Attack analysis device, attack analysis method, and attack analysis program
The attack analysis device filters cyber attack-related objects by using influence source and target objects with a threshold-based attack probability, accurately identifying intrusion routes and contamination ranges, addressing misidentification issues in existing systems.
Patent Information
- Application Number
- JP2024518965
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2022-06-01
- Publication Date
- 2024-11-15
- Estimated Expiration
- 2042-06-01
AI Technical Summary
Existing cyber attack identification systems inaccurately identify intrusion routes and contamination ranges due to inclusion of unrelated data, leading to misidentification of non-affected objects.
An attack analysis device that utilizes an influence source object, influence target object, and attack probability to construct relational data, adding relationships only when the attack probability exceeds a threshold, thereby identifying likely intrusion routes and contamination ranges.
Accurately identifies cyber attack-related objects by filtering out unrelated data, ensuring precise detection of intrusion routes and contamination ranges, suitable for real-time applications.
Smart Images

Figure 0007584708000001 
Figure 0007584708000002 
Figure 0007584708000003
Abstract
Description
[Technical field]
[0001] The present disclosure relates to a technology for identifying a contamination range assumed to be affected by a cyber attack. [Background technology]
[0002] Patent Document 1 describes a technology for identifying the intrusion path and contamination extent of a system that has been subjected to a cyber attack. In Patent Document 1, relationship data between objects is generated from operation data of multiple software programs. Then, when a cyber attack is detected, the intrusion path and contamination extent are identified based on the relationship data. The relational data is a directed graph. The directed graph has objects as nodes. The directed graph has directed edges from an influencing object to an affected object, corresponding to the operation data that affects the objects. When a cyber attack is detected, the relational data is referenced. The range obtained by tracing the directed edges in the reverse direction from the object where the attack was detected is identified as the intrusion route. The range obtained by tracing the edges in the forward direction is identified as the contamination range. [Prior art documents] [Patent documents]
[0003] [Patent Document 1] Patent No. 6987332 Summary of the Invention [Problem to be solved by the invention]
[0004] In Patent Document 1, relationship data between objects is generated from operation data of multiple software programs. At this time, relationship data regarding operations unrelated to cyber attacks is also generated. Therefore, when a cyber attack occurs, tracing the relationship data backward or forward may lead to reaching an object unrelated to the cyber attack. As a result, there is a risk that a part that is not an intrusion route or a contaminated area may be identified as an intrusion route or a contaminated area. The present disclosure aims to make it possible to identify at least one of the intrusion route and the extent of contamination of an object that is likely to be related to a cyber attack. [Means for solving the problem]
[0005] The attack analysis device according to the present disclosure comprises: a data acquisition unit that acquires action data including at least an affected object and an attack probability among an influencing object that is an object that has performed an action that has an effect on another object, an affected object that is an object that is affected by the action performed by the influencing object, and an attack probability that indicates the probability that the action is an attack; a relationship construction unit that constructs the relationship data by adding a relationship between the influencing object and the affected object to the relationship data when the action data acquired by the data acquisition unit includes the influencing object and the attack probability is equal to or greater than a threshold A; an identification unit that, when an attack is detected, identifies at least one of a contamination range assumed to be affected by the detected attack and an intrusion route of the detected attack, based on the relationship data constructed by the relationship construction unit; Equipped with. Effect of the Invention
[0006] In the present disclosure, when the attack certainty is equal to or greater than the threshold A, the relationship data is constructed by adding the relationship between the influencing object and the affected object to the relationship data. This makes it possible to identify only objects that are highly likely to be related to a cyber attack as an intrusion route or a contamination range. [Brief description of the drawings]
[0007] [Figure 1] FIG. 1 is a configuration diagram of an attack analysis device 10 according to a first embodiment. [Diagram 2] FIG. 3 is an explanatory diagram of motion data 31 according to the first embodiment. [Diagram 3]FIG. 4 is an explanatory diagram of a threshold value 41 according to the first embodiment. [Figure 4] FIG. 4 is an explanatory diagram of relationship data 51 according to the first embodiment. [Diagram 5] 4 is a flowchart of a relationship building process according to the first embodiment. [Figure 6] 4 is a flowchart of a specification process according to the first embodiment. [Figure 7] FIG. 13 is a configuration diagram of an attack analysis device 10 according to a first modified example. [Figure 8] FIG. 11 is an explanatory diagram of history data 61 according to the second embodiment. [Figure 9] 11 is a flowchart of a relationship building process according to the second embodiment. [Figure 10] FIG. 11 is a configuration diagram of an attack analysis device 10 according to a third embodiment. [Figure 11] FIG. 13 is an explanatory diagram of motion data 31 according to the third embodiment. [Figure 12] 13 is a flowchart of an attack probability calculation process according to the third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0008] Embodiment 1 ***Configuration Description*** The configuration of an attack analysis device 10 according to the first embodiment will be described with reference to FIG. The attack analysis device 10 is a computer. The attack analysis device 10 includes the following hardware components: a processor 11, a memory 12, and a storage 13. The processor 11 is connected to other hardware components via signal lines and controls these other hardware components.
