A campus network security situation assessment method based on abnormal traffic characteristics
By analyzing and correlating the abnormal traffic characteristics in the campus IPv6 network, the problem of fragmented abnormal behavior caused by changes in terminal temporary addresses was solved, enabling continuous reconstruction of abnormal behavior and accurate assessment of threat propagation, thereby improving the accuracy and reliability of campus network security situation assessment.
Patent Information
- Application Number
- CN202611271769.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2026-08-20
- Publication Date
- 2026-11-10
AI Technical Summary
In an IPv6 network environment, frequent changes in terminal temporary addresses and wireless roaming lead to the dispersion of abnormal traffic behavior. Existing technologies struggle to accurately identify and assess the security posture of campus networks, resulting in abnormal behavior being identified as multiple independent events, which affects the accuracy of threat propagation range and evolution trends.
By acquiring abnormal traffic fragment data, we can perform temporary address drift correlation and cross-address behavior continuity determination, reconstruct abnormal behavior chains, and conduct threat chain inference to restore the continuity and propagation relationship of abnormal behavior and improve the accuracy of security situation assessment.
It effectively reduces the fragmentation and downgrading of anomalies caused by address drift, improves the ability of campus network security situational data to characterize the scope and evolution trend of threat propagation, and enhances the accuracy of identifying and assessing abnormal behavior.
Smart Images

Figure CN123074789A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of computer network security technology, and in particular to a method for assessing campus network security situation based on abnormal traffic characteristics. Background Technology
[0002] As the scale of IPv6 access in campus networks continues to expand, student terminals, teacher terminals, and mobile smart devices frequently access the network across different areas such as teaching buildings, laboratories, libraries, and dormitories. Network security equipment needs to continuously identify abnormal network behaviors such as port scanning, lateral movement, and unusual outbound connections. Existing abnormal traffic detection typically uses source IP address, destination IP address, port, protocol type, and traffic changes as analytical criteria, and aggregates abnormal access records corresponding to the same IP address to form a trajectory of abnormal terminal behavior. In an IPv6 network environment, when terminals use temporary IPv6 addresses for communication, the communication address may change with the address's lifecycle; especially in campus wireless networks, when terminals roam across wireless access points, changes in access location and temporary address replacements may occur simultaneously. Furthermore, during address updates, old and new temporary IPv6 addresses may coexist in a valid state, causing continuous abnormal activity from the same terminal to be distributed across multiple traffic records corresponding to different IPv6 addresses.
[0003] In this situation, existing abnormal aggregation methods based on a single IP address or independent network session are prone to creating behavioral breakpoints at address switching locations, causing continuous abnormal behavior to be identified as multiple independent abnormal events. When multiple terminals roam and update their addresses simultaneously, incorrect associations between traffic from different terminals may also occur. With multiple temporary addresses coexisting, the terminal attribution of abnormal traffic becomes more complex, leading to inaccurate campus network security situation assessments. Summary of the Invention
[0004] To address the aforementioned technical problems, this invention proposes a campus network security situation assessment method based on abnormal traffic characteristics, thereby resolving at least one of the aforementioned technical issues.
[0005] This application provides a campus network security situation assessment method based on abnormal traffic characteristics, the method comprising:
[0006] S1. Obtain abnormal traffic segment data;
[0007] S2. Perform temporary address drift association based on abnormal traffic fragment data to obtain address drift data;
[0008] S3. Based on address drift data, determine the continuity of cross-address behavior to obtain cross-address anomaly correlation data;
[0009] S4. Reconstruct the abnormal behavior chain based on cross-address abnormal correlation data to obtain abnormal behavior chain data;
[0010] S5. Based on the abnormal behavior chain data, perform cross-address threat chain inference to obtain security situation data.
[0011] This invention achieves a step-by-step connection between abnormal traffic fragment identification, temporary IPv6 address drift correlation, cross-address behavior continuity determination, abnormal behavior chain reconstruction, and cross-address threat chain extrapolation. This allows for the re-establishment of continuous correlations for abnormal behaviors in campus networks that have been fragmented due to periodic changes in IPv6 temporary addresses, wireless roaming, and the coexistence of multiple addresses. Compared to methods that rely solely on fixed IP addresses or single sessions for anomaly detection, this solution can continue to track the abnormal access trajectory of the same terminal after an address switch, and reconstruct the behavioral chains of scanning, lateral probing, and abnormal external connections by combining the temporal sequence of preceding and following behaviors. Extrapolating the cross-address threat propagation of the reconstructed abnormal behavior chains can identify the evolutionary relationships of threats across different nodes, network segments, and stages, thereby reducing the discretization and low-risk nature of abnormal events caused by address drift and improving the ability of campus network security situational data to characterize the actual threat propagation range, duration, and evolutionary trends.
[0012] Optionally, S1 includes:
[0013] Obtain raw IPv6 traffic data from the campus network;
[0014] Traffic session aggregation is performed on raw IPv6 traffic data to obtain IPv6 traffic session data.
[0015] Local anomaly deviations are extracted from IPv6 traffic session data to obtain abnormal traffic data.
[0016] The abnormal traffic data is segmented into time-series segments to obtain abnormal traffic segment data.
[0017] This invention performs session-level aggregation on raw IPv6 traffic from the campus network, organizing discrete packets into traffic sessions with complete interactive characteristics based on communication objects, protocol relationships, and temporal continuity. This reduces the interference of single-packet fluctuations on anomaly identification results. Through local anomaly deviation extraction, the current session is compared with normal traffic characteristics under corresponding time periods and similar business conditions. This highlights local anomaly changes such as access frequency, destination node distribution, port changes, and packet rhythm, avoiding the direct identification of normal bursts of traffic like peak course access or concentrated downloads as anomalies. Finally, through time-series segmentation, anomaly sessions that are temporally continuous or behaviorally sequential are grouped into independent anomaly traffic segments.
