Back to results
US 2009/0024549 A1Published application
Methods and Systems for Determining Entropy Metrics for Networks
AssigneeUniversity of South Carolina
InventorJoseph E. Johnson
Life of this patent
Estimating the term…
Abstract
Methods and systems for network monitoring using network metrics which are generalized entropy functions of the Markov monoid matrix M generated by an altered connection matrix C.
Claims29
A method for network analysis comprising:
- receiving network data;
- constructing a C(i, j) matrix comprising a Cc matrix and a Cr matrix;
- normalizing the C(i, j) matrix;
- constructing a Mc and Mr matrix each having at least one row and at least one column;
- determining entropy values for each row and column of the Mc and Mr matrices;
- and
- plotting a spectral curve base on the determined entropy values, wherein the spectral curve indicates network behavior.
4 more elements
A system for network analysis comprising:
- at least two network nodes;
- a network interface in communication with the at least two network nodes;
- a processor coupled to the network interface, the processor configured for performing the steps of:
- receiving network data;
- constructing a C(i, j) matrix comprising a Cc matrix and a Cr matrix;
- normalizing the C(i, j) matrix;
- constructing a Mc and Mr matrix each having at least one row and at least one column;
- determining entropy values for each row and column of the Mc and Mr matrices;
- and
- plotting a spectral curve base on the determined entropy values, wherein the spectral curve indicates network behavior.
7 more elements
A method for network monitoring comprising:
- determining a normal network topology profile for a network based on entropy spectra;
- determining a current network topology profile for a network based on entropy spectra;
- determining a variance between the normal network topology profile and the current network topology profile;
- and
- displaying the determined variance wherein the variance is indicative of abnormal network activity.
2 more elements
Dependencies are inferred from the claim text.