[0009] The processor 11 is an IC that performs processing. IC is an abbreviation for Integrated Circuit. Specific examples of the processor 11 include a CPU, a DSP, and a GPU. CPU is an abbreviation for Central Processing Unit. DSP is an abbreviation for Digital Signal Processor. GPU is an abbreviation for Graphics Processing Unit.
[0010] The memory 12 is a volatile storage device that temporarily stores data. Specific examples of the memory 12 include SRAM and DRAM. SRAM is an abbreviation for Static Random Access Memory. DRAM is an abbreviation for Dynamic Random Access Memory.
[0011] The storage 13 is a non-volatile storage device that stores data. A specific example of the storage 13 is an HDD. HDD is an abbreviation for Hard Disk Drive. The storage 13 may also be a portable recording medium such as an SD (registered trademark) memory card, CompactFlash (registered trademark), NAND flash, a flexible disk, an optical disk, a compact disk, a Blu-ray (registered trademark) disk, or a DVD. SD is an abbreviation for Secure Digital. DVD is an abbreviation for Digital Versatile Disk.
[0012] The attack analysis device 10 includes, as functional components, a data acquisition unit 21, a relationship construction unit 22, and an identification unit 23. The functions of the functional components of the attack analysis device 10 are realized by software. The storage 13 stores programs that realize the functions of each functional component of the attack analysis device 10. These programs are loaded into the memory 12 by the processor 11 and executed by the processor 11. In this way, the functions of each functional component of the attack analysis device 10 are realized.
[0013] 1 shows only one processor 11. However, there may be a plurality of processors 11, and the plurality of processors 11 may cooperate to execute programs that realize the respective functions.
[0014] ***Explanation of Operation*** The operation of the attack analysis device 10 according to the first embodiment will be described with reference to FIGS. The operation procedure of the attack analysis device 10 according to the embodiment 1 corresponds to the attack analysis method according to the embodiment 1. Moreover, the program that realizes the operation of the attack analysis device 10 according to the embodiment 1 corresponds to the attack analysis program according to the embodiment 1.
[0015] The operation of the attack analysis device 10 includes a relationship building process and a determination process. The relationship building process and the determination process will be described after explaining the data and the like used in the relationship building process and the determination process.
[0016] The action data 31 according to the first embodiment will be described with reference to FIG. The operation data 31 indicates an operation when an object in a target system performs an operation that affects other objects in the system. The target system is a system for which at least one of an intrusion route and a contamination range is to be identified. An object is an element used when software is executed. An object is an element such as a process, a file, or a shared memory area.
[0017] The action data 31 includes an influencing object 32, an influenced object 33, and an attack probability . An influencing object 32 is an object that performs an action that affects other objects. An affected object 33 is an object that is affected by an action performed by the influencing object 32. Suppose data is sent from process A to process B. In this case, the influencing object 32 is process A. The affected object 33 is process B. Suppose data is written from process C to file D. In this case, the influencing object 32 is process C. The affected object 33 is file D. Attack likelihood 34 indicates the probability that the action is a cyber attack against the target system.
[0018] The operational data 31 is generated, for example, by an intrusion detection program or a security monitoring program installed on the target system. It is also permissible that the behavior data 31 does not include the influencing object 32. In cases where an intrusion detection program or the like that generates the behavior data 31 cannot identify the influencing source, the behavior data 31 will not include the influencing object 32.
