Analysis device, analysis method, and analysis program
View Patent ↗An analysis device includes processing circuitry configured to generate a graph in which each IP host indicated in communication information is set as a node and communication between the each IP host is set as an edge on a basis of the communication information of a network, calculate a degree of relevance between a predetermined IP host and each IP host on the graph using the generated graph, correct the graph by adding or deleting an edge connecting each IP host on the graph on a basis of a degree of relevance between a predetermined IP host and each IP host on the graph, calculate a degree of relevance between the predetermined IP host and each IP host using the corrected graph, and output a degree of relevance between the predetermined IP host and each IP host.
1. An analysis device comprising:
processing circuitry configured to:
generate, based on data traffic information of a plurality of Internet Protocol (IP) hosts in a network, a graph, wherein the graph comprises a plurality of IP hosts as respective nodes and communication between a first IP host of the plurality of IP hosts and a second IP host of the plurality of IP hosts as an edge;
identify, based on similarity of neighborhoods between a predetermined IP host of the plurality of IP hosts and the first IP host of the plurality of IP hosts on the graph, a first degree of relevance between the predetermined IP host of the plurality of IP hosts and the first IP host of the plurality of IP hosts on the generated graph, wherein the predetermined IP host represents a malicious IP host that cause one or more IP hosts of the plurality of IP hosts to perform relevant malicious operations;
update, based on a first predetermined threshold degree of relevance and the first degree of relevance between the predetermined IP host and the first IP host, the generated graph by modifying an edge connecting the predetermined IP host and the first IP host;
identify, based on the updated graph according to the modified edge, a second degree of relevance between the predetermined IP host and the first IP host; and
identify and extract, based on a second predetermined threshold degree of relevance and the identified second degree of relevance between the predetermined IP host and the first IP host, the first IP host from the network as relevant to the malicious IP host.
2. The analysis device according to claim 1 , wherein the processing circuitry is further configured to update the graph by adding an edge from the predetermined IP host to an IP host having the first degree of relevance equal to or greater than the first predetermined threshold degree of relevance in a case where there is no edge from the predetermined IP host to the IP host having the first degree of relevance equal to or greater than the first predetermined threshold degree of relevance in the graph.
3. The analysis device according to claim 1 , wherein the processing circuitry is further configured to update the graph by deleting the edge in a case where the edge from the predetermined IP host to the first IP host has the first degree of relevance less than the first predetermined threshold degree of relevance in the graph.
4. The analysis device according to claim 1 , wherein the processing circuitry is further configured to calculate the second degree of relevance by a predetermined calculation algorithm.
5. The analysis device according to claim 1 , wherein the processing circuitry is further configured to:
calculate the first degree of relevance between the predetermined IP host and the first IP host using the graph by Deep Walk, and
calculate the second degree of relevance between the predetermined IP host and the first IP host using the updated graph by Personalized PageRank.
6. The analysis device according to claim 1 , wherein
the predetermined IP host represents the malicious IP host designated in advance, the predetermined IP host corresponds to a predetermined node of the graph, and the predetermined host communicates with the first IP host of the plurality of IP hosts over the network.
7. An analysis method performed by an analysis device, the analysis method comprising:
generating, based on data traffic information of a plurality of Internet Protocol (IP) hosts in a network, a graph, wherein the graph comprises a plurality of IP hosts as respective nodes and communication between a first IP host of the plurality of IP hosts and a second IP host of the plurality of IP hosts as an edge;
identifying, based on similarity of neighborhoods between a predetermined IP host of the plurality of IP hosts and the first IP host of the plurality of IP hosts in the graph, a first degree of relevance between the predetermined IP host of the plurality of IP hosts and the first IP host of the plurality of IP hosts on the generated graph, wherein the predetermined IP host represents a malicious IP host that cause one or more IP hosts of the plurality of IP hosts to perform relevant malicious operations;
updating, based on a first predetermined threshold degree of relevance and the first degree of relevance between the predetermined IP host and the first IP host, the generated graph by modifying an edge connecting the predetermined IP host and the first IP host;
identifying, based on the updated graph according to the modified edge, calculating a second degree of relevance between the predetermined IP host and the first IP host; and
identifying and extracting, based on a second predetermined threshold degree of relevance and the identified second degree of relevance between the predetermined IP host and the identified first IP host, the first IP host from the network as relevant to the malicious IP host.
8. A non-transitory computer-readable recording medium storing therein an analysis program that causes a computer to execute a process comprising:
generating, based on data traffic information of a plurality of Internet Protocol (IP) hosts in a network, a graph, wherein the graph comprises a plurality of IP hosts as respective nodes and communication between a first IP host of the plurality of IP hosts and a second IP host of the plurality of IP hosts as an edge;
identifying, based on similarity of neighborhoods between a predetermined IP host of the plurality of IP hosts and the first IP host of the plurality of IP hosts on the graph, a first degree of relevance between the predetermined IP host of the plurality of IP hosts and the first IP host of the plurality of IP hosts on the generated graph, wherein the predetermined IP host represents a malicious IP host that cause one or more IP hosts of the plurality of IP hosts to perform relevant malicious operations;
updating, based on a first predetermined threshold degree of relevance and the first degree of relevance between the predetermined IP host and the first IP host, the generated graph by modifying an edge connecting the predetermined IP host and the first IP host;
identifying, based on the updated graph according to the modified edge, calculating a second degree of relevance between the predetermined IP host and the first IP host; and
identifying and extracting, based on a second predetermined threshold degree of relevance and the identified second a degree of relevance between the predetermined IP host and the identified first IP host, the first IP host from the network as relevant to the malicious IP host.