Inference device, inference method, and program
The reasoning device enhances cyber attack attribution by hierarchically abstracting attack methods and calculating similarity degrees, addressing low estimation accuracy in existing technologies.
Patent Information
- Application Number
- JP2023539222
- Authority / Receiving Office
- JP · JP
- Patent Type
- Patents
- Current Assignee / Owner
- Filing Date
- 2021-08-02
- Publication Date
- 2024-11-19
- Estimated Expiration
- 2041-08-02
AI Technical Summary
Existing technologies for cyber attack attribution lack accuracy in estimating attackers due to limited access to network data and insufficient consideration of attacker characteristics, leading to low estimation accuracy.
A reasoning device and method that performs inference processing on observation logs to abstract attack methods hierarchically, using TTP frameworks and attacker characteristics, and calculates similarity degrees to rank potential attackers based on past attack cases.
Improves estimation accuracy of cyber attackers by providing a ranked assessment of potential attackers based on similarity to past attack characteristics, even in restricted network access scenarios.
Smart Images

Figure 0007586328000003 
Figure 0007586328000004 
Figure 0007586328000005
Abstract
Description
[Technical field]
[0001] The present disclosure relates to techniques for abductive reasoning. [Background technology]
[0002] Abductive reasoning is a method of deriving a valid hypothesis from inference knowledge (rules) given by logical expressions and observed events. For example, in the field of cybersecurity, abductive reasoning can be applied when determining whether an event observed in a computer system is the result of a cyber attack. For example, Patent Document 1 discloses a technology that reflects the temporal context of observations in hypothesis candidates generated by weighted abductive reasoning.
[0003] On the other hand, Patent Document 2 discloses a technology for calculating the similarity between multiple suspicious activity graphs, which are structures that represent the activities of targeted attacks and malware. Patent Document 2 also discloses a technology for associating a predicted attack source with a record of the analysis result of the suspicious activity graph. [Prior art documents] [Patent documents]
[0004] [Patent Document 1] International Publication No. WO2021 / 090497 [Patent Document 2] JP 2016-206943 A Summary of the Invention [Problem to be solved by the invention]
[0005] However, Patent Document 1 does not particularly disclose a viewpoint of performing inference processing using information related to an attacker who carries out a cyber attack. Therefore, the technology disclosed in Patent Document 1 has a problem in that the estimation accuracy of an attacker who carries out a cyber attack is low.
[0006] In addition, the technology disclosed in Patent Document 2 has the problem that in a situation where it is difficult to obtain a suspicious activity graph due to reasons such as restricted access to the network on which the suspicious activity detection device is installed, the accuracy of estimating the predicted source of an attack associated with the record of the analysis results of the suspicious activity graph is low.
[0007] One objective of the present disclosure is to improve the accuracy of estimating attackers who carry out cyber attacks. [Means for solving the problem]
[0008] In one aspect of the disclosure, a reasoning apparatus includes: an inference means for performing inference processing based on an observation log in which information related to a cyber attack is recorded and characteristics of an attacker inferred from past attack cases, thereby obtaining an overall picture of an attack by the attacker, in which the attack method is hierarchically abstracted on the assumption that the attacker has left information equivalent to the observation log; The attacker includes a similarity acquisition means for acquiring a similarity, which is a value indicating the degree of similarity between an overall picture of an attack by the attacker and characteristics of the attacker.
[0009] In another aspect of the present disclosure, a reasoning apparatus includes: an inference means for performing inference processing based on an observation log in which information related to a cyber attack is recorded, thereby acquiring an overall picture of the attack in which the attack method corresponding to the observation log is hierarchically abstracted; a similarity acquisition means for acquiring a similarity, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of an attacker inferred from past attack cases; a display screen generating means for generating a display screen for displaying a ranking in which the similarities are arranged in descending order when a plurality of similarities corresponding to the plurality of attackers are obtained; Equipped with.
[0010] In yet another aspect of the present disclosure, a method of inference includes: By performing inference processing based on the observation log in which information related to the cyber attack is recorded and the characteristics of the attacker inferred from past attack cases, an overall picture of the attack by the attacker is obtained in which the attack method is hierarchically abstracted assuming that the attacker has left information equivalent to the observation log; A similarity is obtained, which is a value indicating the degree of similarity between the overall picture of the attack by the attacker and the characteristics of the attacker.
[0011] In yet another aspect of the present disclosure, a method of inference includes: By performing inference processing based on the observation log in which information related to the cyber attack is recorded, an overall picture of the attack is obtained in which the attack method corresponding to the observation log is hierarchically abstracted; Obtain a similarity score, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of the attacker inferred from past attack cases. death, When a plurality of similarities corresponding to a plurality of attackers are obtained, a display screen is generated for displaying a ranking in which the similarities are arranged in descending order.