[0019] The threshold value 41 according to the first embodiment will be described with reference to FIG. The threshold 41 is information for classifying the actions indicated by the action data 31 into three types, "not an attack", "suspicious attack", and "attack", based on the attack accuracy 34. Here, the threshold 41 indicates that the action is classified as "not an attack" when the attack accuracy 34 is less than threshold A. The threshold 41 indicates that the action is classified as "suspicious attack" when the attack accuracy 34 is equal to or greater than threshold A and less than threshold B. The threshold 41 indicates that the action is classified as "attack" when the attack accuracy 34 is equal to or greater than threshold B. The threshold value 41 is set in the programs of the relationship building unit 22 and the identification unit 23. Alternatively, the threshold value 41 may be set in the storage 13.
[0020] The relationship data 51 according to the first embodiment will be described with reference to FIG. The relationship data 51 is a directed graph consisting of nodes 52 indicating objects and edges 53 each having a direction from an influencing object to an influenced object. In Fig. 4, the nodes 52 are represented by ovals. The edges 53 are represented by arrows. The relationship data 51 is stored in the memory 12 or the storage 13.
[0021] The relationship data 51 is generated from the behavior data 31. In FIG. 4, it can be seen that there is behavior data 31 in which the influencing object 32 is process A and the influenced object 33 is file A. It can be seen that there is behavior data 31 in which the influencing object 32 is file A and the influenced object 33 is process B. It can be seen that there is behavior data 31 in which the influencing object 32 is process B and the influenced object 33 is file B. It can be seen that there is behavior data 31 in which the influencing object 32 is process B and the influenced object 33 is process C. It can be seen that there is behavior data 31 in which the influencing object 32 is process D and the influenced object 33 is process C.
[0022] The relationship building process according to the first embodiment will be described with reference to FIG. (Step S11: Relationship data initialization process) The relationship constructing unit 22 initializes the relationship data 51 to be empty. That is, the relationship constructing unit 22 puts the relationship data 51 into a state in which no nodes 52 or edges 53 exist.
[0023] (Step S12: Operation data input process) The relationship constructing unit 22 receives new action data 31 as an input. Specifically, the data acquiring unit 21 acquires newly generated motion data 31. The motion data 31 acquired here includes at least the influenced object 33 and the attack probability 34 out of the influencing object 32, the influenced object 33, and the attack probability 34. The data acquiring unit 21 inputs the acquired motion data 31 to the relationship building unit 22. The relationship building unit 22 accepts the input motion data 31.
[0024] (Step S13: Data determination process) The relationship building unit 22 determines whether the action data 31 received in step S12 includes the influencing object 32 and whether the attack probability 34 is equal to or greater than the threshold A. If the action data 31 includes the influencing object 32 and the attack certainty 34 is equal to or greater than the threshold A, the relationship building unit 22 advances the process to step S14. If the attack certainty 34 is equal to or greater than the threshold A, it means that the action has been classified as "suspicious attack" or "actual attack." On the other hand, in other cases, the relationship building unit 22 advances the process to step S15.
[0025] (Step S14: Data addition process) The relationship constructing section 22 adds the relationship between the influencing object 32 and the influenced object 33 in the action data 31 received in step S12 to the relationship data 51. In this way, the relationship constructing section 22 constructs the relationship data 51. Specifically, the relationship establishing unit 22 adds directional information indicating the direction from the influencing object 32 to the influenced object 33 to the relationship data 51. Here, the relationship data 51 is a directed graph. Therefore, the relationship establishing unit 22 adds an edge 53 directed from a node 52 corresponding to the influencing object 32 to a node 52 corresponding to the influenced object 33 to the relationship data 51 as directional information. More specifically, the relationship building unit 22 determines whether or not the influencing object 32 exists in the relationship data 51. If the influencing object 32 does not exist in the relationship data 51, the relationship building unit 22 adds a node 52 corresponding to the influencing object 32 to the relationship data 51. Similarly, the relationship building unit 22 determines whether or not the influenced object 33 exists in the relationship data 51. If the influenced object 33 does not exist in the relationship data 51, the relationship building unit 22 adds a node 52 corresponding to the influenced object 33 to the relationship data 51. The relationship building unit 22 adds an edge 53 pointing from the node 52 corresponding to the influencing object 32 to the node corresponding to the influenced object 33.
