A method, apparatus, storage medium, and electronic device for handling network security incidents.
By converting cybersecurity incident records into security observation data and constructing object relationships and evidence graphs, the problem of inaccurate correlation of multi-source data is solved, closed-loop processing of cybersecurity incidents is achieved, and the accuracy and traceability of investigations are improved.
Patent Information
- Application Number
- CN202611543751.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-09-30
- Publication Date
- 2026-11-17
AI Technical Summary
In existing technologies, the investigation of multi-source cybersecurity incidents lacks a unified data association mechanism, which leads to inaccurate evidence association, unclear application status of evidence, and the possibility that security measures may change the status of the investigated object, affecting the accuracy and traceability of the investigation results.
The system converts network security event records from multiple security data sources into security observation data, constructs security objects and object relationships with object versions, generates evidence requirement items and builds an investigation hypothesis verification map, identifies evidence breakpoints through causal tracing, drives investigation tasks to obtain supplementary evidence, and coordinates the execution sequence of disposal actions to achieve closed-loop processing.
It enables multi-source data normalization, evidence sufficiency verification, breakpoint-based supplementary evidence, and versioned evidence rewriting for cybersecurity incidents, improving the accuracy and traceability of the investigation process and reducing the risk of erroneous merging of delayed evidence and damage to the investigation site by handling actions.
Smart Images

Figure CN123123003A_ABST
Abstract
Description
Technical Field
[0001] This specification relates to the field of computer technology, and in particular to a method, apparatus, storage medium, and electronic device for handling network security incidents. Background Technology
[0002] As cyberattacks increasingly exhibit multi-stage, cross-host, and cross-security data source characteristics, cybersecurity incident investigations typically require integrating event records from multiple security data sources, such as endpoint detection and response systems, network traffic analysis systems, security information and event management systems, identity authentication systems, and threat intelligence systems, in order to reconstruct the attack process and determine the scope of the incident's impact.
[0003] In related technologies, event records from different security data sources are typically processed separately by different investigative tools, lacking a unified data correlation mechanism between attack path analysis, evidence retrieval, endpoint forensics, and security actions. When existing evidence is insufficient, evidence from different sources conflicts, or the status of the investigated entity changes, investigators usually need to manually determine the missing evidence and perform further retrieval or forensic operations. This can easily lead to inaccurate correlations between supplementary evidence and existing investigation results, and unclear applicability of evidence. Furthermore, security actions may alter the status of the investigated entity or affect security evidence that has not yet been secured, thereby impacting the accuracy and traceability of cybersecurity incident investigation results. Summary of the Invention
[0004] This specification provides a method, apparatus, storage medium, and electronic device for handling network security incidents, the technical solution of which is as follows: Firstly, this specification provides a method for handling network security incidents, the method comprising: The network security event records from multiple security data sources are converted into security observation data, and security objects with object versions and object relationships between the security objects are constructed based on the security observation data. Evidence requirement items are generated based on the preset evidence roles corresponding to the survey hypothesis data, and a survey hypothesis verification map and evidence sufficiency matrix are constructed based on the evidence requirement items. Causal tracing is performed along the object relationship, and the satisfaction status of each evidence requirement item is determined based on the obtained tracing evidence. For the unmet evidence requirements corresponding to the detected source tracing breakpoints, an investigation task including the expected object version of the target security object is generated. The investigation task is executed to obtain supplementary security evidence. The applicable object version of the supplementary security evidence is determined based on the expected object version and the current object version of the target security object. The evidence sufficiency matrix and the investigation hypothesis verification map are updated based on the supplementary security evidence applicable to the current object version. Based on the evidence sufficiency matrix, the remaining evidence requirements are determined, and the investigation tasks corresponding to the remaining evidence requirements are determined from the generated investigation tasks. Based on the current attack path, affected objects and the remaining evidence requirements obtained by causal tracing, candidate actions are generated to bind the target security object version. According to the impact of the candidate actions on the state change of the unfixed evidence corresponding to the remaining evidence requirements, the execution sequence of the candidate actions and the investigation tasks to be executed is determined. When the current version of the target security object corresponding to the candidate action matches the version of the target security object bound to the candidate action, and the candidate action satisfies the execution sequence, the candidate action is executed through the security device, and investigation conclusion data is generated based on the action execution result, the evidence sufficiency matrix, and the investigation hypothesis verification map.
[0005] Secondly, this specification provides a network security incident handling device, the device comprising: The data processing module is used to convert network security event records from multiple security data sources into security observation data, and to construct security objects with object versions and object relationships between the security objects based on the security observation data. The source tracing investigation module is used to generate evidence requirement items based on the preset evidence roles corresponding to the investigation hypothesis data, and to construct an investigation hypothesis verification map and an evidence sufficiency matrix based on the evidence requirement items, to perform causal tracing along the object relationship, and to determine the satisfaction status of each of the evidence requirement items based on the obtained source tracing evidence. The source tracing investigation module is used to generate an investigation task including the expected object version of the target security object for the unmet evidence requirements corresponding to the detected source tracing breakpoints, execute the investigation task to obtain supplementary security evidence, determine the applicable object version of the supplementary security evidence based on the expected object version and the current object version of the target security object, and update the evidence sufficiency matrix and the investigation hypothesis verification map based on the supplementary security evidence applicable to the current object version. The source tracing and investigation module is used to determine the remaining supplementary evidence requirements based on the evidence sufficiency matrix, and to determine the investigation tasks to be executed corresponding to the remaining supplementary evidence requirements based on the investigation hypothesis verification map. Based on the current attack path, affected objects and the remaining supplementary evidence requirements obtained by the causal source tracing, candidate handling actions bound to the target security object version are generated. According to the impact of the candidate handling actions on the state change of the unfixed evidence corresponding to the remaining supplementary evidence requirements, the execution sequence of the candidate handling actions and the investigation tasks to be executed is determined. The conclusion generation module is used to execute the candidate disposal action through a security device when the current object version of the target security object corresponding to the candidate disposal action matches the target security object version bound to the candidate disposal action and the candidate disposal action satisfies the execution sequence, and to generate investigation conclusion data based on the action execution result, the evidence sufficiency matrix and the investigation hypothesis verification map.
[0006] Thirdly, this specification provides a computer storage medium storing at least one instruction adapted for loading by a processor and executing method steps of one or more embodiments of this specification.
[0007] Fourthly, this specification provides a computer program product storing at least one instruction adapted to be loaded by a processor and to execute the method steps of one or more embodiments of this specification.
[0008] Fifthly, this specification provides an electronic device that may include: a processor and a memory; wherein the memory stores a computer program adapted to be loaded by the processor and to execute the method steps of one or more embodiments of this specification.
[0009] The beneficial effects of the technical solutions provided in some embodiments of this specification include at least the following: In one or more embodiments of this specification, by uniformly converting multi-source network security event records into security observation data and constructing security objects and object relationships with object versions, further evidence requirement items, evidence sufficiency matrices, and investigation hypothesis verification maps are established around the investigation hypotheses. During causal tracing, evidence breakpoints are identified and investigation tasks are driven to obtain supplementary security evidence. Applicability control of supplementary security evidence is achieved by combining object version verification. The execution sequence of investigation tasks and candidate actions is coordinated based on the remaining evidence gaps and the impact of candidate actions on unfixed evidence. Simultaneously, the action results are written back to the investigation chain for cyclical updates. This addresses the problem of... This addresses the problems in related technologies, such as the fragmented correlation of multi-source security data, reliance on manual supplementation for missing evidence, mismatch between supplementary evidence and the changed state of security objects, and the potential for security actions to prematurely destroy unsecured evidence and affect investigation conclusions. It achieves a closed-loop process for cybersecurity incident investigations, encompassing multi-source data normalization, evidence sufficiency verification, targeted supplementation of evidence at breakpoints, versioned evidence rewriting, and security actions under evidence preservation constraints. This improves the accuracy of evidence correlation and applicability judgments during cybersecurity incident investigations, reduces the risk of erroneous merging of delayed evidence and damage to the investigation scene by handling actions, and enhances the continuity, traceability, and reliability of investigation conclusions. Attached Figure Description
[0010] To more clearly illustrate the technical solutions in this specification or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this specification. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0011] Figure 1 This is a flowchart illustrating a network security incident handling method provided in this manual; Figure 2 This is a structural diagram of a network security incident handling device provided in this specification; Figure 3 This is a schematic diagram of the structure of an electronic device provided in this specification. Detailed Implementation
[0012] The technical solutions in this specification will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this specification, and not all embodiments. Based on the embodiments in this specification, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this specification.
[0013] In this specification, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. In this specification, it should be noted that, unless otherwise expressly specified and limited, "comprising" and "including," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices. Those skilled in the art can understand the specific meaning of the above terms in this specification based on the specific circumstances. Furthermore, in this specification, unless otherwise stated, "multiple" means two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent: A alone, A and B simultaneously, or B alone. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship.
[0014] The present specification will now be described in detail with reference to specific embodiments.
[0015] Please see Figure 1This specification provides a flowchart illustrating a network security incident handling method according to one or more embodiments. This method can be implemented using a computer program and can run on a network security incident handling device based on the von Neumann architecture. The computer program can be integrated into an application or run as a standalone utility application. The network security incident handling device can be a service platform.
[0016] Specifically, the network security incident handling method includes: S102: Convert network security event records from multiple security data sources into security observation data, and construct security objects with object versions and object relationships between the security objects based on the security observation data; Multiple security data sources can be understood as one or more security data systems capable of generating, storing, or providing data related to cybersecurity incidents. Illustratively, multiple security data sources can refer to a service platform accessing several security data sources through a multi-source normalization module, including but not limited to one or more of the following: endpoint detection and response data sources, network traffic analysis data sources, security information and incident management data sources, identity authentication data sources, asset management data sources, email security data sources, threat intelligence data sources, and offline evidence data sources.
[0017] Network security incident logs refer to the raw records generated by various security data sources for assets, accounts, processes, files, network connections, or other network behaviors. For example, network security incident logs can be process creation records, file creation records, script execution records, or network connection records collected from the terminal side; they can also be domain name resolution records, network session records, or abnormal communication alarms collected by network security devices; and they can be account login records or remote authentication records generated by identity authentication systems. Because different security data sources may use different data structures, field naming methods, and time recording methods, directly investigating network security incident logs from different security data sources may result in the same security object being represented differently in different data sources, or a unified time and object association relationship not being established between different records. Therefore, this specification proposes first converting network security incident logs from different security data sources into a unified data representation format, namely, security observation data.
[0018] Security observation data (also known as Security Observation) can be understood as a structured representation of the facts of cybersecurity events actually observed from a specific security data source. Optionally, security observation data may include at least the source information of the corresponding cybersecurity event record, the event occurrence time, subject object references, object object references, operational information, and original evidence references. In some embodiments, security observation data may also include information such as platform access time, parsing version, data structure version, access level, data retention deadline, and data quality status.
[0019] The event occurrence time characterizes the actual time when the cybersecurity action occurred, while the platform access time characterizes the time when the corresponding cybersecurity event record was acquired or accessed by the cybersecurity event processing system. By recording the event occurrence time and the platform access time separately, late data, supplementary data, or offline archived data can be distinguished, avoiding time sequence discrepancies that can occur when reconstructing the attack process solely based on the order in which the data entered the platform.
[0020] Citing original evidence is used to establish a link between security observation data and original cybersecurity incident records. For example, citing original evidence can be used to locate corresponding data in original logs, network packets, file samples, or forensic images. Therefore, in subsequent causal attribution processes, it is not necessary to indiscriminately read all original security data; instead, the original security evidence to be verified can be located based on the security observation data corresponding to the candidate attribution path.
[0021] Furthermore, when multiple network security event records are generated from the same underlying source record through different detection rules, different parsing processes, or different upper-layer systems, the source dependencies between different security observation data can be determined, and security observation data that depend on the same underlying source record or the same parsing chain can be grouped into the same source group. The source group is used to subsequently determine whether security evidence has source independence, thereby avoiding the misinterpretation of multiple records generated from repeated parsing of the same underlying event as multiple independent pieces of evidence.
[0022] After converting cybersecurity incident logs into security observation data, the service platform can further identify the security objects involved in the cybersecurity incidents from the security observation data. Security objects can be understood as entities that require status recording, relationship analysis, or evidence verification during cybersecurity incident handling. For example, security objects can be categorized as: physical / network infrastructure objects (Assets, Accounts, Processes, Files, Connections), threat and assessment objects (Threat Indicators (IoC), Vulnerabilities (Attack Techniques), Threat Activities (Threat Activities), and Impacts (Investigation Hypotheses, Evidence Requirements, Evidence Tasks, and Response Actions). Assets can be servers, terminal devices, or network devices; accounts can be user accounts, service accounts, or administrative accounts; processes can be operating system processes or application processes; and files can be executable files, script files, document files, or other data files.
[0023] Specifically, the service platform can identify the corresponding security objects based on the subject object references, object object references, and operation information in the security observation data, and determine the object relationships between different security objects based on the operation information.
[0024] Object relationships are used to characterize the connections between different security objects arising from network security behaviors. For example, object relationships can include relationships such as a process executing another process, a process creating a file, an asset establishing a network connection, an account performing authentication, a process running on an asset, an asset being exploited, and security observation indicating threat activities.
[0025] In one feasible implementation, if a security observation data indicates that an office software process has started a script interpretation process, then the security objects corresponding to the office software process and the script interpretation process can be identified respectively, and a process execution relationship can be established between the two security objects; if another security observation data indicates that the script interpretation process has established a network connection with an external domain name, then the corresponding network connection object or threat indicator object can be further identified, and a connection relationship between the process and the network connection can be established.
[0026] Furthermore, version control can be implemented for security objects and their relationships.
[0027] Object versions are used to characterize the valid data status of a security object or object relationship under a specific investigation state. When the object attributes or object relationships of a security object undergo valid changes, a corresponding updated version can be generated, and the version association between the updated version and the previous version is preserved.
[0028] For example, a target asset might be in a "suspected affected" state at the initial stage of a cybersecurity incident investigation, corresponding to the first object version. Subsequently, after security devices isolate the target asset from the network, its network status changes, generating a second object version and establishing a version association between the second and first object versions. By retaining different object versions, subsequent supplementary security evidence, investigation hypotheses, and candidate actions can be clearly correlated with specific object states, rather than directly overwriting historical investigation information after the target asset's status changes.
[0029] In step S102, network security event records from heterogeneous security data sources can be uniformly converted into security observation data with source, time, object and original evidence association, and further form security objects with version information and object relationships.
[0030] S104: Generate evidence requirement items based on the preset evidence roles corresponding to the survey hypothesis data, and construct a survey hypothesis verification map and evidence sufficiency matrix based on the evidence requirement items. Perform causal tracing along the object relationship, and determine the satisfaction status of each evidence requirement item based on the obtained tracing evidence. The survey hypothesis data is used to characterize unverified security incident explanations for current cybersecurity incidents. Optionally, the survey hypothesis data is used to express candidate explanations to be verified and does not represent confirmed attack facts.
[0031] For example, in the same cybersecurity incident, one or more investigative hypothesis data can be formed. For instance, if an office software process is detected to launch a script interpretation process and further establish communication with an external network address, a first investigative hypothesis can be formed to characterize the candidate attack explanation of "script execution triggered by a malicious document and further command and control communication established"; a second investigative hypothesis can also be formed to characterize the candidate explanation of "the script execution is a normal operational behavior, and the external communication alarm is a false alarm". Subsequently, multiple investigative hypotheses can be verified using security evidence corresponding to different evidentiary roles.
[0032] Predefined evidence roles are used to characterize different categories of evidence needed to verify investigative hypotheses. In some embodiments, predefined evidence roles may include one or more of the following: initial access, execution, persistence, identity or permission use, lateral movement, command and control, data access, scope of influence, and attribution.
[0033] Different evidentiary roles can also have different levels of importance. For example, evidentiary roles that determine whether an investigative hypothesis can be established can be configured as core evidence roles, evidentiary roles used to strengthen the verification of conclusions can be configured as auxiliary evidence roles, and evidentiary roles used to exclude other investigative hypotheses can be configured as exclusionary evidence roles.
[0034] Evidence requirement items are used to characterize the evidentiary requirements that a specific evidentiary role needs to meet to verify a particular survey hypothesis. For each survey hypothesis data and each preset evidentiary role, the service platform 100 can generate corresponding evidence requirement items.
[0035] An evidence requirement can be associated with at least the investigation hypothesis data, evidence roles, the scope of the target security object, and the verification time range, and corresponding fulfillment conditions can be set. These fulfillment conditions may include acceptable evidence source types, independent evidence source requirements, verification time range, the scope of the target security object, and evidence field conditions, etc.
[0036] For example, for the "command and control" evidence role in a certain first investigation hypothesis, it may be required to obtain network communication evidence from at least a predetermined number of independent source groups. Furthermore, the events described in the network communication evidence must occur within the target investigation timeframe, and the network communication evidence should be associated with the target asset or process. Only when the relevant evidence meets these conditions can the corresponding evidence requirement be considered satisfied.
[0037] After generating the evidence requirement items, an evidence sufficiency matrix can be constructed based on the evidence requirement items.
[0038] The sufficiency of evidence matrix is used to structurally record the evidence fulfillment status of each evidence requirement item, using survey hypothesis data and pre-defined evidence roles as different data dimensions. In other words, a combination of survey hypothesis data and a pre-defined evidence role can correspond to a matrix cell, and the corresponding evidence requirement item can be associated with that matrix cell.
[0039] In this specification, the evidence satisfaction status recorded by the matrix unit includes at least a binary status of "with evidence" or "without evidence," and may also include different statuses determined based on the evidence acquisition and verification situation. For example, it may include satisfied status, unsatisfied status, partially satisfied status, conflicting status, source-dependent status, expired status, access-restricted status, and no-verification status, etc.
[0040] Among them, the satisfied state can indicate that the corresponding evidence requirement item has obtained security evidence that meets the preset requirements; the unsatisfied state can indicate that sufficient security evidence has not yet been obtained to complete the verification; the partially satisfied state can indicate that the existing security evidence only covers part of the time range, part of the target object, or part of the evidence fields required by the evidence requirement item; the conflicting state can indicate that security evidence from mutually independent sources gives results that cannot be simultaneously true for the same target fact; the source dependent state can indicate that although the number of evidences meets the requirements, the evidences all depend on the same underlying source and therefore cannot be regarded as mutually independent evidence.
[0041] Furthermore, a hypothesis testing map can be constructed based on the evidence requirements, as explained below: A hypothesis testing graph can be understood as a relational data structure representing the relationships between investigation hypotheses, evidence requirements, and evidence acquisition. A hypothesis testing graph can include nodes corresponding to investigation hypothesis data, nodes corresponding to evidence requirement items, nodes corresponding to existing security evidence, and nodes corresponding to the ability to acquire candidate evidence.
[0042] Specifically, it can establish a requirement verification relationship between survey hypothesis data and evidence requirement items to indicate that a certain survey hypothesis needs to be verified by corresponding evidence requirements; it can establish a support or disproving relationship between evidence requirement items and existing security evidence to indicate the supporting or disproving effect of existing security evidence on the corresponding evidence requirement; and it can also establish an acquisition relationship between evidence requirement items and the ability to acquire candidate evidence to indicate what data acquisition capabilities can be used to further obtain the corresponding security evidence when the current security evidence is insufficient.
[0043] It should be noted that even if a certain evidence requirement has been met, the corresponding evidence requirement can still be retained in the investigation hypothesis verification graph to support subsequent interpretation of investigation conclusions, re-verification after changes in the object version, and playback of historical investigation status.
[0044] After establishing the evidence sufficiency matrix and the hypothesis verification map, the service platform can perform causal tracing along the object relationships established in step S102. In some embodiments, causal tracing can adopt a hierarchical processing approach, including candidate attack stage determination, object relationship expansion, and original security evidence verification, as shown below: First, based on the investigation anchor and the preset evidence role to be verified, candidate attack phases related to the evidence role to be verified can be identified from the security objects and their relationships. The investigation anchor can be a security object that triggers a cybersecurity incident investigation, an abnormal network connection, threat indicators, or security alerts.
[0045] Then, based on the security objects corresponding to the candidate attack phases and the event time range, the process is expanded bidirectionally along the object relationships. For example, the predecessor objects of the current security object can be searched backwards to determine the parent process, script source, download source, identity login session, etc. that caused the current behavior; the successor objects of the current security object can be searched forwards to determine the subsequent file access, network scan, lateral movement, or data access behaviors, thereby forming candidate tracing paths.
