Graph database analysis for network anomaly detection systems

1. A computer-implemented method comprising:

receiving, by a graph analytics module of a data analysis device, log data that includes parameters associated with a computer network event in a computing network;

extracting, by a stream analytics engine of the data analysis device, parameters from the log data, wherein one or more parameters are extracted in real-time;

generating, by the data analysis device, a network event graph based on parameters extracted from the log data, the network event graph comprising data that describes communications between a plurality of nodes in the computing network;

storing the network event graph in a graphics processing unit (“GPU”) accelerated graph database of the data analysis device;

processing a query against the GPU accelerated graph database to obtain risk information from the network event graph about connections that exist between two or more of the nodes;

identifying, using the GPU accelerated graph database and based on queries directed to the network event graph, a type of ordered-connection that exists between a primary node and another node in the computing network;

detecting, using the GPU accelerated graph database and based on the type of ordered-connection, that the other node is a secondary node having a first-order connection to the primary node that indicates the secondary node is affected by a malicious event associated with the computing network; and

generating a node structure that (i) includes data describing the malicious event and (ii) depicts risk information about types of ordered-connections that exist between the primary node and two or more other nodes in the computing network.

2. The method of claim 1 , wherein detecting includes, querying data for the network event graph to detect one or more secondary nodes that have engaged in data communications with at least one primary node.

3. The method of claim 2 , wherein the at least one primary node is directly associated with at least one of an anomalous event or the malicious event.

4. The method of claim 3 , further comprising:

analyzing, by the data analysis device, network communication activities of the one or more secondary nodes to identify at least one tertiary node that has a correlation to the at least one primary node.

5. The method of claim 1 , wherein storing the network event graph comprises:

storing a first graph metric, and a second graph metric, wherein the first graph metric represents analytic measures that are parsed to obtain information about network activity of a first node in the computing network; and the second graph metric represents analytic measures that are parsed to obtain information about network activity of a second, different node in the computing network.

6. The method of claim 5 , further comprising:

storing, in an intermediate storage medium of the data analysis device, the first graph metric and the second graph metric to be stored in the GPU accelerated graph database, wherein storing the first and second graph metrics in the storage medium includes extracting, from the log data, a respective feature that corresponds to the first or second graph metrics to be stored in the GPU accelerated graph database.

7. The method of claim 1 , wherein extracting the parameters from the log data comprises:

using the stream analytics engine to apply an extract, transform, load function to extract the one or more parameters from the log data in real-time.

8. An electronic system comprising:

one or more processing devices;

one or more non-transitory machine-readable storage devices storing instructions that are executable by the one or more processing devices to cause performance of operations comprising:

receiving, by a graph analytics module of a data analysis device, log data that includes parameters associated with a computer network event in a computing network;

extracting, by a stream analytics engine of the data analysis device, parameters from the log data, wherein one or more parameters are extracted in real-time;

generating, by the data analysis device, a network event graph based on the parameters extracted from the log data, the network event graph comprising data that describes communications between a plurality of nodes in the computing network;

storing the network event graph in a graphics processing unit (“GPU”) accelerated graph database of the data analysis device;

processing a query against the GPU accelerated graph database to obtain risk information from the network event graph about connections that exist between two or more of the nodes;

identifying, using the GPU accelerated graph database and based on queries directed to the network event graph, a type of ordered-connection that exists between a primary node and another node in the computing network;

detecting, using the GPU accelerated graph database and based on the type of ordered-connection, that the other node is a secondary node having a first-order connection to the primary node that indicates the secondary node is affected by a malicious event associated with the computing network; and

generating a node structure that (i) includes data describing the malicious event and (ii) depicts risk information about types of ordered-connections that exist between the primary node and two or more other nodes in the computing network.

9. The electronic system of claim 8 , wherein detecting includes, querying data for the network event graph to detect one or more secondary nodes that have engaged in data communications with at least one primary node.