[0018] Optionally, S2 includes:
[0019] Based on abnormal traffic fragment data, roaming coupling drift and multi-address ambiguity resolution are performed to obtain first address drift data and second address drift data, respectively;
[0020] The drift data at the first address and the drift data at the second address are overlapped to obtain the drift overlapped data;
[0021] Behavioral address disambiguation is performed on the drift-overlapping data to obtain drift-related data;
[0022] Drift timing reconstruction is performed based on drift correlation data to obtain address drift data.
[0023] This invention addresses the address attribution uncertainty caused by the coexistence of terminal roaming and temporary addresses in campus IPv6 networks through a layered approach: Address drift relationships are extracted from two dimensions: roaming coupling drift caused by AP handover and multi-address ambiguity caused by the coexistence of multiple temporary IPv6 addresses for the same terminal. This allows address changes under different causes to be independently characterized. By overlapping drifts, the two types of drift relationships are uniformly constrained within the time, access location, and behavioral scope, narrowing down the set of candidate addresses belonging to the same terminal. Behavioral address disambiguation eliminates erroneous associations that are only close in time or access location but have discontinuous abnormal access behaviors, reducing the probability of address misbinding between different terminals. Drift time sequence reconstruction restores the continuous evolutionary order of temporary IPv6 address generation, coexistence, handover, and failure, thereby forming a stable address drift chain.
[0024] Optionally, the roaming coupling drift includes:
[0025] Access handover data is obtained by extracting abnormal traffic fragment data;
[0026] Address replacement adjacency resolution is performed based on roaming handover data to obtain address replacement data;
[0027] Migration path determination is performed on the address replacement data to obtain roaming address association data;
[0028] Based on the roaming address association data, coupled drift reconstruction is performed to obtain the first address drift data.
[0029] This invention analyzes the coupling relationship between "access location change - temporary IPv6 address change" during cross-AP roaming layer by layer, avoiding terminal misassociation caused by simple matching based solely on address appearance time. By extracting access handover data to determine the AP migration sequence corresponding to abnormal traffic segments, spatial constraints on potential terminal roaming are formed. Combined with address replacement adjacency resolution, the time relationship between the cessation of activity of an old address and the start of activity of a new address is included in the same migration window, thereby narrowing down the candidate address replacement range. Through migration path determination, consistency verification is performed on the candidate address relationship with the actual AP topology adjacency relationship, migration direction, and abnormal traffic continuation characteristics, eliminating erroneous address combinations that, although time-proximity, do not conform to the actual roaming path. Through coupled drift reconstruction, the access node migration order and temporary address replacement order are jointly restored to a continuous drift relationship, thereby improving the accuracy of identifying the same terminal address evolution in cross-AP scenarios.
[0030] Optionally, the multiple address attribution ambiguity resolution includes:
[0031] The abnormal traffic fragment data is identified by address coexistence intervals to obtain multi-address coexistence data;
[0032] Based on the multiple access coexistence data, abnormal fragment attribution aggregation is performed to obtain multiple access attribution data;
[0033] Parallel address inheritance determination is performed on the multi-address data to obtain multi-address association data;
[0034] The second address drift data is obtained by resolving the attribution ambiguity based on the multi-address attribution association data.
[0035] This invention addresses the issue of multiple temporary addresses coexisting and abnormal traffic being distributed during IPv6 privacy address updates for the same campus terminal by implementing a step-by-step disambiguation mechanism. First, it identifies the temporal overlap of different temporary IPv6 addresses by identifying address coexistence intervals, establishing validity constraints for determining whether addresses belong to the same terminal. Second, it aggregates abnormal traffic fragments with similar destination nodes, port access order, packet direction, and connection rhythm into candidate multiple access relationships, reducing the probability of irrelevant addresses participating in subsequent judgments. Third, it identifies whether there is session inheritance and abnormal behavior continuation relationship between old and new addresses during coexistence, thus distinguishing between genuine address update inheritance and accidental concurrency between different terminals. Fourth, it determines the true ownership and drift order among temporary IPv6 addresses through disambiguation, reducing abnormal behavior splitting, duplicate counting, and erroneous splicing caused by one-to-many address coexistence, improving the reliability of subsequent cross-address behavior continuity judgment and abnormal behavior chain reconstruction.
[0036] Optionally, the behavior address disambiguation includes:
[0037] Based on drift overlap data, address behavior competition map is constructed to obtain behavior attribution competition map data;
[0038] Inheritance residual fingerprints are extracted based on behavior attribution competition graph data to obtain behavior inheritance fingerprint data;
[0039] Based on behavioral inheritance fingerprint data, counterfactual attribution exclusion is performed to obtain address attribution exclusion data;
[0040] Based on address-based exclusion data, unique inheritance chain convergence is performed to obtain drift-related data.
[0041] This invention constructs address behavior competition relationships and resolves the multi-candidate attribution problem layer by layer, enabling temporary IPv6 addresses in drift-overlapping scenarios to converge from "potential associations" to "unique inheritance." Through address behavior competition mapping, multiple candidate addresses and their behavioral inheritance relationships corresponding to the same abnormal traffic segment are made explicit. Inheritance residual fingerprints are extracted, highlighting the characteristics of continuous destination node access, port hopping, packet rhythm, and direction changes before and after address switching, thus reducing interference caused by shared traffic patterns between different addresses. Counterfactual attribution exclusion is used to examine whether each candidate attribution relationship will cause session timing reversal, behavior sequence breakage, or connection duplication, thereby proactively eliminating erroneous address associations that do not conform to the true behavior inheritance logic. Through unique inheritance chain convergence, the remaining address relationships form a stable and continuous drift association chain along the behavior inheritance direction, effectively reducing the probability of misattribution under conditions of multi-terminal concurrency, address coexistence, and roaming overlap.
[0042] Optionally, S3 includes:
[0043] Cross-address behavior data is obtained by extracting cross-address behavior data based on address drift data.
