Cybersecurity risk score computation based on peer grouping
A computing device computes a risk score for a user using a device based on a peer group identifier. Network activity measures characterize use of the device by the user. For each unique peer group identifier included in netflow records, a mean value is computed of each network activity measure. For each unique IP address and user identifier combination included in the netflow records, the mean value of each network activity measure is selected for a peer group identifier of the user; a risk score is computed by comparing each network activity measure for the unique IP address and user identifier combination to the selected mean value for the respective network activity measure; and when the risk score exceeds a predefined alert threshold, a high risk alert indicator is set indicating that the device is being used in an anomalous manner relative to other devices monitored by the computing device.
1. A non-transitory computer-readable medium having stored thereon computer-readable instructions that when executed by a processor cause a first computing device to:
read a plurality of records, wherein each record includes a user identifier, an Internet protocol (IP) address associated with the user identifier, a peer group identifier determined based on the user identifier, and a plurality of network activity measures, wherein the user identifier identifies a user of a second computing device being monitored by the first computing device, wherein the peer group identifier is predefined and identifies a peer group to which the user is assigned and which is associated with the user identifier, wherein members of the peer group are identified based on an expected common network activity behavior, wherein the IP address identifies the second computing device, and wherein the plurality of network activity measures characterize use of the second computing device by the user determined from a received netflow record;
for each unique peer group identifier included in the read plurality of records, compute a mean value of each of the plurality of network activity measures;
for each unique peer group identifier included in the read plurality of records, compute a standard deviation value of each of the plurality of network activity measures;
for each unique IP address and user identifier combination included in the read plurality of records,
select the computed mean value of each of the plurality of network activity measures for the peer group identifier associated with the user identifier;
select the computed standard deviation value of each of the plurality of network activity measures for the peer group identifier associated with the user identifier;
compute a risk score by computing a number of deviations for the respective network activity measure using the selected mean value and the selected standard deviation value for the respective network activity measure; and
when the computed risk score exceeds a predefined alert threshold,
set a high risk alert indicator indicating that the second computing device is being used in an anomalous manner relative to other computing devices being monitored by the first computing device; and
output the set high risk alert indicator to a display.
2. The non-transitory computer-readable medium of claim 1 , wherein the network activity measures characterize a communication between the second computing device and a third computing device.
3. The non-transitory computer-readable medium of claim 2 , wherein the third computing device is part of an external network relative to the second computing device.
4. The non-transitory computer-readable medium of claim 2 , wherein the third computing device is part of an internal network relative to the second computing device.
5. The non-transitory computer-readable medium of claim 1 , wherein each record further includes a time zone indicator.
6. The non-transitory computer-readable medium of claim 5 , wherein the unique peer group identifier is a combination of the peer group identifier and the time zone indicator.
7. The non-transitory computer-readable medium of claim 1 , wherein the computer-readable instructions further cause the first computing device to compute an activity measure value for each of the plurality of network activity measures for the unique IP address and user identifier combination using the read plurality of records associated with the unique IP address and user identifier combination.
8. The non-transitory computer-readable medium of claim 7 , wherein the number of deviations is computed using N=V−μ/σ, where N is the number of deviations, V is the activity measure for the respective network activity measure, μ is the selected mean value for the respective network activity measure, and σ is the selected standard deviation value for the respective network activity measure.
9. The non-transitory computer-readable medium of claim 1 , wherein the computer-readable instructions further cause the first computing device to aggregate the number of deviations computed for each network activity measure of the plurality of network activity measures from the plurality of records as part of computing the risk score.
10. The non-transitory computer-readable medium of claim 9 , wherein the computer-readable instructions further cause the first computing device to compute a cumulative density function for each network activity measure of the plurality of network activity measures from the aggregated number of deviations as part of computing the risk score.
11. The non-transitory computer-readable medium of claim 10 , wherein the risk score is computed using the cumulative density function for each network activity measure of the plurality of network activity measures.
12. The non-transitory computer-readable medium of claim 10 , wherein the risk score is computed using a probability value selected from the cumulative density function for each network activity measure of the plurality of network activity measures for each unique IP address and user identifier combination.
13. The non-transitory computer-readable medium of claim 10 , wherein the risk score is computed by summing a probability value selected from the cumulative density function for each network activity measure of the plurality of network activity measures for each unique IP address and user identifier combination.
14. The non-transitory computer-readable medium of claim 13 , wherein the risk score is further computed by computing an average of the summed probability values for each unique IP address and user identifier combination.
15. The non-transitory computer-readable medium of claim 14 , wherein the risk score is further computed by computing an inverse of the average of the summed probability values for each unique IP address and user identifier combination.
16. The non-transitory computer-readable medium of claim 10 , wherein the risk score is computed by summing and weighting a probability value selected from the cumulative density function for each network activity measure of the plurality of network activity measures for each unique IP address and user identifier combination.
