Attack correlation detection method based on multi-step attack and related equipment
By obtaining and dividing attack alarm data for multi-step attack scenarios, forming and correcting attack chains, the problem of incomplete multi-step attack association detection in the existing technology is solved, and more accurate and comprehensive attack chain construction is achieved, and the network is improved. Safe detection capability.
Patent Information
- Application Number
- CN202310696059.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-06-12
- Publication Date
- 2024-11-12
- Estimated Expiration
- 2043-06-12
AI Technical Summary
The existing multi-step attack detection and analysis technology is limited to the use of some scenarios. The multi-step attack association detection is not complete and accurate enough, affecting the attack investigation and defense of security administrators.
By obtaining attack alarm data for multi-step attack scenarios, data with the same source IP address and destination IP address are divided into the same group to form an attack chain, and the association relationship is corrected through deep priority traversal and dynamic sliding window strategies to build a complete attack chain.
It realizes comprehensive detection of multi-step attack association relationships, improves the accuracy and completeness of detection, can correctly build attack chains in the network, and enhances network security maintenance capabilities.
Smart Images

Figure CN116781340B_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of multi-step attack detection in a network, and in particular to an attack correlation relationship detection method based on multi-step attacks and related equipment. Background Art
[0002] Existing multi-step attack detection and analysis technologies are mostly limited to certain scenarios, and the detection of multi-step attack correlation relationships is not complete and accurate enough. Incomplete or incorrectly dependent attack correlation relationships will affect security administrators' correct attack investigation and defense. Summary of the invention
[0003] In view of this, the purpose of this application is to propose an attack correlation relationship detection method based on multi-step attacks and related equipment.
[0004] Based on the above objectives, the present application provides an attack correlation relationship detection method based on multi-step attacks, including:
[0005] Obtain attack alarm data for multi-step attack scenarios;
[0006] The attack alarm data with the same source IP address and the same destination IP address are divided into the same group to obtain multiple groups of attack alarm data; wherein each group of attack alarm data corresponds to a group of devices;
[0007] Associate the attack alarm data between the devices in the same group to form an attack chain between the devices in the same group; wherein each group of devices corresponds to at least one attack chain; take the device node without a parent or attack association as the starting node, traverse and search for the associated group whose source IP address is the same as the destination IP address of the current group device, and associate the attack chain of the current group with the attack chain of the associated group to form a spliced attack chain between the devices in the groups; wherein the spliced attack chain is set to multiple, and the spliced attack chain between each group of devices is set to at least one.
[0008] In some embodiments, associating attack alarm data between devices in the same group to form an attack chain between devices in the same group includes:
[0009] For the attack alarm data between the devices in the same group, the attack alarm data of the same type are represented by a group of alarm labels;
[0010] Cluster the attack alarm data of the same type to obtain the alarm labels of the attack alarm data of each cluster, and determine the earliest alarm occurrence time in each cluster as the occurrence time of the corresponding type of attack event, so as to obtain all occurrence times of the same type of attack events between each device in the same group;
[0011] Construct the initial attack chain of attack events of devices in the same group based on the time sequence relationship;
[0012] The intrusion logic of the initial attack chain is corrected based on the dynamic sliding window to remove the attack events in the initial attack chain that have an erroneous dependency relationship with the attack events at a previous moment and the attack events at a next moment, so as to obtain the attack chain between the devices in the group.
[0013] In some embodiments, the intrusion logic of modifying the initial attack chain based on the dynamic sliding window includes:
[0014] In response to determining that the occurrence time of the first attack event outside the sliding window is later than the occurrence time of the last attack event within the sliding window, comparing the chronological relationship of the attack occurrence phases of the first attack event outside the sliding window and all the attack events within the sliding window;
[0015] In response to determining that the attack occurrence phases of all attack events within the sliding window are no earlier than the attack occurrence phase of the first attack event outside the sliding window, the first attack event outside the sliding window is removed; or in response to determining that there is an attack event among all the attack events within the sliding window whose attack occurrence phase is no earlier than the attack occurrence phase of the first attack event outside the sliding window, the attack event is removed; and the size of the sliding window is reduced, and the reduced sliding window is moved forward by one attack phase unit along the progression direction of the attack time.
[0016] In some embodiments, the intrusion logic of modifying the initial attack chain based on the dynamic sliding window includes:
[0017] In response to determining that the occurrence time of the first attack event outside the sliding window is the same as the occurrence time of the last attack event within the sliding window, the sliding window is moved forward by one attack stage unit along the progression direction of the attack time.
[0018] In some embodiments, the association group includes multiple attack chains, and associating the attack chain of the current group with the attack chain of the association group to form a spliced attack chain between devices in the groups includes:
[0019] Remove the attack chains in the associated group whose attack events have an occurrence time and an occurrence stage no earlier than those in the current group, and associate at least one remaining attack chain with the attack chain of the current group;
[0020] Based on the occurrence time and occurrence stage of the attack event, the first position that meets the timing relationship and the intrusion logic in the associated attack chain is found, and the associated attack chain is spliced based on the first position to obtain a spliced attack chain between devices in the groups.
[0021] In some of the embodiments, the method further includes: extending the splicing attack chain or deleting an invalid splicing attack chain based on the circular association.
[0022] In some embodiments, the extending the spliced attack chain based on the circular association includes: for an inter-group device whose source IP address of the current group is the same as the destination IP address of the associated group, and whose destination IP address of the current group is the same as the source IP address of the associated group, searching for a second position in the spliced attack chain and the attack chain before splicing that satisfies the timing relationship and the intrusion logic, and splicing the attack chain before splicing with the spliced attack chain based on the second position;
[0023] The number of attack stages in the invalid splicing attack chain is equal to the number of attack stages in the attack chain before splicing, and the highest attack stage in the invalid splicing attack chain is lower than the highest attack stage in the attack chain before splicing; or the highest attack stage in the invalid splicing attack chain is equal to the highest attack stage in the attack chain before splicing, and the number of attack stages in the invalid splicing attack chain is less than the number of attack stages in the attack chain before splicing.
