Method and apparatus for detecting anomalies in aggregated traffic volume data

1. A method of processing traffic of interest in a network, comprising:

computing, for the traffic of interest, a baseline profile and a threshold using initial aggregated volume data of the traffic of interest, wherein the baseline profile is computed via a processor using exponential smoothing, wherein the aggregated volume data comprises volume data of network traffic passing through a plurality of network elements of the network aggregated over a time interval having a plurality of time periods;

obtaining, by the processor, an aggregated volume count for each particular time period of the plurality of time periods, wherein the aggregated volume count comprises a count of volume of traffic data for each particular period of the plurality of time periods, wherein the aggregated volume count is a count of the network traffic passing through the plurality of network elements of the network;

computing, by the processor, for each particular time period of the plurality of time periods, a difference between the aggregated volume count of the particular time period of the plurality of time periods and a particular value of the baseline profile for the same time period;

triggering, by the processor, an alarm for each difference that is computed and exceeds the threshold;

storing, by the processor, the aggregated volume counts for the plurality of time periods to re-define the aggregated volume data; and

modifying, by the processor, the baseline profile and the threshold in accordance with the aggregated volume data being re-defined, wherein the modifying comprises:

comparing a number of outliers in the aggregated volume data to a second threshold;

updating the baseline profile and the threshold using the aggregated volume data if the number of outliers is less than the second threshold; and

re-computing the baseline profile and the threshold using the aggregated volume data if the number of outliers exceeds the second threshold.

2. The method of claim 1 , wherein the re-computing the baseline profile further comprises re-computing a variance, and updating the variance using the aggregated volume data while ignoring any outliers therein.

3. The method of claim 2 , wherein the re-computing the variance and the updating the variance are performed using the exponential smoothing.

4. The method of claim 1 , wherein the updating the baseline profile and the re-computing the baseline profile are performed using the exponential smoothing.

5. The method of claim 1 , wherein a duration of the time interval comprises a week and a duration of each of the plurality of time periods comprises an hour.

6. An apparatus for processing traffic of interest in a network, comprising:

a hardware processor; and

a tangible computer readable medium storing a plurality of instructions which, when executed by the hardware processor, cause the hardware processor to perform operations, the operations comprising:

computing, for the traffic of interest, a baseline profile and a threshold using initial aggregated volume data of the traffic of interest, wherein the baseline profile is computed using exponential smoothing, wherein the aggregated volume data comprises volume data of network traffic passing through a plurality of network elements of the network aggregated over a time interval having a plurality of time periods;

obtaining an aggregated volume count for each particular time period of the plurality of time periods, wherein the aggregated volume count comprises a count of volume of traffic data for each particular period of the plurality of time periods, wherein the aggregated volume count is a count of the network traffic passing through the plurality of network elements of the network;

computing, for each particular time period of the plurality of time periods, a difference between the aggregated volume count of the particular time period of the plurality of time periods and a particular value of the baseline profile for the same time period;

triggering an alarm for each difference that is computed and exceeds the threshold;

storing the aggregated volume counts for the plurality of time periods to re-define the aggregated volume data; and

modifying the baseline profile and the threshold in accordance with the aggregated volume data being re-defined, wherein the modifying comprises:

comparing a number of outliers in the aggregated volume data to a second threshold;

updating the baseline profile and the threshold using the aggregated volume data if the number of outliers is less than the second threshold; and

re-computing the baseline profile and the threshold using the aggregated volume data if the number of outliers exceeds the second threshold.

7. The apparatus of claim 6 , wherein the re-computing the baseline profile further comprises re-computing a variance, and updating the variance using the aggregated volume data while ignoring any outliers therein.

8. The apparatus of claim 7 , wherein the variance is re-computed and updated using the exponential smoothing.

9. The apparatus of claim 6 , wherein the baseline profile is updated and re-computed using the exponential smoothing.

10. The apparatus of claim 6 , wherein a duration of the time interval comprises a week and a duration of each of the plurality of time periods comprises an hour.

11. A non-transitory computer readable medium to store a plurality of instructions which, when executed by a hardware processor, cause the hardware processor to perform operations of processing traffic of interest in a network, the operations comprising:

computing, for the traffic of interest, a baseline profile and a threshold using initial aggregated volume data of the traffic of interest, wherein the baseline profile is computed using exponential smoothing, wherein the aggregated volume data comprises volume data of network traffic passing through a plurality of network elements of the network aggregated over a time interval having a plurality of time periods;

obtaining an aggregated volume count for each particular time period of the plurality of time periods, wherein the aggregated volume count comprises a count of volume of traffic data for each particular period of the plurality of time periods, wherein the aggregated volume count is a count of the network traffic passing through the plurality of network elements of the network;

computing, for each particular time period of the plurality of time periods, a difference between the aggregated volume count of the particular time period of the plurality of time periods and a particular value of the baseline profile for the same time period;

triggering an alarm for each difference that is computed and exceeds the threshold;

storing the aggregated volume counts for the plurality of time periods to re-define the aggregated volume data; and

modifying the baseline profile and the threshold in accordance with the aggregated volume data being re-defined, wherein the modifying comprises:

comparing a number of outliers in the aggregated volume data to a second threshold;

updating the baseline profile and the threshold using the aggregated volume data if the number of outliers is less than the second threshold; and

re-computing the baseline profile and the threshold using the aggregated volume data if the number of outliers exceeds the second threshold.

12. The non-transitory computer readable medium of claim 11 , wherein the re-computing the baseline profile further comprises re-computing a variance, and updating the variance using the aggregated volume data while ignoring any outliers therein.

13. The non-transitory computer readable medium of claim 12 , wherein the re-computing the variance and the updating the variance are performed using the exponential smoothing.

14. The non-transitory computer readable medium of claim 11 , wherein the updating the baseline profile and the re-computing the baseline profile are performed using the exponential smoothing.