Diagnostic systems for distributed network

• US 8,972,789 B2
• Filed: 08/28/2012
• Issued: 03/03/2015
• Est. Priority Date: 08/28/2012
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Kamble, Keshav, Peethambaram, Nandakumar, Pandey, Vijoy A., Kumbhare, Abhijit Prabhakar
• Primary Examiner(s): Baderman, Scott
• Assistant Examiner(s): KUDIRKA, JOSEPH R
• Application Number:
  US13/597,019
• Publication Number:
  US 20140068338A1
• Time in Patent Office: 917 Days
• Field of Search: 714 25- 27, 714/37, 714/43, 714/47.1, 714/47.2, 714/47.3, 370/216, 370/229
• US Class Current: 714/37
• CPC Class Codes: H04L 41/0654 using network fault recovery ring fault isolation or reconfiguration in loop networks without recovery actions by a network management system H04L12/437
• CPC Class Codes: H04L 41/142 using statistical or mathematical methods
• CPC Class Codes: H04L 41/147 for predicting network behaviour
• CPC Class Codes: H04L 43/16 Threshold monitoring
• CPC Class Codes: H04L 49/1515 Non-blocking multistage, e.g. Clos

Abstract

A diagnostic system provides identification of symptoms in a distributed network and an engine for providing recommended rectification of error sources that correspond to the symptoms. The distributed network may be accessed for current statistics. Symptoms may be identified that correspond to the current statistics. A recommended course of action for the distributed network may be provided based on a predetermined list of courses of actions that correspond to rectifying the performance in the error sources.

Claims

1. A diagnostic system, comprising:

a distributed network of switches; and

a symptom analysis engine in communication with the distributed network of switches, the symptom analysis engine being configured to;

receive current statistics about the distributed network of switches,

identify packet performance related symptoms that correspond to the current statistics,

identify error sources that correspond to the symptoms,

calculate predictive network packet behavior based on simulating courses of action on switches identified as the error sources, and

retrieve a predetermined list of actions rectifying the packet performance related symptoms.

2. The diagnostic system of claim 1, including a relational knowledge database in communication with the symptom analysis engine, the relational knowledge database including a stored list of symptoms and a stored list of error sources corresponding to the identified packet performance related symptoms.

3. The diagnostic system of claim 1, wherein the symptom analysis engine is configured to determine whether the received current statistics exhibit abnormal operating behavior by the distributed network of switches prior to identifying packet performance related symptoms.

4. The diagnostic system of claim 1, including a statistical database in communication with the distributed network of switches, wherein the statistical database is configured to receive the current statistics from the distributed network of switches and provide the current statistics to the symptom analysis engine.

5. The diagnostic system of claim 4, wherein the statistical database periodically receives the current statistics.

6. The diagnostic system of claim 4, wherein the statistical database is refreshed periodically.

7. The diagnostic system of claim 2, wherein the predetermined list of actions is stored in the relational knowledge database.

8. A processor, comprising a computer readable non-transitory memory having computer readable program code embodied therewith, the computer readable program code being configured to:

identify current statistics related to packet performance in a distributed network of switches as symptoms of underperformance in the network;

determine that the current statistics related to packet performance are beyond operating parameters for the distributed network of switches;

identify one or more error sources that correspond to the symptoms;

calculate predictive network packet behavior based on simulating courses of action on switches identified as the one or more error sources; and

provide a recommended course of action for the distributed network of switches based on a course of actions that rectify the one or more error sources.

9. The processor of claim 8, including computer readable program code configured to calculate predictive behavior in the distributed network of switches based on analyzing potential results from rectifying the one or more error sources.

10. The processor of claim 8, including computer readable program code configured to compare the current statistics to archived statistics in determining whether the current statistics are beyond operating parameters.

11. A process of diagnosing faults in a distributed network system of switches, comprising:

retrieving current packet traffic statistics from the distributed network system of switches;

identifying current packet traffic statistics showing underperformance in the distributed network system of switches;

marking current packet traffic statistics showing underperformance as symptoms;

comparing the symptoms to a list of known symptoms;

identifying error sources that correspond to the symptoms;

calculating predictive network packet behavior based on simulating courses of action on switches identified as the error sources; and

providing a recommended rectification of the error sources based on the calculated predictive network packet behavior.

12. The process of claim 11, including storing the current packet traffic statistics to system memory prior to retrieving the current packet traffic statistics.

13. The process of claim 11, including determining whether the current packet traffic statistics show whether the distributed network system of switches is operating abnormally by analyzing various types of statistical data retrieved from the network system.

14. The process of claim 11, including:

archiving previously retrieved statistics about the distributed network system of switches; and

comparing the archived statistics to the current packet traffic statistics.

15. A computer program product for diagnosing underperformance in a distributed network of switches, the computer program product comprising a computer readable storage medium having computer readable program code embodied therewith, the computer readable program code being configured to:

identify packet latency related symptoms in the distributed network of switches;

search a relational knowledge database for error sources in the distributed network of switches corresponding to the identified packet latency related symptoms;

identify one or more error sources in the distributed network of switches corresponding to the packet latency related symptoms;

access a list of potential rectifications to be applied to the distributed network of switches and corresponding to a predetermined list of actions that operate the one or more error sources within acceptable operating parameters;

calculate predictive network packet behavior based on simulating courses of action on switches identified as the one or more error sources; and

provide the list of potential rectifications.

16. The computer program product of claim 15, the computer readable program code being configured to identify the packet latency related symptoms based on comparing current statistical performance data in the distributed network of switches with archived performance statistics.

17. The computer program product of claim 16, the computer readable program code being configured to:

refresh the current statistical performance data periodically;

compare the refreshed current statistical performance data with archived statistical performance data; and

determine whether an abnormal operation in the distributed network of switches is present based on the current statistical performance data operating beyond a threshold parameter when the current statistical performance data is compared to archived statistical performance data.