[0044] The behavior status continuation parsing of cross-address behavior data is performed to obtain cross-address behavior continuation data;
[0045] Behavioral inertia extrapolation is performed based on cross-address behavior continuation data to obtain behavior inertia matching data;
[0046] Breakpoint conflict detection is performed based on behavioral inertia matching data to obtain continuous data across addresses;
[0047] By merging continuous behavior chains based on continuous data across addresses, cross-address anomaly correlation data is obtained.
[0048] This invention restores the continuity of abnormal behavior before and after address drift, enabling the re-establishment of stable behavioral continuity for abnormal activities interrupted by IPv6 temporary address switching. By extracting and locking abnormal traffic segments before and after address drift through cross-address behavior, key features such as access objects, port changes, packet direction, and timing are retained to provide behavioral state continuation analysis, representing the state connection between the end behavior of the previous address and the beginning behavior of the next address. Based on this, behavioral inertia extrapolation is used to predict the trend of abnormal behavior that may continue under conditions where the address has not changed, and this is matched with actual subsequent behavior, thereby improving the ability to judge the continuity of cross-address behavior. Furthermore, by identifying and eliminating erroneous associations with sudden changes in access targets, reversed timing, or abrupt changes in behavior stages through breakpoint conflict discrimination, abnormal behaviors that meet the continuity conditions are finally chained together. This effectively reduces the problems of fragmentation and mis-splitting of abnormal events caused by temporary address changes, providing cross-address abnormal association data with clear continuity for subsequent abnormal behavior chain reconstruction and threat propagation simulation.
[0049] Optionally, S4 includes:
[0050] Based on cross-address anomaly correlation data, node-based anomaly node data is obtained;
[0051] Based on the abnormal node data, behavior-connecting edges are constructed to obtain abnormal behavior topology data;
[0052] Link bifurcation and convergence are performed on the abnormal behavior topology data to obtain abnormal behavior link data;
[0053] Cross-address chain sequence reconstruction is performed based on abnormal behavior chain data to obtain abnormal behavior chain data.
[0054] This invention transforms cross-address anomaly association results into anomaly behavior chains with clear node relationships and propagation order, elevating anomalous activities scattered across multiple temporary IPv6 addresses from "association fragments" to "structured links." Node-based processing unifies different time locations, addresses, and stages of anomalous behavior into anomalous nodes, avoiding interference from inconsistent granularity of original traffic fragments in subsequent link analysis. Behavioral connection edges are constructed based on temporal continuity, target access continuation, and behavioral state transition relationships between nodes, forming a topology that describes the propagation direction of anomalous activities. Link forking and convergence filters multiple candidate successors for the same anomalous node, eliminating invalid branches that do not meet propagation continuity or state evolution rules, reducing path divergence and erroneous splicing during anomaly chain reconstruction. Cross-address chain order reconstruction restores the complete behavioral evolution chain according to the address drift order and the order of occurrence of anomalous events, making the connection between anomalous behaviors across different temporary IPv6 addresses clearer.
[0055] Optionally, S5 includes:
[0056] Threat phase anchoring is performed based on abnormal behavior chain data to obtain threat phase data;
[0057] Based on the threat phase data, the cross-address propagation potential is recursively extrapolated to obtain the threat propagation potential data;
[0058] Threat chain simulation data is obtained by performing link transition simulation based on threat propagation potential data.
[0059] Security situation data is obtained by mapping the situation boundary based on threat chain simulation data.
[0060] This invention elevates the campus network security posture from static anomaly results to a dynamic representation with propagation direction and evolution boundaries by performing phased analysis, propagation state recursion, and link evolution simulation on the reconstructed abnormal behavior chain. Threat stage anchoring determines the positional relationships of different stages in the abnormal behavior chain, such as detection, access, lateral spread, and abnormal external connections, giving each anomaly node a clear threat semantic. Cross-address propagation potential recursion transmits the threat state formed by preceding nodes along the behavioral succession relationship, and dynamically updates the propagation strength based on newly added target nodes, cross-network segment extent, and anomaly persistence, thus reflecting the cumulative effect of the threat across different temporary IPv6 addresses and network areas. Link transition simulation identifies potential paths for the threat to expand from the current stage to subsequent nodes and network layers, avoiding a security posture that remains merely a static statistical representation of anomalies. Posture boundary mapping maps the already covered and potentially extended range of the threat chain to the boundaries of campus network nodes and network segments, more accurately representing the current threatened area, propagation depth, and evolution trend.
[0061] The purpose of this invention is to address the problem of frequent changes in temporary addresses in a campus IPv6 network environment, which makes it easy to fragment abnormal behavior chains. It constructs a progressive security posture assessment process consisting of abnormal traffic fragment acquisition, temporary address drift correlation, cross-address behavior continuity determination, abnormal behavior chain reconstruction, and cross-address threat chain deduction. By performing session aggregation, local anomaly deviation extraction, and anomaly timing segmentation on raw IPv6 traffic, scattered packet data is converged into abnormal traffic fragments with behavioral continuity, reducing the interference of sudden access from normal services on anomaly identification. For the address attribution uncertainty caused by terminal roaming across APs and the parallel existence of multiple temporary IPv6 addresses, roaming coupling drift, multi-address attribution ambiguity, drift overlap, and behavior address disambiguation are used to restore the true timing inheritance relationship between different temporary addresses, preventing the abnormal behavior of the same terminal from being split into multiple independent low-risk events. By employing behavioral state continuation, behavioral inertia extrapolation, and breakpoint conflict discrimination, the continuity of abnormal activities before and after address switching is verified. Furthermore, through nodeization, behavioral continuation edge construction, link fork convergence, and cross-address chain sequence reconstruction, cross-address abnormal fragments are restored into abnormal behavioral chains with clear propagation directions and phase sequences. Through threat phase anchoring, cross-address propagation potential recursion, link transition deduction, and situation boundary mapping, the propagation range, evolution depth, and potential expansion direction of threats across different terminal addresses, network nodes, and network segments are dynamically characterized. This enhances the campus network's ability to identify and correlate covert threats such as low-speed scanning, lateral probing, and continuous abnormal external connections under IPv6 address drift conditions, and enables the obtained security situation data to more accurately reflect the continuity, propagation, and evolutionary trends of real threats. Attached Figure Description
[0062] Other features, objects, and advantages of this application will become more apparent from the following detailed description of the non-limiting embodiments, taken with reference to the accompanying drawings:
[0063] Figure 1 A flowchart illustrating the steps of a campus network security situation assessment method based on abnormal traffic characteristics is shown in one embodiment.