[0012] In yet another aspect of the present disclosure, program teeth, By performing inference processing based on the observation log in which information related to the cyber attack is recorded and the characteristics of the attacker inferred from past attack cases, an overall picture of the attack by the attacker is obtained in which the attack method is hierarchically abstracted assuming that the attacker has left information equivalent to the observation log; and causing a computer to execute a process of acquiring a similarity, which is a value indicating the degree of similarity between an overall picture of an attack by the attacker and characteristics of the attacker. 。
[0013] According to yet another aspect of the present disclosure, a recording medium includes: By performing inference processing based on the observation log in which information related to the cyber attack is recorded, an overall picture of the attack is obtained in which the attack method corresponding to the observation log is hierarchically abstracted; Obtain a similarity score, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of the attacker inferred from past attack cases. death, When a plurality of similarities corresponding to the plurality of attackers are obtained, a display screen is generated for displaying a ranking in which the similarities are arranged in descending order. A program that causes a computer to execute a process is recorded. Effect of the Invention
[0014] According to the present disclosure, it is possible to improve the estimation accuracy regarding the estimation of attackers who carry out cyber attacks. [Brief description of the drawings]
[0015] [Figure 1] FIG. 1 is a block diagram showing the hardware configuration of an inference device according to a first embodiment. [Diagram 2] FIG. 1 is a diagram showing the functional configuration of an inference device according to a first embodiment. [Diagram 3] 4 is a diagram showing an example of an observation log used in the inference process of the inference device according to the first embodiment. [Figure 4] 4 is a diagram showing an example of attacker characteristic information used in the inference process of the inference device according to the first embodiment. [Diagram 5] 4 is a diagram showing an example of attacker characteristic information used in the inference process of the inference device according to the first embodiment. [Figure 6] 3 is a diagram for explaining an example of an inference result obtained by the inference process of the inference device according to the first embodiment. [Figure 7] 3 is a diagram for explaining an example of an inference result obtained by the inference process of the inference device according to the first embodiment. [Figure 8] FIG. 2 is a diagram showing an example of attacker inference information generated by the inference device according to the first embodiment. [Figure 9] FIG. 4 is a diagram for explaining an example of a display screen that is displayed in response to processing by the inference device according to the first embodiment. [Figure 10] 4 is a flowchart for explaining processing performed in the inference device according to the first embodiment. [Figure 11] 10 is a flowchart for explaining processing performed in an inference device according to a modified example of the first embodiment. [Figure 12] FIG. 11 is a block diagram showing the functional configuration of an inference device according to a second embodiment. [Figure 13] 10 is a flowchart for explaining processing performed in an inference device according to a second embodiment. [Figure 14] FIG. 13 is a block diagram showing the functional configuration of an inference device according to a third embodiment. [Figure 15] 13 is a flowchart for explaining processing performed in an inference device according to a third embodiment. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0016] Hereinafter, preferred embodiments of the present disclosure will be described with reference to the drawings.
[0017] <First embodiment> [Hardware configuration] 1 is a block diagram showing a hardware configuration of inference device 100. As shown in the figure, inference device 100 includes an interface (IF) 11, a processor 12, a memory 13, a recording medium 14, a database (DB) 15, a display unit 16, and an input unit 17.
[0018] The IF 11 inputs and outputs data to and from an external device. Specifically, observed information used for inference is input via the IF 11. In addition, attacker inference information (described later) and the like obtained by the inference device 100 are output to the external device via the IF 11.
[0019] The processor 12 is a computer such as a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), and executes a prepared program to control the entire inference device 100. Specifically, the processor 12 executes processes such as inference processing, which will be described later.
[0020] The memory 13 is composed of a ROM (Read Only Memory), a RAM (Random Access Memory), etc. The memory 13 is also used as a working memory while the processor 12 is executing various processes.
[0021] The recording medium 14 is a non-volatile, non-transient recording medium such as a disk-shaped recording medium or a semiconductor memory, and is configured to be detachable from the inference device 100. The recording medium 14 records various programs executed by the processor 12. When the inference device 100 executes various processes, the programs recorded on the recording medium 14 are loaded into the memory 13 and executed by the processor 12.
[0022] The database 15 stores the observation information and the like input through the IF 11. The database 15 also stores the inference results and the like obtained by processing of the attacker estimation unit 22 described below. The database 15 also stores attacker characteristic information described below.
[0023] The display unit 16 is configured with a display device such as a liquid crystal monitor, etc. Furthermore, the display unit 16 displays information such as inference results as necessary.
[0024] The input unit 17 is composed of input devices such as a keyboard, a mouse, and a touch panel.
[0025] [Function configuration] FIG. 2 is a diagram illustrating the functional configuration of the inference device according to the first embodiment.
[0026] The inference device 100 includes an input receiving unit 21, an attacker inferring unit 22, and an information output unit .
[0027] The input receiving unit 21 receives an input of an observation log in which events actually observed in response to attacks on the network domain NDX are recorded.
[0028] Specifically, the aforementioned observation log includes, for example, information in which the contents of events that actually occurred in any of the computers belonging to the network domain NDX and the occurrence times (timestamps) of the events are arranged in chronological order. In other words, the aforementioned observation log records information related to cyber attacks.
[0029] The attacker estimation unit 22 has an inference unit 31, a similarity acquisition unit 32, and an attacker database 33. The attacker estimation unit 22 also performs processing to obtain attacker estimation information, which is information related to an attacker who is estimated to have attacked the network domain NDX, based on an observation log (hereinafter referred to as an observation log LX) received as input by the input receiving unit 21. The attacker estimation unit 22 also outputs the attacker estimation information obtained by the above-mentioned processing to the information output unit 23.