[0046] Furthermore, for the object relationships to be verified in the candidate tracing paths, the corresponding original security evidence can be obtained by referencing the original evidence in the security observation data to verify the object relationships. For example, original logs, network packets, samples, or other original data can be obtained by drilling down only for the relationships entering the candidate tracing paths, without needing to read all the original security data in its entirety.
[0047] Based on the obtained source tracing evidence, information such as the source of the evidence, the time of the event, the associated security object, the object version, and the evidence fields can be combined to determine whether the source tracing evidence meets the corresponding evidence requirements, and update the corresponding evidence requirements and the satisfaction status in the evidence sufficiency matrix.
[0048] For example, when verifying the role of "lateral movement" evidence, if the identity authentication data source records that the target account made an abnormal remote login to the second server, and the network traffic data source records that there is a corresponding east-west network connection between the first server and the second server, and the two pieces of evidence come from independent source groups, then the requirement for the lateral movement evidence can be determined to be satisfied based on the pre-configured conditions.
[0049] Conversely, if multiple alarms, though numerous, are all generated from the same network data packet through different detection rules, then since the corresponding security observations belong to the same source group, it can be determined that the evidence requirement is still in a source-dependent state, rather than being determined to be satisfied simply because the number of evidences has increased.
[0050] Optionally, step S104 is not executed only once in the initial stage of cybersecurity incident handling. After obtaining supplementary security evidence through investigation tasks or changing the status of security objects through candidate actions, the update process for the investigation hypothesis data, evidence requirement items, evidence sufficiency matrix, and investigation hypothesis verification map can be retried. Specifically, when new security observation data, supplementary security evidence, or a new version of a security object is generated, the causal tracing in step S104 can be re-executed based on the updated security object and object relationships, and the satisfaction status of each evidence requirement item can be re-determined, so that the investigation hypothesis verification results are dynamically updated with the addition of new evidence and changes in object status.
[0051] S106: For the unmet evidence requirements corresponding to the detected source tracing breakpoint, generate an investigation task including the expected object version of the target security object, execute the investigation task to obtain supplementary security evidence, determine the applicable object version of the supplementary security evidence based on the expected object version and the current object version of the target security object, and update the evidence sufficiency matrix and the investigation hypothesis verification map based on the supplementary security evidence applicable to the current object version. Source tracing breakpoints are used to characterize the positions where, when verifying candidate source tracing paths, the verification of the corresponding evidence role cannot continue due to insufficient current security evidence or an abnormal current evidence state.
[0052] Indicatively, source tracing breakpoints may include: candidate attack paths lacking preceding security evidence for verifying object relationships; security evidence from independent security data sources conflicting on the same fact; existing security evidence all relying on the same underlying source; original security evidence for verifying object relationships missing or nearing its data retention period; security data required for verification being restricted by permissions; and continued verification requiring the active acquisition of data from the target security object.
[0053] When a source tracing breakpoint is detected, the service platform can locate the evidence requirement item corresponding to that breakpoint. If the corresponding evidence requirement item has not yet been satisfied, it can be identified as an unsatisfied evidence requirement item, and further investigation tasks can be generated based on the unsatisfied evidence requirement item.
[0054] An investigation task is used to transform the abstract concept of "what kind of security evidence is missing" into a data acquisition task that can be performed through a specific security data source. An investigation task may at least include a target security object, the expected version of the target security object, the investigation timeframe, and the target for writing back the results. In some embodiments, an investigation task may also include candidate data sources, expected result structure, retention deadline, maximum forensic intrusion level, access requirements, timeout period, or idempotency information.
[0055] The expected object version is used to characterize the state of the target security object when generating the investigation task. Since there may be a time interval between the generation of the investigation task and the actual acquisition of supplementary security evidence, and the target security object may change its state due to new security events or handling operations during this time interval, it is necessary to associate the investigation task with the corresponding object version to avoid the delayed investigation results directly overwriting the current object state.
[0056] When performing investigation tasks, the appropriate data acquisition method can be selected based on the data status of the candidate data sources. For example, for data that has already been indexed online, an online query can be performed; for data that has been moved to offline or cold storage, an offline archive retrieval can be performed; for data with restricted access permissions, data acquisition can be performed after obtaining the corresponding access permissions; for data that needs to be obtained from the target security object in real time, proactive data acquisition can be performed through endpoint detection components or other security devices.
[0057] If there is no data acquisition channel to complete the investigation task, the investigation task or the corresponding evidence requirement can be recorded as unavailable, but the corresponding evidence requirement cannot be directly updated to a satisfied state.
[0058] After the investigation task is completed, supplementary security evidence can be obtained. Supplementary security evidence refers to security data newly acquired through the investigation task to fill gaps in the current investigation evidence. After obtaining the supplementary security evidence, the investigation results can be converted into unified security observation data and mapped to the corresponding security objects or object relationships according to the processing method in step S102.
[0059] Furthermore, it is necessary to determine the applicable object version for supplementary security evidence based on the relationship between the expected object version corresponding to the investigation task and the current object version of the target security object.
[0060] In one scenario, if the expected object version matches the current object version, it can be determined that the supplementary security evidence is applicable to the current object version.
[0061] In another scenario, even if the expected object version is different from the current object version, but the version change does not involve the object attributes or object relationships used to determine the status of the corresponding evidence requirement, it can still be determined that the supplementary security evidence is applicable to the current object version.
[0062] In another scenario, if the version change between the anticipated object version and the current object version involves object attributes or relationships that the corresponding evidence requirement actually depends on, then supplementary security evidence can be retained and associated with the corresponding historical object version, rather than being directly used for the current object version. Simultaneously, it can be determined whether the same or new evidence gaps still exist for the changed current object version, and supplementary evidence requirements can be generated or updated as needed.
[0063] For example, when the investigation task is generated, the target asset corresponds to version 6. During the investigation, the system only generates an isolation suggestion for the target asset, but has not yet changed the object state related to the current evidence requirements, thus updating the object to version 7. In this case, if the version change does not involve the login relationship to be verified or the network communication relationship, the identity log investigation results initiated for version 6 can still be used for version 7.
[0064] Conversely, if the target process on the target asset has been terminated during the investigation, and the current investigation task is to obtain security evidence related to the real-time memory state of that process, then the supplementary security evidence may only be applicable to the pre-change version of the object. Its historical association should be preserved, and it should be reassessed based on the current version to determine whether the corresponding evidence can or needs to be obtained.
[0065] For supplementary security evidence deemed applicable to the current version of the object, its source independence, coverage, and conflict with existing evidence can be re-examined, and the satisfaction status of the corresponding evidence requirement items can be updated accordingly. The updated status is then written into the evidence sufficiency matrix, and the support, disproving, or other relationships between the corresponding evidence requirement items and the supplementary security evidence in the hypothesis testing graph are updated.
[0066] Through the above processing, the successful execution of an investigation task does not necessarily mean that the corresponding evidence requirements have been met. For example, the investigation task may only obtain a portion of the logs within the required time range. In this case, the evidence requirements can be updated to a partially met state, while the uncovered time range can be retained to continue forming supplementary evidence requirements. Or, for example, although the investigation task returns multiple pieces of security evidence, if they all come from the same source group, the source dependency state can still be maintained.
[0067] In one feasible implementation, after step S106 is completed, the satisfaction status of the corresponding evidence requirement item can be re-determined based on the supplementary security evidence, and the evidence sufficiency evaluation and causal tracing processing in step S104 can be returned to execution. If the supplementary security evidence eliminates the original tracing breakpoint, the candidate tracing path can continue to be expanded along the updated object relationship; if the supplementary security evidence only partially satisfies the corresponding evidence requirement item, or generates new evidence conflicts, source dependencies, version failures, or other unmet conditions, a new tracing breakpoint can be formed or the original evidence requirement item can be updated, and a new investigation task can continue to be generated. Thus, when the evidence requirement item is not yet satisfied and there is still an executable evidence acquisition method, the evidence evaluation, breakpoint identification, investigation, and supplementary evidence write-back process can be cyclically executed between steps S104 and S106.
[0068] S108: Based on the evidence sufficiency matrix, determine the remaining evidence requirements, and determine the investigation tasks to be executed corresponding to the remaining evidence requirements from the generated investigation tasks. Based on the current attack path, affected objects and the remaining evidence requirements obtained by causal tracing, generate candidate actions to bind the target security object version. According to the impact of the candidate actions on the state change of the unfixed evidence corresponding to the remaining evidence requirements, determine the execution sequence of the candidate actions and the investigation tasks to be executed. The remaining evidence requirements can be understood as evidence requirements that, after supplementary evidence collection and re-evaluation of the evidence status in step S106, still do not meet the corresponding evidence satisfaction requirements. For example, this may include evidence requirements that are still in an unsatisfied state, partially satisfied state, conflicting state, source-dependent state, expired state, or access-restricted state.
[0069] The service platform can read the updated evidence sufficiency matrix, determine the evidence gaps that still exist for each investigation hypothesis data under different evidence roles, and determine the remaining supplementary evidence requirements accordingly.
[0070] For investigation tasks that have already been generated in response to these evidence requirements, the pending investigation tasks can be further determined based on the current execution status of the investigation tasks. For example, investigation tasks that have been successfully executed and whose corresponding evidence requirements have been met are no longer considered pending investigation tasks; investigation tasks that are still in the pending, waiting for permissions, queued, executing, or partially completed states can be retained based on the current status of the corresponding evidence requirements.
[0071] At the same time, the service platform can determine the current attack path and affected targets based on the current causal tracing results.
[0072] The current attack path can be understood as the propagation path of attack behaviors supported by existing security evidence or that still require key attention in the current investigation. For example, it can include a sequence of object relationships formed from initial entry, process execution, network communication to lateral movement.
[0073] The affected objects can be security objects that have been affected or may be further affected by the current attack path, such as affected assets, accounts, processes, files, or network connections.
[0074] Furthermore, candidate actions can be generated based on the current attack path, the affected targets, and the remaining required supplementary evidence.
[0075] Candidate actions refer to security response operations that have been generated but not yet actually implemented in response to a current cybersecurity incident. Examples include network isolation, restricting account privileges, blocking network addresses, terminating malicious processes, deleting malicious files, or other security actions.
[0076] Candidate actions can be bound to the corresponding object version of the target security object. For example, for a network isolation action targeting a target asset, the object version of the target asset at the time the action was generated can be recorded; for a termination action targeting a target process, the object version of the target process at the time the action was generated can be recorded. Through this version binding, it can be confirmed again whether the target security object is still in the expected state at the time the action was generated before the candidate action is actually executed.
[0077] Optionally, considering the potential interplay between candidate actions and supplementary evidence collection in cybersecurity incident handling, some candidate actions may alter or delete security evidence that has not yet been fully secured. For example, terminating a process may cause temporary data in the process's memory to disappear, cleaning up malicious files may alter the file context, and restarting the device may cause volatile evidence to disappear.
[0078] Therefore, the service platform can determine the execution sequence between candidate actions and pending investigation tasks based on the impact of candidate actions on the status changes of unfixed evidence corresponding to the remaining supplementary evidence requirements.
[0079] In one implementation, if the current cybersecurity incident does not present a high-risk spread situation that requires immediate containment, and the candidate actions may affect evidence that has not yet been secured, then the investigation task for securing the relevant evidence can be performed first, and then the candidate actions can be performed.
[0080] In another implementation, if the current cybersecurity incident is spreading rapidly and it is necessary to prioritize blocking the spread of the attack, reversible actions can be allowed to be executed first, such as network isolation or access restrictions on the target assets. For irreversible actions that will significantly change or delete evidence that has not yet been secured, they can remain in a pending state until the relevant evidence is secured or the corresponding execution permission is obtained.
[0081] Therefore, this manual does not treat "investigation and evidence collection" and "security response" as two independent sequential processes. Instead, it determines the execution order based on the current evidence requirements and the impact of the handling actions on the evidence status. This ensures that the handling of cybersecurity incidents can both control the risk of attacks and reduce the possibility of losing evidence due to premature execution of destructive security actions.
[0082] If, after one or more iterations of steps S104 and S106, there are still remaining items requiring supplementary evidence, the cybersecurity incident handling process is not yet complete. The service platform can determine whether to continue supplementary evidence processing or proceed with the joint sorting of candidate actions and investigation tasks based on the remaining items, the corresponding investigation tasks to be executed, and the current attack spread status. In some embodiments, when the remaining items still have available evidence sources and the current cybersecurity incident does not require immediate handling, the process can return to step S106 to execute the corresponding investigation task; when the current cybersecurity incident requires security handling, the investigation tasks and candidate actions can be executed serially or in parallel according to the execution sequence determined in step S108. In other words, step S108 can serve as a branching node between continuing the supplementary evidence loop and entering the handling phase, or it can be used to coordinate supplementary evidence processing and security handling simultaneously.
[0083] S110: When the current object version of the target security object corresponding to the candidate disposal action matches the target security object version bound to the candidate disposal action, and the candidate disposal action satisfies the execution sequence, the candidate disposal action is executed through the security device, and investigation conclusion data is generated based on the action execution result, the evidence sufficiency matrix, and the investigation hypothesis verification map.
[0084] Before the candidate action is actually executed, the current version of the target security object can be obtained again, and the current version can be matched with the version of the target security object bound when the candidate action was generated.
[0085] This is because there may be a time interval between the generation of a candidate action and its actual execution. During this period, a new version of the object may be generated due to the writing of new security evidence, the execution of other security actions, or changes in the state of the target system. If the target security object has undergone a state change related to the candidate action, the previously generated candidate action may no longer be applicable to the current object state.
[0086] For example, after generating a process termination action for a suspected malicious process, if the process has already exited and generated a new object state before actual execution, it is not advisable to continue execution directly using the action parameters corresponding to the old version. By performing object version verification again before action execution, security actions based on expired object states can be avoided.
[0087] In addition to object version verification, it can be further determined whether the candidate disposal action meets the execution sequence determined in step S108. For example, when a candidate disposal action is set as a subsequent task of a certain investigation task, the candidate disposal action can only be actually executed after the corresponding investigation task reaches the required execution state; when the candidate disposal action is a reversible disposal action that is allowed to be executed first, it can enter the actual execution stage according to the corresponding execution order.
[0088] When the target security object version matches and the candidate disposal actions meet the execution sequence, the service platform can call the corresponding security device to execute the candidate disposal actions.
[0089] Security devices can be network security devices or security components capable of changing the security status of a target security object. For example, they can be endpoint detection and response devices, network access control devices, firewalls, authentication devices, or other devices capable of performing isolation, blocking, access restriction, or security cleanup operations.
[0090] After executing a candidate action, the safety device can return the corresponding action execution result. The action execution result can be used to characterize whether the candidate action was successful, partially completed, failed, or rolled back, and can also include the status information of the target safety object after execution.
[0091] The service platform can determine the state changes of the target security object based on the action execution results, and generate or update the object version of the target security object according to the aforementioned object version management method. For example, after a network isolation action is successfully executed, the target asset can be updated from its original network-accessible state to an isolated state, and a corresponding updated object version can be generated.
[0092] Since changes in the state of an object may further affect the applicability of previous evidence requirements, the evidence requirements affected by the state changes can be re-determined based on the updated version of the object, and the evidence sufficiency matrix and the hypothesis testing map can be re-evaluated.
[0093] Optionally, the cybersecurity incident handling process does not necessarily end immediately after the candidate action is executed. Since the candidate action may change the object attributes, object relationships, or operational status of the target security object, new security observation data or updated security object versions can be generated based on the action execution results. The affected evidence requirements can then be re-determined based on the updated object version. If it is determined that there are evidence requirements that need to be re-verified due to changes in object status, or if the action execution results generate new attack paths, affected objects, or evidence gaps, the process can return to step S104 to update the evidence sufficiency matrix and investigation hypothesis verification map, and further execute steps S106 to S110 based on the new source tracing results. Thus, the action execution results can re-enter the cybersecurity incident investigation chain, rather than serving as a termination result outside the investigation process.
[0094] In one feasible implementation, investigation termination conditions can be set for cybersecurity incidents. When the core evidence roles corresponding to the investigation hypothesis data have all reached a satisfactory state, evidence conflicts have been explained or eliminated, all necessary investigation tasks have reached a terminated state, and the candidate actions to be performed have brought the target security object to the expected security state, it can be determined that the investigation termination conditions are met, and the execution of steps S104 to S110 is stopped. Finally, based on the final evidence sufficiency matrix, investigation hypothesis verification map, and action execution results, the final investigation conclusion data is generated.
[0095] If the investigation termination conditions are not met, at least some of the steps in steps S104 to S110 can continue to be executed. For example, if there are still supplementary evidence requirements, step S106 is executed to obtain supplementary security evidence; if the investigation hypothesis or candidate attack path changes due to the new evidence, step S104 is returned to re-perform causal tracing and evidence status evaluation; if a new security object status is generated after the candidate action is executed, step S104 is re-executed based on the updated object version; if the action fails, the effect is insufficient, or a new affected object is generated, step S108 can be re-executed to generate or adjust the candidate action, and step S110 can continue to be executed.
[0096] It should be noted that the failure, cancellation, expiration, or unavailability of an investigation task does not indicate that the corresponding evidence requirements have been met. Therefore, the conclusion of an investigation cannot be based solely on the number of pending investigation tasks reaching zero. For core evidence requirements that are still in an unmet, partially met, conflicting, source-dependent, expired, or access-restricted state, their investigation status can be maintained, or appropriate exception handling methods can be determined based on the cybersecurity incident risk and its handling.
[0097] The investigation conclusion data is used to characterize the investigation results of a current cybersecurity incident based on existing security evidence and the results of the response. The investigation conclusion data may include at least the verification results of the investigation hypotheses and the security evidence associated with the verification results; in some embodiments, it may also include the current attack path, the affected security objects, unmet evidence requirements, and the results of the security responses already implemented.
[0098] For example, if current security evidence proves that the target document triggered script execution, the script further established external command and control connections, and subsequent lateral movement behavior was confirmed through independent data sources, the corresponding investigation hypothesis can be established as a verified result supported by evidence. Relevant terminal records, network communication records, and authentication records can then be linked to the investigation conclusion data. If there is still unresolved memory evidence or unconfirmed scope of influence, the corresponding unmet evidence requirements can be retained to characterize the evidence boundaries corresponding to the current investigation conclusion.
[0099] Specifically, the process of generating evidence requirement items based on preset evidence roles corresponding to the investigation hypothesis data is executed to construct an investigation hypothesis verification map and an evidence sufficiency matrix, including: S202: Generate evidence requirement items for each preset evidence role corresponding to the survey hypothesis data, and determine the conditions for meeting the evidence requirement items according to the evidence source type, independent evidence source requirements, verification time range, target security object range and evidence field conditions corresponding to the preset evidence role; Evidence source type is used to characterize the category of data source that can provide the security evidence required for the corresponding evidence role. For example, for the evidence role corresponding to the execution behavior, the evidence source type can be configured to correspond to terminal detection data, process audit data, or script execution records; for the evidence role corresponding to lateral movement, the evidence source type can be configured to correspond to identity authentication data, network connection data, or terminal security data; and for the evidence role corresponding to command and control, the evidence source type can be configured to correspond to network traffic data, domain name resolution data, or terminal network connection data.
[0100] The requirement of independent evidence sources limits the number or composition of mutually independent evidence sources needed for a corresponding evidence requirement to be determined as satisfied. The independence of evidence sources can be determined in conjunction with the source groups in the aforementioned embodiments. That is, even if multiple existing security evidences have been obtained for the same evidence requirement, if these multiple existing security evidences all rely on the same underlying source record or the same parsing chain, they can be considered to belong to the same source group and are not counted repeatedly in the number of mutually independent evidence sources.
[0101] For example, regarding the "lateral movement" evidence role of a certain survey hypothesis, it can be set that at least two independent source groups need to provide corresponding security evidence. If the identity authentication system records that the target account remotely logs into the second asset from the first asset, and the network traffic data source records that a network connection corresponding to the remote login was established between the first and second assets, and the two belong to different source groups, then the requirement for independent evidence sources can be met. If two alarms in the security information and event management system are generated from the same identity authentication log processed by different rules, then although the two alarms appear as different records, they can still be considered as an independent source group for judgment.
[0102] The verification time range is used to define the time interval within which security evidence that can be used to meet the corresponding evidentiary requirements must be covered or fall within. The verification time range can be determined based on the time of the attack corresponding to the investigation hypothesis, the time of the investigation anchor point, and the position of different evidentiary roles in the attack chain.