[0064] Figure 2 A flowchart illustrating the steps of an embodiment of a method for acquiring abnormal traffic fragment data is shown.
[0065] Figure 3 A flowchart illustrating the steps of a temporary address drift association method according to an embodiment is shown.
[0066] Figure 4 A flowchart illustrating the steps of a method for determining cross-address behavior continuity according to an embodiment is shown.
[0067] Figure 5 A flowchart illustrating the steps of an abnormal behavior chain reconstruction method according to one embodiment is shown.
[0068] Figure 6 A flowchart illustrating the steps of a cross-address threat chain deduction method according to one embodiment is shown;
[0069] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0070] The technical method of the present invention will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.
[0071] Furthermore, the accompanying drawings are merely illustrative of the invention and are not necessarily drawn to scale. Functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor methods and / or microcontroller methods.
[0072] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.
[0073] The IPv6 campus network of a university's teaching area, experimental area, and dormitory area was selected as the test environment. Network traffic was continuously collected for 24 hours at the campus network egress, core switching node, and wireless access aggregation node, yielding approximately [data missing]. One IPv6 packet. Set session timeout threshold. Fixed time window The number of messages deviates from the threshold The number of bytes and total bytes deviate from the threshold Duration deviates from threshold Average message interval deviates from threshold Destination IPv6 address switching threshold Abnormal session fragment merging time threshold After processing, 1268 abnormal sessions were extracted from 86420 IPv6 traffic sessions, and 317 abnormal traffic segments were obtained.
[0074] Set address roaming switching time threshold For example, the same terminal connects sequentially. , , (These represent the three wireless access nodes that the terminal sequentially connected to during the period of abnormal behavior), and the corresponding abnormal traffic was generated by temporary IPv6 addresses. , , The time intervals between abnormal segments corresponding to adjacent addresses were 12s and 18s, respectively, both less than 30s; simultaneously, the destination address, destination port, and protocol combination of the abnormal segments corresponding to the three addresses remained continuous. Through roaming coupling drift, multi-address home ambiguity resolution, and behavioral address disambiguation, the records originally identified as three independent address events were restored to [the original text is incomplete and requires further context]. The same address drift chain. 317 abnormal segments formed 94 valid address drift relationships, of which 89 were manually verified to be correctly assigned, and the address drift association accuracy rate reached 94.68%.
[0075] Set cross-address behavior continuation time threshold Number of behavioral inertia backtracking sessions L=3, message interval allowable deviation threshold Regarding the above During the handover process, the average packet intervals of the three abnormal sessions before the handover were 0.42s, 0.39s, and 0.36s, respectively. The predicted average packet interval after the handover was obtained based on the most recent increment. The actual average interval between the first abnormal session is 0.35 seconds, which meets the requirements. Furthermore, the destination access order, protocol type, and message length change direction were continuous, thus classifying it as cross-address continuous behavior. After processing, the original 317 abnormal traffic segments were merged into 76 sets of cross-address abnormal correlation data.
[0076] Set an abnormal behavior link acceptance time threshold. By nodeifying the aforementioned cross-address anomaly-related data and constructing behavior-receiving edges, and after converging the forked links, a total of 41 anomaly behavior chains were reconstructed. One typical behavior chain consists of 9 anomaly nodes, including continuous port probing, concentrated connection to fixed targets, access to multiple campus internal targets, and continuous anomaly communication, spanning 3 temporary IPv6 addresses and 3 wireless access nodes. When using the traditional aggregation method based on a single IPv6 address, this process is divided into 4 independent anomaly events, while this method restores it to a complete anomaly behavior chain.
[0077] Set cross-address threat propagation time threshold Based on a preset threat phase parameter table, the system sequentially anchors the port probing phase, centralized access attempt phase, lateral access phase, and continuous abnormal communication phase, and recursively calculates the propagation potential based on the allowed propagation set. The aforementioned typical abnormal behavior chain starts from the initial propagation potential. Step by step to The impact was determined to involve 3 wireless access areas and 7 internal IPv6 target nodes on campus, lasting for 11 minutes and 36 seconds, thus obtaining the corresponding security situation data.
[0078] On the same test data, a method aggregating anomalous sessions solely based on fixed IPv6 addresses was used as a control. The control method broke down 47 manually confirmed continuous cross-address threat processes into 82 independent events, of which 31 showed broken behavioral chains. Using this method, only 6 events failed to be fully recovered, increasing the complete threat chain identification rate from 34.04% to 87.23%, an increase of 53.19 percentage points. The number of false cross-address anomalies decreased from 17 groups to 5 groups, a reduction of 70.59%. After manual review, the accuracy rate of identifying the address range affected by security posture increased from 76.60% to 93.62%. This demonstrates that this method can effectively recover the continuous cross-address relationships of anomalous behavior in situations involving IPv6 temporary address switching, multiple address coexistence, and wireless roaming, reducing fragmentation of anomalous events and misattribution of addresses, and improving the accuracy of assessing the scope and depth of threat propagation and evolution in campus networks.
