

• US 5,377,196 A
• Filed: 12/30/1993
• Issued: 12/27/1994
• Est. Priority Date: 03/12/1991
• Status: Expired due to Term

• Current Assignee: Agilent Technologies, Inc.
• Original Assignee: Hewlett-Packard Company (HP Inc.)
• Inventors: Smith, Mark S., Godlew, Scott A.
• Primary Examiner(s): Baker, Stephen M.
• Application Number:
  US08/176,499
• Time in Patent Office: 362 Days
• Field of Search: 371/20.1, 395/50, 395/51, 395/53, 395/575, 395/916, 395/917
• US Class Current: 714/712
• CPC Class Codes: H04L 41/16 using machine learning or artificial intelligence
• CPC Class Codes: H04L 43/045 for graphical visualisation of monitoring data
• CPC Class Codes: H04L 43/0847 Transmission error
• CPC Class Codes: H04L 43/16 Threshold monitoring
• CPC Class Codes: H04L 43/18 Protocol analysers
• CPC Class Codes: Y10S 706/917 Communication

Abstract

An expert system for diagnosing a data communication network is discussed. The expert system includes a knowledge base comprising a plurality of rules pertaining to reactive and proactive diagnosis of the network. The expert system also includes a plurality of modules for processing network-related questions, such as a rule module which uses forward rules to hypothesize whether network problems exist, and which uses backward rules to verify whether the hypothesized network problems exist. The expert system further includes a facilitator for selecting modules to process the questions. The facilitator operates by sending a poll request associated with one of the questions to the modules, analyzing estimates from the modules to identify a module which can most efficiently and effectively process the question, and instructing the identified module to process the question. The rule module also operates to proactively diagnose the data communication network by identifying in the knowledge base a rule pertaining to proactive diagnosis, formulating questions based on the rule, and circulating the formulated questions within the expert system.

Claims

1. An expert system for diagnosing a data communication network, the expert system comprising:

(1) a knowledge base comprising a plurality of rules, a first subset of said rules pertaining to reactive diagnosis and a second subset of said rules pertaining to proactive diagnosis, said reactive diagnosis rules corresponding to diagnosis of the data communication network performed by said expert system in response to requests from users, and said proactive diagnosis rules corresponding to diagnosis of the data communication network automatically performed by said expert system without prior requests from users;

(2) a communication medium for exchanging within said expert system questions which pertain to diagnosis of the data communication network;

(3) a plurality of modules for processing said questions, comprising;

(a) a measurement module for enabling a measurement device, coupled to the data communication network, to perform network measurements to collect network data in accordance with a first subset of said questions;

(b) a remote module for obtaining in accordance with a second subset of said questions network data from a remote data communication network, the remote data communication network being external to said expert system and being external to the data communication network, said network data obtained from the remote data communication network pertaining to the interaction between the remote data communication network and the data communication network;

(c) a user question module for interacting with users to obtain user information, said user information being used to generate network data in accordance with a third subset of said questions;

(d) a rule module for processing a fourth subset of said questions in accordance with said rules contained in said knowledge base, said rule module comprising;

(i) means for using forward rules to hypothesize whether network problems exist in the data communication network based on said network data collected by said measurement, remote, and user question modules; and

(ii) means for using backward rules to verify whether said hypothesized network problems exist in the data communication network based on said network data collected by said measurement, remote, and user question modules;

(4) a facilitator for selecting modules to process said questions, comprising;

(a) means for sending a poll request associated with one of said questions to said modules, said modules comprising means for responding to said poll request by generating estimates relating to their respective efficiencies and effectiveness in processing said one of said questions;

(b) means for receiving and analyzing said estimates from said modules to identify a module which can most efficiently and effectively process said one of said questions;

(c) means for instructing said identified module to process said one of said questions;

wherein said rule module further comprises means for proactively diagnosing the data communication network, said proactive diagnosing means comprising;

proactive diagnosis identifying means for identifying in said knowledge base one of said rules pertaining to proactive diagnosis;

means for formulating questions based on said one of said rules pertaining to proactive diagnosis; and

means for circulating said formulated questions within said expert system by placing said formulated questions in said communication medium.

2. The expert system of claim 1, wherein said rules in said knowledge base comprise passive components which specify user input which must be received from users before said passive components can be processed, said rules in said knowledge base also comprising active components which can be processed without any prior user input wherein said proactive diagnosis identifying means comprises:

means for classifying rules having active components as being proactive diagnosis rules; and

means for classifying rules having passive components as being reactive diagnosis rules.

3. The expert system of claim 1, further comprising a user interface for interacting with users, wherein said communication medium comprises:

means for receiving questions and network datums from said modules;

means for transferring a first type of questions and network datums to said facilitator;

means for transferring a second type of questions and network datums to said user interface;

wherein said user interface comprises means for displaying to users said second type of questions and network datums.

4. The expert system of claim 3, further comprising:

means, coupled to said communication medium, for specifying said first type of questions and network datums which said communication medium transfers to said facilitator; and

means, coupled to said communication medium, for specifying said second type of questions and network datums which said communication medium transfers to said user interface.