[0030] The inference unit 31 has an inference engine 31a and an inference rule storage unit 31b. The inference unit 31 performs an inference process based on the observation log LX and attacker characteristic information AFJ (described later) stored in the attacker database 33. The inference unit 31 outputs an inference result obtained by the inference process to the similarity acquisition unit 32.
[0031] The inference engine 31a performs inference processing using the inference rules stored in the inference rule storage unit 31b, thereby estimating an overall picture of attacks on the network domain NDX, assuming that an attacker indicated by the attacker characteristic information AFJ left information equivalent to the observation log LX when attacking the network domain NDX. The inference engine 31a also performs inference processing for each of the multiple attackers indicated by the attacker characteristic information AFJ. That is, the inference result obtained by the inference processing of this embodiment includes an overall picture of multiple attacks on the network domain NDX, assuming that each of the multiple attackers indicated by the attacker characteristic information AFJ left information equivalent to the observation log LX when attacking the network domain NDX independently.
[0032] The inference rule storage unit 31b stores inference rules used in the inference process of the inference engine 31a.
[0033] In addition, the inference engine 31a of this embodiment is assumed to use a technology that reflects the temporal context of observations in hypothesis candidates generated by weighted hypothetical inference, such as that disclosed in International Publication WO2021 / 090497.
[0034] In addition, the inference rule storage unit 31b of this embodiment is assumed to store inference rules that incorporate the perspective of TTP (Tactics, Techniques and Procedures) into the rules (inference knowledge) disclosed in International Publication WO2021 / 090497.
[0035] TTP is a framework that hierarchically abstracts (models) the methods of cyber attacks by attackers. Specifically, for example, in the TTP framework disclosed by MITRE (https: / / attack.mitre.org / ), the methods of cyber attacks by attackers are expressed at three levels of abstraction: "Tactics," "Techniques," and "Procedures." In the following, an example will be described in which the TTP framework disclosed by MITRE is incorporated into inference rules. In the following description, the TTP framework disclosed by MITRE will be abbreviated simply as TTP.
[0036] The similarity acquisition unit 32 acquires a similarity, which is a value indicating the degree of similarity between an overall picture of an attack when one of the multiple attackers indicated by the attacker characteristic information AFJ has attacked the network domain NDX, and the characteristics of the one attacker indicated by the attacker characteristic information AFJ, based on the inference result obtained by the inference unit 31 and the attacker characteristic information AFJ. In addition, the similarity acquisition unit 32 acquires multiple similarities corresponding to each of the multiple attackers indicated by the attacker characteristic information AFJ, generates attacker estimation information in which the acquired multiple similarities are arranged in descending order, and outputs the generated attacker estimation information to the information output unit 23.
[0037] The attacker database 33 stores attacker characteristic information AFJ that indicates the characteristics of each of a plurality of attackers.
[0038] Attacker characteristic information AFJZ, which represents the characteristics of attacker Z, includes information that associates abstraction information, which is information obtained by abstracting, based on TTP, the attack method inferred from the contents of public reports of past attack cases believed to have been carried out by said attacker Z, with an evaluation value, which is a value that evaluates the uniqueness of the attack method represented by said abstraction information. In other words, attacker characteristic information AFJ includes information similar to the attacker characteristic information AFJZ described above, which represents the characteristics of each of multiple attackers.
[0039] In this embodiment, for example, https: / / content.secureworks.com / ~ / media / Files / JP / Reports / Secureworks-Bronze-Butler-Report.ashx?modified=20180419151034 can be used as a public report when creating abstraction information based on TTP.
[0040] The information output unit 23 generates a display screen for displaying the attacker estimation information, etc., output from the attacker estimation unit 22, and outputs the generated display screen to a display device. That is, the information output unit 23 has a function as a display screen generating means. In addition, the information output unit 23 outputs data including the attacker estimation information, etc., output from the attacker estimation unit 22 to an external device.
[0041] [Specific example of processing performed by the inference device] Here, a specific example of the processing performed in the inference device 100 of this embodiment will be described. Fig. 3 is a diagram showing an example of an observation log used in the inference processing of the inference device according to the first embodiment.
[0042] (Entering observation information) The input receiving unit 21 receives an input of an observation log L1 as shown in FIG. 3 as an observation log that was actually observed in response to an attack on the network domain ND1.
[0043] The observation log L1 in Figure 3 records that a session hijacking from PC_123 to PC_456 occurred at time T11, that a reservation for a task requiring administrative privileges occurred at time T21, which was later than time T11, and that a large-scale compressed file was created at time T31, which was later than time T21.
[0044] (Inference processing) The inference unit 31 performs inference processing based on the observation log L1 and the attacker characteristic information AFJ stored in the attacker database 33.
[0045] 4 and 5 are diagrams showing an example of attacker characteristic information used in the inference process of the inference device according to the first embodiment.
[0046] The attacker characteristic information AFJ includes, for example, attacker characteristic information AFJA as shown in FIG. 4 and attacker characteristic information AFJB as shown in FIG.
[0047] The attacker characteristic information AFJA includes abstraction information and an evaluation value that represent the characteristics of the attacker A. In addition, the attacker characteristic information AFJB includes abstraction information and an evaluation value that represent the characteristics of the attacker B.