[0079] Please see Figures 1 to 6 This application provides a campus network security situation assessment method based on abnormal traffic characteristics, the method comprising:
[0080] S1. Obtain abnormal traffic segment data;
[0081] In one embodiment, IPv6 packet records from the campus network egress, core switching node, and wireless access aggregation node are obtained, and the source IPv6 addresses are extracted. Destination IPv6 address Source / Destination Port , The transmission protocol q, timestamp t, message length l, and TCP flags form the raw IPv6 traffic data. As a session key, packets with the same K and adjacent message time intervals satisfy... Messages are grouped into the same session; if the session timeout threshold is exceeded... If a new session is established, then the duration T, number of messages N, total number of bytes B, and average message interval for each session are recorded. The number of uplink and downlink packets is used to obtain IPv6 traffic session data. Sessions within the same access area are then compared using a fixed time window W=300s. For the first... IPv6 traffic session data , , For the ordinal term, calculate the local median. and absolute deviation When any of the following parameters are true: message count, number of bytes, duration, or message interval... ,like , , , Or the number of destination address switches per unit time exceeds the threshold. When this happens, the corresponding session is marked as abnormal / abnormal session, and abnormal traffic data is obtained. Abnormal sessions are arranged in chronological order; if adjacent abnormal sessions meet the following conditions... , This represents the start time of the (j+1)th abnormal session. If the end time of the j-th abnormal session is specified, the sessions are merged into the same segment; otherwise, they are split into new segments. The start and end times of each segment, the IPv6 addresses involved, the session sequence, and the abnormal characteristics are recorded to obtain abnormal traffic segment data.
[0082] S2. Perform temporary address drift association based on abnormal traffic fragment data to obtain address drift data;
[0083] In one embodiment, for IPv6 addresses within abnormal traffic segments, address switching relationships are extracted based on access node identifiers, segment start and end times, and session behavior. If the address corresponding segment end time With address corresponding segment start time satisfy If both access nodes belong to the same roaming path or adjacent wireless coverage areas, an address replacement adjacency relationship is established. A migration path is generated along the continuous replacement relationships to obtain the first address drift data. That is, the established address replacement adjacency relationships are sorted according to their occurrence time. When the starting address of a later address replacement adjacency relationship is the same as the ending address of a previous relationship, the two relationships are connected end-to-end. Abnormal segments carried by multiple IPv6 addresses are retrieved within the same time interval; if the address... , If the activity intervals of the two segments overlap, and at least two of their destination address set, destination port set, and protocol type are identical, they are grouped into the same multiple access candidate group. Then, based on the order in which the abnormal sessions occur, it is determined whether subsequent segments inherit the behavior state of the preceding address, thus obtaining the second address drift data. The first and second address drift data are then overlapped according to their time intervals and the addresses involved. For two drift records... , Let i be the start time of the i-th address drift record. Let be the end time of the i-th address drift record. Let j be the start time of the address drift record. Let j be the end time of the address drift record. When multiple drift records contain at least one identical candidate address, they are merged into the same drift overlap unit to obtain drift overlap data. The identical candidate address refers to an IPv6 address in the first or second address drift data that is determined to have a sequential inheritance relationship with the same abnormal behavior segment. For each overlap unit (i.e., a single local association result formed after overlapping the time interval and candidate addresses of a pair of first and second address drift records), a "candidate address - abnormal behavior" competition relationship is constructed; and a candidate address is removed from each unit. And check whether the remaining addresses can maintain the destination, port, protocol, and time continuation relationship of the original abnormal session; if removed If the aforementioned continuity breaks, but can be maintained when other addresses are removed, then it is determined that... For the inherited addresses of the corresponding behaviors, disambiguation of behavior addresses is performed to obtain drift association data. These are then sorted in ascending order of the first occurrence time of each confirmed inherited address, ensuring that adjacent time intervals do not exceed [a certain value]. The records are connected consecutively to form The address drift sequence is recorded, and the corresponding abnormal segment number, switching time and access node are recorded to obtain address drift data.
[0084] S3. Based on address drift data, determine the continuity of cross-address behavior to obtain cross-address anomaly correlation data;
[0085] In one embodiment, adjacent addresses in the address drift data Extract the corresponding anomaly fragment and construct the behavior state according to the destination IPv6 address, destination port, transport protocol, packet direction, and anomaly type. ,in For the destination address, For destination port, Protocol type, For message direction, This is an anomaly type, yielding cross-address behavior data. A continuation judgment is made regarding the behavior state before and after the address switch. When the behavior corresponding to adjacent addresses meets... , And the destination addresses are the same or belong to the same set of destination addresses, and at the same time If the behavior of the subsequent address inheriting from the previous address is confirmed, then it is determined that the behavior is inherited by the subsequent address; otherwise, it is marked as a point to be judged, and cross-address behavior continuation data is obtained. For the point to be judged, the change trend is calculated based on the packet interval, packet length, and destination access order of the L=3 consecutive abnormal sessions before the handover, such as the packet interval increment. , The average message interval for the k-th abnormal session. The prediction interval is obtained by extrapolating the average packet interval of the (k-1)th abnormal session based on the first session after the most recent consecutive incremental handover. , The predicted average packet interval for the first abnormal session after the handover. This is the allowable deviation threshold for message interval inertial matching, such as 0.05s. If the actual interval of subsequent messages falls within the allowable deviation range... If the destination access order remains continuous and the direction of message length change remains consistent with that before the handover, then behavioral inertia matching data is obtained. This represents the actual average message interval of the first abnormal session after the address switch. Conflict detection is performed on the inertial matching results in the behavioral inertial matching data; if a protocol mutation occurs before or after the switch point, the destination set is completely separated, or the time interval... If the breakpoint is not continuous, it is determined that the breakpoint is discontinuous; otherwise, it is retained as a cross-address continuous candidate data. According to the address drift sequence, the continuous candidate records in the cross-address continuous candidate data that share consecutive abnormal segments and have no conflicting breakpoints are merged sequentially to form... It records the address switching time, associated abnormal segments, and continuity determination results to obtain cross-address abnormal association data. Let i be the i-th cross-address anomalous behavior unit, representing the IPv6 address. Its abnormal behavior state The correspondence.
