System and method for automatic determination of thresholds in network management

• US 5,751,964 A
• Filed: 09/12/1995
• Issued: 05/12/1998
• Est. Priority Date: 09/12/1995
• Status: Expired due to Term

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Spitz, David John, Ordanic, Zvonimir
• Primary Examiner(s): Shin, Christopher B.
• Application Number:
  US08/526,829
• Time in Patent Office: 973 Days
• Field of Search: 395/182, 395/117, 395/161, 364/413, 364/514, 340/146, 340/825, 370/229, 370/241
• US Class Current: 709/224
• CPC Class Codes: H04L 41/0213 Standardised network management protocols, e.g. simple network management protocol [SNMP]
• CPC Class Codes: H04L 41/12 Discovery or management of network topologies
• CPC Class Codes: H04L 41/142 using statistical or mathematical methods
• CPC Class Codes: H04L 41/20 Network management software packages
• CPC Class Codes: H04L 41/22 comprising specially adapted graphical user interfaces [GUI]
• CPC Class Codes: H04L 43/16 Threshold monitoring

Abstract

Method and system for automatically determining and updating thresholds based on collected data samples for counter variables being monitored by a network management application. Historical data is accumulated and used to determine the mean and standard deviation of the monitored counter variables based on the aggregated sample values and a threshold factor is applied to the standard deviation and the resulting value added to the mean to establish the threshold value. The threshold value, which is adaptively updated, is used to determine whether the subsequently sampled values of the monitored counter variables are within a normal range, or that a potential problem exists requiring intervention by a network operator.

Claims

1. A method for automatically determining and adaptively updating thresholds without user intervention in a communications network including a network management station running a network management application and having at least one processor, said network containing a plurality of network nodes, said method comprising the steps of:

selecting at least one variable for each of said plurality of network nodes that is to have a threshold value determined automatically and adaptively updated;

receiving a plurality of data sample values for said at least one variable that is to have said threshold value determined automatically;

accumulating based on each data sample value received, a first sum of each data sample value received and a second sum of the square of each data sample value received and then discarding said each data sample received;

determining a mean value and a standard deviation for said at least one variable based on said first sum, said second sum and a total number of data sample values received;

setting at least one threshold value for said at least one variable by scaling said standard deviation by a threshold factor and adding the result to said mean value; and

graphically displaying a meter depicting said data sample value received and said at least one threshold value.

2. The method of claim 1 wherein said communications network is SNMP-enabled and includes at least one network router node connected by a communications link to said network management station.

3. The method of claim 1 wherein said communications network is SNMP-enabled and includes at least one network bridge node connected by a communications link to said network management station by a network bridge node.

4. The method of claim 1 wherein a first lower threshold value for said at least one variable is determined by scaling using a first threshold factor and a second, higher threshold value is determined by scaling using a second threshold factor.

5. The method of claim 4 wherein said displayed meter is a bar graph representation of said at least one variable and includes a minimum value, a maximum value, said first threshold value, said second threshold value and a current data sample value.

6. The method of claim 1 further comprising the step of polling each of said plurality of network nodes for an update on all thresholded variable values monitored at said each node and transmitting said all thresholded variable values to said network management station for node status determination.

7. The method of claim 1 wherein said graphically displayed data sample value is color encoded based on a comparison with said at least one threshold value for said at least one variable.

8. The method of claim 1 further comprising the step of disabling automatic thresholding and maintaining said at least one threshold value at a current setting if said second sum exceeds a maximum number that can be represented by said at least one processor at said network management station.

9. The method of claim 1 further comprising executing a script of programming instructions if the data sample value received exceeds said at least one threshold value.

10. A system for automatically determining and adaptively updating thresholds without user intervention for at least one selected variable at a network node in a communications network including a network management station running a network management application and having a processor, said network containing a plurality of network nodes, said system comprising:

a transmitter at said network management station for sending a polling request to each of said plurality of network nodes to sample a plurality of variables at said each network node that are thresholded and monitored;

a receiver at said network management station for receiving a response to said polling request from each of said plurality of network nodes, said response including data sample values for each of said plurality of variables that are thresholded;

accumulating means in said network management application for said at least one selected variable for storing and updating a first sum of each data sample value received, a second sum of the square of each data sample value received, and a total number of data sample values received, said each data sample value then being discarded;

statistics generating means in said network management application for determining a mean value and a standard deviation for said at least one selected variable based on said first sum, said second sum and said total number of data sample values received;

threshold setting means in said network management application for scaling said standard deviation and combining the result with said mean value to set and adaptively update said at least one threshold value; and

a display for graphically presenting to a user a meter depicting for said at least one selected variable each said data sample value received and said at least one threshold value.

11. The system of claim 10 wherein said communications network is SNMP-enabled and includes at least one network router connected by a communications link to said network management station.

12. The system of claim 10 wherein said communications network is SNMP-enabled and includes at least one network bridge connected by a communications link to said network management station.

13. The system of claim 10 wherein said display represents said meter associated with said at least one selected variable with a bar graph, said bar graph depicting a minimum value, a maximum value, a current data sample value and said at least one threshold value.

14. The system of claim 10 further comprising means in said network management application for disabling automatic thresholding and maintaining said at least one threshold value at a current setting if said second sum exceeds a maximum number that can be represented by said processor at said network management station.