[0103] For example, if the current investigation anchor point is that the target asset detects abnormal external connection behavior at time T, then for the "execution" evidence role prior to the abnormal external connection, a verification time range within a preset time range before time T can be set; for the "lateral movement" or "data access" evidence role after the abnormal external connection, a corresponding verification time range starting from time T and continuing forward can be set. The specific verification time range can be configured according to the current cybersecurity incident investigation scenario, and this manual does not impose any limitations on it.
[0104] The target security object scope is used to define the set of security objects or object levels that can be verified by the corresponding evidence requirement. For example, the target security object scope can be defined as a specific asset, a specific account, a specific process, a specific file, or a set of objects consisting of multiple associated security objects. In some embodiments, it can also be defined as a target security object scope associated with a certain object version, thereby establishing a correspondence between the evidence requirement and the specific object state that needs to be verified.
[0105] Evidence field conditions are used to limit the fields that existing security evidence must have or satisfy. Evidence field conditions can be set separately for different evidence roles, and may include matching conditions for fields such as target account identifier, source asset identifier, target asset identifier, process identifier, network address, port, file identifier, operation type, or event result.
[0106] Furthermore, the "satisfaction conditions" can be understood as a set of structured conditions used to determine whether sufficient security evidence has been obtained for the corresponding evidence requirement. These conditions can be composed of the evidence source type, independent evidence source requirements, verification time range, target security object range, and evidence field conditions.
[0107] In one feasible implementation, a set of conditions for each evidence requirement can be formed. The set of conditions records at least the types of evidence sources that are allowed to be used, the number of independent source groups required to be reached, the time range for the use of secure evidence, the scope of the target secure objects that need to be verified, and the field conditions that the secure evidence needs to meet.
[0108] In the subsequent evidence verification process, the service platform can match existing security evidence or supplementary security evidence obtained through investigation tasks with the conditions to determine whether the corresponding security evidence can be used to support the corresponding evidence requirement and the current state of satisfaction of the corresponding evidence requirement.
[0109] As an illustration, a first evidence requirement can be generated for the "command and control" evidence role in the investigation hypothesis data. The conditions for meeting the first evidence requirement may include: the security evidence originates from network traffic data or terminal network connection data; it includes at least a preset number of independent source groups; the event time falls within the target communication time range corresponding to the investigation hypothesis; the associated objects include target assets and target processes; and the target network address, connection direction, and connection result fields in the security evidence meet the corresponding verification requirements.
[0110] If only one network connection record that meets the time and object scope requirements is subsequently obtained, but the number of independent source groups has not yet reached the required number, the first evidence requirement can be kept in the corresponding incompletely satisfied state according to the state determination method of the aforementioned embodiment, and the evidence role will not be deemed to have been fully verified simply because there is one network connection record.
[0111] Furthermore, different evidentiary roles can be assigned different role levels based on their importance in verifying the investigation hypotheses. For example, evidentiary roles used to directly verify key attack behaviors can be designated as core roles, evidentiary roles that enhance the credibility of the hypotheses but are not essential can be designated as auxiliary roles, and evidentiary roles used to exclude other competing hypotheses can be designated as exclusionary roles. Role levels can serve as one of the bases for subsequently determining the conditions for ending the investigation and prioritizing investigative tasks.
[0112] S204: Using the survey hypothesis data and the preset evidence roles as the data dimensions of the initial evidence sufficiency matrix, determine the matrix unit corresponding to each of the evidence requirement items in the initial evidence sufficiency matrix, and establish the association between the evidence requirement items and the corresponding matrix units in the initial evidence sufficiency matrix to obtain the evidence sufficiency matrix; The initial evidence sufficiency matrix can be understood as a two-dimensional data organization structure established at the current stage of the investigation based on the investigation hypothesis data and its corresponding preset evidence roles, which is yet to be written or continuously updated with evidence verification results.
[0113] Data dimensions are used to determine the organizational position of evidence requirement items in the evidence sufficiency matrix. In this embodiment, the investigation hypothesis data can be used as the first data dimension, and the preset evidence role as the second data dimension. Thus, a combination of an investigation hypothesis data and a preset evidence role can correspond to a matrix unit. The matrix unit is used to centrally record the verification information of the evidence requirement items corresponding to a specific investigation hypothesis data under a specific evidence role. A matrix unit can be associated with one evidence requirement item; when the evidence requirements are split or refined, a matrix unit can also be associated with multiple evidence requirement items with parent-child relationships.
[0114] For example, if there are two investigation hypotheses H1 and H2 in a current cybersecurity incident, and the pre-defined evidence roles include initial access, execution, persistence, lateral movement, and command and control, then investigation hypotheses H1 and H2 can be arranged along the first data dimension, and the pre-defined evidence roles can be arranged along the second data dimension, thereby forming multiple matrix units.
[0115] For example, the intersection of the investigation hypothesis H1 and the "execution" evidence role corresponds to the first matrix unit, and the evidence requirement item generated in step S202 for the "execution" evidence role of the investigation hypothesis H1 can be associated with the first matrix unit; the intersection of the investigation hypothesis H1 and the "command and control" evidence role corresponds to the second matrix unit, and the corresponding "command and control" evidence requirement item is associated with the second matrix unit.
[0116] In some embodiments, even if a certain predefined evidence role does not currently have existing security evidence that can support the investigation hypothesis, a corresponding matrix unit can be established for the investigation hypothesis and the evidence role, and corresponding evidence requirement items can be generated.
[0117] Furthermore, the service platform can associate the satisfaction conditions determined in step S202 with the verification results corresponding to the existing security evidence to the matrix unit, and determine the current satisfaction status of the matrix unit.
[0118] For example, for a matrix cell corresponding to the "lateral movement" evidence role in the investigation hypothesis H1, it can be associated with the lateral movement evidence requirement item, the conditions for satisfying the evidence requirement item, existing supporting security evidence, existing counter-evidence, and the current satisfaction status. When new security evidence is subsequently obtained through investigation tasks, the satisfaction status of the evidence requirement item corresponding to the matrix cell can be re-determined based on the new security evidence, without having to change the data organization of the entire matrix.
[0119] In one feasible implementation, when existing evidence only covers a portion of the target security object range or a portion of the verification time range required by the evidence requirement, the association of the existing security evidence with the covered portion can be retained, and corresponding supplementary evidence requirements can be retained or generated for the uncovered portions. These supplementary evidence requirements can continue to be associated with the original matrix unit, enabling the matrix unit to record the current evidence coverage and remaining evidence gaps.
[0120] For example, a certain lateral movement evidence requirement needs to verify the remote identity usage from the first asset to the second asset within a time range T1 to T3. If only identity authentication records for the T1 to T2 range are currently available, then existing security evidence can be associated with this matrix unit, recording the covered area, while retaining the supplementary evidence requirement for the T2 to T3 range. This allows for targeted supplementary evidence collection for the remaining time range during subsequent investigations, instead of re-acquiring all security data for the entire verification time range.
[0121] It should be noted that the evidence sufficiency matrix can be continuously updated throughout the cybersecurity incident investigation process, without needing to be rebuilt each time supplementary security evidence is obtained. When new security evidence is obtained in subsequent steps or the object version changes, the matrix cell can be directly located based on the corresponding evidence requirement item, and the supporting evidence, rebuttal evidence, fulfillment status, or remaining supplementary evidence requirements associated with that matrix cell can be updated, thereby forming a continuously evolving evidence sufficiency record.
[0122] S206: The survey hypothesis data, the evidence requirement items, the existing security evidence and the ability to obtain candidate evidence are used as nodes in the initial survey hypothesis verification graph, and the node relationships in the survey hypothesis verification graph are established to generate the survey hypothesis verification graph.
[0123] The initial investigation hypothesis verification map can be understood as a relational data structure established based on the investigation hypotheses, evidence requirements, existing security evidence, and available evidence acquisition capabilities in the current cybersecurity incident. It is used to represent what kind of security evidence is needed for different investigation hypotheses, what evidence requirements can be verified by the current security evidence, and what capabilities can be used to further obtain the required security evidence when evidence is missing.
[0124] Candidate evidence acquisition capability is used to characterize the evidence acquisition capability (i.e., the means or channels used to acquire evidence) or data acquisition channel of the current network security incident handling system that can be used to obtain specific types of security evidence. For example, it may include online log query capability (channel), offline archive retrieval capability (channel), restricted data access capability (channel), terminal active forensics capability (channel), network traffic query capability (channel), deep forensics capability (channel), etc.
[0125] It should be noted that the ability to acquire candidate evidence is not the same as the already generated investigation task. The ability to acquire candidate evidence describes the basic computer processing capabilities that "can obtain the required evidence from where and by how," and it is implemented by relying on a certain computer processing channel. The investigation task, on the other hand, is an execution task generated after a specific evidence gap is detected, based on the evidence requirements and the ability to acquire candidate evidence, and is geared towards a specific target security object and a specific verification scope.
[0126] Specifically, each survey hypothesis data can be mapped to a corresponding survey hypothesis node, each evidence requirement item can be mapped to a corresponding evidence requirement node, existing security evidence can be mapped to a corresponding evidence node, and the current candidate evidence acquisition capability that can be used to obtain security evidence can be mapped to a corresponding (computer processing) capability node.
[0127] Then, a graph relationship is established between nodes based on the business relationships corresponding to each node.
[0128] For example, a requirement verification relationship can be established between the investigation hypothesis node and the evidence requirement node based on the investigation hypothesis data corresponding to the evidence requirement item, which indicates that the corresponding investigation hypothesis needs to be verified through the evidence requirement.
[0129] Based on the verification function of existing security evidence for the evidence requirement item, an evidentiary relationship can be established between the evidence requirement node and the existing security evidence node. This evidentiary relationship can include supporting evidence relationships and contradictory evidence relationships. A supporting evidence relationship indicates that the existing security evidence supports the target fact required by the corresponding evidence requirement item; a contradictory evidence relationship indicates that the existing security evidence contradicts the target fact required to be verified by the corresponding evidence requirement item or is used to refute the corresponding investigation hypothesis.
[0130] Furthermore, based on the satisfaction conditions determined in step S202, an acquisition relationship can be established between the evidence requirement items and the candidate evidence acquisition capabilities that can obtain the secure evidence required to satisfy the satisfaction conditions.
[0131] For example, if a certain evidence requirement requires obtaining identity authentication evidence of a target asset within a specific verification time range, then the online log query capability that can query identity authentication logs can be identified as a candidate evidence acquisition capability, and an acquisition relationship can be established between the corresponding evidence requirement node and the capability node; if the corresponding identity authentication logs have been migrated from online storage to offline archiving, an acquisition relationship can also be established with the offline archive retrieval capability.
[0132] In one feasible implementation, the same evidence requirement can be associated with multiple candidate evidence acquisition capabilities. For example, for evidence requirements regarding whether a target process has established an external network connection, the terminal network connection record query capability and the network traffic query capability can be associated simultaneously. Subsequently, when generating specific investigation tasks, the appropriate data acquisition channel can be selected from multiple candidate evidence acquisition capabilities based on data online status, data access permissions, data retention status, and the permissible level of intrusion for evidence collection.
[0133] Furthermore, the investigation hypothesis verification graph can continuously receive subsequently generated nodes. For example, after detecting an evidence gap and generating an investigation task, the investigation task can be added to the investigation hypothesis verification graph as a supplementary task node to the corresponding evidence requirement node; after the investigation task is executed and supplementary security evidence is generated, the evidence node corresponding to the supplementary security evidence can be added to the investigation hypothesis verification graph, and a supporting evidence relationship or a rebuttal relationship can be established between it and the corresponding evidence requirement node.
[0134] In other words, the investigation hypothesis verification map is not only used to preserve the investigation hypotheses that are ultimately valid, but also to continuously express the verification link between "investigation hypothesis - evidence requirement - existing evidence - evidence acquisition capability - subsequent investigation results" during the investigation process.
[0135] For example, for investigation hypothesis H1, execution evidence requirement items, lateral movement evidence requirement items, and command and control evidence requirement items can be generated separately. If existing process creation records can support the execution evidence requirement item, a supporting evidence relationship can be established between the execution evidence requirement node and the corresponding process creation evidence node. If there is currently no sufficient security evidence for the lateral movement evidence requirement item, but the system has the ability to query authentication logs and network traffic, an acquisition relationship can be established between the lateral movement evidence requirement node and the two candidate evidence acquisition capability nodes. Subsequently, if new security evidence is obtained from the authentication log query, the new security evidence can be added to the graph as an evidence node, and a corresponding supporting or disproving evidence relationship can be established based on its verification results.
[0136] It should be noted that for evidence requirements that have already met the requirements, the relationships between the corresponding evidence requirement nodes and their security evidence nodes can be retained instead of being deleted from the investigation hypothesis verification graph. Therefore, when subsequent security object versions change, additional security evidence is added, or the investigation hypothesis needs to be re-verified, the corresponding evidence requirement items can be re-evaluated based on the existing node relationships. Furthermore, the security evidence upon which the investigation conclusions are based can be traced through these node relationships.
[0137] In one feasible implementation, the establishment of node relationships in the survey hypothesis verification graph described in S206 can be carried out in the following manner: Step C2: Based on the survey hypothesis data corresponding to each of the evidence requirement items, establish the requirement verification relationship between the node corresponding to the survey hypothesis data and the node corresponding to the evidence requirement item; The requirement validation relationship is used to characterize the validation dependency between survey hypothesis data and evidence requirement items. In other words, it is used to indicate which evidence requirement items are needed to complete the validation of the corresponding survey hypothesis.
[0138] As an illustration, when generating evidence requirement items, the service platform can record corresponding investigation hypothesis identifiers for each evidence requirement item, thereby establishing a correspondence between the evidence requirement items and the investigation hypothesis data. When constructing the investigation hypothesis verification graph, the investigation hypothesis data corresponding to each evidence requirement item can be determined based on the investigation hypothesis identifiers, and a requirement verification relationship can be established between the nodes corresponding to the investigation hypothesis data and the nodes corresponding to the evidence requirement items.
[0139] For example, for the first survey hypothesis data H1, execution evidence requirement item R1, lateral movement evidence requirement item R2, and command and control evidence requirement item R3 can be generated respectively. Then, a requirement verification relationship can be established between the node corresponding to the survey hypothesis data H1 and the nodes corresponding to the evidence requirement items R1, R2, and R3 respectively, to indicate that the survey hypothesis data H1 needs to be verified through the above-mentioned multiple evidence requirement items.
[0140] In some embodiments, different evidence requirements can also have different evidence role levels. For example, evidence requirements directly related to the validity of the investigation hypothesis can be configured as core evidence requirements, evidence requirements used to enhance the credibility of the investigation hypothesis can be configured as auxiliary evidence requirements, and evidence requirements used to exclude other investigation hypotheses can be configured as exclusionary evidence requirements. The service platform can associate the corresponding role levels with the requirement verification relationship so that it can subsequently determine the verification status of the investigation hypothesis and the investigation termination conditions based on the importance of different evidence requirements.
[0141] Step C4: Based on the support or negation of the existing security evidence for the corresponding evidence requirement item, establish a supporting evidence relationship or a rebuttal relationship between the node corresponding to the evidence requirement item and the node corresponding to the existing security evidence; Supporting evidence relationships are used to characterize that existing security evidence can support the target fact that the corresponding evidence requirement requires verification; contradictory evidence relationships are used to characterize that existing security evidence can negate, weaken, or conflict with the target fact that the corresponding evidence requirement requires verification.
[0142] As an illustration, the service platform can match existing security evidence with the conditions for fulfilling the corresponding evidence requirements, and determine the verification effect of the existing security evidence on the evidence requirements based on the matching results.
[0143] When the existing security evidence meets the requirements of the corresponding evidence requirement item in terms of evidence source type, event time, target security object scope, and evidence fields, and the target facts reflected by the existing security evidence are consistent with the target facts required to be verified by the evidence requirement item, it can be determined that the existing security evidence has a supporting role for the evidence requirement item, and a supporting evidence relationship is established between the node corresponding to the evidence requirement item and the node corresponding to the existing security evidence.
[0144] For example, a certain evidence requirement item is used to verify whether a target account remotely logs into a second asset from a first asset within a target time frame. If the identity authentication log records the corresponding account, source asset, target asset, authentication time, and authentication success result, and the security evidence satisfies the verification time frame and object scope of the evidence requirement item, then a supporting evidence relationship can be established between the evidence requirement item node and the evidence node corresponding to the identity authentication log.
[0145] When the target facts reflected by existing security evidence are contrary to the target facts required to be verified by the evidence requirement, or when existing security evidence can prove that the corresponding attack did not occur, a rebuttal relationship can be established between the node corresponding to the evidence requirement and the node corresponding to the existing security evidence.
[0146] For example, if a certain evidence requirement is used to verify whether a target process establishes a network connection with a specified external network address within a preset time range, and both the network traffic record and the terminal network connection record indicate that the target process does not have the network connection within the corresponding time range, then the network traffic record or the terminal network connection record can be used as evidence of contradiction, and a corresponding evidence of contradiction relationship can be established.
[0147] Furthermore, the same evidence requirement can be associated with both supporting and contradictory evidence. When supporting and contradictory evidence for the same objective fact comes from independent source groups, and the corresponding results cannot be simultaneously true, the supporting and contradictory evidence relationships can be retained in the investigation hypothesis testing map. In the subsequent evidence status evaluation process, the corresponding evidence requirement item will be identified as a conflict state, rather than directly deleting one type of security evidence.
[0148] Therefore, the investigation hypothesis verification map can not only record a certain investigation hypothesis supported by security evidence, but also preserve refutation or conflicting evidence at the same time, so that subsequent investigation conclusions can be verified and traced based on the complete correlation between supporting evidence and counter-evidence.
[0149] Step C6: Based on the conditions for satisfying each of the evidence requirements, determine the candidate evidence acquisition capabilities that can obtain the evidence required by the conditions, and establish the evidence acquisition relationship between the node corresponding to the evidence requirement and the node corresponding to the determined candidate evidence acquisition capability.
[0150] Evidence acquisition relationship is used to characterize what candidate evidence acquisition capabilities can be used to obtain supplementary security evidence that meets the requirements of the evidence requirement when sufficient security evidence has not yet been obtained for the corresponding evidence requirement.
[0151] As an illustration, the service platform can read the conditions that meet each evidence requirement and, based on the evidence source type, verification time range, target security object range, and evidence field conditions, query the candidate evidence acquisition capabilities that have been registered in the current network security incident handling system.
[0152] For example, when a certain evidence requirement calls for obtaining the identity authentication records of a target asset within a specific time range, the online log query capability that can query the identity authentication logs can be identified as a candidate evidence acquisition capability; if the required identity authentication records have already been stored offline, the offline archive retrieval capability can also be identified as a candidate evidence acquisition capability.
[0153] For example, when a certain evidence requirement demands the acquisition of the real-time running status or memory information of the target process, and the existing log data cannot meet the corresponding evidence requirement, the evidence acquisition capability that can perform active data acquisition through the terminal security component can be identified as a candidate evidence acquisition capability, and an evidence acquisition relationship can be established between the corresponding evidence requirement node and the candidate evidence acquisition capability node.
[0154] In some embodiments, an evidence requirement can simultaneously establish evidence acquisition relationships with multiple candidate evidence acquisition capabilities. For example, for network communication-related evidence requirements, evidence acquisition relationships can be simultaneously established with terminal network connection query capabilities, network traffic query capabilities, and offline traffic archive retrieval capabilities. When generating specific investigation tasks subsequently, the actual data acquisition channels to be used can be further determined based on the current online data status, data access permissions, data retention status, and degree of intrusion for each candidate evidence acquisition capability.
[0155] Furthermore, when no available candidate evidence acquisition capability can be determined based on the fulfillment conditions of the evidence requirement item, the absence of an available evidence acquisition path for that evidence requirement item can be recorded in the investigation hypothesis verification graph, but this does not necessarily mean that the evidence requirement item is satisfied. The candidate evidence acquisition capability can be re-determined subsequently based on the secure data source access status or changes in access permissions.