[0024] The embodiment of the present application also provides an attack correlation relationship detection device based on multi-step attacks, including:
[0025] An acquisition module is used to obtain attack alarm data of multi-step attack scenarios;
[0026] A grouping module, used to group the attack alarm data with the same source IP address and the same destination IP address into the same group, so as to obtain multiple groups of attack alarm data; wherein each group of attack alarm data corresponds to a group of devices;
[0027] An association module associates the attack alarm data between the devices in the same group to form an attack chain between the devices in the same group; wherein each group of devices corresponds to at least one attack chain; taking a device node without a parent or an attack association as a starting node, traversing to find an association group whose source IP address is the same as the destination IP address of the current group device, and associating the attack chain of the current group with the attack chain of the associated group to form a spliced attack chain between devices between groups; wherein the spliced attack chain is set to multiple, and the spliced attack chain between each group of devices is set to at least one.
[0028] An embodiment of the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements any of the methods described above when executing the program.
[0029] An embodiment of the present application also provides a non-transitory computer-readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to execute any of the methods described above.
[0030] An embodiment of the present application further provides a computer program product, comprising computer program instructions, which, when executed on a computer, enable the computer to execute any of the methods described above.
[0031] As can be seen from the above, the attack association relationship detection method based on multi-step attack provided by the present application obtains attack alarm data of multi-step attack scenarios; divides attack alarm data with the same source IP address and the same destination IP address into the same group to obtain multiple groups of attack alarm data; associates the attack alarm data between devices in the same group to form an attack chain between devices in the group; takes the device node with no parent or no attack association as the starting node, and performs depth-first traversal to find the association group with the same source IP address as the destination IP address of the current group device, and associates the attack chain of the current group with the attack chain of the association group to form a spliced attack chain between devices between groups; can comprehensively consider the attack association between devices in the same group and the attack association between devices in different groups, dig out a relatively complete attack chain, and correct the erroneous intrusion logic, and construct a complete and correct attack association relationship of multi-step attacks. BRIEF DESCRIPTION OF THE DRAWINGS
[0032] In order to more clearly illustrate the technical solutions in the present application or related technologies, the drawings required for use in the embodiments or related technical descriptions are briefly introduced below. Obviously, the drawings described below are only embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.
[0033] Figure 1 A flowchart of an attack correlation relationship detection method based on multi-step attacks according to an embodiment of the present application;
[0034] Figure 2 A schematic diagram of an attack correlation relationship detection method based on multi-step attacks according to an embodiment of the present application;
[0035] Figure 3 A schematic diagram of a process of forming an attack chain between devices in a group according to an embodiment of the present application;
[0036] Figure 4 A schematic diagram of a sliding window strategy according to an embodiment of the present application;
[0037] Figure 5 A schematic diagram of inter-group broken links and associations in an embodiment of the present application;
[0038] Figure 6 The letters in the embodiment of the present application represent corresponding alarm information situations;
[0039] Figure 7A schematic diagram of actual alarm details of an attack scenario in an embodiment of the present application;
[0040] Figure 8 A schematic diagram of the relationship between the actual alarm details of the current part of the embodiment of the present application;
[0041] Fig. 9 This is a schematic diagram of the attack association relationship that only considers the association between different groups of devices and devices in the embodiment of the present application;
[0042] Fig.10 This is a schematic diagram of an attack association relationship that only considers associations between devices in the same group in an embodiment of the present application;
[0043] Fig.11 A schematic diagram of the attack association relationship in an embodiment of the present application;
[0044] Fig.12 This is a schematic diagram of attack associations extracted from the LLDOS1.0 data set according to an embodiment of the present application;
[0045] Fig.13 Schematic diagram of an attack correlation relationship detection device based on multi-step attacks according to an embodiment of the present application;
[0046] Fig.14 A schematic diagram of the hardware structure of the electronic device provided in this embodiment is shown. DETAILED DESCRIPTION
[0047] In order to make the objectives, technical solutions and advantages of the present application more clearly understood, the present application is further described in detail below in combination with specific embodiments and with reference to the accompanying drawings.
[0048] It should be noted that, unless otherwise defined, the technical terms or scientific terms used in the embodiments of the present application should be understood by people with ordinary skills in the field to which the present application belongs. The "first", "second" and similar words used in the embodiments of the present application do not represent any order, quantity or importance, but are only used to distinguish different components. "Including" or "comprising" and similar words mean that the elements or objects appearing in front of the word cover the elements or objects listed after the word and their equivalents, without excluding other elements or objects. "Connect" or "connected" and similar words are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect.
[0049] In the related art, a method for attack association in power grid scenarios is based on the assumption that only one type of continuous and identical attack occurs between devices in the same group, only considering the attack association relationship between devices, and automatically extracting multi-step attack events based on a small amount of prior knowledge. By analyzing the alarm data of the power monitoring system, the attack tree is constructed using the correlation of the device IP, and the attack tree is converted into an attack chain after aggregation processing; finally, the attack chain is pruned and denoised through negative causal associations and non-secondary events to form the final attack chain and visual attack graph. In an IDS alarm association analysis based on intrusion behavior, based on the alarm information of the same group of IPs, only the attack association relationship between devices in the same group is mined. This method aggregates the attack actions between the same source IP and the destination IP according to the destination port and attack type, and divides the attack session according to the time constraint between the actions. A pruning algorithm that iteratively deletes adjacent and repeated behavior connections is used to reduce the impact of redundancy and noise, and finally the pruned attack sessions are merged into an attack association graph. In a similarity-based alarm association, alarms are associated based on attribute similarity. The attribute characteristics in the alarm traffic are analyzed by a similarity method to define the association relationship between alarms, and the association relationship can be extracted without using predefined knowledge and prerequisites.
[0050] In the above-mentioned attack association method for power grid scenarios, the attack association relationship between devices in the same group is not considered, and it is limited to some attack scenarios that meet the assumption, so the applicability is not high. Moreover, this method only extracts multi-step attack events between devices, and cannot detect the complete attack association relationship in the network, and cannot effectively evaluate the multi-step attack situation of the entire network, which is not conducive to the maintenance of network security in the later stage. In the IDS alarm association analysis based on intrusion behavior, the association relationship between the multi-step attack steps occurring on multiple devices is ignored. This method is not satisfied with the attack scenario where the attack steps are distributed among multiple devices, the detected attack association relationship is incomplete, and the accuracy of attack prediction is low. In the alarm association based on similarity, although the detected alarm association conforms to the attack occurrence stages in the actual attack scenario, it is only a simple operation at the attribute level, ignoring the intrusion logic between alarms, and the association result is different from the actual situation.