[0048] "Lateral Movement" in the abstraction information of the Attacker Characteristic Information AFJA corresponds to an abstract expression of the attack method inferred from the phrase "Perform lateral intrusion" contained in a public report of a past attack case believed to have been carried out by Attacker A. In addition, the aforementioned "Lateral Movement" corresponds to "Tactic," which is set as the highest level of abstraction in TTP.
[0049] "Impact" in the abstraction information of the Attacker Characteristic Information AFJA corresponds to an abstract expression of the attack method inferred from the phrase "attacking the control of the core system" contained in a public report of a past attack case believed to have been carried out by Attacker A. In addition, the aforementioned "Impact" corresponds to "Tactic," which is set as the highest level of abstraction in TTP.
[0050] "File Deletion" from the abstraction information of the Attacker Characteristic Information AFJA corresponds to an abstract expression of the attack method inferred from the phrase "We will delete the evidence after achieving the objective" contained in a public report of a past attack case believed to have been carried out by Attacker A. In addition, the aforementioned "File Deletion" corresponds to "Technique," which is set as the second highest level of abstraction in TTPs.
[0051] "Lateral Movement," one of the abstract information in the Attacker Characteristic Information AFJB, corresponds to an abstract representation of the attack method inferred from the phrase "spread the infection to other devices" contained in a public report of a past attack case thought to have been carried out by Attacker B.
[0052] "Privilege Escalation" from the abstract information of the Attacker Characteristic Information AFJB corresponds to an abstract expression of the attack method inferred from the phrase "stealing administrative privileges of the authentication server" contained in a public report of a past attack case believed to have been carried out by Attacker B. In addition, the aforementioned "Privilege Escalation" corresponds to "Tactic," which is set as the highest level of abstraction in TTP.
[0053] Among the abstraction information of the attacker characteristic information AFJB, "Data from Network Shared Drive ⇒ Data Compressed" corresponds to an abstract expression of the attack method inferred from the phrase "Information will be collected and compressed files will be generated" contained in a public report of a past attack case believed to have been carried out by attacker B. In addition, the aforementioned "Data from Network Shared Drive" and "Data Compressed" correspond to "Technique," which is set as the second highest level of abstraction in TTP.
[0054] In this embodiment, it is sufficient that abstraction information corresponding to the contents of one public report includes one or more pieces of information corresponding to any one of "Tactics," "Techniques," and "Procedures."
[0055] The evaluation values included in the attacker characteristic information AFJA and AFJB are values calculated using the following formula (1). Note that in the following formula (1), EV represents the evaluation value, AGN represents the total number of attackers included in the attacker characteristic information AFJ, ATN represents the number of attackers who have the same abstraction information in the attacker characteristic information AFJ, and ln represents the natural logarithm.
[0056]
number
[0057] 6 and 7 are diagrams for explaining an example of an inference result obtained by the inference process of the inference device according to the first embodiment.
[0058] The inference unit 31 performs inference processing based on the observation log L1 and the attacker characteristic information AFJA, thereby obtaining, for example, an inference result HRA as shown in FIG. 6 as an inference result when it is assumed that the attack on the network domain ND1 is by attacker A. Also, the inference unit 31 performs inference processing based on the observation log L1 and the attacker characteristic information AFJB, thereby obtaining, for example, an inference result HRB as shown in FIG. 7 as an inference result when it is assumed that the attack on the network domain ND1 is by attacker B. Also, the inference unit 31 outputs the inference results HRA and HRB obtained by the inference processing to the similarity acquisition unit 32.
[0059] The inference result HRA corresponds to an overall picture of an attack by attacker A, which is a hierarchical abstraction of the attack methods against network domain ND1, assuming that attacker A left information equivalent to observation log L1 when attacking network domain ND1.
[0060] The inference result HRB corresponds to an overall picture of an attack by attacker B, which is a hierarchical abstraction of the attack methods against network domain ND1, assuming that attacker A left information equivalent to observation log L1 when attacking network domain ND1.
[0061] The abstraction information ("Tactics" and "Techniques") included in each of the inference results HRA and HRB is appropriately selected by the inference engine 31a. The occurrence time in each of the inference results HRA and HRB is set as the time corresponding to the occurrence time of the observation log L1. The logs in each of the inference results HRA and HRB correspond to "Procedures" which is set to the third highest (lowest) abstraction level in the TTP. The abstraction information and logs shown in the column of time T22 in the inference result HRB in FIG. 7 correspond to the attack method added by the inference engine 31a as a hypothesis based on the observation log L1 and the attacker characteristic information AFJB. Time T22 corresponds to a time later than time T21 and earlier than time T31.
[0062] In this embodiment, the inference engine 31a appropriately selects what kind of abstraction information and logs to add as hypotheses. Also, in this embodiment, the inference engine 31a appropriately selects where to add the abstraction information and logs as hypotheses.