[0156] In this specification, based on steps C2 to C6, by establishing a requirement verification relationship between the investigation hypothesis data and the evidence requirement items, establishing a supporting evidence relationship or a rebuttal relationship between the evidence requirement items and existing security evidence, and establishing an evidence acquisition relationship between the evidence requirement items and the ability to acquire candidate evidence, a complete verification link can be formed in the investigation hypothesis verification graph, from the investigation hypothesis, evidence requirements, existing evidence to the subsequent evidence acquisition path. This enables the service platform to directly locate the evidence gap and available supplementary evidence capabilities corresponding to a specific investigation hypothesis based on the graph node relationships.
[0157] Optionally, the specific execution of causal tracing along the object relationships, and the determination of the satisfaction status of each of the evidence requirements based on the obtained tracing evidence, includes: S302: Based on the preset evidence role corresponding to the survey hypothesis data, determine the candidate attack phase corresponding to the preset evidence role in the security object; The candidate attack phase characterizes the range of objects relevant to the corresponding attack phase, selected from multiple security objects and their behavioral relationships involved in a cybersecurity incident based on the currently unverified evidence role. The candidate attack phase limits the starting range of subsequent object relationship searches to avoid indiscriminately expanding to all security objects involved in the cybersecurity incident.
[0158] Indicatively, the service platform can obtain the investigation hypothesis data and preset evidence roles corresponding to the current evidence requirement item to be verified, and determine the security objects that are semantically related to the attack behavior from the security objects and object relationships constructed in step S102 based on the attack behavior represented by the preset evidence role.
[0159] For example, when the default evidence role is execution behavior, security objects related to process startup, script execution, file loading, or command execution can be identified; when the default evidence role is persistence, security objects related to startup items, services, scheduled tasks, or other security objects that can affect the continuous running state can be identified; when the default evidence role is lateral movement, security objects related to account authentication, remote connections, and access behavior between different assets can be identified; when the default evidence role is command and control, security objects related to target processes, target assets, and external network connections can be identified.
[0160] Furthermore, the security objects can be filtered based on the investigation anchor points corresponding to the investigation hypothesis data. The investigation anchor points can be anomalous objects or behaviors that trigger the current cybersecurity incident investigation, such as anomalous processes, anomalous accounts, anomalous network connections, target assets, or threat indicators. The service platform can filter security objects that match the currently preset evidence roles from those that have object relationships with the investigation anchor points, thereby determining candidate attack phases.
[0161] In one feasible implementation, the degree of matching between a security object and a preset evidence role can be determined based on the object type, object attributes, object relationship type with other security objects, and event time of the security object, and the security object that meets the semantic requirements of the corresponding role can be used as the security object corresponding to the candidate attack stage.
[0162] For example, if the current investigation assumes that the data is used to verify the attack process of "establishing external communication and further lateral movement after the execution of an abnormal script on the target asset", and if the preset evidence role to be verified is lateral movement, the service platform can identify security objects related to the target account, source asset, target asset, and remote connection behavior from the constructed security objects, and identify these security objects as the candidate attack phase corresponding to lateral movement, without having to expand other security objects unrelated to file persistence or data access at this phase.
[0163] Therefore, semantic filtering can be performed on the scope of attack behaviors to be searched based on the investigation hypothesis data and the preset evidence roles, so that subsequent causal tracing can be expanded around the attack stage corresponding to the current evidence requirement item.
[0164] S304: Based on the security objects and event time ranges corresponding to the candidate attack stages, search backward along the object relationships for the predecessor security objects of the security objects, and search forward along the object relationships for the successor security objects generated or affected by the security objects, so as to form candidate tracing paths; Precursor security objects are used to characterize security objects that precede the current security object in the current object relationship chain in terms of time or causal relationship, and may form a causal relationship with the current security object's generation, state change, or behavior.
[0165] Subsequent security objects are used to represent security objects that are further generated, triggered, or affected by the behavior or state of the current security object in the current object relationship chain.
[0166] Candidate tracing paths are used to characterize a sequence of object relationships to be verified, starting from the security object corresponding to the candidate attack stage and connected by one or more object relationships that meet the search criteria. At this stage, candidate tracing paths are used to limit the range of object relationships that require further verification of original security evidence, and do not necessarily indicate that the path has been determined as the actual attack path.
[0167] Schematic, the service platform 100 can use the security object corresponding to the candidate attack stage determined in step S302 as the current search object, and determine the event time range by combining the verification time range of the corresponding evidence requirement item. Subsequently, it performs reverse search and forward search along the established object relationship.
[0168] During the reverse search process, one can start from the current security object and search for predecessor security objects along object relationships that can characterize the cause, triggering source, or preceding behavior. For example, for an abnormal process, one can search backwards to find the parent process that started the abnormal process; for an abnormal script, one can search backwards to find the process or file that downloaded or created the script; for abnormal remote authentication, one can search backwards to find the account, source asset, or corresponding login session used for authentication.
[0169] During a forward search, you can start with the current secure object and search for subsequent secure objects along the object relationships that can characterize the results or subsequent effects of the current secure object. For example, for an abnormal process, you can continue to search for files created by that abnormal process, other processes started by that abnormal process, and network connections established by that process; for an asset that has been abnormally logged into, you can continue to search for subsequent remote connections initiated by that asset, file accesses, or the impact on other assets.
[0170] When expanding object relationships, the object relationships to be expanded can be filtered according to the current preset evidence role and the time range of the event corresponding to the evidence requirement item. For example, if the event time corresponding to a certain object relationship is clearly outside the verification time range of the current evidence requirement item, the object relationship can be excluded from the current candidate tracing path; if the behavior type represented by the object relationship does not match the current preset evidence role, the expansion along the object relationship can also be stopped.
[0171] Furthermore, when searching along object relationships, the object versions can be considered to determine whether different security objects have a comparable state basis. For example, when an object relationship is formed on a historical version of a target security object, and the current fact to be verified only applies to the changed object version, it can be determined whether this historical relationship should continue to be retained in the candidate tracing path based on specific evidence requirements. This reduces the possibility of directly splicing behavioral relationships formed under different object states into the same attack path.
[0172] In one feasible implementation, for each object relationship expansion, the matching status between the object relationship and the current preset evidence role can be verified, whether the event time corresponding to the object relationship is within the corresponding event time range, whether the object version of the associated security object meets the current verification requirements, and whether the object relationship has corresponding security observation data support. Object relationships that meet the corresponding search requirements and their associated security objects can be added to the candidate tracing path and used as the basis for the next round of reverse or forward search.
[0173] For example, for an abnormal process P1 on target asset A, a reverse search can sequentially identify the office software process P0 that started P1 and the target file F1 opened by P0; a forward search can identify the script process P2 started by P1, and further identify the connection established between P2 and an external network address. When the event times of the above object relationships are all within the current investigation time range, and the relationship types are consistent with the current investigation hypotheses, a candidate tracing path can be formed consisting of the target file F1, the office software process P0, the abnormal process P1, the script process P2, and the external network connection.
[0174] S306: For the object relationship to be verified in the candidate tracing path, obtain the corresponding original security evidence by referencing the original evidence in the security observation data associated with the object relationship to be verified, and determine the original security evidence as the tracing evidence; The object relationships to be verified are used to characterize the object relationships in the candidate tracing paths that have been established based on security observation data, but need to be further verified by combining the underlying original security data to confirm their authenticity or effectiveness.
[0175] Original security evidence can be understood as the underlying security records actually stored by the security data source and used to form the corresponding security observation data. For example, it may include terminal event records, authentication records, network connection records, traffic records, original alarm records, file forensics data, or other underlying data that can directly support the corresponding object relationships.
[0176] As an illustration, when generating security observation data in step S102, the original evidence reference between the security observation data and the original network security event record has been established. Therefore, in step S306, the service platform 100 does not need to perform a full search of all the original data in multiple security data sources again. Instead, it can determine the relationship of the object to be verified in each of the candidate tracing paths formed in step S304 and obtain the security observation data used to establish the relationship of the object.
[0177] Furthermore, original evidence references can be read from the security observation data, and the original security evidence in the corresponding security data source can be located based on the original evidence references. The corresponding original security evidence can then be obtained to verify the relationship between the objects to be verified.
[0178] For example, if the candidate tracing path contains the object relationship of "process P1 creates file F1", and the object relationship is established based on a certain terminal security observation data, the original evidence reference can be read from the security observation data to locate the original process file operation record in the terminal detection data source, and the original process file operation record can be used as the tracing evidence to verify the object relationship.
[0179] For example, if the candidate tracing path contains the object relationship of "account A logs into asset Y from asset X", the corresponding original identity authentication record can be obtained by referencing the original evidence in the identity authentication security observation data. If there is also network security observation data corresponding to the remote login, the corresponding original network connection record can also be obtained for subsequent judgment of the independence of evidence source and the satisfaction of evidence status.
[0180] In one feasible implementation, for each object relationship to be verified, it can be checked at least whether there is an original evidence reference that can locate the original security evidence. When the corresponding original security evidence can be obtained based on the original evidence reference, the obtained original security evidence can be determined as the source evidence corresponding to the object relationship to be verified; when there is no original evidence reference, the original security evidence has exceeded the retention period, the corresponding security data is currently restricted from access, or the original security evidence can only be obtained through active evidence collection, this situation can be used as the basis for subsequent source tracing breakpoint detection and evidence requirement updates.
[0181] Furthermore, if an object to be verified is associated with multiple security observation data, the original security evidence corresponding to each security observation data can be obtained separately, and the data source and source group information corresponding to each original security evidence can be retained, so as to determine in step S308 whether the different traceability evidences belong to mutually independent evidence sources.
[0182] Step S306 allows for further drilling down to the underlying raw security data of the object-level candidate tracing path formed in step S304, thereby enabling the transformation from object relationship association to raw evidence verification.
[0183] S308: Based on the evidence source, event time, associated security object, object version and evidence fields corresponding to the traceability evidence, determine the degree to which the traceability evidence meets the corresponding evidence requirement item, and determine the satisfaction status of the evidence requirement item based on the degree to which it meets the requirement item.
[0184] After obtaining the source tracing evidence corresponding to the candidate source tracing path, the source tracing evidence can be matched item by item according to the pre-determined conditions for each evidence requirement item in step S202.
[0185] As an illustration, one can first determine whether the source of the traceable evidence belongs to the type of evidence source allowed for the corresponding evidence requirement based on its source of evidence, and then combine this with the aforementioned source group to determine whether different traceable evidence meets the requirements for an independent source of evidence.
[0186] Furthermore, the event time of the source tracing evidence can be matched with the verification time range corresponding to the evidence requirement to determine whether the source tracing evidence falls within the time interval allowed for verifying the target fact. For source tracing evidence whose event time does not fall within the corresponding verification time range, it can be retained as investigation data but not used to meet the current evidence requirement.
[0187] Furthermore, it can be determined whether the traceability evidence belongs to the target security object range corresponding to the evidence requirement based on the security object associated with it. For example, if a certain evidence requirement is used to verify the remote identity use of a target account on a target asset, then traceability evidence that is only associated with other accounts or other unrelated assets is not used to satisfy that evidence requirement.
[0188] Service platform 100 can also determine whether the traceability evidence is applicable to the object state that needs to be verified in the current evidence requirement item based on the object version of the associated security object. If the object version corresponding to the traceability evidence is the same as the version of the object to be verified, or if the version change does not affect the object attributes or object relationships on which the corresponding evidence requirement item depends, it can continue to be used for the judgment of the current evidence requirement item; if the relevant object attributes or object relationships have changed and affected the current verification facts, the correspondence between the traceability evidence and the historical object version can be retained, and direct use with the changed current object version should be avoided.
[0189] Furthermore, the evidence fields in the source tracing evidence can be matched with the evidence field conditions corresponding to the evidence requirement items. For example, it can be determined whether fields such as account identifier, source asset identifier, target asset identifier, process identifier, file identifier, network address, operation type, and behavior result meet the requirements of the current target fact to be verified.
[0190] After completing the above matching, we can obtain a comprehensive picture of the satisfaction of the source evidence and the corresponding evidence requirements.
[0191] For example, a certain requirement for lateral movement evidence includes the following: the evidence event time is within the time range T1 to T2; the associated security objects include source asset A, target asset B, and target account U; and at least two independent evidence sources are required. If the first source evidence obtained is a successful remote login record generated by an identity authentication data source, and the second source evidence is the corresponding remote connection record from A to B generated by a network traffic data source, and both pieces of evidence belong to different source groups, the event time is within T1 to T2, and the associated objects are consistent with the target objects, then the source evidence can be determined to meet the independent source, time range, object range, and field condition requirements of the corresponding evidence requirement, and the corresponding evidence requirement can be determined to be satisfied.
[0192] If the obtained traceability evidence only covers a portion of the time frame, a portion of the target security object, or a portion of the target facts required by the evidence requirement, then it can be determined that the evidence requirement is partially satisfied, and the evidence requirement corresponding to the uncovered portion will continue to be retained.
[0193] If multiple pieces of source evidence have been obtained, but the corresponding evidence depends on the same underlying source record or the same parsing chain, then even if the number of evidences meets the preset requirements, the corresponding evidence requirement item can still be determined as a source-dependent state according to the independent evidence source requirements.
[0194] If tracing evidence from independent sources produces results that cannot be simultaneously established for the same objective fact, then the corresponding evidence requirements are in conflict.
[0195] If valid traceability evidence that meets the corresponding verification time range, target security object range, or evidence field conditions has not yet been obtained, the corresponding evidence requirement item can be kept in an unmet state, and further detection can be carried out in conjunction with subsequent implementation examples to determine whether a traceability breakpoint that needs to be supplemented is formed.
[0196] Additionally, when the event time of the source tracing evidence exceeds the corresponding freshness requirement, or when the associated security object undergoes changes in object attributes directly related to the current evidence assessment, the corresponding evidence requirement can be determined as invalid. It should be noted that not every change in the version of a security object will invalidate existing source tracing evidence; rather, it allows for further assessment of whether the changed object attributes or relationships are actually part of the verification content upon which the current evidence requirement depends.
[0197] Furthermore, once step S308 determines the satisfaction status of the new evidence requirement item, the corresponding matrix cells in the evidence sufficiency matrix and the node relationships between the evidence requirement item and the source evidence in the investigation hypothesis verification graph can be updated simultaneously. If the updated satisfaction status indication still indicates situations such as missing evidence, partial coverage, evidence conflict, source dependence, invalid evidence, or unavailable original evidence, then the subsequent source breakpoint detection and investigation task generation process can proceed.
[0198] Indicatively, steps S302 to S308 can be repeated as new security observation data, supplementary security evidence, or changes in the version of the security object occur. When subsequent investigation tasks obtain new supplementary security evidence, the supplementary security evidence can be reintegrated into the evidence set corresponding to the current candidate tracing path, and step S308 can be executed again to re-determine the satisfaction status of the evidence requirements. When new security evidence changes the original candidate attack stage or object relationship, steps S302 and S304 can be re-executed to form an updated candidate tracing path, thereby continuously updating the causal tracing results as the security evidence and the status of the security object change during the investigation.
[0199] In this specification, based on steps S302 to S308, candidate attack stages are determined according to the preset evidence roles corresponding to the investigation hypothesis. Based on the security objects corresponding to the candidate attack stages, a bidirectional search is performed along the object relationships to form candidate tracing paths. Then, the original evidence references are used to perform low-level original security evidence verification on the object relationships to be verified in the candidate tracing paths. The satisfaction status of evidence requirements is determined by combining evidence source, event time, associated objects, object version, and evidence fields. This can form a progressively converging tracing process for semantic filtering of investigation hypotheses, object-level causal relationship expansion, and original security evidence verification, reducing the amount of irrelevant data processing generated by directly performing full retrieval of multi-source original security data.
[0200] In one feasible implementation, the determination of the satisfaction status of the evidence requirement item based on the satisfaction status described in S308 can be carried out in the following manner: Step D2: For the relationships between objects to be verified in the candidate tracing paths, verify the integrity of the predecessor security evidence, the event time of the tracing evidence, the object version of the associated security object, the independence of the evidence source, and the availability of the original security evidence to obtain the verification results; The completeness of preceding security evidence is used to characterize whether the security evidence used to explain or support the formation of the relationship between the objects to be verified is complete. For example, for the relationship between the target process and the script process to be verified, it can be further determined whether there is evidence of a parent process, file source, or other preceding behavior that explains the origin of the target process's own startup. If the current candidate tracing path only records that the script process has been started, but lacks preceding evidence that can explain what object or behavior triggered it, then it can be considered that there is a gap in preceding security evidence at the corresponding position.
[0201] The availability of original security evidence characterizes whether the underlying security data required to verify the relationship between the objects to be verified can be obtained through original evidence citation or the corresponding candidate evidence acquisition capability. The availability of original security evidence can be affected by factors such as data retention period, access permissions, data storage status, and whether proactive evidence collection is required.
[0202] As an illustration, the service platform can determine the relationships of the objects to be verified one by one for the candidate tracing paths formed in step S304, and perform multi-dimensional verification for each relationship of the objects to be verified.
[0203] First, based on the object relationships and tracing evidence preceding the current object relationship in the candidate tracing path, determine whether there is any preceding security evidence that can support the reason for the formation of the current object relationship. If there is an evidence gap in the preceding link of the current object relationship, the corresponding preceding evidence missing information can be recorded in the verification result.
[0204] Next, the event time of the tracing evidence corresponding to the relationship between the objects to be verified is compared with the verification time range corresponding to the current evidence requirement to determine whether the tracing evidence is within the time range allowed for verifying the target fact. For tracing evidence that exceeds the corresponding time range or no longer meets the current evidence time requirements, the corresponding time verification result can be recorded in the verification results.
[0205] Furthermore, the version of the security object associated with the traceability evidence can be obtained, and combined with the object status that needs to be verified in the current evidence requirement, it can be determined whether the traceability evidence is comparable to the current object version. For example, when the traceability evidence is associated with a historical object version, and the corresponding object attributes or object relationships have changed in subsequent versions that are directly related to the current verification facts, it can be determined that the traceability evidence cannot be directly used for verification under the current object version.
[0206] The service platform can determine the independence of evidence sources based on the source group to which each piece of traceability evidence belongs. For example, although multiple pieces of traceability evidence may be generated by different alarm rules, if they are all based on the same underlying log record or the same parsing chain, then it can be determined that the multiple pieces of traceability evidence do not meet the requirement of independent evidence source.
[0207] Furthermore, it can be determined whether the original security evidence is still accessible based on the original evidence citations and the current status of the corresponding security data source. For example, it can be determined whether the original security evidence is still within the data retention period, whether there are usable original evidence citations, whether the current account has access permissions, and whether active forensic evidence collection through the terminal is required to further obtain the corresponding security data.
[0208] Based on the above verification, a verification result can be generated for the relationship between the objects to be verified. The verification result can be used to indicate whether there are situations such as missing precursor evidence, conflicting independent evidence, source dependence, expiring original evidence, missing original evidence citations, restricted access, or need for active evidence collection, and to provide subsequent steps D4 to D10 with information to determine the status of the corresponding evidence requirement item and the type of tracing breakpoint.
[0209] Step D4: When the verification result indicates that the relationship between the objects to be verified lacks corresponding predecessor security evidence, the corresponding evidence requirement item is determined to be unmet, and the corresponding position is determined as the source breakpoint of the predecessor evidence missing type. The type of source tracing breakpoint with missing precursor evidence is used to characterize the current candidate source tracing path that has located a certain security object or object relationship, but lacks security evidence that can continue to explain the cause of its formation or verify its preceding attack behavior, thus preventing the current attack chain from continuing forward to complete the evidence verification.
[0210] As an illustration, the service platform can determine whether the preceding relationship corresponding to the current object relationship to be verified has the support of the corresponding original security evidence based on the verification result obtained in step D2.
[0211] For example, if the target script process P2 has been identified as having established a connection with an external network address in the candidate tracing path, and tracing evidence related to the network connection can be obtained, but the preceding security evidence that can prove which parent process started P2 or which file triggered its execution cannot be obtained, then it can be determined that there is a lack of preceding evidence at the preceding position of P2 in the current path.
[0212] In this case, the evidence requirement item corresponding to the preceding behavior can be determined as unmet, and the location of the missing preceding security evidence can be recorded as a source breakpoint of the preceding evidence missing type.
[0213] Furthermore, the source tracing breakpoint can be associated with the corresponding evidence requirement, associated security object, target time range, and existing security evidence, so that subsequent investigation tasks targeting missing precursor evidence can be directly generated based on the source tracing breakpoint.
[0214] For example, for the source breakpoint of the missing parent process startup record, it is possible to further determine the process creation record, script execution record or related file source record of the target asset within the corresponding time range, so that the subsequent re-certification process can be executed around the specific missing position, rather than re-executing the full log query of the entire network security incident.