[0051] Since multi-step attacks are multiple steps between one or more attackers, the above-mentioned technologies that only consider the attack correlation between devices and ignore the attack correlation between devices in the same group, and only consider the attack correlation between devices in the same group and ignore the attack correlation between devices are not comprehensive enough. In addition, due to the disorder and irrelevant behavior of the attacks, the attack correlation restored through alarm correlation is prone to ignore the intrusion logical relationship, and the alarm correlation is not accurate enough. Therefore, there are problems in the relevant technologies that the applicable scenarios are limited and the constructed attack correlation is incomplete and incorrect.
[0052] Based on this, the embodiment of the present application provides an attack event and association relationship detection solution applicable to all attack scenarios. By correcting the erroneous intrusion logic in the association relationship, a relatively complete and correct attack chain in the current network is excavated, and a complete and correct attack association relationship of multi-step attacks is constructed. This can, to a certain extent, solve the problems of the limitations of the above-mentioned applicable scenarios and the incomplete and incorrect constructed attack associations.
[0053] Figure 1 A flowchart of an exemplary attack correlation relationship detection method based on multi-step attacks according to an embodiment of the present application is shown.
[0054] like Figure 1 As shown, the attack correlation relationship detection method based on multi-step attacks may include:
[0055] Step S100, acquiring attack alarm data of a multi-step attack scenario;
[0056] Step S200, classifying the attack alarm data with the same source IP address and the same destination IP address into the same group, to obtain multiple groups of attack alarm data; wherein each group of attack alarm data corresponds to a group of devices;
[0057] Step S300, associating the attack alarm data between the devices in the same group, forming an attack chain between the devices in the same group, and obtaining attack chains of multiple devices in the group; wherein each group of devices corresponds to at least one attack chain; taking the device node without a parent or without an attack association as the starting node, depth-first traversal is performed to find the associated group whose source IP address is the same as the destination IP address of the current group device, and associating the attack chain of the current group with the attack chain of the associated group to form a spliced attack chain between devices between groups; wherein the spliced attack chain is set to multiple, and the spliced attack chain between each group of devices between groups is set to at least one.
[0058] The attack association relationship detection method based on multi-step attack provided in the embodiment of the present application obtains attack alarm data of multi-step attack scenarios; divides attack alarm data with the same source IP address and the same destination IP address into the same group to obtain multiple groups of attack alarm data; associates the attack alarm data between devices in the same group to form an attack chain between devices in the group; takes a device node with no parent or no attack association as the starting node, performs depth-first traversal to find an association group with the same source IP address as the destination IP address of a device in the current group, and associates the attack chain of the current group with the attack chain of the association group to form a spliced attack chain between devices in the groups; can comprehensively consider the attack association between devices in the same group and the attack association between devices in different groups, dig out a relatively complete attack chain, correct the erroneous intrusion logic, and construct a complete and correct attack association relationship of multi-step attacks.
[0059] In some embodiments, see Figure 2 In step S100, the attack warning data of the multi-step attack scenario may be pre-processed attack warning data in a unified format. The acquisition of the attack warning data of the multi-step attack scenario may generally include:
[0060] Get the original attack alarm data of multi-step attack scenarios.
[0061] Extract the necessary field information of the original attack alarm data, and reconstruct the necessary field information to obtain the attack alarm data in a unified format. The necessary field information can be understood as the necessary field information for analyzing the attack event. For example, it can include the serial number, source IP address, destination IP address, protocol, alarm information, and alarm occurrence time of the data log.
[0062] Attack alarm data irrelevant to the multi-step attack scenario in the attack alarm data in the unified format is removed to obtain the attack alarm data of the multi-step attack scenario. The attack alarm data may be in the form of an alarm log. By removing irrelevant attack alarm data, noise data may be removed, the amount of data analyzed may be reduced, and the analysis efficiency may be improved. Among them, a One-Hot method may be used to represent the alarm record, and non-zero records may be retained, so that alarm records irrelevant to the multi-step attack scenario may be removed.
[0063] In some embodiments, continue to refer to Figure 2 In step S200, the attack alarm data is grouped according to the device IP, the same source IP address and the same destination IP address, and all devices can be grouped into groups.
[0064] In some embodiments, continue to refer to Figure 2 In step S300, the attack association is divided into two parts, including the attack association between devices in the same group and the attack association between devices in different groups (ie, devices between groups).
[0065] In some embodiments, Figure 3 As shown, it can be understood that, for the attack chain between the devices in the same group, the number of correct and complete attack chains between the devices in the same group includes at least one, for example, one or two or more. There is only one group of devices in the same group, and the attack chain of the devices in multiple groups is relative to the multiple groups of devices. That is, the attack chain of the devices in multiple groups is the attack chain of multiple groups of attack alarm data. The said associating the attack alarm data between the devices in the same group to form the attack chain between the devices in the group includes:
[0066] S311, for the attack alarm data between the devices in the same group, the same type of attack alarm data is represented by a group of alarm labels. The label information in a group of alarm labels may include the source IP address, the destination IP address, the protocol, the alarm type, and the earliest occurrence time of the type of attack. This is to clearly display the alarms occurring between the devices in each group at each attack stage.
[0067] S312, cluster the attack alarm data of the same type, obtain the alarm label of each cluster of attack alarm data, determine the earliest alarm occurrence time in each cluster as the occurrence time of the corresponding type of attack event, so as to obtain all the occurrence times of the same type of attack events between the devices in the same group. The clustering method can use K-means clustering or elbow rule to cluster the same type of alarms, extract the source IP address, destination IP address, protocol and alarm type of the cluster. And use the earliest alarm occurrence time in the cluster as the occurrence time of the attack event. The attack events between the devices in the same group can represent all the attacks that occurred successively between the devices in the group.
[0068] S313, constructing an initial attack chain of attack events of devices in the same group according to the time sequence relationship. The attack events between devices in the same group can be connected according to the time sequence relationship such as the sequence of attack events and the sequence of attack events to construct the initial attack chain between devices in the same group.
[0069] S314, based on the dynamic sliding window, the intrusion logic of the initial attack chain is corrected to remove the attack events in the initial attack chain that have an incorrect dependency relationship with the attack events at the previous moment and the attack events at the next moment, and obtain the attack chain between the devices in the group. The incorrect dependency relationship can be understood as the existence of an incorrect dependency relationship in the chronological stages of the attack events and the chronological time of the attack events. In this way, a correct and complete attack chain between the devices in the same group can be obtained.
