Methods And System For Modeling Network Traffic

• US 20090138420A1
• Filed: 11/28/2007
• Published: 05/28/2009
• Est. Priority Date: 11/28/2007
• Status: Active Grant

• Current Assignee: Curators of the University of Missouri (University of Missouri)
• Original Assignee: The Boeing Co.
• Inventors: Dagli, Cihan H., Swift, Douglas K.
• Granted Patent: US 8,352,392 B2
• Time in Patent Office: Days
• US Class Current: 706/21
• CPC Class Codes: H04L 41/145 involving simulating, designing, planning or modelling of a network
• CPC Class Codes: H04L 41/147 for predicting network behaviour

Abstract

A method and system are provided for modeling network traffic in which an artificial neural network architecture is utilized in order to intelligently and adaptively model the capacity of a network. Initially, the network traffic is decomposed into a plurality of categories, such as individual users, application usage or common usage groups. Inputs to the artificial neural network are then defined such that a respective combination of inputs permits prediction of bandwidth capacity needs for that input condition. Outputs of the artificial neural network are representative of the network traffic associated with the respective inputs. For example, a plurality of bandwidth profiles associated with respective categories may be defined. An artificial neural network may then be constructed and trained with those bandwidth profiles and then utilized to relate predict future bandwidth needs for the network.

Claims

1. A method for modeling network traffic comprising:

decomposing the network traffic into a plurality of categories for training an artificial neural network;

defining a plurality of inputs which permit prediction of network traffic to be determined from a respective combination of inputs;

constructing the artificial neural network based on the categories of network traffic such that the artificial neural network is configured to evaluate the plurality of inputs and develop a plurality of outputs that represent a prediction of bandwidth capacity for the related inputs, wherein the plurality of outputs generated by the artificial neural network provide a model of the network traffic.

2. A method according to claim 1 wherein decomposing the network traffic comprises decomposing the network traffic into a plurality of categories selected from the group consisting of individual users, application usage and common usage groups.

3. A method according to claim 1 wherein constructing the artificial neural network comprises training the artificial neural network with a plurality of predefined inputs and predefined outputs, wherein training the artificial neural network comprises:

comparing the plurality of outputs from the artificial neural network with a plurality of anticipated predefined outputs based upon the plurality of predefined inputs; and

adapting the artificial neural network to compensate for differences between the plurality of outputs and the plurality of anticipated predefined outputs.

4. A method according to claim 3 wherein the artificial neural network comprises a plurality of intermediate nodes having associated transfer functions and a plurality of connections having associated weights interconnecting the inputs, outputs and intermediate nodes, and wherein adapting the artificial neural network comprises altering the weights associated with respective connections.

5. A method according to claim 4 wherein adapting the artificial neural network comprises changing the number of intermediate nodes disposed between the inputs and the outputs.

6. A method for modeling network traffic comprising:

providing an artificial neural network to evaluate a plurality of inputs and produce a plurality of outputs, wherein each output is representative of a predicted bandwidth profile related to the respective inputs;

receiving a plurality of attributes representative of the network traffic at the inputs of the artificial neural network;

processing the plurality of attributes with the artificial neural network in order to generate the plurality of outputs; and

generating a model of the network traffic based upon the plurality of outputs.

7. A method according to claim 6 further comprising defining an anticipated network capacity to at least equal an aggregate of the plurality of outputs.

8. A method according to claim 6 further comprising decomposing the network traffic into a plurality of categories selected from the group consisting of individual users, application usage and common usage groups.

9. A method according to claim 6 wherein providing the artificial neural network comprises training the artificial neural network with a plurality of predefined inputs and predefined outputs, wherein training the artificial neural network comprises:

comparing the plurality of outputs from the artificial neural network with a plurality of anticipated predefined outputs based upon the plurality of predefined inputs; and

adapting the artificial neural network to compensate for differences between the plurality of outputs and the plurality of anticipated predefined outputs.

10. A method according to claim 9 wherein the artificial neural network comprises a plurality of intermediate nodes having associated transfer functions and a plurality of connections having associated weights interconnecting the inputs, outputs and intermediate nodes, and wherein adapting the artificial neural network comprises altering the weights associated with respective connections.

11. A method according to claim 10 wherein adapting the artificial neural network comprises changing the number of intermediate nodes disposed between the inputs and the outputs and the number of layers in which they are arranged.

12. A system for modeling network traffic comprising:

a memory device configured to store a representation of an artificial neural network which relates a plurality of inputs to a plurality of outputs, wherein each output is representative of a predicted capacity bandwidth profile; and

a processor configured to receive a plurality of attributes representative of the network traffic as inputs to the artificial neural network, the processor also configured to process the plurality of attributes with the artificial neural network in order to generate the plurality of outputs, and the processor further configured to generate a model of the network traffic based upon the plurality of outputs.

13. A system according to claim 12 wherein the processor is further configured to define an anticipated network capacity to at least equal an aggregate of the plurality of outputs.

14. A system according to claim 12 wherein each output is representative of a prediction which was derived from a respective category selected from the group consisting of individual users, application usage and common usage groups.

15. A system according to claim 12 wherein the plurality of outputs comprise bandwidth profiles associated with respective categories.

16. A system according to claim 12 wherein the processor is further configured to train the artificial neural network with a plurality of predefined inputs and predefined outputs, wherein the processor is configured to train the artificial neural network by:

comparing the plurality of outputs from the artificial neural network with a plurality of anticipated predefined outputs based upon the plurality of predefined inputs; and

adapting the artificial neural network to compensate for differences between the plurality of outputs and the plurality of anticipated predefined outputs.

17. A system according to claim 16 wherein the artificial neural network comprises a plurality of intermediate nodes having associated transfer functions and a plurality of connections having associated weights interconnecting the inputs, outputs and intermediate nodes, and wherein the processor is configured to adapt the artificial neural network by altering the weights associated with respective connections.

18. A system according to claim 14 wherein the processor is further configured to adapt the artificial neural network by changing the number of intermediate nodes disposed between the inputs and the outputs.