Traffic anomaly analysis for the detection of aberrant network code
View Patent ↗A method for detecting nodes in an enterprise network infected with aberrant code is presented in which traffic conversation information representative of traffic conversation in the enterprise network over an analysis period is obtained. Analysis of the obtained traffic conversation information identifies suspected infected nodes in the enterprise network that exhibit behavior outside of the normal behavior associated with the one or more traffic conversation factors. Anomaly analysis may be performed on traffic conversation information associated with the suspected infected nodes to identify any existing infected nodes in the enterprise network.
1. A method, comprising:
obtaining a first plurality of values representative of data transported via respective ones of a plurality of traffic conversations within an enterprise network during an analysis period;
computing, via a processor, a statistical attribute of the first plurality of values;
computing, via the processor, a threshold based on the computed statistical attribute;
comparing the first plurality of values to the computed threshold to identify a subset of nodes of the enterprise network suspected of including aberrant code; and
identifying from the subset of nodes a first node as including the aberrant code based on a second plurality of values characterizing destinations contacted by respective ones of the subset of nodes.
2. The method of claim 1 , wherein the threshold represents at least one of a number of attempted contacts, a number of frames per conversation, or a variation in conversation size.
3. The method of claim 1 , further comprising analyzing a portion of the first plurality of values associated with the first node to confirm presence of the aberrant code on the first node.
4. The method of claim 1 , wherein the second plurality of values represents at least one of a first number of attempted contacts with destination hosts in an adjacent address space, a second number of attempted contacts with bogons, or a third number of attempted contacts with destination hosts having unresolvable destination names.
5. An article of manufacture comprising a tangible computer-readable storage medium excluding propagating signals and storing machine-accessible instructions that, when executed, cause a machine to at least:
obtain a first plurality of values representative of data transported via respective ones of a plurality of traffic conversations within an enterprise network during an analysis period;
compute a statistical attribute of the first plurality of values;
compute a threshold based on the computed statistical attribute;
compare the first plurality of values to the computed threshold to identify a subset of nodes of the enterprise network suspected of including aberrant code; and
identify from the subset of nodes a first node as including the aberrant code based on a second plurality of values characterizing destinations contacted by the first node.
6. The article of manufacture of claim 5 , wherein the threshold represents at least one of a number of attempted contacts, a number of frames per conversation, or a variation in conversation size.
7. The article of manufacture of claim 5 , wherein the second plurality of values represent at least one of a first number of attempted contacts with destination hosts in an adjacent address space, a second number of attempted contacts with bogons, or a third number of attempted contacts with destination hosts having unresolvable destination names.
8. An apparatus, comprising:
a data collector to obtain a first plurality of values representative of data transported via respective ones of a plurality of traffic conversations within an enterprise network during an analysis period;
a statistical analyzer to compute a statistical attribute of the first plurality of values;
a source profiler to compute a threshold based on the computed statistical attribute, and compare the first plurality of values to the computed threshold to identify a subset of nodes of the enterprise network suspected of including aberrant code; and
a destination analyzer to identify from the subset of nodes a first node as including the aberrant code based on a second plurality of values characterizing destinations contacted by respective ones of the subset of nodes, wherein at least one of the data collector, the statistical analyzer, the source profiler or the destination analyzer is implemented with hardware.
9. The apparatus of claim 8 , further comprising an anomaly analyzer to analyze a portion of the first plurality of values associated with the first node to confirm presence of the aberrant code on the first node.
10. The apparatus of claim 8 , wherein each of the second plurality of values represents at least one of a first number of attempted contacts with destination hosts in an adjacent address space, a second number of attempted contacts with bogons, or a third number of attempted contacts with destination hosts having unresolvable destination names.
11. The method of claim 1 , wherein the statistical attribute comprises at least one of a mean or a standard deviation of the first plurality of values.
12. The article of manufacture of claim 5 , wherein the statistical attribute comprises at least one of a mean or a standard deviation of the first plurality of values.
13. The apparatus of claim 8 , wherein the statistical attribute comprises at least one of a mean or a standard deviation of the first plurality of values.
14. The apparatus of claim 8 , wherein the threshold represents at least one of a number of attempted contacts, a number of frames per conversation or a variation in conversation size.
15. The method of claim 4 , further comprising:
setting a weight associated with the first node to an initial amount;
increasing the weight by a first amount when the second number is non-zero;
increasing the weight by a second amount when the first number exceeds a second threshold; and
increasing the weight by a third amount when the third number exceeds a third threshold.
16. The method of claim 1 , further comprising identifying the first node as having the largest associated one of the second plurality of values.