[0063] (Obtaining similarity) The similarity acquisition unit 32 acquires a similarity DSA, which is a value indicating the degree of similarity between the overall picture of the attack by attacker A and the characteristics of attacker A indicated by the attacker characteristic information AFJA, based on the inference result HRA obtained by the inference unit 31 and the attacker characteristic information AFJA. Also, the similarity acquisition unit 32 acquires a similarity DSB, which is a value indicating the degree of similarity between the overall picture of the attack by attacker B and the characteristics of attacker B indicated by the attacker characteristic information AFJB, based on the inference result HRB obtained by the inference unit 31 and the attacker characteristic information AFJB.
[0064] The similarities DSA and DSB can be obtained, for example, by calculating a cosine similarity based on the idea of Term Trequency-Inverse Document Frequency (TF-IDF).
[0065] Here, a specific method for obtaining the similarity DSB will be described.
[0066] The similarity acquisition unit 32 compares the abstraction information included in the attacker characteristic information AFJB with the abstraction information included in the inference result HRB to determine whether the attack method included in the attacker characteristic information AFJB appears in the attack method of the inference result HRB. Then, according to the processing of the similarity acquisition unit 32, it is determined that "Lateral Movement", "Privilege Escalation", and "Data from Network Shared Drive⇒Data Compressed" have appeared in the inference result HRB.
[0067] The similarity acquisition unit 32 obtains the calculation result (0, ln(2), ln(2)) by calculating the element-by-element product of the evaluation value (0, ln(2), ln(2)) contained in the attacker characteristic information AFJB and (1, 1, 1), in which the presence or absence of the occurrence specified as described above is written as 1 if it occurs and 0 if it does not occur.
[0068] The similarity acquisition unit 32 calculates the cosine similarity using the above-mentioned evaluation value (0, ln(2), ln(2)) and the above-mentioned calculation result (0, ln(2), ln(2)) as vectors, thereby acquiring the similarity DSB.
[0069] The cosine similarity can be calculated by the following formula (2). For example, in the following formula, by applying the evaluation value (0, ln(2), ln(2)) to vector X and the calculation result (0, ln(2), ln(2)) to vector Y, it is possible to obtain 1 as the value of the similarity DSB, which corresponds to cos(X,Y).
[0070]
number
[0071] In this embodiment, when cos(X, Y) has no solution due to the denominator value of the above formula (2) being 0, the similarity is set to 0. Therefore, the similarity acquisition unit 32 in this embodiment acquires 0 as the value of the similarity DSA.
[0072] (Display of estimated attacker information, etc.) The similarity acquisition unit 32 generates attacker estimation information ASJ in which the similarity DSA corresponding to the attacker A and the similarity DSB corresponding to the attacker B are arranged in descending order, and outputs the generated attacker estimation information ASJ to the information output unit 23. Fig. 8 is a diagram showing an example of the attacker estimation information generated by the inference device according to the first embodiment.
[0073] The attacker estimation information ASJ is generated, for example, as information as shown in Fig. 8. Specifically, the attacker estimation information ASJ is generated as a ranking table in which the attackers (their corresponding names) included in the attacker characteristic information AFJ are arranged in order of likelihood of having attacked the network domain ND1.
[0074] The information output unit 23 generates a display screen for displaying a ranking table included in the attacker estimation information ASJ output from the attacker estimation unit 22, and outputs the generated display screen to a display device. That is, when multiple similarities corresponding to multiple attackers are obtained, the information output unit 23 generates a display screen for displaying a ranking in which the similarities are arranged in descending order.
[0075] According to this embodiment, for example, when a plurality of past attack cases are stored in the attacker database 33, the similarity acquisition unit 32 may acquire an attack case similar to the characteristics of attacker A or B from among the plurality of past attack cases when generating the attacker estimation information ASJ, and add the acquired attack case to the above-mentioned ranking table. That is, according to this embodiment, the similarity acquisition unit 32 may acquire other past attack cases similar to the characteristics of attacker A or attacker B.
[0076] According to this embodiment, the similarity acquisition unit 32 may calculate a probability value PSA indicating the degree of possibility (reliability) that the attacker A actually performed an attack on the network domain ND1 and a probability value PSB indicating the degree of possibility (reliability) that the attacker B actually performed an attack on the network domain ND1 based on the similarities DSA and DSB. Specifically, the similarity acquisition unit 32 may calculate each probability value such that the sum of the probability values PSA and PSB is 100% by applying a softmax function to the similarities DSA and DSB. Also, according to this embodiment, the similarity acquisition unit 32 may add the probability values PSA and PSB to the ranking table described above when generating the attacker estimation information ASJ.
[0077] According to this embodiment, the attacker characteristic information AFJ may include, for example, an IoC (Indicator of Compromise) such as a hash value of a specific malware and an IP address used in a cyber attack. The IoC may be rephrased as trace information indicating a trace of an attack made by an attacker.
[0078] According to this embodiment, the inference unit 31 may perform an inference process using the attacker characteristic information AFJ and IoC. Also, according to this embodiment, when the similarity acquisition unit 32 generates the attacker estimation information ASJ, the IoC may be added to the ranking table described above.
[0079] According to this embodiment, the information output unit 23 may generate a display screen JDG as shown in Fig. 9 based on, for example, the inference results HRA and HRB, the attacker characteristic information AFJA and AFJB, and the probability values PSA and PSB. In other words, the information output unit 23 may generate a display screen for displaying an overall picture of an attack by an attacker. Fig. 9 is a diagram for explaining an example of a display screen displayed in response to the processing of the inference device according to the first embodiment.