[0026] (Step S15: Output process) The relationship building unit 22 outputs the action data 31 received in step S12 to the identification unit 23. Then, the relationship building unit 22 returns the process to step S12 and waits for the next action data 31 to be input.
[0027] The identification process according to the first embodiment will be described with reference to FIG. (Step S21: Operation data input process) The identification unit 23 receives as input the new action data 31. Specifically, the identification unit 23 receives the action data 31 output in step S15.
[0028] (Step S22: Data determination process) The identification unit 23 determines whether the attack probability 34 in the action data 31 received in step S21 is equal to or greater than a threshold value B. If the attack certainty 34 is equal to or greater than the threshold B, the identification unit 23 proceeds to step S23. The fact that the attack certainty 34 is equal to or greater than the threshold B means that the action has been classified as "an attack." In other words, the fact that the attack certainty 34 is equal to or greater than the threshold B means that an attack on the target system has been detected. Otherwise, the identification unit 23 returns the process to step S21 and waits for the input of the next action data 31.
[0029] (Step S23: Intrusion route identification process) The identification unit 23 refers to the relationship data 51 constructed by the relationship construction unit 22 and identifies the intrusion route of the attack indicated by the operation data 31 received in step S21. Specifically, the identification unit 23 identifies an object in which a motion indicating an attack indicated by the motion data 31 received in step S21 has been detected. Here, the identification unit 23 identifies the affected object 33 in the motion data 31 as the object in which a motion indicating an attack has been detected. The identification unit 23 identifies an intrusion route by tracing backward from the identified object as a starting point the direction indicated by the directional information included in the relationship data 51. The identification unit 23 identifies the range reached by tracing backward from the starting point as the intrusion route. For example, suppose that process B in Fig. 4 is an object for which an action indicating an attack has been detected. In this case, edge 53, which is directional information, is traced in the reverse direction, and process A, file A, and process B are identified as the intrusion route.
[0030] (Step S24: Contamination range identification process) The identifying unit 23 refers to the relationship data 51 constructed by the relationship constructing unit 22, and identifies the contamination range assumed to be affected by the attack indicated by the operation data 31 received in step S21. Specifically, the identification unit 23 identifies an object in which a motion indicating an attack indicated in the motion data 31 received in step S21 has been detected. Here, the identification unit 23 identifies the affected object 33 in the motion data 31 as the object in which a motion indicating an attack has been detected. The identification unit 23 identifies the contamination range by tracing forward from the identified object as a starting point in the direction indicated by the direction information included in the relationship data 51. The identification unit 23 identifies the range reached by tracing forward from the starting point as the contamination range. For example, assume that process B in Fig. 4 is an object in which an action indicating an attack has been detected. In this case, edge 53, which is directional information, is traced in the forward direction, and process B, file B, process C, and process D are identified as the contamination range.
[0031] (Step S25: Data generation process) The identification unit 23 generates output data indicating the intrusion route identified in step S23 and the contamination range identified in step S24. Specifically, the identification unit 23 generates data including a set of objects included in the intrusion path and a set of objects included in the contaminated range as output data. The identification unit 23 may further include the edge 53 traced in step S23 and the edge 53 traced in step S24 in the output data. Then, the specification unit 23 returns the process to step S21 and waits for the next action data 31 to be input.
[0032] ***Advantages of the First Embodiment*** As described above, only when the action is classified as "suspicious attack" or "actual attack", the attack analysis device 10 according to the first embodiment adds the influence relationship between objects due to the action indicated by the action data 31 to the relationship data 51. In other words, only when the probability of a cyber attack is equal to or higher than a certain level, the attack analysis device 10 according to the first embodiment adds the influence relationship between objects due to the action indicated by the action data 31 to the relationship data 51. This makes it possible to specify only objects that are highly likely to be related to a cyber attack as an intrusion route or a contamination range. In other words, it makes it possible to avoid specifying objects that are unlikely to be related to a cyber attack as an intrusion route or a contamination range.
[0033] Furthermore, the attack analysis device 10 according to the first embodiment constructs the relationship data 51 every time the operation data 31 is input. Then, when the operation of the operation data 31 is determined to be an "attack", the attack analysis device 10 according to the first embodiment can immediately identify the intrusion route and the contamination range. This makes it possible to apply the technology to control systems that require real-time performance.
