Analysis device, analysis system, analysis method, and analysis program
An analysis device includes processing circuitry configured to acquire, from each network traffic sensor that monitors communication of an Internet of Things (IoT) device, a normal communication model that is used for monitoring the communication and indicates a characteristic of normal communication of the IoT device, cluster a normal communication model group of a same feature among acquired normal communication model groups, calculate a majority cluster that is a cluster having a largest number of normal communication models by using a result of the clustering, and calculate an average model of the normal communication model group belonging to the majority cluster, and notify the network traffic sensor serving as an acquisition source of the normal communication model of attribution information indicating whether or not the normal communication model belongs to the majority cluster and the average model.
1. An analysis device comprising:
processing circuitry configured to:
obtain normal communications models from a plurality of network traffic sensors respectively corresponding to a plurality of Internet of Things (IoT) devices, a respective network traffic sensor monitoring communications of a corresponding IoT device, and a respective normal communications model (i) being used to monitor the communications of the corresponding IoT device and (ii) indicating a characteristic of normal communications of the corresponding IoT device;
identify, from the obtained normal communications models, normal communications models that share a communications feature;
cluster the identified normal communications models that share the communications feature to generate a plurality of clusters of normal communications models;
determine a majority cluster having a largest number of normal communications models using a result of the clustering;
generate an average model of the normal communications models belonging to the majority cluster; and
notify the respective network traffic sensor of (i) attribution information indicating whether the respective normal communications model belongs to the majority cluster and (ii) the average model.
2. The analysis device according to claim 1 , wherein the processing circuitry is further configured to:
notify the respective network traffic sensor belonging to the majority cluster of the attribution information, and
notify the respective network traffic sensor not belonging to the majority cluster of the attribution information and the average model.
3. The analysis device according to claim 1 , wherein the processing circuitry is further configured to:
in a case where a normal communications model whose parameter has been updated from the respective network traffic sensor is obtained,
identify, from the obtained normal communications models including the normal communications model whose parameter has been updated, the normal communications models that share the communications feature,
cluster again the identified normal communications models that share the communications feature to regenerate the plurality of clusters of normal communications models,
redetermine the majority cluster, and
regenerate the average model of the normal communications models belonging to the redetermined majority cluster, and
in a case where any one of the normal communications models is out of the redetermined majority cluster,
notify the respective network traffic sensor serving as an acquisition source of the normal communications model that is out of the redetermined majority cluster of the attribution information indicating that the respective normal communications model is out of the redetermined majority cluster and the regenerated average model.
4. The analysis device according to claim 1 , wherein
in a case where notification indicating a failure in learning the respective normal communications model of the corresponding IoT device is obtained from the respective network traffic sensor,
notify the respective network traffic sensor of an average model of normal communications models belonging to one of the plurality of clusters having a same communications feature as the respective normal communications model whose learning has failed.
5. The analysis device according to claim 1 , wherein the communications feature includes at least one of a number of transmission packets per certain time period, a number of destination IP addresses, a destination port number, or a number of bytes of packets in the communication of the IoT device.
6. An analysis method executed by an analysis device, the analysis method comprising:
obtaining normal communications models from a plurality of network traffic sensors respectively corresponding to a plurality of Internet of Things (IoT) devices, a respective network traffic sensor monitoring communications of a corresponding IoT device, and a respective normal communications model (i) being used to monitor the communications of the corresponding IoT device and (ii) indicating a characteristic of normal communications of the corresponding IoT device;
identifying, from the obtained normal communications models, normal communications models that share a communications feature;
clustering the identified normal communications models that share the communications feature to generate a plurality of clusters of normal communications models;
determining a majority cluster having a largest number of normal communications models using a result of the clustering:
generating an average model of the normal communications models belonging to the majority cluster; and
notifying the respective network traffic sensor of (i) attribution information indicating whether the respective normal communications model belongs to the majority cluster and (ii) the average model.
7. A non-transitory computer-readable recording medium storing therein an analysis program that causes a computer to execute a process comprising:
obtaining normal communications models from a plurality of network traffic sensors respectively corresponding to a plurality of Internet of Things (IoT) devices, a respective network traffic sensor monitoring communications of a corresponding IoT device, and a respective normal communications model (i) being used to monitor the communications of the corresponding IoT device and (ii) indicating a characteristic of normal communications of the corresponding IoT device;
identifying, from the obtained normal communications models, normal communications models that share a communications feature;
clustering the identified normal communications models that share the communications feature to generate a plurality of clusters of normal communications models;
determining a majority cluster having a largest number of normal communications models using a result of the clustering:
generating an average model of the normal communications models belonging to the majority cluster; and
notifying the respective network traffic sensor of (i) attribution information indicating whether the respective normal communications model belongs to the majority cluster and (ii) the average model.