[0215] Step D6: When the verification result indicates that the source evidence from mutually independent sources has mutually exclusive results for the same target fact, the corresponding evidence requirement item is determined to be in conflict state, and the corresponding position is determined to be the source breakpoint of the evidence conflict type. Mutually exclusive results are used to characterize two or more sources of evidence from independent sources that provide verification results for the same objective fact that cannot be simultaneously true.
[0216] The source breakpoint of the evidence conflict type is used to characterize that the current candidate source path is not lacking in security evidence, but rather that there are multiple security evidences with independent source basis, but the security evidences give inconsistent results for the same fact, which makes it impossible to directly determine the position of whether the corresponding evidence requirement is met.
[0217] As an illustration, the service platform can first determine whether multiple pieces of tracing evidence for the same target fact come from independent sources based on the source group information. After confirming that the sources are independent, it can then compare the verification results given by each piece of tracing evidence for the target fact.
[0218] For example, the first piece of evidence comes from an identity authentication data source, recording that target account U successfully logged into asset B at a target time; the second piece of evidence comes from another independent security data source, recording that the same target account U did not have a valid authentication session for asset B at the corresponding time. If the objects, time ranges, and target facts of the two pieces of security evidence are all related, and the two cannot be simultaneously true, then it can be determined that there is a conflict of evidence regarding the target fact.
[0219] In this case, instead of directly selecting one piece of security evidence as the final conclusion, and instead of simply determining the corresponding evidence requirement as either satisfied or unsatisfied, the corresponding evidence requirement can be determined as a conflict state, and the location corresponding to the conflicting target fact can be determined as the source breakpoint of the evidence conflict type.
[0220] Furthermore, supporting and contradictory evidentiary relationships can be preserved, and the need to obtain independent third-party evidence, re-acquire original logs, or verify object versions and time bases from different security data sources can be determined based on the cause of the conflict. This allows for further resolution of the conflict through subsequent investigation tasks.
[0221] Step D8: When the verification result indicates that the source of evidence used to verify the relationship between the objects to be verified does not meet the independence requirement, the corresponding evidence requirement item is determined as a source dependency state, and the corresponding position is determined as a source dependency type source breakpoint; Source-dependent breakpoints are used to characterize the locations where a certain amount of source-tracing evidence has been obtained, but the source-tracing evidence has the same underlying data source or the same parsing chain dependency, and cannot meet the independent evidence source requirement preset by the current evidence requirement item.
[0222] For example, a certain evidence requirement demands security evidence from at least two independent sources to verify whether target process P1 executes target file F1. The current service platform receives the first endpoint security alert and the second security information and event management alert; both records indicate that P1 executes F1.
[0223] Further verification based on the source group identifier revealed that both alarms originated from the same terminal process log forwarded through different detection rules or platforms. Therefore, although there are two pieces of evidence in form, they essentially rely on the same underlying source record. At this point, it can be determined that while the number of pieces of evidence meets the preset requirement, the independence of the evidence source does not meet the condition.
[0224] Based on this, the service platform can determine the corresponding evidence requirement as a source dependency state and determine the corresponding object relationship or target fact location as a source dependency type tracing breakpoint.
[0225] Furthermore, based on this source tracing breakpoint, the ability to obtain candidate evidence that can provide independent evidence can be redefined. For example, if both existing pieces of evidence come from terminal logs, network traffic data, identity authentication data, or other independent security data sources can be selected as subsequent investigation directions, thereby increasing the number of truly independent evidence sources.
[0226] The state of source dependence differs from the state of evidence conflict. A state of source dependence indicates that the verification results of existing evidence are consistent, but the sources of that evidence are not sufficiently independent. A state of evidence conflict, on the other hand, indicates that independent sources of evidence yield mutually exclusive results regarding the same objective fact. These two states can respectively form different types of source tracing breakpoints and correspond to different subsequent supplementary evidence strategies.
[0227] Step D10: When the verification result indicates that the original security evidence has reached the critical condition of the preset retention period, the relationship between the objects to be verified lacks original evidence references, the data access required for verification is restricted, or the data of the target security object needs to be actively obtained for further verification, update the satisfaction status of the corresponding evidence requirement item and determine the corresponding type of tracing breakpoint.
[0228] The preset retention period threshold condition is used to characterize the original security evidence, which, although still existing or accessible, has reached a critical state where its remaining retention time has reached the point where it needs to be prioritized for fixation or retrieval.
[0229] As an illustration, it can be determined whether the preset retention period threshold has been reached based on the data retention policy of the corresponding secure data source, the current time, and the retention deadline of the original secure evidence. When the original secure evidence is about to be archived, overwritten, or deleted, the corresponding evidence requirement item can be updated to a state that requires priority for supplementary evidence or has an impending expiration risk, and the corresponding evidence retention impending expiration type tracing breakpoint can be determined.
[0230] When the object relationship to be verified lacks original evidence references, it can be understood that the current security observation data or object relationship already exists, but its underlying original security record cannot be directly located through the corresponding security observation data. In this case, although the object relationship has been formed in the current graph structure, there is a lack of original evidence links that can further verify the object relationship. The corresponding evidence requirement item can be kept in an unmet or unverified state, and this position can be identified as a source breakpoint of the missing original evidence reference type.
[0231] When access to the data required for verification is restricted, it can be understood that the original security evidence actually exists, but the current network security incident handling process cannot directly obtain the original security evidence due to permissions, sensitive data access control, or other access constraints.
[0232] For example, if a certain identity authentication record or sensitive asset audit record is only accessible after a preset authorization, the corresponding evidence requirement item can be updated to an access-restricted state, and the corresponding location can be identified as a source breakpoint for the data access-restricted type. Subsequently, the appropriate permission application or approval process can be triggered around this source breakpoint, instead of incorrectly determining that the corresponding evidence requirement item does not contain security evidence.
[0233] For cases where continued verification requires actively obtaining data from the target security object, it can be understood that existing offline logs, online indexes, or archived data cannot meet the corresponding evidence requirements, and it is necessary to obtain the current status data directly from the target security object through endpoint security components or other security devices.
[0234] For example, if the current investigation needs to verify the real-time memory status of the target process, but the existing logs only record the existence of the process and cannot obtain the network connection parameters or loaded module information in memory, then it can be determined that further verification requires actively obtaining the data of the target security object, and the corresponding position is determined as the source tracing breakpoint of the active evidence collection type.
[0235] Furthermore, the different types of source tracing breakpoints in step D10 can correspond to different evidence requirement statuses and different subsequent evidence acquisition methods. For example, if the original security evidence reaches the retention period threshold, the priority of the corresponding evidence requirement or investigation task can be increased; if the original evidence citation is missing, other candidate data sources can be searched or the original record can be relocated from the associated security object; if data access is restricted, the authorized acquisition path can be entered; if it is necessary to actively acquire the target security object's data, an active investigation task with the corresponding maximum evidence intrusion level can be generated.
[0236] It should be noted that the various situations listed in step D10 can occur individually or simultaneously. For example, if a piece of original security evidence is both at the critical point of its retention period and cannot be directly obtained due to access restrictions, then the risk of the evidence expiring and the access restriction status can be recorded simultaneously. The execution priority of subsequent investigation tasks can then be determined by combining the urgency of the event, the importance of the evidence, and the data retention period.
[0237] In this specification, based on steps D2 to D10, the relationships between objects to be verified in the candidate tracing paths are checked from the dimensions of completeness of preceding evidence, time validity, applicability of object versions, independence of evidence sources, and availability of original evidence. Different abnormal verification results are mapped to different types of tracing breakpoints, such as missing preceding evidence, evidence conflict, source dependence, evidence retention nearing expiration, missing original evidence citation, restricted data access, and active evidence collection. This makes the specific reasons why the evidence verification in the candidate attack path cannot be completed explicit and associated with the corresponding evidence requirements. This provides a basis for generating targeted investigation tasks based on different breakpoint types, reduces indiscriminate supplementary evidence collection, and improves the accuracy of evidence gap location and supplementary evidence collection in the process of cybersecurity incident investigation.
[0238] In one feasible implementation, the step of generating an investigation task including an expected version of the target security object for the unmet evidence requirement item corresponding to the detected source tracing breakpoint, and executing the investigation task to obtain supplementary security evidence, includes: S402: Based on the unmet evidence requirements, determine the target security object, the expected version of the target security object, the investigation time range, the candidate data source, the expected result structure, the retention deadline, the maximum evidence collection intrusion level, and the result write-back target, so as to generate the corresponding investigation task. The investigation time range is used to characterize the time interval within which security data needs to be queried or obtained when performing the current investigation task. The investigation time range can be determined based on the verification time range of the corresponding evidence requirement item, the event time corresponding to the source tracing breakpoint, and the related event time in the candidate source tracing path. This ensures that the investigation task obtains security data within the time range related to the current evidence gap, rather than indiscriminately querying all historical security data in the candidate data source.
[0239] Candidate data sources are used to characterize data sources that may preserve or provide secure evidence to meet currently unmet evidence requirements. Candidate data sources can be determined based on the evidence acquisition relationship between evidence requirements in the aforementioned survey hypothesis verification map and the ability to acquire candidate evidence.
[0240] For example, if the evidence requirement that is not currently met is used to verify whether the target account has engaged in remote authentication behavior within a specific time range, the candidate data source may include the identity authentication data source, the endpoint security data source of the target asset, and the network traffic data source that can record relevant network connection behavior; if the evidence requirement that is not currently met is used to verify the running status of the target process at a specific time, the candidate data source may include the endpoint detection data of the target asset or a security device that can obtain the corresponding running status data through proactive evidence collection.
[0241] The expected result structure is used to characterize the data content or data fields that the security data expected to be obtained after the investigation task is completed. The expected result structure can be determined based on the evidence field conditions of the corresponding evidence requirement item.
[0242] For example, for investigation tasks related to remote authentication, the expected result structure could require information such as account identifier, source asset identifier, target asset identifier, authentication time, and authentication result. For investigation tasks related to process network communication, the expected result structure could require information such as process identifier, target network address, connection time, and connection status. By setting the expected result structure in the investigation task, the subsequently returned investigation results can be directly matched and verified with the corresponding evidence requirements.
[0243] The retention deadline is used to characterize the expected retention period of the security data to be acquired in the corresponding candidate data source, or the time constraint for obtaining evidence before that time. The service platform can determine the retention deadline by combining the retention status of the original security evidence corresponding to the tracing breakpoint.
[0244] For example, when step D10 has determined that a certain original security evidence has reached the critical condition of the preset retention period, the retention deadline of the corresponding data can be written into the investigation task, so that subsequent task scheduling can identify the risk that the data in the investigation task will be overwritten, deleted or migrated.
[0245] The maximum forensic intrusion level limits the maximum state impact allowed on the target security object during the execution of an investigation task. This forensic intrusion level can be used to distinguish between different acquisition methods, such as simply querying existing log data and actively accessing the target security object to obtain real-time data.
[0246] For example, when the maximum level of intrusion for forensic investigation is low, investigation can be completed using existing online logs, historical network traffic, or offline archived data; when the corresponding evidence requirements cannot be met by existing data, and a higher degree of proactive forensics is allowed, process status, memory status, or other data that needs to be actively collected can be obtained from the target security object through endpoint security components.
[0247] The result write-back objective is used to characterize the location of survey data that needs to be associated with or updated after the completion of the investigation task. The result write-back objective may include at least one of the following: corresponding evidence requirement items, corresponding matrix units in the evidence sufficiency matrix, and corresponding evidence requirement nodes in the investigation hypothesis verification map.
[0248] As an illustration, after detecting that a certain lateral movement evidence requirement is not met, the service platform 100 can read the target security object, object version, verification time range, evidence source type, and evidence field conditions corresponding to the evidence requirement. Based on this, it can determine the target asset A and the target account U as the target security objects, determine the current corresponding object version as the expected object version, determine the attack behavior time range to be verified as the investigation time range, determine the identity authentication data source and network traffic data source as candidate data sources, and determine the corresponding expected result structure, retention deadline, maximum forensic intrusion level, and result write-back target, thereby generating an investigation task to supplement the verification of the lateral movement behavior.
[0249] By incorporating the above information into the investigation task, the subsequent evidence acquisition process can be clearly linked to specific evidence gaps, target security objects, and object versions.
[0250] S404: Determine the data acquisition channel corresponding to the investigation task based on the online status, data access permissions, data retention status, and the maximum evidence intrusion level of the candidate data source; The online status of data is used to characterize whether the target security data in the candidate data source is currently in an online storage state that can be directly queried or accessed.
[0251] Data access permissions are used to characterize whether the current network security incident handling process has the authority to access the corresponding candidate data source or target security data. When the corresponding security data is sensitive data or data subject to access control, the current access permissions can be used to determine whether the corresponding data acquisition method can be used directly.
[0252] Data retention status is used to characterize whether the target security data is still within the scope of effective retention. For example, it can include online retention, offline archiving, nearing the retention period, or being unavailable.
[0253] A data acquisition channel is used to characterize the specific acquisition path for obtaining corresponding security data from candidate data sources when actually performing an investigation task. In some embodiments, a data acquisition channel may include an online index acquisition channel, an offline archive acquisition channel, a restricted data authorization acquisition channel, and an active forensics acquisition channel, etc.
[0254] As an illustration, the service platform 100 can sequentially or in combination determine the online status of candidate data sources, data access permissions, data retention status, and maximum level of intrusion for evidence collection.
[0255] When the target security data is still online and the corresponding data access permissions are currently available, the online index acquisition channel can be determined. For example, one can directly query online stored terminal event records, authentication records, or network traffic index data.
[0256] When the target security data is no longer online but is still stored in offline archives or cold storage, the offline archive retrieval channel can be determined, and the corresponding security data can be retrieved through offline data retrieval.
[0257] When the target security data actually exists but access is restricted, the restricted data authorization acquisition channel can be determined, so that the investigation task can continue to be executed after obtaining the corresponding data access permissions.
[0258] When existing online or offline archived data cannot meet the corresponding evidence requirements, and it is necessary to obtain current status data directly from the target security object, the decision on whether to use an active evidence acquisition channel can be made based on the maximum evidence intrusion level. Only when the evidence intrusion level corresponding to the active evidence acquisition method does not exceed the maximum evidence intrusion level will the active evidence acquisition method be determined as a data acquisition channel usable for the current investigation task.
[0259] For example, for an investigation task that requires obtaining the current running status of a target process, if the existing terminal logs cannot provide the corresponding data, but the maximum forensic intrusion level allows for the active collection of process information from the target asset, then the active forensic data acquisition channel can be determined through the terminal security component; if the current maximum forensic intrusion level does not allow for the execution of this active collection method, then this acquisition channel will not be determined as an available data acquisition channel for the current task.
[0260] Furthermore, when multiple data acquisition channels exist for an investigation task, the actual acquisition channel to be executed can be determined by combining the retention deadline for the corresponding security data and the status of the candidate data sources. For example, a channel that can complete the acquisition of security data before the data retention deadline can be selected.
[0261] Step S404 can further transform the candidate data sources determined in step S402 into evidence acquisition paths that can actually be executed.
[0262] S406: Execute the investigation task through the data acquisition channel to obtain supplementary security evidence corresponding to the unmet evidence requirements; As an illustration, after determining the data acquisition channel, the service platform can obtain security data from the corresponding candidate data source according to the target security object, expected object version, investigation time range, and expected result structure recorded in the investigation task.
[0263] If an online index retrieval channel is used, the corresponding online security data can be queried based on the target security object and the investigation time range, and data that meets the expected result structure can be extracted from the query results.
[0264] If an offline archiving acquisition channel is used, the corresponding archived data can be located based on the target security object and time range, and security records related to the current evidence requirement can be recovered or read from the archived data.
[0265] If a restricted data access channel is used, data queries can be performed after obtaining the corresponding data access permissions, and the results can be associated with the current investigation task.
[0266] If an active evidence collection channel is used, the required security data can be obtained from the target security object specified in the investigation task through the corresponding security device or terminal security component.
[0267] After obtaining the investigation results, they can be processed according to the unified data structure of the aforementioned security observation data, and an original evidence reference can be established between them and the original investigation results. Simultaneously, supplementary security evidence obtained can be associated with the corresponding cybersecurity incident investigation context based on the investigation identifier, evidence requirement identifier, target security object, and expected object version in the investigation task.
[0268] Furthermore, the completion of the investigation task does not mean that the corresponding evidentiary requirements have been met. The investigation results still need to be re-verified in subsequent processing, taking into account the expected version of the object, the current version of the object, the independence of the evidence source, the scope of evidence coverage, and any conflicts in evidence.
[0269] For example, if the investigation successfully obtains security data, but the data only covers a portion of the required investigation timeframe, the corresponding evidence requirements can still be partially met; if the new security evidence obtained during the investigation comes from the same source group as the existing security evidence, the requirement for an independent source of evidence may still not be met.
[0270] S408: When there is no data acquisition channel that satisfies the investigation task, record the unavailability status of the corresponding evidence requirement item and keep the evidence requirement item in an unsatisfied state.
[0271] The "unavailable" status indicates that although the current investigation task corresponds to a clear evidence gap, based on existing candidate data sources, data access permissions, data retention status, and the permissible level of forensic intrusion, there is currently no data acquisition channel that can actually obtain the required secure evidence.
[0272] For example, if the target security data has exceeded the data retention period and has not been archived, and the target security object is offline, making it impossible to obtain the corresponding data again through proactive evidence collection, then it can be determined that there is currently no data acquisition channel that meets the requirements of the investigation task.
[0273] For example, verifying the target facts requires active forensics from the target security object, but the maximum allowed level of forensic intrusion is lower than the level of forensic intrusion required by this active forensics method, and there is no other online or offline data that can provide corresponding security evidence, then it can be determined that there is currently no available data acquisition channel for this investigation task.
[0274] Once it is determined that there is no data acquisition channel that satisfies the investigation task, the unavailability status corresponding to the investigation task can be recorded, and the correlation between the unavailability status and the corresponding evidence requirement item can be recorded simultaneously.
[0275] Understandably, the "unavailable" status reflects that "the required security evidence cannot be obtained at present," and does not indicate that the corresponding evidence requirement has been verified. Therefore, the corresponding evidence requirement should not be updated to a "satisfied" status simply because the evidence cannot be obtained.
[0276] For example, a core evidence requirement may be to obtain the process memory data of the target asset at the time of a cybersecurity incident. However, the target asset has been restarted and memory fixation was not completed beforehand. In this case, even if there is no longer a data acquisition channel to re-obtain the historical memory state, the corresponding evidence requirement should remain in an unmet state, and the reason for its unavailability should be recorded. Instead of converting "unable to obtain evidence" into "the corresponding target fact has been confirmed", the requirement should not be changed to "the corresponding target fact has been confirmed".
[0277] Optionally, the unavailability of evidence does not necessarily mean the termination of the entire cybersecurity incident investigation process. The corresponding evidence requirements can still be retained in the evidence sufficiency matrix and the investigation hypothesis verification map to characterize the current evidence boundaries for the investigation conclusions.
[0278] When subsequent candidate data sources come online again, access permissions change, new data archives are accessed, or new candidate evidence acquisition capabilities emerge, the data acquisition channel for evidence requirements that are currently unavailable can be reassessed, and the corresponding investigation tasks can be regenerated or continued when a new data acquisition channel can be determined.
[0279] Optionally, when a cybersecurity incident poses a high security risk and corresponding evidence remains unavailable for an extended period, the incident can be handled by combining the candidate action determination logic from subsequent embodiments, rather than waiting indefinitely for unobtainable evidence. In this case, the investigation status of unmet evidence requirements and unobtainable evidence can be retained, and its impact on the current investigation conclusion can be recorded in the relevant logs.
[0280] In this specification, based on steps S402 to S408, the specific evidence gaps identified in the aforementioned causal tracing process can be transformed into targeted evidence acquisition tasks that can be actually executed. At the same time, when there is no available data acquisition channel, the corresponding evidence requirement item remains unmet, thereby avoiding the misinterpretation of evidence unavailability as evidence being met, and providing a data foundation for subsequent object version verification, supplementary security evidence write-back, and evidence boundary preservation in investigation conclusions.