10. The electronic system of claim 9 , wherein the at least one primary node is directly associated with at least one of an anomalous event or the malicious event.

11. The electronic system of claim 10 , wherein the operations further comprise:

analyzing, by the data analysis device, network communication activities of the one or more secondary nodes to identify at least one tertiary node that has a correlation to the at least one primary node.

12. The electronic system of claim 8 , wherein storing the network event graph comprises:

storing a first graph metric, and a second graph metric, wherein the first graph metric represents analytic measures that are parsed to obtain information about network activity of a first node in the computing network; and the second graph metric represents analytic measures that are parsed to obtain information about network activity of a second, different node in the computing network.

13. The electronic system of claim 12 , wherein the operations further comprise:

storing, in an intermediate storage medium of the data analysis device, the first graph metric and the second graph metric to be stored in the GPU accelerated graph database, wherein storing the first and second graph metrics in the storage medium includes extracting, from the log data, a respective feature that corresponds to the first or second graph metrics to be stored in the GPU accelerated graph database.

14. The electronic system of claim 8 , wherein extracting the parameters from the log data comprises:

using the stream analytics engine to apply an extract, transform, load function to extract the one or more parameters from the log data in real-time.

15. One or more non-transitory machine-readable storage devices storing instructions that are executable by one or more processing devices to cause performance of operations comprising:

receiving, by a graph analytics module of a data analysis device, log data that includes parameters associated with a computer network event in a computing network;

extracting, by a stream analytics engine of the data analysis device, parameters from the log data, wherein one or more parameters are extracted in real-time;

generating, by the data analysis device, a network event graph based on the parameters extracted from the log data, the network event graph comprising data that describes communications between a plurality of nodes in the computing network;

storing the network event graph in a graphics processing unit (“GPU”) accelerated graph database of the data analysis device;

processing a query against the GPU accelerated graph database to obtain risk information from the network event graph about connections that exist between two or more of the nodes;

identifying, using the GPU accelerated graph database and based on queries directed to the network event graph, a type of ordered-connection that exists between a primary node and another node in the computing network;

detecting, using the GPU accelerated graph database and based on the type of ordered-connection, that the other node is a secondary node having a first-order connection to the primary node that indicates the secondary node is affected by a malicious event associated with the computing network; and

generating a node structure that (i) includes data describing the malicious event and (ii) depicts risk information about types of ordered-connections that exist between the primary node and two or more other nodes in the computing network.

16. The non-transitory machine-readable storage devices of claim 15 , wherein detecting includes, querying data for the network event graph to detect one or more secondary nodes that have engaged in data communications with at least one primary node.

17. The non-transitory machine-readable storage devices of claim 16 , wherein the at least one primary node is directly associated with at least one of an anomalous event or the malicious event.

18. The non-transitory machine-readable storage devices of claim 17 , wherein the operations further comprise:

analyzing, by the data analysis device, network communication activities of the one or more secondary nodes to identify at least one tertiary node that has a correlation to the at least one primary node.

19. The non-transitory machine-readable storage devices of claim 15 , wherein storing the network event graph comprises:

storing a first graph metric, and a second graph metric, wherein the first graph metric represents analytic measures that are parsed to obtain information about network activity of a first node in the computing network; and the second graph metric represents analytic measures that are parsed to obtain information about network activity of a second, different node in the computing network.

20. The non-transitory machine-readable storage devices of claim 15 , wherein extracting the parameters from the log data comprises:

using the stream analytics engine to apply an extract, transform, load function to extract the one or more parameters from the log data in real-time.

21. The method of claim 1 , further comprising:

receiving the query for processing against the GPU accelerated graph database, the query being received using a query terminal that communicates with the GPU accelerated graph database.

22. The method of claim 21 , wherein:

the node structure is generated based on a respective risk score for each of the two or more nodes; and

the risk information about the types of ordered connections indicate whether each of the two or more other nodes are affected by the malicious event by way of a direct connection to the primary node or an indirect connection to the primary node.