Agent based router monitoring, diagnostic and maintenance

• US 7,284,162 B2
• Filed: 07/26/2004
• Issued: 10/16/2007
• Est. Priority Date: 04/07/2004
• Status: Active Grant

• Current Assignee: RPX Corporation
• Original Assignee: Alcatel SA (Nokia Corporation)
• Inventors: McBride, Brian, Dragnea, Raluca, Levy, Thomas, Marce, Olivier, Rabinovitch, Peter
• Primary Examiner(s): LE, DIEU MINH T
• Application Number:
  US10/898,222
• Publication Number:
  US 20050240835A1
• Time in Patent Office: 1,177 Days
• Field of Search: 714/47, 714/48, 714/4, 709/202, 709/223, 709/224
• US Class Current: 714/47.2
• CPC Class Codes: H04L 41/046 comprising network management agents or mobile agents therefor
• CPC Class Codes: H04L 41/0631 using root cause analysis; using analysis of correlation between notifications, alarms or events based on decision criteria, e.g. hierarchy, tree or time analysis

Abstract

A distributed, multistage agent for monitoring, diagnosis and maintenance of network devices comprises an input/output interface for receiving input monitoring and diagnostic data relevant to operation of a first neighboring agent and transmitting output monitoring and diagnostic data relevant to operation of the host network device to a second neighboring agent. An agent process analyzes the input data and generates processed data characterizing operation of the host network device, using a local database which stores expected performance data on operation of said network device. The processed data is correlated with the expected data any error is signaled to the network operator. The agents may be designed as a multiple stage process, the stages being triggered by the data output by an earlier stage only when needed.

Claims

1. An agent based system for monitoring, diagnostic and maintenance of a network device said agent based system being resident on the network device, comprising:

an input/output interface for receiving input monitoring and diagnostic data relevant to operation of a first neighboring agent and transmitting output monitoring and diagnostic data relevant to operation of said network device to a second neighboring agent;

an agent process for analyzing said input data and generating processed data characterizing operation of said network device;

a local database for storing expected performance data on operation of said network device; and

a correlation unit for correlating said processed data with expected data to detect a basic error in said operation of said network device, and assembling said output data.

2. The agent of claim 1, further comprising a next hop selector for choosing said second neighboring agent.

3. The agent of claim 1, further comprising an error generator for generating an error message whenever said processed data is different from said expected data.

4. The agent of claim 3, wherein said error generator provides additional information on the particulars of an error.

5. The agent of claim 1, further comprising an agent request queue at the input side of said agent processor that organizes said input data in a queue according to a preset rule.

6. The agent of claim 1, further comprising means for storing a plurality of rules for defining the correct operation of said network device.

7. The agent of claim 1, further comprising means for storing a plurality of policies applicable to said particular network device.

8. The agent of claim 1, wherein said local database further collects statistics for enabling said correlation unit to detect certain systematic abnormalities in the operation of said network device.

9. The agent of claim 1, wherein said input-output interface comprises a receiver for receiving said input data from said first agent and a transmitter for transmitting said output data to said second agent.

10. The agent of claim 1, wherein said input-output interface comprises decryption means for decrypting said input data received from said first agent and encryption means for encoding said output data before transmission to said second agent.

11. The agent of claim 2, wherein said next hop selector includes random path selection mechanism.

12. The agent of claim 2, wherein said next hop selector includes a random multiple path selection mechanism.

13. An agent based method for monitoring performance of a network device, comprising:

a) providing a network device with an embedded agent, connected to a first neighboring agent over an incoming transmission path for receiving input monitoring and diagnostic data relevant to operation of said first neighboring agent;

b) establishing an outgoing transmission path from said network device to a second neighboring agent for transmitting output monitoring and diagnostic data relevant to operation of said network device;

c) at said embedded agent, analyzing said input data and generating processed data characterizing operation of said network device;

d) correlating said processed data with expected data available at said network device to detect a basic error in said operation of said network device, and assembling said output data.

14. The method of claim 13, wherein step c) comprises:

arranging said input data in a queue at the input of an agent process;

verifying said input data against one or more rules for defining the correct operation of said network device;

verifying said input data against one or more policies applicable to said particular network device; and

generating said processed data to provide information about operation of said network device in the context of operation of a first neighboring network device hosting said first neighboring agent.

15. The method of claim 13, wherein step d) comprises:

storing expected performance data on operation of said network device; and

correlating said processed data with expected data and assembling said output data.

16. An embedded agent for monitoring operation of a network device comprising:

a first stage for determining basic errors in operation of said network device based on a plurality of thresholds for each type and level of errors; and

a second stage specialized for determining complex errors in operation of said network device,

wherein said first stage runs continuously, and said second stage is triggered by said first stage whenever one of said thresholds is violated.