[0070] In some embodiments, Figure 4 As shown. The sliding window has an initial size. During the correction process, the size of the sliding window can be adjusted or maintained unchanged according to the specific situation. The size of the sliding window can be adapted to the attack phase unit of the attack event, such as one attack phase unit or two attack phase units. Generally, the intrusion logic for correcting the initial attack chain based on the dynamic sliding window may include:
[0071] In the first case, in response to determining the time of occurrence of the first attack event outside the sliding window (e.g., time 3 ) and the time of occurrence of the last attack event in the sliding window (e.g. time 2 ) are the same (for example, time 3 =time2 ), the sliding window is moved forward (for example, to the right) by one attack stage unit along the process direction of the attack time.
[0072] In the second case, in response to determining the time of occurrence of the first attack event outside the sliding window (e.g., time 3 ) is later than the last attack event in the sliding window (e.g. time 2 ) occurs (e.g. time 3 >time 2 ), compare the first attack event outside the sliding window (e.g. Attack 3 ) and all attack events within the sliding window (e.g. Attack 1 and Attack 2 )’s attack phases:
[0073] In response to determining all attack events (e.g., Attack 1 and Attack 2 ) is no earlier than the first attack event outside the sliding window (e.g. Attack 3 ) of the attack phase (e.g. Attack 1 and Attack 2 ≥Attack 3 ), that is, all attack events within the sliding window (such as Attack 1 and Attack 2 ) occurs in the first attack event outside the sliding window (e.g. Attack 3 ) after the attack occurs, remove the first attack event (e.g. Attack 3 ).
[0074] In the third case, in response to determining the time of occurrence of the first attack event outside the sliding window (e.g., time 3 ) is later than the last attack event in the sliding window (e.g. time 2 ) occurs (e.g. time 3 >time 2 ), compare the first attack event outside the sliding window (e.g. Attack 3 ) and all attack events within the sliding window (e.g. Attack 1 and Attack 2 )’s attack phases:
[0075] In response to determining all attack events (e.g., Attack 1 and Attack 2 ) contains an attack event that occurs no earlier than the first attack event outside the sliding window (e.g., Attack 3 ) of the attack phase (e.g. Attack 1 ≥Attack 3 , Attack 2 <Attack 3 ), the attack event (e.g. Attack 1 ) is removed; and the size of the sliding window is reduced, and the reduced sliding window is moved forward by one attack stage unit along the process direction of the attack time.
[0076] In this way, based on the dynamic sliding window strategy, it is possible to determine whether the previous and next attacks in the attack chain meet the dependency relationship of the attack stages, remove redundant nodes, disconnect attack associations that do not meet the dependency relationship, and build a complete and correct attack chain.
[0077] In some embodiments, in the attack association between different groups of devices (i.e., inter-group devices), no parents means that among all devices, there is no device whose destination IP address is the same as the source IP address of the device node. No attack association means that among all attack events, there is no attack phase and attack event that is earlier than the attack phase and attack event of the device node. Specifically, starting from the starting node, a depth-first traversal method can be used to find a device (i.e., a subsequent device) whose source IP address is the same as the destination IP address of the starting node, and the device is determined to be an associated device of the starting node device, and then the current group attack chain corresponding to the current group device is associated with the attack chain of the associated group corresponding to the associated device, and the attack association between these inter-group devices is deeply mined to form a spliced attack chain between the inter-group devices.
[0078] In some embodiments, the associated group includes multiple attack chains, and associating the attack chain of the current group with the attack chain of the associated group to form a spliced attack chain between devices in the groups includes:
[0079] Remove the attack chains in the associated group whose attack events have an occurrence time and an occurrence stage no earlier than those in the current group, and associate at least one remaining attack chain with the attack chain of the current group;
[0080] Based on the time and stage of the attack event, find the first position in the associated attack chain that meets the timing relationship and intrusion logic, and splice the associated attack chain based on the first position to obtain the spliced attack chain between the devices in the group. When splicing, you can first break the chain and then splice it. Figure 5 As shown, it can be at the first position (for example, the alert in attack chain 2)2 With alert 4 Then, attack chain 1 is spliced with the broken attack chain 2. In this way, the attack associations between different groups can be linked together to construct the attack association relationship between the entire network.
[0081] In practical applications, the splicing attack chain between devices in different groups can be shown in Table 1.
[0082] Table 1 Attack associations between devices
[0083]
[0084] Taking the attack association between devices in Table 1 as an example, the source IP of the latter group (i.e., group 2) of devices is the same as the destination IP of the former group (i.e., group 1) of devices, and the two groups of devices are called associated devices. By analyzing the attack association relationship between the two groups of devices, it can be seen that the occurrence time of attack phase D in the second group of attack associations is later than the occurrence time of attack phase B, and attack phase D is located after attack phase B. The association relationship between attack phase D and the first attack chain (i.e., the attack association of group 1) satisfies the temporal relationship and the attack phase sequence. The second attack chain (i.e., the attack association of group 2) is broken and spliced with the first attack chain. The final spliced attack chain (i.e., spliced attack association) between different groups of devices is A2->A3->B->D.
[0085] At this point, the attack correlation relationship detection method based on multi-step attacks in the embodiment of the present application extracts attack events according to the clustering algorithm, and mines the correlation relationships within each group in combination with the timing relationship; then, the associated devices are found according to the IP association, and according to the timing relationship and the attack stage dependency, the appropriate location of the attack correlation relationship connection between the two groups of associated devices is found and associated, and the attack correlation between the entire network is completed through depth-first traversal iteration, and finally a relatively complete spliced attack chain is formed.
[0086] In some embodiments, it may also include extending the spliced attack chain or deleting an invalid spliced attack chain based on the cyclic association, that is, optimizing the attack association according to the cyclic association of the device and the validity of the device association to obtain the simplest and complete attack association relationship.