[0086] S4. Reconstruct the abnormal behavior chain based on cross-address abnormal correlation data to obtain abnormal behavior chain data;
[0087] In one embodiment, each consecutive abnormal behavior segment in the cross-address abnormal correlation data is node-based, and the node is represented as... ,in, The IPv6 address that carries this abnormal behavior, For behavioral states, , These represent the start and end times of the behavior; nodes are divided according to the boundaries of the abnormal segments to obtain abnormal node data. For any node... , Perform behavior continuation determination; when The time interval satisfies Furthermore, a directed edge is established when the two parties have a previously confirmed connection relationship in terms of destination address, destination port, protocol type, or abnormal behavior state. This yields abnormal behavior topology data. For cases where multiple successor nodes exist for the same node, link forking and convergence are performed; nodes with time intervals exceeding [a certain threshold] are removed. Branches that conflict with the preceding behavior state; for the remaining candidate branches, compare their starting times and retain those that satisfy the condition. Furthermore, it can continue to connect to the branches of subsequent abnormal nodes; if different branches correspond to independent abnormal behaviors, they are retained as independent links to obtain abnormal behavior link data. Starting from a node without a predecessor, it traverses along the directed edges in ascending order of node start time; combining the address drift sequence, it rearranges nodes with different IPv6 addresses but continuous behavior in sequence. Record the node's corresponding address, abnormal behavior, connection relationship, and occurrence sequence to obtain abnormal behavior chain data.
[0088] S5. Based on the abnormal behavior chain data, perform cross-address threat chain inference to obtain security situation data.
[0089] In one embodiment, the abnormal behavior types, access objects, and occurrence order of each node in the abnormal behavior chain are analyzed. Based on a preset threat stage parameter table, behaviors such as port probing, centralized access attempts, lateral access, and continuous abnormal communication are anchored to their corresponding threat stages, and the nodes are represented as... ,in For IPv6 addresses, The threat phase, To determine the time of occurrence, threat stage data is obtained. A pre-defined threat stage parameter table is established based on campus network security management rules, historical confirmed anomaly handling records, and pre-configured behavior-stage correspondences by network security operations personnel. This table maps different anomaly behavior types to corresponding threat stages. Cross-site propagation potential is recursively calculated along the anomaly behavior chain; let nodes... The spread of the trend is The initial node satisfies When adjacent nodes belong to a previously confirmed cross-address contiguous chain and satisfy the following conditions: =30s, and the phase relationship Belongs to the pre-defined allowed propagation set season Otherwise, terminate the current propagation chain and make the new chain's first node... This yields threat propagation potential data, which includes a pre-defined set of permissible propagation. It is a set of stage transition relationships pre-established based on the allowed sequential evolution relationships between each threat stage in the aforementioned threat stage parameter table. For example, if the preset threat stages include the port detection stage in sequence... Centralized access trial phase Lateral access phase and continuous abnormal communication phase If the order of the smaller item indicates a precedence relationship, then set... ,in To allow the evolution from port probing to centralized access attempts, To allow for the evolution from centralized access to lateral access, To allow the evolution from lateral access to continuous anomalous communication, , To represent the permissible cross-threat stage transitions, link transitions are extrapolated for nodes with continuously increasing propagation potential; if... ,and for If the subsequent phase or the phase that allows skipping stages is followed, then a threat jump edge is established. If a stage reversal, time limit exceeding, or no continuity exists in the original abnormal behavior chain, the transition is terminated, and threat chain projection data is obtained. This includes the start and end times of each valid threat chain, the set of IPv6 addresses involved, the highest threat stage, and the maximum propagation potential. Establish situation boundaries; merge threat chains with overlapping timelines and shared anomalous nodes into a single security situation region, outputting its affected address range, duration, current threat stage, and propagation depth to obtain security situation data. and allowed propagation set These are preset parameters. All are derived from abnormal behavior chains and stage anchoring. It is generated by recursion.
[0090] Optionally, S1 includes:
[0091] S11. Obtain raw IPv6 traffic data from the campus network;
[0092] In one embodiment, IPv6 packets are collected at the campus network egress, core switching node, and wireless access aggregation node, and the source IPv6 address is extracted from each packet. Destination IPv6 address Source port Destination port The data includes the transmission protocol q, collection timestamp t, message length l, and message direction, arranged in ascending order of timestamp to form raw IPv6 traffic data.
[0093] S12. Perform traffic session aggregation based on the raw IPv6 traffic data to obtain IPv6 traffic session data;
[0094] In one embodiment, with As a session identifier, adjacent messages with the same session identifier will be considered. Packets are grouped into the same IPv6 session; when the time interval exceeds the preset session timeout threshold. The current session is terminated and a new session is established. For each session, the following statistics are collected: duration T, number of packets N, total number of bytes B, and average packet interval. The destination address and uplink / downlink packet sequences are used to obtain IPv6 traffic session data.
[0095] S13. Extract local anomaly deviations based on IPv6 traffic session data to obtain abnormal traffic data;
[0096] In one embodiment, sessions are divided according to a preset local time window W, and sessions within the same window and the same access area are treated as a local comparison set. This comparison is applied to any of the following characteristics: number of packets, total number of bytes, duration, and average packet interval. Calculate the corresponding local median and deviation When any feature satisfies Or the number of destination IPv6 address switches per unit time exceeds the preset address switch threshold. When this happens, the corresponding session is marked as an abnormal session, and its address, time, and abnormal characteristics are retained to obtain abnormal traffic data.
[0097] S14. Perform time-series segmentation on the abnormal traffic data to obtain abnormal traffic segment data.