[0080] In the left display area of the display screen JDG of Fig. 9, a selection field SLR in which either attacker A or attacker B can be selected, a selection frame SLW indicating that attacker B has been selected, a probability value PSB (73%), and a probability value PSA (27%) are displayed. In addition, in the selection field SLR, attacker B corresponding to the probability value PSB is displayed at the top, and attacker A corresponding to the probability value PSA is displayed below attacker B. In addition, in the right display area of the display screen JDG of Fig. 9, an overall picture ATB of the attack by attacker B selected in the selection field SLR is displayed.
[0081] According to the overall picture ATB, elements belonging to "Tactic" in FIG. 7 are connected by arrows pointing in the direction corresponding to the order of occurrence times in the figure. Also, according to the overall picture ATB, "Log", "Technique", and "Tactic" at one occurrence time in FIG. 7 are connected by arrows pointing in the direction corresponding to this order. Also, according to the overall picture ATB, an element added as a hypothesis in the inference result HRB (an element corresponding to the column of time T22 in FIG. 7) is displayed in a manner different from other elements. Also, according to the overall picture ATB, an attack method common to both the attacker characteristic information AFJB and the inference result HRB is highlighted.
[0082] The overall picture ATB may be displayed in color as long as it is possible to identify the elements added as hypotheses in the inference result HRB, and the attack methods common to both the attacker characteristic information AFJB and the inference result HRB. The overall picture ATB may also include information indicating IoC.
[0083] (Processing flow) FIG. 10 is a flowchart illustrating the processing performed in the inference device according to the first embodiment.
[0084] First, the input receiving unit 21 receives an input of an observation log L1 in which events and the like actually observed in response to attacks on the network domain ND1 are recorded (step S11).
[0085] Next, the inference unit 31 obtains inference results HRA and HRB by performing inference processing based on the observation log L1 input in step S11 and the attacker characteristic information AFJA and AFJB stored in the attacker database 33 (step S12).
[0086] Next, the similarity acquisition unit 32 acquires a similarity DSA based on the inference result HRA and the attacker characteristic information AFJA obtained in step S12 (step S13). Also, the similarity acquisition unit 32 acquires a similarity DSB based on the inference result HRB and the attacker characteristic information AFJB obtained in step S12 (step S13).
[0087] The similarity acquisition unit 32 generates attacker estimation information ASJ by arranging the similarities DSA and DSB obtained in step S13 in descending order, and outputs the generated attacker estimation information ASJ to the information output unit 23 (step S14). Then, in response to such processing by the similarity acquisition unit 32, a ranking table included in the attacker estimation information ASJ is displayed on the display device.
[0088] As described above, according to this embodiment, an inference process is performed based on an observation log in which information related to a cyber attack is recorded and the characteristics of an attacker inferred from past attack cases, and it is possible to estimate the degree to which the overall picture of the attack obtained by the inference process resembles the characteristics of the attacker. Also, as described above, according to this embodiment, even if a series of attack processes includes an attack method that does not leave clear traces, such as the creation of a large-scale compressed file, it is possible to obtain information that is useful for estimating the attacker who actually performed the series of attack processes. Therefore, according to this embodiment, it is possible to improve the estimation accuracy of the attacker who performs the cyber attack.
[0089] (Modification) Below, a modified example of the above embodiment will be described. In the following, for simplicity, a specific description of the parts to which the above-mentioned processes can be applied will be omitted as appropriate. Fig. 11 is a flowchart for explaining the process performed in an inference device according to a modified example of the first embodiment.
[0090] The inference unit 31 of this embodiment may perform an inference process for estimating an overall picture ATX of attacks against the network domain ND1 based on the observation log L1 received by the input receiving unit 21. The similarity acquisition unit 32 of this embodiment may acquire similarities DSA and DSB using the overall picture ATX and the attacker characteristic information AFJA and AFJB. A specific example of such a process will be described below.
[0091] First, the input receiving unit 21 receives an input of an observation log L1 in which events and the like actually observed in response to attacks on the network domain ND1 are recorded (step S21).
[0092] Next, the inference unit 31 performs inference processing based on the observation log L1 input in step S21, and obtains an inference result HRX corresponding to the overall attack picture ATX in which the attack method corresponding to the observation log L1 is hierarchically abstracted (step S22).
[0093] Here, the inference result HRX is obtained as one that does not include hypotheses and is as close as possible to the observation log L1. Therefore, in step S22, the inference result HRX similar to the inference result HRA in FIG.
[0094] Next, the similarity acquisition unit 32 acquires a similarity DSC, which is a value indicating the degree of similarity between the overall image ATX and the characteristics of the attacker A indicated by the attacker characteristic information AFJA, based on the inference result HRX obtained in step S22 and the attacker characteristic information AFJA (step S23). Also, the similarity acquisition unit 32 acquires a similarity DSD, which is a value indicating the degree of similarity between the overall image ATX and the characteristics of the attacker B indicated by the attacker characteristic information AFJB, based on the inference result HRX obtained in step S22 and the attacker characteristic information AFJB (step S23).