[0087] In some embodiments, the extended splicing attack chain based on cyclic association includes: for the inter-group devices whose source IP address of the current group is the same as the destination IP address of the associated group, and whose destination IP address of the current group is the same as the source IP address of the associated group, searching for the second position in the spliced attack chain and the attack chain before splicing that satisfies the timing relationship and intrusion logic, and splicing the attack chain before splicing with the spliced attack chain based on the second position. Among them, the inter-group devices whose source IP address of the current group is the same as the destination IP address of the associated group, and whose destination IP address of the current group is the same as the source IP address of the associated group can be referred to as ABA-type devices. In this way, the attack association between ABA-type associated devices can be optimized by cyclic association, the timing relationship and intrusion logic of the attack chain between such ABA-type devices can be cyclically judged, and the qualified attack events can be connected multiple times to improve the integrity of the attack chain in the attack scenario.
[0088] In practical applications, the splicing attack chain between devices in groups after circular association can be shown in Table 2.
[0089] Table 2 ABA type device attack association
[0090]
[0091] Taking a group of ABA type devices in Table 2 as an example, the attack association relationship constructed between the former devices (i.e. the former group of devices) is A2->A3->D1, and the attack association relationship constructed between the latter devices (i.e. the latter group of devices) is A1->A3->B->C. When the devices between the groups are associated, the attacks between the two groups of devices are unidirectionally associated by judging the timing relationship and the attack sequence, and the attack association (i.e. the spliced attack chain) A2->A3->B->C can be obtained. However, after analysis, it can be seen that the attack D1 between the former devices occurs later than the attack C between the latter devices, which satisfies the timing relationship and the attack sequence. Therefore, the actual intrusion logic should be A2->A3->B->C->D1. In this way, the spliced attack chain can be spliced at the second position, and the aforementioned spliced attack chain can be extended to obtain a complete attack chain after circular association.
[0092] In some embodiments, an invalid splicing attack chain may include a splicing attack chain in which the highest attack stage of the attack event after splicing is lower than the attack chain before splicing. It can be understood that the number of attack stages in the invalid splicing attack chain is equal to the number of attack stages in the attack chain before splicing, and the highest attack stage in the invalid splicing attack chain is lower than the highest attack stage in the attack chain before splicing. For example, two groups of attack association relationships (attack chains before splicing) are A->B->C->F and C->E, respectively, where the F attack occurs earlier than the E attack. After associating the two groups of attack chains according to the timing and attack sequence conditions, the resulting splicing attack chain is A->B->C->E. At this time, the highest attack stage detected in the splicing attack chain is reduced from the F stage before splicing to the E stage, and the number of attack stages is the same as the number of attack stages extracted before splicing. Therefore, after deleting this type of invalid splicing attack chain, the accuracy and completeness of the attack analysis can be increased, which is conducive to the administrator to correctly judge the severity and urgency of the attack.
[0093] In some embodiments, an invalid splicing attack chain may include an attack event after splicing in which the highest attack stage remains unchanged, but the attack stages included are reduced or the same. It can be understood that the highest attack stage in the invalid splicing attack chain is equal to the highest attack stage in the attack chain before splicing, and the number of attack stages in the invalid splicing attack chain is less than the number of attack stages in the attack chain before splicing. For example, two groups of attack association relationships (attack chains before splicing) are A->B->D->E->F, B->C->F, where the D attack occurs earlier than the C attack, then after associating them according to the timing and attack sequence conditions, the resulting splicing attack chain is A->B->C->F. At this time, in the splicing attack chain, the highest attack stage remains unchanged, and it is the F stage before and after splicing, but the attack chain before splicing contains 5 attack stages, and the splicing attack chain obtained after splicing only contains 4 attack stages, and the number of attack stages in the splicing attack chain is reduced. Therefore, after deleting this invalid splicing attack chain, the integrity of the attack association relationship in the attack chain can be avoided from decreasing.
[0094] In this way, the embodiment of the present application can optimize the spliced attack chain by deleting the spliced attack chain whose number of attacks has not changed and whose highest attack stage has decreased after the inter-group association, and the spliced attack chain whose highest attack stage has not changed and whose attack stage has decreased or remained the same after the inter-group association, filter out the inter-group associations that cannot improve the association effect, avoid invalid inter-group attack associations, avoid increasing the difficulty of later attack analysis, avoid reducing the integrity of the attack chain, and finally form a complete, simplest and meaningful attack chain.
[0095] The effects of the embodiments of the present application are further analyzed and explained below through application examples.
[0096] Application Examples
[0097] The present invention is further described in detail using DARPA2000LLDOS 1.0 as a data set. Letters are used to simply represent the alarm information in the data set, such as Figure 6 shown.
[0098] by Figure 7 Taking the actual attack data between devices shown in as an example, the attack association in the current attack scenario is analyzed. It can be seen that there are 8 different attack types and 5 different attack stages in the current attack scenario that conform to the actual multi-step attack sequence between the 3 associated devices, and the highest attack stage is E. The actual association relationship between the devices is as follows Figure 8 shown.
[0099] Fig. 9 The attack association relationship detected by the attack association method that only considers different groups of devices and devices. It can be seen that in the above-mentioned multi-step attack scenario, this method only detects 3 different attack types and 3 attack stages, and its attack coverage rate is only 60%. In addition, the attack in the above attack scenario only occurs between 3 devices, but the detection results of this method associate 4 devices, and there is 1 redundant associated device. The effective proportion of devices in this scenario is only 75%. The redundant devices will cause the associated devices to be inaccurate, which will affect the security personnel in the later stage to quickly and accurately locate the attacking devices, which is not conducive to the later maintenance.
[0100] Fig.10 The experimental results of the attack associations between three groups of devices in the actual data are shown when only the attack associations between devices in the same group are considered. It can be seen that the detection results of this scenario contain at most 6 different attacks and 3 attack stages, and its attack coverage rate is only 40%. In addition, only one attack association in this part of the detection results can correctly extract the highest attack stage E in this part, and its attack severity accuracy is only 33%.
[0101] Fig.11The attack association relationship detected by the analysis method of the embodiment of the present application can detect all corresponding attack types and all attack stages between the three associated devices, can correctly extract the highest attack stage E of the current scene, and the association relationship conforms to the actual intrusion logic order. In addition, the scene releases attacks back and forth twice between devices 172.16.112.10 and 202.77.162.213. The present application uses a cyclic connection optimization method to deeply mine the attack associations between devices in the same group, thereby ensuring the integrity of the attack associations in this case. It can overcome the problem that the above two association methods only associate one-way once, which will miss some attacks and the extracted attack associations are incomplete. It can also overcome the problem that the attack association method that only considers the attack association between devices in different groups is limited to the assumption that only continuous and identical attacks occur between devices in the same group, and only one attack is extracted between devices in the same group, ignoring the problem of attack associations between devices in the same group. It can also overcome the problem that the attack association extracted by the method that only considers the attack association between devices in the same group lacks the attack relationship between associated devices.