17. A computing device comprising:
a processor; and
a non-transitory computer-readable medium operably coupled to the processor, the computer-readable medium having computer-readable instructions stored thereon that, when executed by the processor, cause the computing device to
read a plurality of records, wherein each record includes a user identifier, an Internet protocol (IP) address associated with the user identifier, a peer group identifier determined based on the user identifier, and a plurality of network activity measures, wherein the user identifier identifies a user of a second computing device being monitored by the first computing device, wherein the peer group identifier is predefined and identifies a peer group to which the user is assigned and which is associated with the user identifier, wherein members of the peer group are identified based on an expected common network activity behavior, wherein the IP address identifies the second computing device, and wherein the plurality of network activity measures characterize use of the second computing device by the user determined from a received netflow record;
for each unique peer group identifier included in the read plurality of records, compute a mean value of each of the plurality of network activity measures;
for each unique peer group identifier included in the read plurality of records, compute a standard deviation value of each of the plurality of network activity measures;
for each unique IP address and user identifier combination included in the read plurality of records,
select the computed mean value of each of the plurality of network activity measures for the peer group identifier associated with the user identifier;
select the computed standard deviation value of each of the plurality of network activity measures for the peer group identifier associated with the user identifier;
compute a risk score by computing a number of deviations for the respective network activity measure using the selected mean value and the selected standard deviation value for the respective network activity measure; and
when the computed risk score exceeds a predefined alert threshold,
set a high risk alert indicator indicating that the second computing device is being used in an anomalous manner relative to other computing devices being monitored by the first computing device; and
output the set high risk alert indicator to a display.
18. A method of computing a risk score for a user using a device based on a peer group identifier for the user, the method comprising:
reading, by a computing device, a plurality of records, wherein each record includes a user identifier, an Internet protocol (IP) address associated with the user identifier, a peer group identifier determined based on the user identifier, and a plurality of network activity measures, wherein the user identifier identifies a user of a second computing device being monitored by the first computing device, wherein the peer group identifier is predefined and identifies a peer group to which the user is assigned and which is associated with the user identifier, wherein members of the peer group are identified based on an expected common network activity behavior, wherein the IP address identifies the second computing device, and wherein the plurality of network activity measures characterize use of the second computing device by the user determined from a received netflow record;
for each unique peer group identifier included in the read plurality of records, computing, by the computing device, a mean value of each of the plurality of network activity measures;
for each unique peer group identifier included in the read plurality of records, computing, by the computing device, a standard deviation value of each of the plurality of network activity measures;
for each unique IP address and user identifier combination included in the read plurality of records,
selecting, by the computing device, the computed mean value of each of the plurality of network activity measures for the peer group identifier associated with the user identifier;
selecting, by the computing device, the computed standard deviation value of each of the plurality of network activity measures for the peer group identifier associated with the user identifier;
computing, by the computing device, a risk score by computing a number of deviations for the respective network activity measure using the selected mean value and the selected standard deviation value for the respective network activity measure; and
when the computed risk score exceeds a predefined alert threshold,
setting, by the computing device, a high risk alert indicator indicating that the second computing device is being used in an anomalous manner relative to other computing devices being monitored by the first computing device; and
outputting, by the computing device, the set high risk alert indicator to a display.
19. The method of claim 18 , further comprising computing, by the computing device, an activity measure value for each of the plurality of network activity measures for the unique IP address and user identifier combination using the read plurality of records associated with the unique IP address and user identifier combination.
20. The method of claim 19 , wherein the number of deviations is computed using N=V−μ/σ, where N is the number of deviations, V is the activity measure for the respective network activity measure, μ is the selected mean value for the respective network activity measure, and σ is the selected standard deviation value for the respective network activity measure.
21. The method of claim 18 , further comprising aggregating, by the computing device, the number of deviations computed for each network activity measure of the plurality of network activity measures from the plurality of records as part of computing the risk score.
22. The method of claim 21 , further comprising computing, by the computing device, a cumulative density function for each network activity measure of the plurality of network activity measures from the aggregated number of deviations as part of computing the risk score.
23. The method of claim 22 , wherein the risk score is computed using the cumulative density function for each network activity measure of the plurality of network activity measures.
24. The method of claim 22 , wherein the risk score is computed using a probability value selected from the cumulative density function for each network activity measure of the plurality of network activity measures for each unique IP address and user identifier combination.
25. The method of claim 22 , wherein the risk score is computed by summing a probability value selected from the cumulative density function for each network activity measure of the plurality of network activity measures for each unique IP address and user identifier combination.
26. The method of claim 25 , wherein the risk score is further computed by computing an average of the summed probability values for each unique IP address and user identifier combination.
27. The method of claim 26 , wherein the risk score is further computed by computing an inverse of the average of the summed probability values for each unique IP address and user identifier combination.
28. The method of claim 22 , wherein the risk score is computed by summing and weighting a probability value selected from the cumulative density function for each network activity measure of the plurality of network activity measures for each unique IP address and user identifier combination.