[0095] In step S23, the similarity can be obtained using the same method as described above. Therefore, in step S23, 0 is obtained as the value of the similarity DSC, and 1 / √2 is obtained as the value of the similarity DSD.
[0096] The similarity acquisition unit 32 generates attacker estimation information ASK by arranging the similarities DSC and DSD obtained in step S23 in descending order, and outputs the generated attacker estimation information ASK to the information output unit 23 (step S24). Then, in response to such processing by the similarity acquisition unit 32, a ranking table included in the attacker estimation information ASK is displayed on the display device. Note that the attacker estimation information ASK is generated as the same as the attacker estimation information ASJ exemplified in FIG. 8.
[0097] As described above, according to this modification, an inference process is performed based on an observation log in which information related to a cyber attack is recorded, and it is possible to estimate the degree to which the overall picture of the attack obtained by the inference process is similar to the characteristics of an attacker inferred from past attack cases. Also, as described above, according to this modification, even if a series of attack processes includes an attack method that does not leave clear traces, such as the creation of a large-scale compressed file, it is possible to obtain information that is useful for estimating the attacker who actually performed the series of attack processes. Therefore, according to this modification, it is possible to improve the estimation accuracy of the attacker who performs the cyber attack.
[0098] <Second embodiment> FIG. 12 is a block diagram showing the functional configuration of an inference device according to the second embodiment.
[0099] Inference device 100A according to this embodiment has the same hardware configuration as inference device 100. Inference device 100A also has inference means 41 and similarity obtaining means .
[0100] FIG. 13 is a flowchart illustrating the processing performed in the inference device according to the second embodiment.
[0101] The inference means 41 performs inference processing based on the observation log in which information related to the cyber attack is recorded and the characteristics of the attacker inferred from past attack cases, thereby obtaining an overall picture of the attack by the attacker, in which the attack methods are hierarchically abstracted assuming that the attacker has left information equivalent to the observation log (step S41).
[0102] The similarity obtaining means 42 obtains a similarity, which is a value indicating the degree of similarity between the overall picture of an attack by an attacker and the characteristics of the attacker (step S42).
[0103] According to this embodiment, it is possible to improve the estimation accuracy regarding the estimation of attackers who carry out cyber attacks.
[0104] <Third embodiment> FIG. 14 is a block diagram showing the functional configuration of an inference device according to the third embodiment.
[0105] An inference device 100B according to this embodiment has the same hardware configuration as that of inference device 100. Inference device 100B further includes inference means 51 and similarity obtaining means 52.
[0106] FIG. 15 is a flowchart illustrating processing performed in the inference device according to the third embodiment.
[0107] The inference means 51 performs inference processing based on an observation log in which information related to a cyber attack is recorded, thereby obtaining an overall picture of the attack in which the attack methods corresponding to the observation log are hierarchically abstracted (step S51).
[0108] The similarity obtaining means 52 obtains a similarity, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of an attacker inferred from past attack cases (step S52).
[0109] According to this embodiment, it is possible to improve the estimation accuracy regarding the estimation of attackers who carry out cyber attacks.
[0110] A part or all of the above-described embodiments can be described as, but is not limited to, the following supplementary notes.
[0111] (Appendix 1) an inference means for performing inference processing based on an observation log in which information related to a cyber attack is recorded and characteristics of an attacker inferred from past attack cases, thereby obtaining an overall picture of an attack by the attacker, in which the attack method is hierarchically abstracted on the assumption that the attacker has left information equivalent to the observation log; a similarity acquisition means for acquiring a similarity which is a value indicating the degree of similarity between an overall picture of an attack by the attacker and characteristics of the attacker; An inference device comprising:
[0112] (Appendix 2) an inference means for performing inference processing based on an observation log in which information related to a cyber attack is recorded, thereby acquiring an overall picture of the attack in which the attack method corresponding to the observation log is hierarchically abstracted; a similarity acquisition means for acquiring a similarity, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of an attacker inferred from past attack cases; An inference device comprising:
[0113] (Appendix 3) The inference device of appendix 1 or 2, wherein the similarity acquisition means further acquires other past attack cases that are similar to the characteristics of the attacker.
[0114] (Appendix 4) 3. The inference device of claim 1 or 2, wherein the similarity acquisition means calculates a probability value indicating the degree of possibility that the attacker actually carried out an attack based on the similarity.
[0115] (Appendix 5) 3. The inference device of claim 1 or 2, wherein the characteristics of the attacker include trace information indicating traces of an attack performed by the attacker.
[0116] (Appendix 6) 3. The inference device of claim 1 or 2, further comprising a display screen generation means for generating a display screen for displaying a ranking of each similarity in descending order when a plurality of similarities corresponding to each of a plurality of attackers are obtained.
[0117] (Appendix 7) 2. The inference device of claim 1, further comprising a display screen generating means for generating a display screen for displaying an overall picture of an attack by the attacker.
[0118] (Appendix 8) By performing inference processing based on the observation log in which information related to the cyber attack is recorded and the characteristics of the attacker inferred from past attack cases, an overall picture of the attack by the attacker is obtained in which the attack method is hierarchically abstracted assuming that the attacker has left information equivalent to the observation log; An inference method for obtaining a similarity, which is a value indicating the degree of similarity between an overall picture of an attack by the attacker and the characteristics of the attacker.