[0102] The embodiment of the present application extracts all attack associations in the DDOS LLDOS1.0 attack scenario as follows Fig.12 As shown, after calculation, the results obtained by this solution do not contain redundant device associations, and the attack coverage rate and attack severity accuracy can reach 100%. Comparing the detection results of this application with the attack association method that only considers the attack association between devices in the same group and the method that only considers the attack association between devices in different groups, it can be seen that the attack coverage rate of this application is increased by 13.28% and the average accuracy of attack severity is increased by 14.07% compared with the attack association method that only considers the attack association between devices in the same group; the attack coverage rate is increased by 19.4% compared with the method that only considers the attack association between devices in different groups, and the proportion of effective device associations is increased by 27.21%.
[0103] Therefore, the present application can comprehensively consider the association relationship between attack events of the same group of devices, different groups of devices and devices. First, the present application extracts attack events according to the clustering algorithm, and mines the association relationship within each group in combination with the time sequence relationship; and based on the dynamic sliding window strategy, it determines whether the previous and next attacks of the attack chain meet the attack stage dependency relationship, can remove redundant nodes, disconnect the attack association that does not meet the dependency relationship, and build a complete and correct attack chain. Then, according to the IP association, the associated devices are found, and according to the time sequence relationship and the attack stage dependency relationship, the appropriate position of the attack association relationship connection between the two groups of associated devices is found and associated, and the attack association between the entire network is completed through depth-first traversal iteration, and finally a relatively complete splicing attack chain is formed. Finally, through the optimization strategy of cyclic association and effective association, the association relationship between the missing cyclic associated devices is supplemented and the invalid redundant device association is pruned. Specifically, the cyclic association optimization strategy can be used to cyclically associate the attack association between the associated device groups whose source IP and destination IP of the former device correspond to the destination IP and source IP of the latter device respectively, to solve the problem of missing the attack association due to only associating once, and improve the integrity of the attack association. The effective association optimization strategy determines whether the associated devices improve the integrity and correctness of the attack chain, removes redundant device associations, and increases the effective proportion of devices, which helps administrators to accurately locate attack devices in the later stage.
[0104] It should be noted that the method of the embodiment of the present application can be performed by a single device, such as a computer or server. The method of this embodiment can also be applied to a distributed scenario and completed by multiple devices cooperating with each other. In the case of such a distributed scenario, one of the multiple devices can only perform one or more steps in the method of the embodiment of the present application, and the multiple devices will interact with each other to complete the described method.
[0105] It should be noted that the above describes some embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the above embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0106] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides an attack correlation relationship detection device based on multi-step attacks.
[0107] refer to Fig.13 , the attack correlation relationship detection device based on multi-step attack comprises:
[0108] An acquisition module 410, used to acquire attack alarm data of a multi-step attack scenario;
[0109] The grouping module 420 is used to group the attack alarm data with the same source IP address and the same destination IP address into the same group, so as to obtain multiple groups of attack alarm data; wherein each group of attack alarm data corresponds to a group of devices;
[0110] The association module 430 associates the attack alarm data between the devices in the same group to form an attack chain between the devices in the same group; wherein each group of devices corresponds to at least one attack chain; taking the device node without a parent or without an attack association as the starting node, traverses and searches for the association group whose source IP address is the same as the destination IP address of the current group device, and associates the attack chain of the current group with the attack chain of the associated group to form a spliced attack chain between the devices between the groups; wherein the spliced attack chain is set to multiple, and the spliced attack chain between each group of devices is set to at least one.
[0111] In some embodiments, associating attack alarm data between devices in the same group to form an attack chain between devices in the same group includes:
[0112] For the attack alarm data between the devices in the same group, the attack alarm data of the same type are represented by a group of alarm labels;
[0113] Cluster the attack alarm data of the same type to obtain the alarm labels of the attack alarm data of each cluster, and determine the earliest alarm occurrence time in each cluster as the occurrence time of the corresponding type of attack event, so as to obtain all occurrence times of the same type of attack events between each device in the same group;
[0114] Construct the initial attack chain of attack events of devices in the same group based on the time sequence relationship;
[0115] The intrusion logic of the initial attack chain is corrected based on the dynamic sliding window to remove the attack events in the initial attack chain that have an erroneous dependency relationship with the attack events at a previous moment and the attack events at a next moment, so as to obtain the attack chain between the devices in the group.
[0116] In some embodiments, the intrusion logic of modifying the initial attack chain based on the dynamic sliding window includes:
[0117] In response to determining that the occurrence time of the first attack event outside the sliding window is later than the occurrence time of the last attack event within the sliding window, comparing the chronological relationship of the attack occurrence phases of the first attack event outside the sliding window and all the attack events within the sliding window;
[0118] In response to determining that the attack occurrence phases of all attack events within the sliding window are no earlier than the attack occurrence phase of the first attack event outside the sliding window, the first attack event outside the sliding window is removed; or in response to determining that there is an attack event among all the attack events within the sliding window whose attack occurrence phase is no earlier than the attack occurrence phase of the first attack event outside the sliding window, the attack event is removed; and the size of the sliding window is reduced, and the reduced sliding window is moved forward by one attack phase unit along the progression direction of the attack time.
[0119] In some embodiments, the intrusion logic of modifying the initial attack chain based on the dynamic sliding window includes:
[0120] In response to determining that the occurrence time of the first attack event outside the sliding window is the same as the occurrence time of the last attack event within the sliding window, the sliding window is moved forward by one attack stage unit along the progression direction of the attack time.
[0121] In some embodiments, the association group includes multiple attack chains, and associating the attack chain of the current group with the attack chain of the association group to form a spliced attack chain between devices in the groups includes:
[0122] Remove the attack chains in the associated group whose attack events have an occurrence time and an occurrence stage no earlier than those in the current group, and associate at least one remaining attack chain with the attack chain of the current group;
[0123] Based on the occurrence time and occurrence stage of the attack event, the first position that meets the timing relationship and the intrusion logic in the associated attack chain is found, and the associated attack chain is spliced based on the first position to obtain a spliced attack chain between devices in the groups.