[0281] In one feasible implementation, the specific execution of determining the applicable object version of the supplementary security evidence based on the expected object version and the current object version of the target security object, and updating the evidence sufficiency matrix and the investigation hypothesis verification map based on the supplementary security evidence applicable to the current object version, can refer to the following method: S502: Compare the expected object version corresponding to the investigation task with the current object version of the target security object; As an illustration, after obtaining supplementary security evidence through the aforementioned investigation task, the service platform can read the expected object version recorded in the investigation task and obtain the current object version corresponding to the target security object at the current investigation time, and compare the two.
[0282] The expected object version can be the object version corresponding to the target security object when the investigation task is generated. The current object version can be the latest object version corresponding to the target security object when the investigation task is completed and supplementary security evidence is returned.
[0283] Because there may be a time interval between the generation, issuance, execution, and return of results of an investigation task, the target security object may generate a new version due to new security events, security equipment taking action, or changes in object attributes during this interval. Therefore, before using supplementary security evidence to update the current investigation status, it is advisable to compare the expected object version with the current object version to determine whether the supplementary security evidence can be directly applied to the current status of the target security object.
[0284] S504: When the expected object version matches the current object version, the current object version is determined as the applicable object version for the supplementary security evidence; The applicable object version is used to characterize the security object status version that can be used to verify the corresponding evidence requirements when supplementary security evidence is available.
[0285] Schematic, when the comparison result of step S502 shows that the expected object version corresponding to the investigation task is consistent with the current object version of the target security object, it can be considered that the target security object has not undergone any state changes that need to be distinguished by object version updates from the generation of the investigation task to the return of supplementary security evidence.
[0286] At this point, the current object version can be identified as the applicable object version for supplementary security evidence, and an association relationship can be established between the supplementary security evidence and the current object version.
[0287] For example, after generating an investigation task for object version V3 of target account U, remote login records of target account U within a specified time range are obtained through an identity authentication data source. When supplementary security evidence is returned, target account U is still in object version V3, so object version V3 can be determined as the applicable object version for the supplementary security evidence, and the remote login records can be further used to determine the satisfaction status of the corresponding evidence requirements.
[0288] Through the above processing, the supplementary security evidence returned by the investigation can be bound to the corresponding object state, avoiding the neglect of object state changes based solely on the target security object identifier.
[0289] S506: When the expected object version does not match the current object version, determine whether the version change between the expected object version and the current object version involves object attributes or object relationships used to determine the satisfaction status of the corresponding evidence requirement item; The discrepancy between the expected object version and the current object version does not necessarily mean that supplementary security evidence cannot be used with the current object version. A new object version of the target security object may arise due to changes in object attributes or relationships unrelated to the current evidence requirements. Therefore, it is possible to further determine whether the specific version changes affect the current evidence assessment.
[0290] As an illustration, the service platform can obtain the version association between the expected object version and the current object version, and read the changes in object attributes or object relationships recorded in the version association.
[0291] Based on the conditions for fulfilling the corresponding evidence requirement and the current verification objective, determine the object attributes or object relationships on which the determination of the fulfillment status of the evidence requirement depends, and match them with the version change content.
[0292] For example, a certain evidence requirement is used to verify whether the target account U remotely logs into asset B from asset A within a target time range. The object attributes or object relationships used to determine whether the evidence requirement is satisfied may include account identifier, source asset, target asset, authentication relationship, and event time, etc.
[0293] If the reason for updating target asset B from object version V1 to object version V2 is solely due to a change in the asset tag or other attributes unrelated to the current authentication behavior, then the version change can be determined as a basis for judgment that does not involve the current evidence requirement.
[0294] Conversely, if the object version change is due to the target account U being disabled, the authentication session being terminated, or the authentication relationship corresponding to the target asset B changing, and the above changes directly affect the target facts that the current evidence requirement needs to verify, then it can be determined that the version change involves the object attributes or object relationships used to determine the satisfaction status of the corresponding evidence requirement.
[0295] Furthermore, the service platform can make judgments only on object attributes or object relationships that have actually changed from the expected object version to the current object version, without having to deem all previously obtained supplementary security evidence invalid simply because the object version number has changed.
[0296] S508: When the version change does not involve object attributes and object relationships used to determine the satisfaction status of the evidence requirement item, the current object version is determined as the applicable object version of the supplementary security evidence; when the version change involves object attributes or object relationships used to determine the satisfaction status of the evidence requirement item, the supplementary security evidence is associated with the corresponding historical object version, and it is determined whether to generate the corresponding supplementary evidence requirement item based on the current object version. Historical object versions are used to represent object states that were formed and preserved prior to the current object version. Historical object versions are used to save security evidence, object attributes, and object relationships corresponding to the respective historical states, to support subsequent investigation playback and the tracing of the scope of evidence application.
[0297] When step S506 determines that the version change does not affect the basis for judging the current evidence requirement, although the object version of the target security object has changed, the change will not change the verification role of the supplementary security evidence for the current evidence requirement.
[0298] At this point, the current object version can be determined as the applicable object version of the supplementary security evidence, and the supplementary security evidence can be allowed to participate in the subsequent judgment of the current evidence requirement status.
[0299] For example, the investigation task is generated for version V5 of asset A. When the investigation result is returned, asset A has been updated to V6. However, only asset description information unrelated to the current investigation has been added between V5 and V6. The current evidence requirement is used to verify whether asset A has established a specific network connection within the target time range. Therefore, it can be determined that the above version change does not affect the applicability of the network connection evidence, and V6 is determined as the applicable version of the supplementary security evidence.
[0300] When version changes involve object attributes or object relationships upon which the current evidence requirement depends, it is not advisable to directly use supplementary security evidence obtained based on historical states for the current object state. In this case, the supplementary security evidence can be associated with its actual corresponding historical object version, while retaining the verification function of the supplementary security evidence for historical investigation states.
[0301] For example, a memory forensics task is generated for object version V1 of target process P1. During the execution of the forensics task, target process P1 has been terminated, and the object version has been updated to V2. If the returned supplementary security evidence reflects the memory content of process P1 in state V1, this supplementary security evidence can be associated with the historical object version V1, rather than being directly used as the current evidence in state V2.
[0302] Furthermore, the service platform can determine whether there are still evidence gaps in the current object status that require re-acquiring security evidence, based on the changed version of the current object and the verification objectives of the original evidence requirements.
[0303] The "Supplementary Evidence Requirement" item is used to characterize the need for evidence acquisition that cannot be directly overridden by the original supplementary security evidence due to changes in the state of the target security object, and therefore needs to be re-formed or retained for the current version of the object. For example, if the original evidence requirement item needs to verify the current network connection status of the target process P1, but the network connection relationship of P1 has changed during the investigation, then the original supplementary security evidence can be retained for historical verification of version V1, while a corresponding supplementary evidence requirement item is regenerated for the current version V2 to obtain network connection evidence under the V2 state.
[0304] In one feasible implementation, if the change in the state of the current object version has made it no longer necessary to verify the target facts corresponding to the original evidence requirement, then no new supplementary evidence requirement item can be generated; if the current investigation hypothesis still requires verification of the same or related evidence roles for the changed object state, then a new supplementary evidence requirement item can be generated for the current object version.
[0305] By employing the above methods, we can preserve the validity of historical security evidence while preventing supplementary security evidence obtained in the historical object state from being incorrectly used in the current object version.
[0306] S510: Perform evidence source independence, evidence coverage and evidence conflict verification on the supplementary security evidence applicable to the current object version, and update the satisfaction status of the corresponding evidence requirement items according to the verification results, so as to update the evidence sufficiency matrix and the investigation hypothesis verification map.
[0307] Evidence coverage is used to characterize the scope of supplementary security evidence that can actually cover the time frame, target security object scope, or evidence fact scope required for verification of the evidence requirement.
[0308] As an illustration, after determining through step S504 or step S508 that the supplementary security evidence is applicable to the current object version, the service platform can add the supplementary security evidence to the evidence set corresponding to the current evidence requirement item and re-verify the validity of the evidence.
[0309] First, determine whether the supplementary security evidence meets the independent evidence source requirement with the existing security evidence based on the source group information corresponding to it. For example, if a certain evidence requirement requires at least two independent source groups to provide supporting evidence, the number of independent evidence sources can be increased if the supplementary security evidence and the existing evidence come from different underlying data sources; however, if the supplementary security evidence is returned by another system but actually depends on the same underlying source record as the existing security evidence, it cannot be counted twice in the number of independent sources.
[0310] Furthermore, the scope of supplementary security evidence coverage can be determined. For example, if a certain evidence requirement asks to verify the identity authentication behavior of target account U within the time range T1 to T4, and the existing security evidence only covers T1 to T2, while the supplementary security evidence covers T2 to T4, then the two can be combined to determine that the current evidence set has covered the complete verification time range.
[0311] If the supplementary security evidence only covers T2 to T3, then the current evidence set still does not cover the evidence requirements corresponding to T3 to T4. In this case, the corresponding uncovered parts can continue to be kept as evidence requirements to be supplemented.
[0312] A similar approach can be used to determine the scope of the target security object. For example, if the evidence requirement asks to verify the impact of assets A, B, and C, but the supplementary security evidence only covers assets A and B, then the portion corresponding to asset C can still be retained as part of the scope of supplementary evidence.
[0313] Furthermore, the supplementary security evidence can be compared with the verification results of existing security evidence for the same objective fact to determine whether there is a conflict of evidence.
[0314] When newly added security evidence is consistent with existing supporting evidence and meets the requirements for evidence source independence and evidence coverage, the corresponding evidence requirement item can be updated to a satisfied state based on the current conditions for fulfilling the evidence requirement item. When the added security evidence only covers a portion of the verification time range, a portion of the target security objects, or a portion of the evidence fields, the corresponding evidence requirement item can be updated to a partially satisfied state, while retaining the remaining uncovered scope. When the newly added security evidence increases the amount of evidence but still cannot meet the requirement for independent evidence source, the corresponding evidence requirement item can be maintained or updated to a source-dependent state. When the newly added security evidence and existing security evidence from another independent source produce mutually exclusive results regarding the same target fact, the corresponding evidence requirement item can be updated to a conflicting state.
[0315] Furthermore, the corresponding matrix element in the evidence sufficiency matrix can be located based on the updated evidence requirement status, and the supporting evidence, counter-evidence, coverage, and satisfaction status associated with the matrix element can be updated.
[0316] Meanwhile, supplementary security evidence can be added to the corresponding evidence node in the investigation hypothesis verification graph, and a supporting evidence relationship or a rebuttal relationship can be established between the evidence node and the evidence requirement node based on the verification effect of the supplementary security evidence on the evidence requirement item. If the supplementary evidence requirement item is regenerated due to changes in the object version, a corresponding evidence requirement node can be added to the investigation hypothesis verification graph, and a correlation relationship can be established between it and the corresponding investigation hypothesis data and the ability to obtain candidate evidence.
[0317] Optionally, after steps S502 to S510 complete the applicability judgment and status write-back of the supplementary security evidence object version, the aforementioned causal tracing and evidence requirement status evaluation process can be retried. If the updated evidence sufficiency matrix still contains evidence requirement items such as unmet, partially met, conflicting, or source-dependent items, the corresponding tracing breakpoints can be identified and new investigation tasks can be generated. If the supplementary security evidence causes an update to the security object or object relationship, candidate tracing paths can be re-formed based on the updated object version. Therefore, the write-back of supplementary security evidence is not a one-time static update, but can continue to enter the aforementioned investigation cycle until the corresponding investigation termination conditions are met or there is no further executable evidence acquisition path.
[0318] In this specification, based on steps S502 to S510, the risk of mismatch between evidence and object status caused by delays in investigation results is reduced, and the evidence sufficiency matrix and investigation hypothesis verification map can be continuously and accurately updated with the object status and supplementary security evidence.
[0319] Optionally, the specific execution sequence of the candidate disposal action and the investigation task to be performed, based on the impact of the change in the state of the unfixed evidence corresponding to the remaining supplementary evidence requirements, can be determined by referring to the following method: S602: Determine the level of damage to the unsecured evidence by the candidate action and the reversibility of the candidate action; Unsecured evidence is used to characterize security evidence that corresponds to the current remaining evidence requirements but has not yet been collected, saved, or persisted. Unsecured evidence may include runtime data, temporary data, or security data that is susceptible to changes in the state of the target security object, or data that has not yet been extracted and saved from the original data source.
[0320] The destructiveness level characterizes the degree to which the candidate action modifies, overwrites, deletes, or has an irreversible impact on the unsecured evidence. Reversibility characterizes whether, after the candidate action is executed, the target security object can be restored to its state before the action was executed, or to a state equivalent to that before the action, through undoing, restoration, or compensation operations.
[0321] As an illustration, the service platform can obtain the action type, target security object, expected execution effect, and possible changes in object state after the action is executed for each candidate disposal action, and further determine whether the state changes will affect the corresponding unfixed evidence. For example, for a network isolation candidate disposal action, its execution can restrict the target asset from continuing to communicate with the external network, but it usually does not directly delete local logs, process running information, or file data in the target asset. Therefore, it can be determined that its destructive level to the corresponding unfixed evidence is relatively low, and the network connectivity of the target asset can be restored after the isolation is lifted, thus the candidate disposal action can be determined to be reversible. As another example, for a candidate disposal action to terminate a target process, if there are still remaining supplementary evidence items that need to be obtained from the process, such as memory information, process handle information, or runtime network connection information, terminating the target process may cause the corresponding runtime evidence to disappear. Therefore, it can be determined that this action has a high destructive effect on the corresponding unfixed evidence. For example, if the current evidence requirement still needs to be verified based on the target file itself, file attributes, or file associations, then executing the candidate action may result in the inability to obtain the corresponding evidence again. It can be determined that the candidate action has an irreversible impact on the corresponding unfixed evidence.
[0322] In one feasible implementation, the level of destructiveness can be classified into different levels based on the changes in the state of the target security object before and after the execution of the candidate disposal action. These levels include those that do not affect unsecured evidence, alter unsecured evidence but can be recovered, and alter or delete unsecured evidence that is difficult to recover. The specific number of levels and the classification method can be set according to the actual security strategy of the network security incident handling system, and this specification does not limit them.
[0323] It should be noted that the level of destructiveness and reversibility can be determined separately. Even if a candidate action can restore the operational state of the target security object through compensation operations, it may have already resulted in the loss of some volatile evidence. Conversely, although a candidate action may not easily revoke the change in the state of the target security object, if it is unrelated to the current remaining supplementary evidence requirements, its destructiveness to the currently unsecured evidence can be determined to be low. Therefore, the level of destructiveness and reversibility can be combined to determine the subsequent execution order constraints.
[0324] S604: Based on the remaining supplementary evidence requirements that need to be fixed before the candidate disposal action is executed, determine the evidence dependency items corresponding to the candidate disposal action; Evidence dependency terms are used to characterize the remaining supplementary evidence requirements that need to be secured before a candidate action can be executed. In other words, evidence dependency terms are used to establish the constraint relationship between a candidate action and the evidence requirements it may affect.
[0325] As an illustration, the service platform 100 can obtain the target security object corresponding to the current candidate disposal action, and read the evidence requirements corresponding to the relationship between the target security object, associated object, or related object in the remaining supplementary evidence requirements.
[0326] Furthermore, based on the level of destructiveness determined in step S602, it can be determined whether the execution of the candidate action will change or delete the unfixed evidence required to satisfy the remaining supplementary evidence requirements. If it will have a corresponding impact, then the remaining supplementary evidence requirements can be identified as the evidence dependency item corresponding to the candidate action.
[0327] For example, if there is a candidate action for the target process P1, namely "terminating process P1," and the remaining supplementary evidence requirement R1 requests access to the network connection information currently in P1's memory, then since terminating P1 may result in the loss of the corresponding memory data, R1 can be identified as an evidentiary dependency for the candidate action. As another example, if the current candidate action is to delete the target file F1, and the remaining supplementary evidence requirement R2 requests further verification of F1's file content, file attributes, or source relationship, then R2 can be identified as an evidentiary dependency for the deletion action.
[0328] In one feasible implementation, a candidate action may correspond to one or more evidentiary dependencies; the same evidentiary requirement may also become an evidentiary dependency for multiple candidate actions simultaneously. For example, the memory forensics requirement on a target asset may be affected by both the "isolation and restart" action and the "termination of related processes" action, and the evidentiary requirement can establish evidentiary dependencies with the aforementioned multiple candidate actions respectively.
[0329] Furthermore, the evidence dependencies can be associated with the corresponding investigation tasks to be executed, so that it can be determined directly based on the evidence dependencies which investigation tasks need to be completed first before the candidate action is executed.
[0330] Through the above processing, the abstract action that may affect the evidence can be further transformed into a clear constraint chain of candidate action - evidence dependency - investigation task to be performed.
[0331] S606: Determine the corresponding containment level based on the scope of impact and spread status of the cybersecurity incident; Containment level is used to characterize the handling level of a current cybersecurity incident by classifying the urgency of security measures after comprehensively considering the scope of affected objects and the state of attack spread.
[0332] As an example, the service platform can determine the scope of security objects affected by a cybersecurity incident based on the current causal tracing results. For example, it can determine the number of affected assets, the number of affected accounts, or the confirmed lateral movement range.
[0333] Simultaneously, the attack propagation status can be determined based on information such as the number of newly affected security objects and the number of new lateral movement relationships within a preset time range, reflecting whether the cybersecurity incident is still spreading. For example, when new affected assets or new lateral movement relationships continuously appear within a short period of time, it can be determined that the current cybersecurity incident has a high degree of propagation; when no new affected objects appear or the attack path remains basically stable for a period of time, it can be determined that the current cybersecurity incident has a low degree of propagation.
[0334] Furthermore, the corresponding containment level can be determined by combining the criticality level of the affected security object and the type of threat impact. The containment level can include monitoring only, delayed containment, and immediate containment.
[0335] Among them, "monitoring only" can be used to indicate that the current attack spread is low and the scope of impact is limited, and it is not necessary to change the status of the target security object immediately; "delayed containment" can be used to indicate that there is a certain security risk, but it is allowed to carry out security measures after the necessary evidence is secured; "immediate containment" can be used to indicate that the current attack is spreading rapidly, the affected object is highly critical or the threat is significant, and it is necessary to prioritize the implementation of security measures to block the spread of the attack.
[0336] For example, if only a single low-critical test asset exhibits abnormal behavior and no new lateral movement relationships are discovered, the corresponding cybersecurity incident can be classified as a monitoring-only incident. If multiple assets have been confirmed to be affected, but the attack spreads slowly and important operational evidence remains unconfirmed, it can be classified as a delayed containment incident. If new affected critical assets and lateral movement relationships continue to emerge within a preset timeframe, the current cybersecurity incident can be determined to have a high risk of spread and classified as an immediate containment incident.
[0337] It should be noted that the specific rules for determining the containment level can be configured based on the criticality of the assets, the degree of attack spread, and the type of threat impact in the actual application scenario, and this manual does not impose any restrictions on this.
[0338] S608: Based on the level of destructiveness, the reversibility, the evidence dependency, and the level of containment, determine the execution order constraint between the candidate disposal action and the corresponding investigation task to be executed, and generate the execution sequence based on the execution order constraint.
[0339] Execution order constraints are used to characterize the sequential or priority execution relationship that must be satisfied between candidate actions and one or more investigation tasks to be executed.
[0340] As an illustration, the service platform can combine the level of destructiveness and reversibility determined in step S602, the evidence dependencies determined in step S604, and the level of containment determined in step S606 to jointly rank candidate disposal actions and investigation tasks to be performed.
[0341] When the containment level is monitored only, candidate actions that would change the state of the target security object can remain in a pending state, and security monitoring or investigation tasks corresponding to the remaining supplementary evidence requirements can continue to be performed to supplement the security evidence required for the current investigation.
[0342] When the containment level is delayed containment, if the candidate action has evidence dependencies for which evidence has not yet been fixed, and step S602 determines that the candidate action may change or delete the corresponding unfixed evidence, then an execution order constraint of "the investigation task to be performed takes precedence over the candidate action" can be established.
[0343] For example, for a target process P1, if there is both a memory forensics task T1 and a candidate action A1 to terminate P1, and the evidence requirement item corresponding to T1 is an evidence dependency item of A1, then it can be determined that T1 should be executed first to complete the required memory evidence fixation, and then A1 should be executed to terminate the target process.
[0344] In another feasible implementation, when the containment level is immediate containment, the execution order can be further determined by combining the reversibility of the candidate disposal actions and the level of destructiveness.
[0345] Candidate actions that are reversible and intended to prevent further spread of the attack can be prioritized. For example, network isolation of affected assets can be prioritized to limit the attack's continued spread as quickly as possible.