[0119] (Appendix 9) By performing inference processing based on the observation log in which information related to the cyber attack is recorded, an overall picture of the attack is obtained in which the attack method corresponding to the observation log is hierarchically abstracted; An inference method for obtaining a similarity, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of an attacker inferred from past attack cases.
[0120] (Appendix 10) By performing inference processing based on the observation log in which information related to the cyber attack is recorded and the characteristics of the attacker inferred from past attack cases, an overall picture of the attack by the attacker is obtained in which the attack method is hierarchically abstracted assuming that the attacker has left information equivalent to the observation log; A recording medium having a program recorded thereon that causes a computer to execute a process of obtaining a similarity, which is a value indicating the degree of similarity between the overall picture of an attack by the attacker and the characteristics of the attacker.
[0121] (Appendix 11) By performing inference processing based on the observation log in which information related to the cyber attack is recorded, an overall picture of the attack is obtained in which the attack method corresponding to the observation log is hierarchically abstracted; A recording medium having recorded thereon a program that causes a computer to execute a process of obtaining a similarity, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of the attacker inferred from past attack cases.
[0122] Although the present disclosure has been described above with reference to the embodiments and examples, the present disclosure is not limited to the above-mentioned embodiments and examples. Various modifications that can be understood by a person skilled in the art can be made to the configuration and details of the present disclosure within the scope of the present disclosure. [Explanation of symbols]
[0123] 12 processors 21 Input reception section 22 Attacker Estimation Department 23 Information output section 31 Reasoning part 32 Similarity acquisition unit 33 Attacker Database
Claims
1. an inference means for performing inference processing based on an observation log in which information related to a cyber attack is recorded and characteristics of an attacker inferred from past attack cases, thereby obtaining an overall picture of an attack by the attacker, in which the attack method is hierarchically abstracted on the assumption that the attacker has left information equivalent to the observation log; a similarity acquisition means for acquiring a similarity which is a value indicating the degree of similarity between an overall picture of an attack by the attacker and characteristics of the attacker; An inference device comprising:
2. an inference means for performing inference processing based on an observation log in which information related to a cyber attack is recorded, thereby acquiring an overall picture of the attack in which the attack method corresponding to the observation log is hierarchically abstracted; a similarity acquisition means for acquiring a similarity, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of an attacker inferred from past attack cases; a display screen generating means for generating a display screen for displaying a ranking in which the similarities are arranged in descending order when a plurality of similarities corresponding to the plurality of attackers are obtained; An inference device comprising:
3. The inference device according to claim 1 or 2, wherein the similarity acquisition means further acquires other past attack cases that are similar to the characteristics of the attacker.
4. 3. The inference device according to claim 1, wherein the similarity obtaining means calculates a probability value representing a degree of possibility that the attacker has actually carried out an attack, based on the similarity.
5. The inference device according to claim 1, further comprising a display screen generating means for generating a display screen for displaying a ranking of each similarity in descending order when a plurality of similarities corresponding to each of a plurality of attackers are obtained.
6. 2. The inference device according to claim 1, further comprising a display screen generating means for generating a display screen for displaying an overall picture of an attack by said attacker.
7. By performing inference processing based on the observation log in which information related to the cyber attack is recorded and the characteristics of the attacker inferred from past attack cases, an overall picture of the attack by the attacker is obtained in which the attack method is hierarchically abstracted assuming that the attacker has left information equivalent to the observation log; An inference method for obtaining a similarity, which is a value indicating the degree of similarity between an overall picture of an attack by the attacker and the characteristics of the attacker.
8. By performing inference processing based on the observation log in which information related to the cyber attack is recorded, an overall picture of the attack is obtained in which the attack method corresponding to the observation log is hierarchically abstracted; A similarity is obtained, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of the attacker inferred from past attack cases; and An inference method for generating a display screen for displaying a ranking of the similarities arranged in descending order when a plurality of similarities corresponding to each of a plurality of attackers are obtained.
9. By performing inference processing based on the observation log in which information related to the cyber attack is recorded and the characteristics of the attacker inferred from past attack cases, an overall picture of the attack by the attacker is obtained in which the attack method is hierarchically abstracted assuming that the attacker has left information equivalent to the observation log; A program that causes a computer to execute a process of acquiring a similarity, which is a value indicating the degree of similarity between an overall picture of an attack by the attacker and the characteristics of the attacker.
10. By performing inference processing based on the observation log in which information related to the cyber attack is recorded, an overall picture of the attack is obtained in which the attack method corresponding to the observation log is hierarchically abstracted; A similarity is obtained, which is a value indicating the degree of similarity between the overall picture of the attack and the characteristics of the attacker inferred from past attack cases; and A program that causes a computer to execute a process of generating a display screen for displaying a ranking of each similarity in descending order when a plurality of similarities corresponding to each of a plurality of attackers are obtained.
Citation Information
Patent Citations
Cyber attack analyzer and cyber attack analytic method
JP2016206943A
Management device and management method
JP2018005282A
Action plan estimation device, action plan estimation method, and computer-readable recording medium
WO2020161780A1
Inference-making device, inference-making method, and computer-readable recording medium
WO2021090497A1