[0034] ***Other configurations*** <Variation 1> In the first embodiment, each functional component is realized by software. However, as a first modification, each functional component may be realized by hardware. The following describes the first modification in terms of differences from the first embodiment.
[0035] The configuration of an attack analysis device 10 according to the first modification will be described with reference to FIG. When each functional component is realized by hardware, the attack analysis device 10 includes an electronic circuit 14 instead of the processor 11, the memory 12, and the storage 13. The electronic circuit 14 is a dedicated circuit for realizing the functions of each functional component, the memory 12, and the storage 13.
[0036] The electronic circuit 14 may be a single circuit, a composite circuit, a programmed processor, a parallel programmed processor, a logic IC, a GA, an ASIC, or an FPGA. GA is an abbreviation for Gate Array. ASIC is an abbreviation for Application Specific Integrated Circuit. FPGA is an abbreviation for Field-Programmable Gate Array. Each functional component may be realized by one electronic circuit 14, or each functional component may be distributed across multiple electronic circuits 14.
[0037] <Variation 2> As a second modification, some of the functional components may be realized by hardware, and other functional components may be realized by software.
[0038] The processor 11, the memory 12, the storage 13, and the electronic circuit 14 are collectively referred to as a processing circuit. In other words, the functions of the respective functional components are realized by the processing circuit.
[0039] Embodiment 2 The second embodiment differs from the first embodiment in that, even for action data 31 with attack likelihood 34 below threshold A, if the attack likelihood 34 is determined to be higher than the standard in relation to other action data 31, the influence relationship between objects is added to the relationship data 51. In the second embodiment, this difference will be explained, and explanations of the same points will be omitted.
[0040] ***Explanation of Operation*** The operation of the attack analysis device 10 according to the second embodiment will be described with reference to FIGS. The operation procedure of the attack analysis device 10 according to the embodiment 2 corresponds to the attack analysis method according to the embodiment 2. Moreover, the program that realizes the operation of the attack analysis device 10 according to the embodiment 2 corresponds to the attack analysis program according to the embodiment 2.
[0041] The relationship building process will be described after explaining the data and the like used in the relationship building process. The identification process is the same as in the first embodiment.
[0042] With reference to FIG. 8, the history data 61 according to the second embodiment will be described. The history data 61 is the action data 31 that has an attack probability 34 less than the threshold A and includes an influencing object 32, among the action data 31 that has been input in the past.
[0043] The history data 61 includes an influencing object 62 , an influenced object 63 , and time information 64 . The influencing object 62 is the influencing object 32 in the action data 31 that is the basis of the history data 61. The influenced object 63 is the influenced object 33 in the action data 31 that is the basis of the history data 61. The time information 64 indicates the time when the action indicated by the action data 31 that is the basis of the history data 61 was performed.
[0044] The relationship building process according to the second embodiment will be described with reference to FIG. The process of step S31 is the same as the process of step S11 in Fig. 5. The processes of steps S33 to S35 are the same as the processes of steps S12 to S14 in Fig. 5. The process of step S39 is the same as the process of step S15 in Fig. 5. However, in step S34, if the action data 31 includes the influencing object 32 and the attack probability 34 is equal to or greater than the threshold A, the relationship building unit 22 advances the process to step S35. On the other hand, in other cases, the relationship building unit 22 advances the process to step S36.
[0045] (Step S32: History data initialization process) The relationship building unit 22 initializes the history data 61 to be empty. That is, the history data 61 is put into a state in which no data exists.
[0046] (Step S36: Second data addition process) The relationship building unit 22 extracts past data, which is the action data 31 included in the history data 61, whose attack accuracy 34 is determined to be higher than the standard based on the relationship with the newly acquired action data 31 in step S33. The relationship building unit 22 may determine that the attack accuracy 34 of the past data is higher than the standard in the following cases (1) and (2). (1) A case where condition 1 is satisfied and condition 2 is satisfied. Condition 1 is a condition that the attack probability 34 in the newly acquired action data 31 is equal to or greater than threshold A. Condition 2 is a condition that the difference between the action time of the newly acquired action data 31 and the action time of the past data is within a reference time. The current time may be considered as the action time of the newly acquired action data 31. Note that condition 1 may be omitted. (2) A case where condition 1 is satisfied and condition 3 is satisfied. Condition 1 is as described above. Condition 3 is a condition that at least one of the influencing object and the influenced object matches between the newly acquired action data 31 and the past data. Condition 3 is satisfied in either of the following cases (a) and (b). (a) The influencing object 32 in the newly acquired action data 31 matches either the influencing object 62 or the influenced object 63 in the past data. (b) The influenced object 33 in the newly acquired action data 31 matches either the influencing object 62 or the influenced object 63 in the past data. Note that condition 1 may be omitted.