[0098] In one embodiment, abnormal traffic data is sorted in ascending order according to the start time of the abnormal session. For adjacent abnormal sessions, the end time of the previous session is calculated. Start time of the next session The interval between; when If the two are in a state of flux, they are merged into the same abnormal traffic segment; otherwise, the previous session is used as the end point of the current segment, and a new abnormal traffic segment is created. The start and end times, involved IPv6 addresses, abnormal session sequence, and abnormal characteristics of each segment are recorded to obtain abnormal traffic segment data.
[0099] Optionally, S2 includes:
[0100] S21. Based on the abnormal traffic fragment data, perform roaming coupling drift and multi-address ambiguity resolution to obtain the first address drift data and the second address drift data respectively;
[0101] In one embodiment, the abnormal traffic segments are sorted according to their start and end times, IPv6 addresses, and access node identifiers. For addresses... End time of corresponding segment With address The start time of the corresponding segment When satisfied And when two access nodes belong to the same roaming path or adjacent wireless coverage areas, The roaming coupling drift relationship is identified, forming the first address drift data; at the same time, multiple IPv6 addresses with overlapping time intervals are identified in parallel. When the abnormal segments are consistent in at least two of the destination address, destination port and protocol type, they are assigned to the same multi-access candidate group, and the second address drift data is formed according to the order of occurrence of abnormal behavior.
[0102] S22. Perform drift overlap on the first address drift data and the second address drift data to obtain drift overlap data;
[0103] In one embodiment, the address sets and valid time intervals of the two types of drift records are extracted respectively. For the first drift record... Second drift record ,when and ( First drift record The effective time interval For the second drift record The effective time interval First drift record The set of candidate IPv6 addresses involved in the process. For the second drift record When dealing with the candidate IPv6 address set involved in the process, the two records are merged into the same drift overlap unit, and the common address, candidate address, corresponding abnormal segment and occurrence time sequence are retained to obtain drift overlap data.
[0104] S23. Perform behavioral address disambiguation on the drift-overlapping data to obtain drift-related data;
[0105] In one embodiment, for each drift overlap unit, an attribution relationship is established between the candidate IPv6 address and the corresponding abnormal behavior; candidate addresses are then removed sequentially. Check if the remaining addresses can still maintain the destination address, destination port, protocol type, and time continuity between the abnormal segments. If removed... If the original continuity relationship is broken, but the continuity relationship still holds when other candidate addresses are removed, then... The actual inherited address of the abnormal behavior is determined, and the ownership conflict of multiple addresses corresponding to the same behavior is eliminated accordingly to obtain drift association data.
[0106] S24. Based on the drift correlation data, perform drift timing reconstruction to obtain address drift data.
[0107] In one embodiment, drift-related records are arranged in ascending order according to the first occurrence time of the abnormal segment corresponding to each confirmed inherited address; for adjacent records, if the address switching time interval still meets the requirement... If a successor drift relationship is established, then the current drift sequence is terminated and a new drift sequence is established, thus forming... The address drift chain is recorded, and the abnormal segments, access nodes and switching times corresponding to each address are recorded to obtain address drift data.
[0108] Optionally, the roaming coupling drift includes:
[0109] Access handover data is obtained by extracting abnormal traffic fragment data;
[0110] In one embodiment, the access nodes involved in the same continuous abnormal behavior are sorted according to the occurrence time of the abnormal traffic segments; when the preceding access node of an adjacent segment... With the next access node They are different, and both belong to adjacent coverage areas in the campus wireless network, while their corresponding segments satisfy... At that time, Once a valid access handover is identified, the IPv6 address, access node, and handover time before and after the handover are recorded to obtain roaming handover data.
[0111] Address replacement adjacency resolution is performed based on roaming handover data to obtain address replacement data;
[0112] In one embodiment, for each valid access handover, the IPv6 addresses used by the abnormal traffic in the terminal segment before the handover are compared. The IPv6 address used by the first abnormal traffic segment after the switchover ;when When the abnormal segments of the two are consistent in at least two of the destination address, destination port, and transport protocol, an agreement is established. The address replacement adjacency relationship is established; if the address has not changed, only the access handover record is retained and no address replacement edge is established, thus obtaining the address replacement data.
[0113] Migration path determination is performed on the address replacement data to obtain roaming address association data;
[0114] In one embodiment, the edges are arranged in ascending order of their occurrence time, and when two consecutive edges are replaced... and Shared intermediate address The corresponding access nodes are sequentially adjacent, and the time interval between adjacent replacement events does not exceed [a certain value]. When the time comes, the two replacement edges are concatenated; the above judgments are executed sequentially to form Candidate migration paths are obtained, and roaming address association data is acquired.
[0115] Based on the roaming address association data, coupled drift reconstruction is performed to obtain the first address drift data.
[0116] In one embodiment, candidate migration paths are aligned one-to-one with the corresponding access node migration sequences, so that the addresses... Corresponding access node , For the preceding wireless access node corresponding to the i-th abnormal traffic segment, The subsequent wireless access node is the one corresponding to the next abnormal traffic segment. If the address replacement order, access handover order, and abnormal segment time order are consistent, the path is determined as a complete roaming coupling drift process. For positions where time limits are exceeded or access nodes are discontinuous, the path is truncated and formed into independent drift paths. Finally, the address sequence, access node sequence, corresponding abnormal segment, and handover time are recorded to obtain the first address drift data.
[0117] Optionally, the multiple address attribution ambiguity resolution includes:
[0118] The abnormal traffic fragment data is identified by address coexistence intervals to obtain multi-address coexistence data;
[0119] In one embodiment, IPv6 address activity intervals are compared according to the start and end times of abnormal traffic segments. For addresses... and Their activity ranges are respectively and ,when When the two addresses coexist, the system determines that they exist. Multiple IPv6 addresses with continuous interval intersections are grouped into the same coexistence group, and the common time intervals and corresponding abnormal traffic segments are recorded to obtain multi-address coexistence data.