[0124] Some of the embodiments further include an optimization module for extending the splicing attack chain or deleting invalid splicing attack chains based on the cyclic association.
[0125] In some embodiments, the extending the spliced attack chain based on the circular association includes: for an inter-group device whose source IP address of the current group is the same as the destination IP address of the associated group, and whose destination IP address of the current group is the same as the source IP address of the associated group, searching for a second position in the spliced attack chain and the attack chain before splicing that satisfies the timing relationship and the intrusion logic, and splicing the attack chain before splicing with the spliced attack chain based on the second position;
[0126] The number of attack stages in the invalid splicing attack chain is equal to the number of attack stages in the attack chain before splicing, and the highest attack stage in the invalid splicing attack chain is lower than the highest attack stage in the attack chain before splicing; or the highest attack stage in the invalid splicing attack chain is equal to the highest attack stage in the attack chain before splicing, and the number of attack stages in the invalid splicing attack chain is less than the number of attack stages in the attack chain before splicing.
[0127] For the convenience of description, the above device is described in terms of functions divided into various modules. Of course, when implementing the present application, the functions of each module can be implemented in the same or multiple software and / or hardware.
[0128] The device of the above embodiment is used to implement the corresponding attack correlation relationship detection method based on multi-step attacks in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.
[0129] Based on the same inventive concept, corresponding to any of the above-mentioned embodiments and methods, the present application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the attack correlation relationship detection method based on multi-step attacks described in any of the above embodiments is implemented.
[0130] Fig.14 A more specific schematic diagram of the hardware structure of an electronic device provided in this embodiment is shown, and the device may include: a processor 1010, a memory 1020, an input / output interface 1030, a communication interface 1040, and a bus 1050. The processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040 are connected to each other through the bus 1050 in the device.
[0131] The processor 1010 can be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this specification.
[0132] The memory 1020 may be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 1020 may store an operating system and other application programs. When the technical solutions provided in the embodiments of this specification are implemented by software or firmware, the relevant program codes are stored in the memory 1020 and are called and executed by the processor 1010.
[0133] The input / output interface 1030 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.
[0134] The communication interface 1040 is used to connect a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).
[0135] The bus 1050 includes a path that transmits information between the various components of the device (eg, the processor 1010, the memory 1020, the input / output interface 1030, and the communication interface 1040).
[0136] It should be noted that, although the above device only shows the processor 1010, the memory 1020, the input / output interface 1030, the communication interface 1040 and the bus 1050, in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include the components necessary for implementing the embodiments of the present specification, and does not necessarily include all the components shown in the figure.
[0137] The electronic device of the above embodiment is used to implement the corresponding attack association relationship detection method based on multi-step attack in any of the above embodiments, and has the beneficial effects of the corresponding method embodiment, which will not be repeated here.
[0138] Based on the same inventive concept, corresponding to any of the above-mentioned embodiment methods, the present application also provides a non-transitory computer-readable storage medium, wherein the non-transitory computer-readable storage medium stores computer instructions, and the computer instructions are used to enable the computer to execute the attack association relationship detection method based on multi-step attacks as described in any of the above embodiments.
[0139] The computer-readable medium of this embodiment includes permanent and non-permanent, removable and non-removable media, and information storage can be implemented by any method or technology. Information can be computer-readable instructions, data structures, modules of programs, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, read-only compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device.
[0140] The computer instructions stored in the storage medium of the above embodiment are used to enable the computer to execute the attack association relationship detection method based on multi-step attacks as described in any of the above embodiments, and have the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0141] Based on the same inventive concept, corresponding to the attack correlation relationship detection method based on multi-step attacks described in any of the above embodiments, the present disclosure also provides a computer program product, which includes computer program instructions. In some embodiments, the computer program instructions can be executed by one or more processors of a computer so that the computer and / or the processor execute the attack correlation relationship detection method based on multi-step attacks. Corresponding to the execution subject corresponding to each step in each embodiment of the attack correlation relationship detection method based on multi-step attacks, the processor that executes the corresponding step can belong to the corresponding execution subject.
[0142] The computer program product of the above embodiment is used to enable the computer and / or the processor to execute the attack association relationship detection method based on multi-step attacks as described in any of the above embodiments, and has the beneficial effects of the corresponding method embodiments, which will not be repeated here.
[0143] Those skilled in the art should understand that the discussion of any of the above embodiments is merely illustrative and is not intended to imply that the scope of the present application (including the claims) is limited to these examples. In line with the concept of the present application, the technical features in the above embodiments or different embodiments may be combined, the steps may be implemented in any order, and there are many other variations of the different aspects of the embodiments of the present application as described above, which are not provided in detail for the sake of simplicity.
[0144] In addition, to simplify the description and discussion, and in order not to make the embodiments of the present application difficult to understand, the known power supply / ground connection with the integrated circuit (IC) chip and other components may or may not be shown in the provided drawings. In addition, the device can be shown in the form of a block diagram to avoid making the embodiments of the present application difficult to understand, and this also takes into account the fact that the details of the implementation of these block diagram devices are highly dependent on the platform to be implemented in the embodiments of the present application (that is, these details should be fully within the scope of understanding of those skilled in the art). In the case of elaborating specific details (e.g., circuits) to describe exemplary embodiments of the present application, it is obvious to those skilled in the art that the embodiments of the present application can be implemented without these specific details or when these specific details are changed. Therefore, these descriptions should be considered to be illustrative rather than restrictive.
[0145] Although the present application has been described in conjunction with specific embodiments of the present application, many replacements, modifications and variations of these embodiments will be apparent to those skilled in the art from the foregoing description. For example, other memory architectures (e.g., dynamic RAM (DRAM)) may use the embodiments discussed.
[0146] The embodiments of the present application are intended to cover all such substitutions, modifications and variations that fall within the broad scope of the appended claims. Therefore, any omissions, modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the embodiments of the present application should be included in the scope of protection of the present application.