[0047] The relationship building unit 22 may use (1) and (2) in combination to determine that the attack likelihood 34 of the past data is higher than the standard. In other words, the relationship building unit 22 may determine that the attack likelihood 34 of the past data is higher than the standard when condition 1 is satisfied, condition 2 is satisfied, and condition 3 is satisfied.
[0048] The relationship constructing unit 22 adds the relationship between the influencing object 62 and the influenced object 63 in the extracted history data 61 to the relationship data 51. Specifically, similar to the process of step S14 in Fig. 5, the relationship constructing unit 22 adds the edge 53 directed from the node 52 corresponding to the influencing object 62 to the node 52 corresponding to the influenced object 63 to the relationship data 51 as directional information.
[0049] (Step S37: Influence source determination process) The relationship building unit 22 determines whether or not the influence source object 62 is included in the action data 31 newly acquired in step S33. If the influence source object 62 is included, the relationship building section 22 advances the process to step S38. Otherwise, the relationship building section 22 advances the process to step S39.
[0050] (Step S38: History addition process) The relationship building unit 22 adds the newly acquired action data 31 to the history data 61 in step S33. Specifically, the relationship building unit 22 creates a new record in the history data 61. Then, the relationship building unit 22 writes the influencing object 32 and the influenced object 33 in the action data 31 to the influencing object 62 and the influenced object 63 in the history data 61. The relationship building unit 22 writes the action time of the action data 31 to the time information 64 in the history data 61.
[0051] ***Effects of the second embodiment*** As described above, the attack analysis device 10 according to the second embodiment adds the influence relationship between objects to the relationship data 51 when it is determined that the attack likelihood 34 is higher than the standard based on the relationship with other action data 31. There are some actions that cannot be judged as "attack" or "suspicious attack" based on the action data 31 alone. Even for such actions, it is possible to construct the relationship data 51 by considering them to be similar to "suspicious attack" actions based on the presence or absence of a relationship with "attack" or "suspicious attack" actions.
[0052] Embodiment 3 The third embodiment differs from the first and second embodiments in that the attack probability 34 is calculated when the attack probability 34 is not included in the action data 31. In the third embodiment, this difference will be explained, and explanations of the same points will be omitted.
[0053] ***Configuration Description*** The configuration of an attack analysis device 10 according to the third embodiment will be described with reference to FIG. 1 in that the attack analysis device 10 includes an attack likelihood calculation unit 24 as a functional component. The function of the attack likelihood calculation unit 24 is realized by software or hardware, like the other functional components.
[0054] ***Explanation of Operation*** The operation of the attack analysis device 10 according to the third embodiment will be described with reference to FIGS. The operation procedure of the attack analysis device 10 according to the embodiment 3 corresponds to the attack analysis method according to the embodiment 3. Moreover, the program that realizes the operation of the attack analysis device 10 according to the embodiment 3 corresponds to the attack analysis program according to the embodiment 3.
[0055] The operation of the attack analysis device 10 includes a relationship building process, an identification process, and an attack likelihood calculation process. The attack likelihood calculation process will be described after explaining the data and the like used in the attack likelihood calculation process.
[0056] The action data 31 according to the third embodiment will be described with reference to FIG. The action data 31 includes action information 35 in addition to an influencing object 32, an influenced object 33, and an attack probability 34. The operation information 35 indicates the type of operation and the operation content. For example, the type of operation is data transmission by socket communication. In this case, the operation content includes a port number, a data size, and transmission data. As another example, the type of operation is writing to a file. In this case, the operation content includes a file path, a write size, a write position, and write data.
[0057] The attack probability calculation process according to the third embodiment will be described with reference to FIG.