[0346] For candidate actions that rely on evidence that has not yet been secured and whose execution would have a highly destructive impact on the corresponding unsecured evidence, they can remain in a pending state until the corresponding evidence relies on evidence is secured, or until an execution permit is obtained to execute the candidate action ahead of schedule.
[0347] For example, in an immediate containment scenario, reversible network isolation of the target asset can be performed first, while memory forensics of the target process can be retained. Before the memory evidence is secured, actions that could lead to irrecoverable loss of evidence, such as deleting target files or terminating critical target processes, should not be taken. If the attack continues to spread or other security risks prevent further waiting, corresponding candidate actions can be taken after recording the corresponding evidentiary gaps and obtaining execution permission.
[0348] Furthermore, the service platform 100 can generate an execution sequence for the current network security incident based on the aforementioned multiple execution order constraints. The execution sequence can represent the sequential order or allowed parallel relationships between multiple pending investigation tasks and multiple candidate handling actions.
[0349] For example, regarding the current cybersecurity incident, it can be determined that the first investigation task T1 and the second investigation task T2 can be executed in parallel. Candidate action A1 needs to be executed after T1 is completed, reversible candidate action A2 can be executed first, and highly destructive candidate action A3 needs to be executed after both T1 and T2 are completed. The service platform can form the execution sequence corresponding to the current cybersecurity incident based on the above execution order constraints for use in subsequent action execution phases.
[0350] It should be noted that the execution sequence can be dynamically updated as the investigation status of a cybersecurity incident changes. When new supplementary security evidence enables a certain evidence dependency to reach a satisfied state, the waiting constraint between the corresponding candidate action and the investigation task can be released; when new causal tracing results discover new affected objects or the degree of attack spread increases, the containment level can also be redefined and the corresponding execution sequence can be regenerated.
[0351] In this specification, based on steps S602 to S608, by determining the level of damage and reversibility of candidate actions to unsecured evidence, an evidentiary dependency relationship is established between candidate actions and remaining supplementary evidence requirements. Furthermore, the containment level is determined by combining the impact scope and spread status of the cybersecurity incident. Then, based on the level of damage, reversibility, evidentiary dependency, and containment level, execution order constraints are generated between candidate actions and investigation tasks to be performed. This can suppress the spread of attacks while reducing the possibility of prematurely changing or deleting key unsecured evidence, enabling the cybersecurity incident handling process to balance attack containment efficiency and the integrity of investigation evidence.
[0352] In one feasible implementation, specifically executing the step of determining the corresponding containment level based on the impact scope and spread status of the cybersecurity incident, and determining the execution order constraint based on the level of damage, the reversibility, the evidentiary dependencies, and the containment level, can refer to the following approach: Step E2: Count the number of newly affected security objects and the number of newly added lateral movement object relationships within a preset time range, and determine the degree of attack spread based on the number of newly affected security objects and the number of newly added lateral movement object relationships; The attack diffusion level is used to characterize the extent to which a cybersecurity incident continues to spread or expand from an affected security object to other security objects within a preset time range; the preset time range can be set according to the processing cycle of the cybersecurity incident, the monitoring granularity, or the rate of change of attack behavior.
[0353] The number of newly affected security objects can be determined based on the aforementioned causal tracing results and changes in the status of the security objects.
[0354] As an example, assets, accounts, processes, or other security objects that are first identified as being affected by the current cybersecurity incident within the current preset time frame can be counted as newly affected security objects. Security objects that have already been identified as affected within the previous time frame can be excluded from the current count of newly affected security objects.
[0355] The number of newly added lateral movement object relationships can be used to reflect the continued propagation of attacks between different security objects. For example, when a remote authentication relationship, remote execution relationship, or other object relationship that can characterize the spread of an attack from one security object to another is newly identified within the current preset time frame from the first affected asset to the second asset, the corresponding relationship can be included in the number of newly added lateral movement object relationships.
[0356] As an illustration, the service platform can, based on candidate tracing paths and object relationships verified by tracing evidence, count the number of newly affected security objects and the number of newly added lateral movement object relationships within a current preset time range. Furthermore, the attack spread level can be determined based on these two metrics: corresponding judgment ranges can be set for the number of newly affected security objects and the number of newly added lateral movement object relationships, and the attack spread level can be determined based on the range in which they fall. For example, if no or only a small number of newly affected security objects appear within the preset time range, and no new lateral movement object relationships appear, the attack spread level can be determined to be low; when the number of newly affected security objects or the number of newly added lateral movement object relationships continuously increases, the attack spread level can be determined to be high.
[0357] Furthermore, for critical business areas or highly critical assets, the quantitative requirements for determining a high degree of diffusion can be reduced. For example, even if the number of newly affected security objects is currently small, but these newly affected objects include critical servers or critical identity accounts, the corresponding risk weight can be increased in subsequent containment level assessments.
[0358] Step E4: Based on the criticality level of the affected security object, the extent of attack spread, and the type of threat impact, determine the containment level corresponding to the network security incident. The containment level includes monitoring only, delayed containment, and immediate containment. Criticality levels characterize the importance of affected security objects to business operations, data security, or network infrastructure. For example, a criticality level can be determined based on the importance of the services carried by the security object, its asset category, or pre-configured asset levels.
[0359] Threat impact type is used to characterize the category of security impact that a current cybersecurity incident has caused or may cause. For example, it can include different impact types such as affecting the availability, integrity, confidentiality, or control status of the target security object. The specific threat impact type can be determined based on the aforementioned causal attribution results, investigation hypothesis data, and obtained security evidence.
[0360] As an example, the service platform can use the criticality level of the affected security object, the degree of attack spread determined in step E2, and the type of threat impact as parameters for determining the containment level, and determine the corresponding containment level according to the preset containment strategy.
[0361] If the monitoring is only used to characterize the scope and extent of the impact of the current cybersecurity incident and has not yet reached the point where it is necessary to immediately change the status of the target security object, the investigation, evidence collection, and security status monitoring can continue.
[0362] Delayed containment is used to characterize a current cybersecurity incident as having a risk that requires security measures, but under the current risk conditions, it is still permissible to prioritize the fixation of evidence corresponding to key unfixed evidence before executing candidate measures that may affect the unfixed evidence.
[0363] Immediate containment is used to characterize the scope, spread, or threat impact of a current cybersecurity incident, indicating that the attack has reached a level requiring priority to limit its continued propagation. Corresponding candidate actions to block the spread of the attack can be prioritized.
[0364] For example, when the criticality level of the affected security object is low, the attack spread is low, and the current threat impact is limited, the containment level can be set to monitoring only; when the affected security object is highly critical, but the current attack spread is low, or the current event has already had a certain security impact but allows for the short-term acquisition of key evidence, the containment level can be set to delayed containment; when a highly critical security object is affected, and the attack spread is high, or the current threat impact has continued to expand, the containment level can be set to immediate containment.
[0365] In some embodiments, a correspondence between criticality level, attack spread level, threat impact type, and different containment levels can be pre-established. After obtaining the above parameters of the current network security incident, the containment level is determined through the correspondence. The correspondence can be configured according to the actual network security strategy.
[0366] It should be noted that the containment level does not remain constant throughout the entire cybersecurity incident handling process. As new security evidence emerges, the number of affected security objects changes, or the attack spread status changes, steps E2 and E4 can be re-executed to update the containment level corresponding to the current cybersecurity incident.
[0367] Step E6: When the containment level is monitoring only, keep the candidate actions for changing the state of the target security object in the pending state, and maintain security monitoring of the network security event; As an example, when the current network security incident is determined to be at the containment level of only monitoring, candidate actions that would change the running state, access state, or data state of the target security object may be temporarily suspended.
[0368] As an example, candidate actions such as network isolation, account disabling, process termination, file deletion, or other actions that would change the state of the target security object can be kept in a pending state, and continuous security monitoring of the security objects, object relationships, and network behaviors involved in the network security incident can be maintained.
[0369] During the monitoring period, the service platform can continue to execute the investigation tasks corresponding to the remaining supplementary evidence requirements, and update the evidence sufficiency matrix and investigation hypothesis verification map based on the newly obtained supplementary security evidence.
[0370] Optionally, during the security monitoring process, if the statistical results show an increase in the number of newly affected security objects or the emergence of new lateral movement object relationships, steps E2 to E4 can be re-executed to update the attack spread level and containment level.
[0371] If the redefined containment level changes from monitoring only to delayed containment or immediate containment, then the current monitoring only processing state can be exited, and the execution order constraints of candidate disposal actions and pending investigation tasks can be redefined according to the corresponding containment level.
[0372] Step E8: When the containment level is delayed containment, the candidate disposal action for evidence dependency items that have not yet completed evidence fixation and the destructive level has reached the preset destructive level is set to be executed after the corresponding investigation task to be performed; The preset damage level is used to determine whether the impact of the candidate action on the unsecured evidence is sufficient to restrict its execution order. The preset damage level can be configured according to the damage level determined in step S602 above.
[0373] When the containment level is delayed containment, it is possible to further examine whether the evidence dependencies corresponding to each candidate action have been secured.
[0374] If a candidate action does not have any evidentiary dependencies, or if the corresponding evidentiary dependencies have been secured by evidence, then the execution order of the candidate action can no longer be restricted based on the evidentiary dependencies.
[0375] If a candidate action involves evidence dependencies for which evidence has not yet been secured, and the destructive level of the candidate action reaches a preset level, then executing the candidate action first may prevent the subsequent acquisition of the necessary security evidence. In this case, the investigation task corresponding to the evidence dependency can be set as a pre-processing task of the candidate action.
[0376] For example, a target process P1 has a pending memory forensics task T1, and a candidate action A1 to terminate process P1. Since executing A1 may cause the disappearance of the running memory data corresponding to P1, it can be determined that the destructive level of A1 reaches the preset destruction level, and the evidence requirement item corresponding to T1 is an evidence dependency item of A1. Under the current containment level of delayed containment, T1 can be set to execute before A1, that is, the relevant evidence of the target process P1 is fixed first, and then the action to terminate process P1 is executed. As another example, a target file F1 is determined to need to be cleaned, but it is still necessary to obtain and fix the file content and associated attributes of F1 through investigation task T2. In this case, the candidate action to delete F1 can be kept in a waiting state until T2 is completed.
[0377] In one feasible implementation, when a candidate action has multiple evidentiary dependencies, the investigation tasks corresponding to each of the multiple evidentiary dependencies can all be treated as pre-processing tasks of the candidate action. For multiple investigation tasks that do not have mutual dependencies, they can be executed in parallel, and the waiting constraint on the corresponding candidate action can be released after the relevant evidence is secured.
[0378] Step E10: When the containment level is immediate containment, the isolation-type candidate disposal action that is reversible and used to contain the network security incident is set as the priority to be executed, and the cleanup-type candidate disposal action that has not yet completed the evidence fixation of the evidence dependency and the damage level reaches the preset damage level is kept in the pending execution state until the corresponding unfixed evidence is fixed or the execution permission of the cleanup-type candidate disposal action is obtained.
[0379] Isolation-type candidate actions are used to restrict the target security object from continuing to participate in attack propagation or engaging in abnormal interactions with other security objects without prioritizing destructive cleanup. Examples include actions that can inhibit the further spread of an attack, such as network isolation of affected assets, restricting access to target accounts, or limiting specific network communications.
[0380] Cleanup-type candidate actions are used to eliminate or remove identified threat objects or attack remnants, and their execution may directly change or delete the data state of the corresponding security object. For example, it may include terminating related processes, deleting malicious files, or performing other cleanup operations that may cause unsecured evidence to disappear.
[0381] When the containment level is immediate containment, the service platform can identify reversible isolation-type candidate actions from the candidate actions that can limit the further spread of the cybersecurity incident, and set them for priority execution. For example, when it is detected that an affected asset is continuously establishing lateral connections to other assets, network isolation can be prioritized for that asset. Since the isolation action is mainly used to block further propagation, and the isolation can be lifted later according to investigation or business needs, it can be prioritized even if all key security evidence has not yet been secured. For cleanup-type candidate actions where evidence dependencies have not been secured and the destructive level has reached a preset destructive level, they can remain in a pending state. For example, after network isolation is completed, a malicious process on the target asset is still running, and there is currently a memory forensics task targeting the malicious process. In this case, network isolation can be used to limit the further spread of the attack, but the target process is not immediately terminated. After the corresponding memory evidence is secured, process termination or other cleanup actions can be executed.
[0382] Furthermore, when the attack risk continues to increase, business security requirements, or other security conditions prevent waiting for the relevant unsecured evidence to be secured, the waiting constraint can be lifted after obtaining permission to execute the corresponding cleanup candidate disposal action, and the cleanup candidate disposal action can be executed.
[0383] The execution permission can be used to indicate that, when it is clearly known that the corresponding unsecured evidence may be affected, priority is given to implementing corresponding cleanup actions. The execution permission can be associated with corresponding candidate actions, evidentiary dependencies, and security evidence that has not yet been secured, in order to determine the cause of the corresponding evidentiary gap during subsequent investigations. It should be noted that obtaining execution permission does not mean that the corresponding evidentiary requirements have been met. If the premature execution of cleanup candidate actions results in the inability to obtain corresponding security evidence, the unmet or unobtainable status of the relevant evidentiary requirements should still be maintained based on the actual evidence acquisition situation.
[0384] Furthermore, under immediate containment status, reversible isolation candidate actions and pending investigation tasks unaffected by the isolation actions can be executed in parallel based on actual processing capacity. For example, while isolating the target asset, historical security evidence can continue to be obtained from a centralized log system or other independent security data sources, thereby continuing to supplement evidence while limiting the spread of the attack. After the spread of the cybersecurity incident is brought under control, steps E2 and E4 can be re-executed. When the updated attack spread level decreases and the containment level is adjusted from immediate containment to delayed containment or only monitoring, the execution order constraints of the remaining candidate actions and pending investigation tasks can be redefined accordingly.
[0385] In this specification, based on steps E2 to E10, network security incidents are classified into monitoring-only, delayed containment, and immediate containment. Under different containment levels, the order of candidate action actions and investigation tasks to be performed is controlled, which allows the action sequence to be dynamically adjusted according to the attack spread status. When the risk is low, the target state is maintained to continue the investigation and evidence collection. When delayed action is allowed, priority is given to securing security evidence that may be damaged. When immediate containment is required, reversible isolation is given priority and cleanup actions with the risk of evidence destruction are postponed. This balances attack spread control and preservation of critical security evidence, reducing the risk of irrecoverable loss of investigative evidence due to premature destructive actions.
[0386] In one feasible implementation, the specific execution of the candidate treatment action via the security device, and the generation of investigation conclusion data based on the action execution result, the evidence sufficiency matrix, and the investigation hypothesis verification map, can be carried out in the following manner: Step F2: Obtain the version of the target security object bound to the candidate disposal action and the current version of the target security object before the action is executed, and perform version verification between the target security object version and the current object version; The current object version of the target security object mentioned before the action is executed can be understood as the latest object version of the target security object in the network security incident handling system before the candidate action is actually issued to the security device.
[0387] As an illustration, the service platform can read the target security object identifier and target security object version pre-bound in the candidate disposal action, query the current object version based on the target security object identifier, and then compare whether the target security object version bound to the candidate disposal action is consistent with the current object version.
[0388] When both are consistent, it can be determined that the target object state upon which the candidate action is based is consistent with the currently recorded target object state, thus confirming that the version verification has passed. When they are inconsistent, it can be determined that the target security object has undergone a state change after the candidate action was generated but before the action was actually executed. In this case, instead of directly invoking the security device to execute the original candidate action, it is not necessary to re-determine whether the candidate action is still applicable based on the updated target security object state, and regenerate or adjust the candidate action if necessary. By performing object version verification again before the action is actually executed, the security action can maintain a correspondence with the object state upon which the action was generated, reducing the risk of executing outdated actions due to changes in the target security object state during the investigation.
[0389] Step F4: When the version verification passes, the preconditions of the candidate action are verified, and when the preconditions are met and the candidate action meets the execution sequence, the corresponding security device is invoked to execute the candidate action. Action preconditions are used to characterize one or more conditions that must be met before a candidate action is allowed to be actually executed. The action preconditions can be set according to the action type of the candidate action, the status of the target security object, the execution permission, the approval requirements, and the execution capabilities of the corresponding security equipment.
[0390] For example, for network isolation candidate actions, the prerequisites may include the target asset being currently manageable by the corresponding security device, the required permissions for isolation being valid, and the corresponding security device having isolation capabilities. For account restriction candidate actions, the prerequisites may include the target account being still operable and the corresponding identity security device having the corresponding account control capabilities. For cleanup candidate actions, the prerequisites may also include whether the aforementioned evidentiary dependencies have been secured, or whether permission has been obtained to execute the corresponding cleanup action in advance.
[0391] After version verification, the service platform can verify each precondition associated with a candidate action. If any (necessary) action precondition is not met, the security device can be temporarily suspended from executing the candidate action, and the corresponding precondition failure information can be recorded for subsequent reassessment or adjustment of the action plan. When all action preconditions are met, it is further determined whether the candidate action meets the execution sequence determined in step S608 or steps E6 to E10. For example, if a candidate action is set to be executed after a certain investigation task, it is necessary to determine whether the corresponding investigation task has reached a state where the waiting constraint can be released; if the candidate action is a reversible isolation action that is allowed to be executed first under the immediate containment state, the corresponding security device can be called first after its preconditions are met.
[0392] When the preconditions for the action are met and the execution sequence is also met, the service platform 100 can determine the security device that can execute the candidate action based on the action type, target security object and action parameters corresponding to the candidate action, and send the corresponding action execution request to the security device.
[0393] For example, network isolation actions targeting a target asset can be executed by calling security devices with network access control or endpoint isolation capabilities; access restriction actions targeting a target account can be executed by calling corresponding authentication or access control devices; and termination actions targeting a target process can be executed by calling endpoint security devices on the target asset.
[0394] Furthermore, during the execution of an action, information such as the action version, target security object, execution parameters, execution permissions, approval status, timeout conditions, retry conditions, and result write-back target can be recorded for candidate actions, so as to trace the action execution process and action results in the future.
[0395] Step F6: Obtain the action execution result returned by the security device, determine the state change of the target security object based on the action execution result, and update the object version of the target security object based on the state change; After a candidate action is executed by a security device, the corresponding action execution result can be returned. This result can be used to characterize whether the candidate action was actually executed, whether the execution was successful, and whether the target security object has reached the expected handling state. The result may include states such as successful execution, partial execution, execution failure, or rollback, and may further include information such as the action execution time, the target object's state after execution, and the reason for failure.
[0396] When the action execution result indicates that the candidate action was successfully executed, the service platform 100 can determine the state change of the target security object before and after the action execution based on the expected effect of the candidate action and the actual state returned by the security device.
[0397] For example, if target asset A is in a normal network connection state before the action is executed, and the security device returns that target asset A has entered an isolated state after the network isolation action is executed, then it can be determined that the network access status of target asset A has changed. As another example, if the security device returns that the access permissions of target account U have been restricted after the target account U executes an access restriction action, then it can be determined that the corresponding account's permission attributes have changed. Furthermore, if the candidate disposal action to terminate target process P1 is successfully executed, it can be determined that P1's running status has changed from running to terminated.
[0398] Based on the determined state changes, a new object version can be generated for the target security object according to the aforementioned object version management method, and a version association can be established between the updated object version and the object version before the update. For example, if the target asset A corresponds to object version V4 before network isolation, after successful isolation, object version V5 can be generated, and the updated network state of target asset A can be recorded in V5, while retaining the version relationship between V5 and V4.
[0399] If the action execution result indicates that the candidate disposal action is only partially completed, a new object version can be generated based on the actual changes in the object's state, and the incomplete portion can be retained in the action execution result. If the action execution fails and the target security object's state remains unchanged, the original object version can be kept unchanged, and the action execution failure result can be recorded. If a rollback operation is further performed after the action execution, restoring the target security object to its previous state or forming a new recovery state, a corresponding object version can be generated based on the actual state change, instead of directly deleting the historical object version already formed during the action execution process.
[0400] In this way, the result of each action that changes the state of the target security object can be associated with the corresponding object version, thereby providing a state basis for subsequent reassessment of the applicability of evidence and replay of the investigation process.
[0401] Step F8: Determine the evidence requirement items affected by the state change based on the updated object version, update the evidence sufficiency matrix and the investigation hypothesis verification graph, and generate the investigation conclusion data based on the updated evidence sufficiency matrix, the investigation hypothesis verification graph and the action execution results.