5. The expert system of claim 3, further comprising:

means for storing a plurality of data descriptors each relating to a network problem, comprising for each of said dam descriptors;

first data descriptor storage means for storing information identifying a particular network problem,

second data descriptor storage means for storing information describing characteristics of said particular network problem,

third data descriptor storage means for storing information describing a solution to said particular network problem,

fourth data descriptor storage means for storing information specifying a frequency of occurrence of said particular network problem, and

fifth data descriptor storage means for storing information specifying a severity of impact on the network of said particular network problem;

wherein said user interface comprises;

means for retrieving information from said first, second, third, fourth, and fifth data descriptor storage means in response to information requests from users, and

means for displaying said retrieved information.

6. The expert system of claim 3, further comprising a blackboard coupled to the facilitator for storing questions and network data, wherein said facilitator comprises:

means for storing in said blackboard one of said first type of questions transferred to said facilitator from said communication medium;

means for identifying in said blackboard any network data related to said one of said first type of questions;

means for retrieving from said blackboard said identified network data, if any;

means for identifying which of said modules generated said one of said first type of questions; and

means for transferring said retrieved network data, if any, to said identified module such that said identified module can process said one of said first type of questions.

7. The expert system of claim 6, wherein said facilitator further comprises:

means for prioritizing questions stored in said blackboard according to frequency and severity to thereby identify a highest priority question;

means, coupled to said user interface, for determining whether a user override specifying a user question exists;

means for selecting said highest priority question to process if said user override does not exist; and

means for selecting said user question to process if said user override exists.

8. The expert system of claim 3, further comprising a blackboard coupled to the facilitator for storing questions and network data, wherein said facilitator comprises:

means for identifying in said blackboard any network data related to one of said first type of network datums transferred to said facilitator from said communication medium;

means for retrieving from said blackboard said identified network data, if any;

means for generating a cumulative answer as a function of said one of said first type of network datums and said retrieved network data; and

means for storing said cumulative answer in said blackboard.

9. The expert system of claim 8, wherein said facilitator further comprises:

means for identifying in said blackboard a question related to said cumulative answer;

means for identifying which of said modules generated said identified question; and

means for transferring said cumulative answer to said identified module such that said identified module can process said identified question.

10. A method of diagnosing a data communication network the method for use in an expert system comprising a knowledge base comprising a plurality of rules, a first subset of the rules pertaining to reactive diagnosis and a second subset of the rules pertaining to proactive diagnosis, the reactive diagnosis rules corresponding to diagnosis of the data communication network performed by the expert system in response to requests from users, and the proactive diagnosis rules corresponding to diagnosis of the data communication network automatically performed by the expert system without prior requests from users, the expert system also comprising a plurality of modules for processing questions pertaining to diagnosis of the data communication network, the method comprising the steps of:

(1) proactively diagnosing tile data communication network, comprising the steps of;

(a) identifying in the knowledge base one of the rules pertaining to proactive diagnosis,

(b) formulating questions based on said one of the rules pertaining to proactive diagnosis;

(2) selecting modules to process said formulated questions, comprising the steps of;

(a) sending a poll request associated with one of said formulated questions to the modules;

(b) receiving from the modules estimates relating to their respective efficiencies and effectiveness in processing said one of said formulated questions;

(c) analyzing said estimates from the modules to identify a module which can most efficiently and effectively process said one of said formulated questions;

(d) instructing a measurement module to process said one of said formulated questions if the measurement module is identified as being the most efficient and effective module at processing said one of said formulated questions, the measurement module comprising means for enabling a measurement device, coupled to the data communication network, to perform network measurements to collect network data in accordance with said one of said formulated questions;

(e) instructing a remote module to process said one of said formulated questions if the remote module is identified as being the most efficient and effective module at processing said one of said formulated questions, the remote module comprising means for obtaining in accordance with said one of said formulated questions network data from a remote data communication network, the remote data communication network being external to the expert system and being external to the data communication network, said network data obtained from the remote data communication network pertaining to the interaction between the remote data communication network and the data communication network; and

(f) instructing a rule module to process said one of said formulated questions if the rule module is identified as being the most efficient and effective module at processing said one of said formulated questions, the rule module comprising means for using forward rules from the knowledge base to hypothesize whether network problems exist in the data communication network based on said network data collected by the modules, and means for using backward rules to verify whether said hypothesized network problems exist in the data communication network based on said network data collected by the modules.

11. The method of claim 10 in which the rules in the knowledge base comprise passive components which specify user input which must be received from users before the passive components can be processed, the rules in the knowledge base also comprising active components which can be processed without any prior user input, wherein step (1)(a) comprises the steps of:

classifying rules having active components as being proactive diagnosis rules; and

classifying rules having passive components as being reactive diagnosis rules.

12. The method of claim 10 wherein the expert system also includes a blackboard for storing questions and network data, further comprising the steps of:

identifying in said blackboard any network data related to said one of said formulated questions;

retrieving from said blackboard said identified network data, if any; and

transferring said retrieved network data, if any, to said module selected in step (2) to process said one of said formulated questions.

13. The method of claim 12, further comprising the steps of:

prioritizing questions stored in the blackboard according to frequency and severity to thereby identify a highest priority question;

determining whether a user override specifying a user question exists;

selecting said highest priority question to process if said user override does not exist; and

selecting said user question to process if said user override exists.