[0058] (Step S41: Operation data input process) The attack probability calculation unit 24 receives as input the new action data 31. The specific processing is the same as that in step S12 in FIG.
[0059] (Step S42: Data determination process) The attack probability calculation unit 24 determines whether the action data 31 received in step S41 includes the attack probability 34 or not. If the action data 31 does not include the attack accuracy calculation unit 24, the attack accuracy calculation unit 24 advances the process to step S43. On the other hand, in other cases, the attack accuracy calculation unit 24 advances the process to step S44.
[0060] (Step S43: Accuracy calculation process) The attack accuracy calculation unit 24 calculates the attack accuracy 34 from the action content indicated by the action information 35 in the action data 31 received in step S41. The attack accuracy calculation unit 24 adds the calculated attack accuracy 34 to the action data 31. Specifically, the attack accuracy calculation unit 24 sets a value of the attack accuracy 34 for each judgment criterion in advance. The judgment criterion is determined based on the action content. The attack accuracy calculation unit 24 identifies the judgment criterion that is satisfied by the action content in the action data 31. The attack accuracy calculation unit 24 identifies the value of the attack accuracy 34 that corresponds to the identified judgment criterion.
[0061] A specific example of the judgment criteria will be described. Assume that the type of operation is data transmission via socket communication. In this case, the following criteria can be considered. Criterion 1: The port number is one that could be used in an attack. Criterion 2: The data size is different from the data size used in normal communication. Criterion 3: The content of the transmitted data matches or is similar to known attack data. Assume that the operation type is file writing. In this case, the following criteria can be considered for judgment: Criterion 1: Writing is being performed to a file that is not written to during normal operation. Criterion 2: The write size is different from the write size during normal operation.
[0062] (Step S44: Output process) The attack likelihood calculation unit 24 outputs the action data 31 to the relationship construction unit 22. Then, the attack likelihood calculation unit 24 returns the process to step S41 and waits for the input of the next action data 31.
[0063] The relationship building process and the identification process are the same as those in the first and second embodiments. However, in the action data input process in the relationship building process (step S12 in FIG. 5 and step S33 in FIG. 9), the action data 31 output in step S44 is accepted.
[0064] ***Effects of the Third Embodiment*** As described above, the attack analysis device 10 according to the third embodiment calculates the attack likelihood 34 when the action data 31 does not include the attack likelihood 34. As a result, even if the attack probability 34 is not specified in advance, it is possible to obtain the same effects as those of the first and second embodiments.
[0065] In addition, the word "part" in the above description may be read as a "circuit," "step," "procedure," "processing," or "processing circuit."
[0066] The above describes the embodiments and modifications of the present disclosure. Some of these embodiments and modifications may be combined and implemented. Also, one or some of them may be partially implemented. Note that the present disclosure is not limited to the above embodiments and modifications, and various modifications are possible as necessary. [Explanation of symbols]
[0067] 10 attack analysis device, 11 processor, 12 memory, 13 storage, 14 electronic circuit, 21 data acquisition unit, 22 relationship construction unit, 23 identification unit, 24 attack probability calculation unit, 31 operation data, 32 source object, 33 affected object, 34 attack probability, 35 operation information, 41 threshold, 51 relationship data, 52 node, 53 edge, 61 history data, 62 source object, 63 affected object, 64 time information.
Claims
1. a data acquisition unit which sequentially acquires operation data including at least an affected object and an attack probability among an influencing object which is an object of a target system that has performed an operation that affects other objects, an affected object which is an object of the target system that is affected by the operation performed by the influencing object, and an attack probability which indicates the probability that the operation is a cyber attack against the target system; a relationship construction unit which uses data of an initial state as initial relationship data and repeatedly adds a relationship between the influencing object and the affected object to the relationship data when the action data acquired by the data acquisition unit includes the influencing object and the attack probability is equal to or greater than a threshold A; an identification unit that uses the operation data as an input, determines that a cyber-attack on the target system has been detected if the attack certainty is equal to or higher than a threshold B that is higher than the threshold A, and identifies at least one of a contamination range assumed to be affected by the detected cyber-attack and an intrusion route of the detected cyber-attack based on relationship data finally obtained by repeating the process of adding relationships by the relationship construction unit; An attack analysis device comprising:
2. the relationship construction unit adds direction information indicating a direction from the influencing object to the influenced object as a relationship between the influencing object and the influenced object, The identification unit identifies the contamination range by tracing forward a direction indicated by the direction information included in the finally obtained relationship data starting from an object in which a behavior indicating a cyber attack has been detected. The attack analysis apparatus of claim 1 .