Claims
1. A method for detecting attack correlation relationships based on multi-step attacks, characterized in that: include: Obtain attack alarm data for multi-step attack scenarios; The attack alarm data with the same source IP address and the same destination IP address are divided into the same group to obtain multiple groups of attack alarm data; wherein each group of attack alarm data corresponds to a group of devices; Correlate the attack alarm data between the devices in the same group to form an attack chain between the devices in the same group, and obtain attack chains of multiple devices in the group; wherein each group of devices corresponds to at least one attack chain; take the device node without a parent or attack association as the starting node, traverse and search for the association group whose source IP address is the same as the destination IP address of the current group device, and associate the attack chain of the current group with the attack chain of the association group to form a spliced attack chain between devices in the groups; wherein the spliced attack chain is set to multiple, and the spliced attack chain between each group of devices is set to at least one; The associating attack alarm data between devices in the same group to form an attack chain between the devices in the same group includes: For the attack alarm data between the devices in the same group, the attack alarm data of the same type are represented by a group of alarm labels; Cluster the attack alarm data of the same type to obtain the alarm labels of the attack alarm data of each cluster, and determine the earliest alarm occurrence time in each cluster as the occurrence time of the corresponding type of attack event, so as to obtain all occurrence times of the same type of attack events between each device in the same group; Construct the initial attack chain of attack events of devices in the same group based on the time sequence relationship; Correcting the intrusion logic of the initial attack chain based on a dynamic sliding window to remove attack events in the initial attack chain that have an erroneous dependency relationship with an attack event at a previous moment and an attack event at a next moment, thereby obtaining an attack chain between devices in the group; The associated group includes multiple attack chains, and associating the attack chain of the current group with the attack chain of the associated group to form a spliced attack chain between devices in the groups includes: Remove the attack chains in the associated group whose attack events have an occurrence time and an occurrence stage no earlier than those in the current group, and associate at least one remaining attack chain with the attack chain of the current group; Based on the occurrence time and occurrence stage of the attack event, the first position that meets the timing relationship and the intrusion logic in the associated attack chain is found, and the associated attack chain is spliced based on the first position to obtain a spliced attack chain between devices in the groups.
2. The attack correlation detection method based on multi-step attack according to claim 1 is characterized in that: The intrusion logic of correcting the initial attack chain based on the dynamic sliding window includes: In response to determining that the occurrence time of the first attack event outside the sliding window is later than the occurrence time of the last attack event within the sliding window, comparing the chronological relationship of the attack occurrence phases of the first attack event outside the sliding window and all the attack events within the sliding window; In response to determining that the attack occurrence phases of all attack events within the sliding window are no earlier than the attack occurrence phase of the first attack event outside the sliding window, the first attack event outside the sliding window is removed; or in response to determining that there is an attack event among all the attack events within the sliding window whose attack occurrence phase is no earlier than the attack occurrence phase of the first attack event outside the sliding window, the attack event is removed; and the size of the sliding window is reduced, and the reduced sliding window is moved forward by one attack phase unit along the progression direction of the attack time.
3. The attack correlation detection method based on multi-step attack according to claim 1 is characterized in that: The intrusion logic of correcting the initial attack chain based on the dynamic sliding window includes: In response to determining that the occurrence time of the first attack event outside the sliding window is the same as the occurrence time of the last attack event within the sliding window, the sliding window is moved forward by one attack stage unit along the progression direction of the attack time.
4. The attack correlation detection method based on multi-step attack according to claim 1 is characterized in that: Also includes: Extend the splicing attack chain or delete the invalid splicing attack chain based on the circular association.
5. The attack correlation detection method based on multi-step attack according to claim 4 is characterized in that: The method of extending the spliced attack chain based on the circular association includes: for the inter-group device whose source IP address of the current group is the same as the destination IP address of the associated group, and whose destination IP address of the current group is the same as the source IP address of the associated group, searching for a second position in the spliced attack chain and the attack chain before splicing that satisfies the timing relationship and the intrusion logic, and splicing the attack chain before splicing with the spliced attack chain based on the second position; The number of attack stages in the invalid splicing attack chain is equal to the number of attack stages in the attack chain before splicing, and the highest attack stage in the invalid splicing attack chain is lower than the highest attack stage in the attack chain before splicing; or the highest attack stage in the invalid splicing attack chain is equal to the highest attack stage in the attack chain before splicing, and the number of attack stages in the invalid splicing attack chain is less than the number of attack stages in the attack chain before splicing.
6. An attack correlation detection device based on multi-step attack, characterized in that: include: An acquisition module is used to obtain attack alarm data of multi-step attack scenarios; A grouping module, used to group the attack alarm data with the same source IP address and the same destination IP address into the same group, so as to obtain multiple groups of attack alarm data; wherein each group of attack alarm data corresponds to a group of devices; An association module associates the attack alarm data between the devices in the same group to form an attack chain between the devices in the same group; wherein each group of devices corresponds to at least one attack chain; a device node without a parent or an attack association is used as a starting node, and a traversal is performed to find an association group whose source IP address is the same as the destination IP address of the current group device, and the attack chain of the current group is associated with the attack chain of the associated group to form a spliced attack chain between the devices in the groups; wherein the spliced attack chain is set to be multiple, and the spliced attack chain between the devices in each group is set to be at least one; The associating attack alarm data between devices in the same group to form an attack chain between the devices in the same group includes: For the attack alarm data between the devices in the same group, the attack alarm data of the same type are represented by a group of alarm labels; Cluster the attack alarm data of the same type to obtain the alarm labels of the attack alarm data of each cluster, and determine the earliest alarm occurrence time in each cluster as the occurrence time of the corresponding type of attack event, so as to obtain all occurrence times of the same type of attack events between each device in the same group; Construct the initial attack chain of attack events of devices in the same group based on the time sequence relationship; Correcting the intrusion logic of the initial attack chain based on a dynamic sliding window to remove attack events in the initial attack chain that have an erroneous dependency relationship with an attack event at a previous moment and an attack event at a next moment, thereby obtaining an attack chain between devices in the group; The associated group includes multiple attack chains, and associating the attack chain of the current group with the attack chain of the associated group to form a spliced attack chain between devices in the groups includes: Remove the attack chains in the associated group whose attack events have an occurrence time and an occurrence stage no earlier than those in the current group, and associate at least one remaining attack chain with the attack chain of the current group; Based on the occurrence time and occurrence stage of the attack event, the first position that meets the timing relationship and the intrusion logic in the associated attack chain is found, and the associated attack chain is spliced based on the first position to obtain a spliced attack chain between devices in the groups.
7. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the method according to any one of claims 1 to 5 when executing the program.
8. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause a computer to execute the method according to any one of claims 1 to 5.