[0402] As an illustration, after updating the object version of the target security object based on the action execution result, the service platform can determine which evidence requirements are affected by the state change based on the object attributes, object relationships, and associated security objects corresponding to this state change.
[0403] Specifically, the object version before the update can be compared with the object version after the update to determine the changed object attributes or object relationships, and further determine whether the satisfaction conditions of each evidence requirement item depend on the changed object attributes or object relationships.
[0404] For example, if a candidate action updates target asset A from a normal networked state to an isolated state, and a certain evidentiary requirement is used to verify whether target asset A established an abnormal connection with an external network address before isolation, the network connection evidence that has been fixed and associated with the historical object version can still be used to verify the historical target facts; if another evidentiary requirement is used to verify whether target asset A still has external connection behavior after isolation, the corresponding evidentiary requirement state can be re-determined for the updated object version.
[0405] For evidence requirements affected by changes in status, the aforementioned object version applicability judgment, evidence source independence verification, evidence coverage verification, and evidence conflict verification can be re-executed, and the corresponding satisfaction status can be updated based on the re-evaluation results.
[0406] Furthermore, based on the updated evidence requirement status, the corresponding matrix cell in the evidence sufficiency matrix can be located, and the status of satisfaction, supporting evidence, counter-evidence, remaining supplementary evidence requirements, or other relevant information in the matrix cell can be updated.
[0407] Simultaneously, the investigation hypothesis verification graph can be updated. For example, the results of action execution or new security observation data generated by action execution can be added to the graph as new investigation-related data; for supplementary evidence requirements regenerated due to changes in object status, corresponding evidence requirement nodes can be added, and the relationship between them and investigation hypothesis nodes, existing security evidence nodes, and candidate evidence acquisition capability nodes can be re-established.
[0408] After completing the above updates, investigation conclusion data can be generated based on the updated evidence sufficiency matrix, investigation hypothesis verification map, and action execution results.
[0409] The data from the investigation findings can be used to characterize the investigation status and processing results of current cybersecurity incidents after evidence verification, supplementary evidence collection, and security measures.
[0410] Indicatively, the investigation conclusion data may include the results of the verification of investigation hypotheses supported by security evidence, corresponding key security evidence, current attack paths, affected security objects, candidate actions that have been executed and their results, evidence requirements that are still not met, and situations where evidence is unavailable.
[0411] If there are still evidence requirements that are in conflict, partially satisfied, dependent on sources, or unavailable, then the corresponding evidence boundaries can be retained in the investigation conclusion data, rather than expressing the current investigation results as a definitive result with no evidence gaps.
[0412] Optionally, after step F8 is completed, it can be determined whether the aforementioned investigation termination conditions are met based on the updated evidence sufficiency matrix and the investigation hypothesis verification map.
[0413] When all the core evidence roles meet the corresponding requirements, the evidence conflicts that need to be explained have been eliminated or explained, the investigation tasks that must be performed have entered the corresponding termination state, and the necessary candidate actions have enabled the target security object to reach the expected security state, the current loop processing between steps S104 and S110 can be ended, and the currently generated investigation conclusion data can be used as the final investigation conclusion data of the corresponding cybersecurity incident.
[0414] If there are still evidence requirements that need to be supplemented or re-verified, the process can return to step S104 to re-evaluate the evidence requirements and trace the causal origin. When a change in status creates a new tracing breakpoint, step S106 can be executed to generate a new investigation task. When the attack path, affected objects, or remaining evidence requirements change, step S108 can be re-executed to update the candidate actions and execution sequence. Thus, the results of the security device's actions can re-enter the evidence verification and causal origination process, forming a closed-loop update between the security action results and the investigation status.
[0415] In this specification, based on steps F2 to F8, security actions can be avoided based on expired target states, and the changes in object states resulting from the actions can continue to participate in the determination of evidence validity and the verification of investigation hypotheses. Furthermore, investigation conclusion data is generated based on the updated evidence state and the results of action execution, thereby improving the consistency between the results of cybersecurity incident handling and the state of investigation evidence, as well as the traceability of investigation conclusions.
[0416] The following will combine Figure 2 This manual provides a detailed description of the network security incident handling device provided. It should be noted that... Figure 2 The network security incident handling device shown is used to execute the instructions in this manual. Figure 1The methods of the embodiments shown are illustrated only in connection with this specification for ease of explanation. For specific technical details not disclosed, please refer to this specification. Figure 1 The example shown.
[0417] Please see Figure 2 This document illustrates a structural diagram of the network security incident handling device 1. This network security incident handling device 1 can be implemented as all or part of a device through software, hardware, or a combination of both. According to some embodiments, the network security incident handling device 1 includes: a data processing module 11, used to convert network security incident records from multiple security data sources into security observation data, and construct security objects with object versions and object relationships between the security objects based on the security observation data; a source tracing investigation module 12, used to generate evidence requirement items based on preset evidence roles corresponding to investigation hypothesis data, and construct an investigation hypothesis verification map and an evidence sufficiency matrix based on the evidence requirement items, perform causal tracing along the object relationships, and determine the satisfaction status of each evidence requirement item based on the obtained source tracing evidence; the source tracing investigation module 12 is used to generate an investigation task including the expected object version of the target security object for unsatisfied evidence requirement items corresponding to detected source tracing breakpoints, execute the investigation task to obtain supplementary security evidence, and so on. The applicable object version of the supplementary security evidence is determined based on the expected object version and the current object version of the target security object. The evidence sufficiency matrix and the investigation hypothesis verification map are updated based on the supplementary security evidence applicable to the current object version. The source tracing investigation module 12 is used to determine the remaining supplementary evidence requirements based on the evidence sufficiency matrix and to determine the investigation tasks to be executed corresponding to the remaining supplementary evidence requirements based on the investigation hypothesis verification map. Based on the current attack path, affected objects and the remaining supplementary evidence requirements obtained by the causal source tracing, candidate actions bound to the target security object version are generated. Based on the impact of the candidate actions on the state change of the unfixed evidence corresponding to the remaining supplementary evidence requirements, the execution sequence of the candidate actions and the investigation tasks to be executed is determined. The conclusion generation module 13 is used to execute the candidate disposal action through a security device when the current object version of the target security object corresponding to the candidate disposal action matches the target security object version bound to the candidate disposal action and the candidate disposal action satisfies the execution sequence, and to generate investigation conclusion data based on the action execution result, the evidence sufficiency matrix and the investigation hypothesis verification map.
[0418] It should be noted that the network security incident handling device provided in the above embodiments is only illustrated by the division of the above functional modules when executing the network security incident handling method. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device can be divided into different functional modules to complete all or part of the functions described above. In addition, the network security incident handling device and the network security incident handling method embodiments provided in the above embodiments belong to the same concept, and the implementation process is detailed in the method embodiments, which will not be repeated here. The serial numbers in this specification are for descriptive purposes only and do not represent the superiority or inferiority of the embodiments. This specification also provides a computer storage medium that can store multiple instructions, which are adapted to be loaded and executed by a processor as described above. Figure 1 The network security incident handling method described in the illustrated embodiment can be found in the following documentation for its specific execution process. Figure 1 Detailed descriptions of the illustrated embodiments are omitted here. This specification also provides a computer program product storing at least one instruction, which is loaded and executed by the processor as described above. Figure 1 The network security incident handling method described in the illustrated embodiment can be found in the following documentation for its specific execution process. Figure 1 The specific details of the illustrated embodiments will not be elaborated here.
[0419] Please refer to Figure 3This is a structural block diagram of an electronic device provided in an embodiment of this specification. The electronic device in this specification may include one or more of the following components: a processor 1010, a memory 1020, an input device 1030, an output device 1040, and a bus 1050. The processor 1010, memory 1020, input device 1030, and output device 1040 can be connected via the bus 1050. The processor 1010 may include one or more processing cores. The processor 1010 connects various parts within the electronic device using various interfaces and lines, and performs various functions and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory 1020, and by calling data stored in the memory 1020. Optionally, the processor 1010 may be implemented using at least one hardware form selected from Digital Signal Processing (DSP), Field-Programmable Gate Array (FPGA), and Programmable Logic Array (PLA). Processor 1010 may integrate one or more of the following: a central processing unit (CPU), a graphics processing unit (GPU), and a modem. The CPU primarily handles the operating system, user interface, and applications; the GPU is responsible for rendering and drawing displayed content; and the modem handles wireless communication. It is understood that the modem may not be integrated into processor 1010 and may be implemented separately through a communication chip. Memory 1020 may include random access memory (RAM) or read-only memory (ROM). Optionally, memory 1020 may include a non-transitory computer-readable medium. Memory 1020 can be used to store instructions, programs, code, code sets, or instruction sets. Input device 1030 is used to receive input instructions or data, and input device 1030 includes, but is not limited to, a keyboard, mouse, camera, microphone, or touch device. Output device 1040 is used to output instructions or data, and output device 1040 includes, but is not limited to, display devices and speakers. In the embodiments described in this specification, the input device 1030 can be a temperature sensor for acquiring the operating temperature of the electronic device. The output device 1040 can be a speaker for outputting audio signals. Furthermore, those skilled in the art will understand that the structure of the electronic device shown in the above figures does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown, or combine certain components, or have different component arrangements.For example, electronic devices also include radio frequency circuits, input units, sensors, audio circuits, Wireless Fidelity (WIFI) modules, power supplies, Bluetooth modules, and other components, which will not be elaborated here.
[0420] In the embodiments of this specification, the executing entity for each step can be the electronic device described above. Optionally, the executing entity for each step is the operating system of the electronic device. The operating system can be Android, iOS, or other operating systems; this specification does not limit this. Figure 3 In the electronic device, the processor 1010 can be used to call a program stored in the memory 1020 and execute it to implement the network security incident handling method as described in the various method embodiments of this specification.
Claims
1. A method for handling network security incidents, characterized in that, The method includes: The network security event records from multiple security data sources are converted into security observation data, and security objects with object versions and object relationships between the security objects are constructed based on the security observation data. Evidence requirement items are generated based on the preset evidence roles corresponding to the survey hypothesis data, and a survey hypothesis verification map and evidence sufficiency matrix are constructed based on the evidence requirement items. Causal tracing is performed along the object relationship, and the satisfaction status of each evidence requirement item is determined based on the obtained tracing evidence. For the unmet evidence requirements corresponding to the detected source tracing breakpoints, an investigation task including the expected object version of the target security object is generated. The investigation task is executed to obtain supplementary security evidence. The applicable object version of the supplementary security evidence is determined based on the expected object version and the current object version of the target security object. The evidence sufficiency matrix and the investigation hypothesis verification map are updated based on the supplementary security evidence applicable to the current object version. Based on the evidence sufficiency matrix, the remaining evidence requirements are determined, and the investigation tasks corresponding to the remaining evidence requirements are determined from the generated investigation tasks. Based on the current attack path, affected objects and the remaining evidence requirements obtained by causal tracing, candidate actions are generated to bind the target security object version. According to the impact of the candidate actions on the state change of the unfixed evidence corresponding to the remaining evidence requirements, the execution sequence of the candidate actions and the investigation tasks to be executed is determined. When the current version of the target security object corresponding to the candidate action matches the version of the target security object bound to the candidate action, and the candidate action satisfies the execution sequence, the candidate action is executed through the security device, and investigation conclusion data is generated based on the action execution result, the evidence sufficiency matrix, and the investigation hypothesis verification map.
2. The method according to claim 1, characterized in that, The process of generating evidence requirement items based on preset evidence roles corresponding to investigation hypothesis data to construct an investigation hypothesis verification map and an evidence sufficiency matrix includes: Evidence requirement items are generated for each preset evidence role corresponding to the survey hypothesis data, and the conditions for satisfying the evidence requirement items are determined according to the evidence source type, independent evidence source requirements, verification time range, target security object range and evidence field conditions corresponding to the preset evidence role. Using the survey hypothesis data and the preset evidence roles as the data dimensions of the initial evidence sufficiency matrix, the matrix unit corresponding to each of the evidence requirement items in the initial evidence sufficiency matrix is determined, and the association between the evidence requirement items and the corresponding matrix units is established in the initial evidence sufficiency matrix to obtain the evidence sufficiency matrix. The survey hypothesis data, the evidence requirements, the existing security evidence, and the ability to acquire candidate evidence are used as nodes in the initial survey hypothesis verification graph. The node relationships in the survey hypothesis verification graph are established to generate the survey hypothesis verification graph.
3. The method according to claim 2, characterized in that, The process of establishing the node relationships in the survey hypothesis verification graph includes: Based on the survey hypothesis data corresponding to each of the evidence requirement items, establish the requirement verification relationship between the node corresponding to the survey hypothesis data and the node corresponding to the evidence requirement item. Based on the support or negation of the existing security evidence for the corresponding evidence requirement item, establish a supporting evidence relationship or a rebuttal relationship between the node corresponding to the evidence requirement item and the node corresponding to the existing security evidence. Based on the conditions for satisfying each of the evidence requirements, candidate evidence acquisition capabilities that can obtain the evidence required by the conditions are determined, and evidence acquisition relationships are established between the nodes corresponding to the evidence requirements and the nodes corresponding to the determined candidate evidence acquisition capabilities.
4. The method according to claim 1, characterized in that, The causal tracing along the object relationships, and determining the satisfaction status of each of the evidence requirements based on the obtained tracing evidence, includes: Based on the preset evidence roles corresponding to the survey hypothesis data, candidate attack phases corresponding to the preset evidence roles are determined in the security object; Based on the security objects and event time ranges corresponding to the candidate attack stages, the predecessor security objects of the security objects are searched in reverse along the object relationships, and the successor security objects generated or affected by the security objects are searched in forward along the object relationships to form candidate tracing paths. For the object relationship to be verified in the candidate tracing path, the corresponding original security evidence is obtained by referencing the original evidence in the security observation data associated with the object relationship to be verified, and the original security evidence is determined as the tracing evidence. Based on the evidence source, event time, associated security object, object version, and evidence fields corresponding to the traceability evidence, determine the degree to which the traceability evidence meets the corresponding evidence requirement, and determine the satisfaction status of the evidence requirement based on the degree to which it meets the requirement.
5. The method according to claim 4, characterized in that, Determining the satisfaction status of the evidence requirement item based on the satisfaction status includes: For the relationships between objects to be verified in the candidate tracing paths, the integrity of the preceding security evidence, the event time of the tracing evidence, the object version of the associated security object, the independence of the evidence source, and the availability of the original security evidence are verified to obtain the verification results. When the verification result indicates that the relationship between the objects to be verified lacks corresponding predecessor security evidence, the corresponding evidence requirement item is determined to be unmet, and the corresponding position is determined as the source breakpoint of the predecessor evidence missing type. When the verification result indicates that the source evidence from mutually independent sources has mutually exclusive results for the same target fact, the corresponding evidence requirement item is determined to be in conflict state, and the corresponding position is determined to be the source breakpoint of the evidence conflict type. When the verification result indicates that the source of evidence used to verify the relationship between the objects to be verified does not meet the independence requirement, the corresponding evidence requirement item is determined to be in a source-dependent state, and the corresponding position is determined to be a source-dependent type source breakpoint. When the verification result indicates that the original security evidence has reached the critical condition of the preset retention period, the relationship between the objects to be verified lacks original evidence references, the data access required for verification is restricted, or the data of the target security object needs to be actively obtained for further verification, the satisfaction status of the corresponding evidence requirement item is updated, and the corresponding type of tracing breakpoint is determined.
6. The method according to claim 1, characterized in that, The process involves generating an investigation task for the unmet evidence requirements corresponding to the detected source tracing breakpoints, including the expected object version of the target security object, and executing the investigation task to obtain supplementary security evidence, including: Based on the unmet evidence requirements, the target security object, the expected version of the target security object, the investigation time range, the candidate data source, the expected result structure, the retention deadline, the maximum evidence intrusion level, and the result write-back target are determined to generate the corresponding investigation task. Based on the online status, access permissions, retention status, and maximum intrusion level of the candidate data sources, the data acquisition channel corresponding to the investigation task is determined. The investigation task is performed through the data acquisition channel to obtain supplementary security evidence corresponding to the unmet evidence requirements. When there is no data acquisition channel that satisfies the investigation task, the unavailability status of the corresponding evidence requirement item is recorded, and the evidence requirement item is kept in an unsatisfied state.
7. The method according to claim 1, characterized in that, The step of determining the applicable object version of the supplementary security evidence based on the expected object version and the current object version of the target security object, and updating the evidence sufficiency matrix and the investigation hypothesis verification map based on the supplementary security evidence applicable to the current object version, includes: Compare the expected object version corresponding to the investigation task with the current object version of the target security object; When the expected object version matches the current object version, the current object version is determined as the applicable object version for the supplementary security evidence; When the expected object version does not match the current object version, determine whether the version change between the expected object version and the current object version involves object attributes or object relationships used to determine the satisfaction status of the corresponding evidence requirement item; When the version change does not involve object attributes and object relationships used to determine the satisfaction status of the evidence requirement, the current object version is determined as the applicable object version of the supplementary security evidence; when the version change involves object attributes or object relationships used to determine the satisfaction status of the evidence requirement, the supplementary security evidence is associated with the corresponding historical object version, and it is determined whether to generate the corresponding supplementary evidence requirement based on the current object version. The supplementary security evidence applicable to the current object version is verified for independence of evidence source, scope of evidence coverage and evidence conflict, and the satisfaction status of the corresponding evidence requirement items is updated according to the verification results, so as to update the evidence sufficiency matrix and the investigation hypothesis verification map.
8. The method according to claim 1, characterized in that, The step of determining the execution sequence of the candidate disposal action and the investigation task to be performed based on the impact of the candidate disposal action on the status change of the unfixed evidence corresponding to the remaining supplementary evidence requirements includes: Determine the level of destructiveness of the candidate actions to the unsecured evidence and the reversibility of the candidate actions; Based on the remaining supplementary evidence requirements that need to be fixed before the candidate disposal action is executed, the evidence dependency items corresponding to the candidate disposal action are determined. The corresponding containment level is determined based on the scope of impact and spread of the cybersecurity incident. Based on the level of destructiveness, the reversibility, the evidence dependency, and the level of containment, the execution order constraints between the candidate disposal actions and the corresponding investigation tasks to be executed are determined, and the execution sequence is generated based on the execution order constraints.
9. The method according to claim 8, characterized in that, The step of determining the corresponding containment level based on the impact scope and spread status of the cybersecurity incident, and determining the execution order constraint based on the level of destructiveness, the reversibility, the evidentiary dependencies, and the containment level, includes: The number of newly affected security objects and the number of newly lateral movement object relationships are counted within a preset time range, and the degree of attack spread is determined based on the number of newly affected security objects and the number of newly lateral movement object relationships. Based on the criticality level of the affected security object, the extent of attack spread, and the type of threat impact, the containment level corresponding to the network security incident is determined. The containment level includes monitoring only, delayed containment, and immediate containment. When the containment level is monitoring only, the candidate actions to change the state of the target security object are kept in a pending state, and security monitoring of the network security incident is maintained. When the containment level is a delayed containment, the candidate action for evidence dependency items that have not yet completed evidence fixation and whose destructive level has reached a preset destructive level is set to be executed after the corresponding investigation task to be performed. When the containment level is immediate containment, the isolation-type candidate actions that are reversible and used to contain the network security incident are set as priority actions, and the cleanup-type candidate actions that have not yet been secured and whose destructive level has reached the preset destructive level are kept in a pending state until the corresponding unsecured evidence is secured or the execution permission for the cleanup-type candidate actions is obtained.
10. The method according to claim 1, characterized in that, The step of executing the candidate treatment action through a security device and generating investigation conclusion data based on the action execution result, the evidence sufficiency matrix, and the investigation hypothesis verification map includes: Obtain the version of the target security object bound to the candidate disposal action and the current version of the target security object before the action is executed, and perform version verification between the target security object version and the current object version; When the version verification passes, the preconditions of the candidate disposal action are verified, and when the preconditions are met and the candidate disposal action meets the execution sequence, the corresponding security device is invoked to execute the candidate disposal action. Obtain the action execution result returned by the security device, determine the state change of the target security object based on the action execution result, and update the object version of the target security object based on the state change; Based on the updated object version, determine the evidence requirement items affected by the state change, update the evidence sufficiency matrix and the investigation hypothesis verification graph, and generate the investigation conclusion data based on the updated evidence sufficiency matrix, the investigation hypothesis verification graph, and the action execution results.