3. the relationship construction unit adds direction information indicating a direction from the influencing object to the influenced object as a relationship between the influencing object and the influenced object, The identification unit identifies an intrusion route by tracing in a reverse direction the direction information included in the finally obtained relationship data, starting from an object in which a behavior indicating a cyber attack has been detected. The attack analysis apparatus of claim 1 .
4. The identification unit receives the operation data as an input, and when the attack certainty is equal to or greater than a threshold B that is higher than the threshold A, identifies the intrusion route as an object in which an operation indicating a detected cyber attack has been detected, the affected object. The attack analysis device according to claim 3 .
5. The relationship construction unit adds a relationship between the influencing object and the affected object in past data that is determined to have an attack probability higher than a standard based on a relationship with the newly acquired action data, among past data that is the action data having an attack probability lower than the threshold A. The attack analysis apparatus of claim 1 .
6. The relationship construction unit determines that the attack certainty of the past data is higher than a reference value when the attack certainty in the newly acquired action data is equal to or higher than the threshold value A and a difference between an action time of the newly acquired action data and an action time of the past data is within a reference time. The attack analysis device according to claim 5 .
7. The relationship construction unit determines that the attack certainty of the past data is higher than a standard when the attack certainty in the newly acquired action data is equal to or higher than the threshold A and at least one of the influencing object and the affected object matches between the newly acquired action data and the past data. The attack analysis device according to claim 5 .
8. The action data includes action information indicating action content, The attack analysis device further comprises: an attack probability calculation unit that calculates the attack probability from the action content indicated by the action information The attack analysis device of claim 1 .
9. A data acquisition unit of an attack analysis device sequentially acquires operation data including at least an affected object and an attack probability indicating the probability that the operation is a cyber attack against the target system, out of an influencing object, which is an object of a target system that performs an operation that affects other objects, an affected object, which is an object of the target system that is affected by the operation performed by the influencing object, and an attack probability indicating the probability that the operation is a cyber attack against the target system; a relationship construction unit of the attack analysis device repeatedly adds a relationship between the influencing object and the affected object to the relationship data when the action data includes the influencing object and the attack certainty is equal to or greater than a threshold A, using the data in an initial state as initial relationship data; An attack analysis method in which an identification unit of the attack analysis device inputs the operational data, and if the attack certainty is equal to or greater than a threshold B that is higher than the threshold A, determines that a cyber-attack on the target system has been detected, and identifies at least one of the contamination range assumed to be affected by the detected cyber-attack and the intrusion route of the detected cyber-attack based on the relationship data finally obtained by repeatedly adding relationships.
10. a data acquisition process for sequentially acquiring operation data including at least an affected object and an attack probability among an influencing object which is an object of a target system that has performed an operation that affects other objects, an affected object which is an object of the target system that is affected by the operation performed by the influencing object, and an attack probability that indicates the probability that the operation is a cyber attack against the target system; a relationship construction process for repeatedly adding a relationship between the influencing object and the affected object to the relationship data when the action data acquired by the data acquisition process includes the influencing object and the attack likelihood is equal to or greater than a threshold A, using data in an initial state as initial relationship data; an identification process in which, using the operation data as an input, if the attack certainty is equal to or higher than a threshold B that is higher than the threshold A, it is determined that a cyber-attack has been detected against the target system, and, based on relationship data finally obtained by repeating the process of adding relationships through the relationship construction process, it is determined that at least one of a contamination range assumed to be affected by the detected cyber-attack and an intrusion route of the detected cyber-attack is identified; An attack analysis program that causes a computer to function as an attack analysis device that performs the above steps.
Citation Information
Patent Citations
Range of influence setting device for program and range of influence specification method using the same and extraction method for program with ripple of influence
JP2012212281A
Change influence analyzer
JP2020021427A
Information security support system and information security support method
JP2021099589A
Contamination range identification device and contamination range identification program
JP6987332B1
Network attack detection method and device
US20180212986A1
