Cognitive information security using a behavior recognition system
Embodiments presented herein describe a method for processing streams of data of one or more networked computer systems. According to one embodiment of the present disclosure, an ordered stream of normalized vectors corresponding to information security data obtained from one or more sensors monitoring a computer network is received. A neuro-linguistic model of the information security data is generated by clustering the ordered stream of vectors and assigning a letter to each cluster, outputting an ordered sequence of letters based on a mapping of the ordered stream of normalized vectors to the clusters, building a dictionary of words from of the ordered output of letters, outputting an ordered stream of words based on the ordered output of letters, and generating a plurality of phrases based on the ordered output of words.
1. A method, comprising:
receiving, at a processor, a first ordered stream of normalized values associated with information security data for a computer network;
generating a stable neuro-linguistic model by:
clustering the first ordered stream of normalized values to form a plurality of clusters,
outputting an ordered sequence of letters based on the plurality of clusters,
building a dictionary of words based on the ordered sequence of letters, each word from the dictionary of words having a length from a plurality of lengths, each length from the plurality of lengths being less than or equal to a specified maximum letter length,
outputting an ordered stream of words based on the ordered sequence of letters, and
generating a plurality of phrases based on the ordered stream of words;
detecting, using a codelet template, a pattern in a linguistic representation of the information security data, the linguistic representation of the information security data including at least one of the ordered sequence of letters, the ordered stream of words, or the plurality of phrases;
receiving a second ordered stream of normalized values associated with the information security data for the computer network; and
issuing an alert based on the stable neuro-linguistic model, the pattern, and the second ordered stream of normalized values.
2. The method of claim 1 , wherein the information security data includes one of security log data, disk mount data, physical access data, disk input/output log data, internet protocol (IP) address log data, security tool data, memory usage data, routing data, packet traffic data, datagram traffic data, segment traffic data, port traffic data, or Simple Network Management Protocol (SNMP) trap data.
3. The method of claim 1 , wherein each normalized value from the first ordered stream of normalized values is associated with a value of the information security data, the value for that normalized value from the first ordered stream of normalized values being normalized to a value within a range of 0 to 1, inclusive.
4. The method of claim 1 , wherein the alert is indicative of a detected anomaly.
5. The method of claim 1 , wherein issuing the alert includes sending alert data to an output device for display thereon.
6. A non-transitory computer-readable storage medium storing processor-executable instructions to:
receive a first ordered stream of values associated with information security data for a computer network;
generate a stable neuro-linguistic model by:
clustering the first ordered stream of values to form a plurality of clusters,
outputting an ordered sequence of symbols based on the plurality of clusters,
outputting an ordered stream of symbol combinations based on the ordered sequence of symbols, and
generating a plurality of phrases based on the ordered stream of symbol combinations;
detect a pattern in a linguistic representation of the information security data, the linguistic representation of the information security data including at least one of the ordered sequence of symbols, the ordered stream of symbol combinations, or the plurality of phrases;
receive a second ordered stream of values associated with the information security data for the computer network, the second ordered stream of values including Simple Network Management Protocol (SNMP) trap data; and
issue an alert to at least one compute device within the computer network based on the stable neuro-linguistic model, the pattern, and the second ordered stream of values.
7. The non-transitory computer-readable storage medium of claim 6 , wherein the first ordered stream of values is received from a first compute device of the computer network, and the second ordered stream of values is received from a second compute device of the computer network, the second compute device different from the first compute device.
8. The non-transitory computer-readable storage medium of claim 6 , further comprising building a dictionary of words based on the ordered sequence of symbols, each word from the dictionary of words having a length from a plurality of lengths, each length from the plurality of lengths being less than or equal to a specified maximum length.
9. The non-transitory computer-readable storage medium of claim 6 , wherein each value from the first ordered stream of values is associated with a value of the information security data.
10. The non-transitory computer-readable storage medium of claim 6 , wherein the alert is indicative of a detected anomaly.
11. The non-transitory computer-readable storage medium of claim 6 , wherein issuing the alert includes sending alert data to an output device for display thereon.
12. The non-transitory computer-readable storage medium of claim 6 , wherein the processor-executable instructions to generate the stable neuro-linguistic model include instructions to generate a syntax based on probabilistic relationships between the symbol combinations.
13. A system, comprising:
a processor; and
a memory storing instructions executable by the processor to:
receive, at the processor, a first ordered stream of values associated with information security data for a computer network;
generate a stable neuro-linguistic model by:
clustering the first ordered stream of values to form a plurality of clusters,
outputting an ordered sequence of letters based on the plurality of clusters,
building a dictionary of words based on the ordered sequence of letters, each word from the dictionary of words having a length from a plurality of lengths, each length from the plurality of lengths being less than or equal to a specified maximum letter length,
outputting an ordered stream of words based on the ordered sequence of letters, and
generating a plurality of phrases based on the ordered stream of words;
detect, using a codelet template, a pattern in a linguistic representation of the information security data, the linguistic representation of the information security data including at least one of the ordered sequence of letters, the ordered stream of words, or the plurality of phrases;
receive a second ordered stream of values associated with the information security data for the computer network; and
issue an alert based on the stable neuro-linguistic model, the pattern, and the second ordered stream of values.
14. The system of claim 13 , wherein the information security data includes one of security log data, disk mount data, physical access data, disk input/output log data, internet protocol (IP) address log data, security tool data, memory usage data, routing data, packet traffic data, datagram traffic data, segment traffic data, port traffic data, or Simple Network Management Protocol (SNMP) trap data.
15. The system of claim 13 , wherein each value from the first ordered stream of values is associated with a value of the information security data.
16. The system of claim 13 , wherein the alert is indicative of a detected anomaly.
17. The system of claim 13 , wherein issuing the alert includes sending alert data to an output device for display thereon.
18. The system of claim 13 , wherein the processor-executable instructions to generate the stable neuro-linguistic model include instructions to generate a syntax based on probabilistic relationships between the words.