[0120] Based on the multiple access coexistence data, abnormal fragment attribution aggregation is performed to obtain multiple access attribution data;
[0121] In one embodiment, for each coexisting group, the destination IPv6 address, destination port, transport protocol, and occurrence time of each abnormal fragment are extracted; when two fragments are identical in at least two of the destination address, destination port, and transport protocol, they are assigned to the same abnormal behavior attribution set and arranged in ascending order of occurrence time to form a "coexisting address - abnormal fragment sequence" correspondence, thus obtaining multi-address attribution data.
[0122] Parallel address inheritance determination is performed on the multi-address data to obtain multi-address association data;
[0123] In one embodiment, the terminal abnormal segment before the address enters the coexistence interval is taken as the preceding segment. The behavioral continuity relationship between the segment carried by each coexisting address and the preceding segment is checked. If the segment corresponding to a certain address continues the destination address, destination port, and protocol combination of the preceding segment, and its occurrence time is after the preceding segment, then an inheritance relationship from the preceding address to the coexisting address is established. For cases where multiple addresses meet the inheritance conditions, the number of consecutively inherited abnormal segments is recorded to obtain multi-address attribution association data.
[0124] The second address drift data is obtained by resolving the attribution ambiguity based on the multi-address attribution association data.
[0125] In one embodiment, when the same preceding address corresponds to multiple inherited addresses, the address with the most consecutive inherited abnormal segments is preferentially retained; if the number of consecutive segments is the same, the interval between the end time of the preceding segment and the start time of the first inherited segment of each candidate address is compared, and the candidate address with the smallest time interval is retained. This determines the preceding and following address inheritance relationship, and the segments are connected in chronological order to form an address migration sequence under multiple access conditions, resulting in the second address drift data.
[0126] Optionally, the behavior address disambiguation includes:
[0127] Based on drift overlap data, address behavior competition map is constructed to obtain behavior attribution competition map data;
[0128] In one embodiment, candidate IPv6 addresses in the drift overlap unit are used as address nodes, and corresponding abnormal traffic segments are used as behavior nodes; when a candidate address With abnormal fragments ( For the r-th anomalous traffic segment, at least two of its destination IPv6 address, destination port, and transport protocol are consistent with the preceding anomalous behavior, and Within the valid active range corresponding to this address (i.e.) When the time range of the abnormal flow segment actually carried in the current drift overlap unit is considered, a competition edge is established. The bipartite competition graph G=(A,B,E) is formed by all candidate addresses, nodes with abnormal behavior, and competing edges, thus obtaining the behavior attribution competition graph data.
[0129] Inheritance residual fingerprints are extracted based on behavior attribution competition graph data to obtain behavior inheritance fingerprint data;
[0130] In one embodiment, for each contention edge, the average packet interval, average packet length, destination address, destination port, and protocol differences between the candidate address carrying behavior and the preceding behavior are extracted to construct... ,in Candidate address With abnormal fragments The behavior between them inherits residual fingerprints. Candidate address The average message interval corresponding to abnormal behavior Abnormal fragment The corresponding average message interval for preceding abnormal behaviors, Candidate address The average message length corresponding to abnormal behavior Abnormal fragment The average message length of the corresponding preceding abnormal behavior, This represents the destination IPv6 address consistency residual, set to 0 if the destination IPv6 addresses match, and 1 if they do not. This represents the destination port consistency residual, which is 0 if the destination ports are consistent and 1 if they are inconsistent. This represents the transport protocol consistency residual, which is 0 when the transport protocol is consistent and 1 when it is inconsistent. This is the IPv6 address sequence number. This refers to the sequence number of the abnormal traffic segment. Segments of the same candidate address are connected in the order they occurred. This forms the behavioral inheritance residual sequence for that address, thus obtaining the behavioral inheritance fingerprint data.
[0131] Based on behavioral inheritance fingerprint data, counterfactual attribution exclusion is performed to obtain address attribution exclusion data;
[0132] In one embodiment, candidate addresses are assumed sequentially. The corresponding abnormal behavior is not carried out, and it is removed from the competition graph. And its competing edges; if, after deletion, other candidate addresses still exist that can maintain the destination address, destination port, protocol, and sequential relationship of the preceding and following segments according to the original abnormal segment sequence, then... Mark as an outlierable address; if the corresponding abnormal behavior cannot form a continuous path with the preceding and following segments after deletion, then retain it. As necessary inheritance candidates. In the case of multiple necessary candidates, their residual fingerprints are compared, and candidates whose residual components are not greater than other candidates and whose at least one component is smaller are retained to obtain address attribution exclusion data.
[0133] Based on address-based exclusion data, unique inheritance chain convergence is performed to obtain drift-related data.
[0134] In one embodiment, address attribution exclusion data is processed in ascending order according to the time of occurrence of the abnormal segment. Starting from the previously determined addresses, at each action node, only candidate addresses that have not been excluded and can continue to establish a succession relationship with the next abnormal segment are retained. When multiple candidates still exist, priority is given to retaining the one with the smaller inherited residual fingerprint (i.e., compared first). , , If the value is 0, compare the message interval residual, then compare the message length residual. If they are still the same, compare the candidates (the time of occurrence of the first inherited segment). If the residuals are completely identical, retain the address with the earliest occurrence time of the first inherited abnormal segment. Repeat the above convergence process for each node until each abnormal behavior corresponds to only one inherited address, forming... The unique address inheritance relationship is used to obtain drift-related data.
[0135] Optionally, S3 includes:
[0136] S31. Based on address drift data, perform cross-address behavior extraction to obtain cross-address behavior data;
[0137] In one embodiment, according to the address drift sequence Retrieve the abnormal segments corresponding to each address one by one to extract the destination IPv6 address. Destination port Transmission Protocol Message direction and abnormal types Construct behavioral states And organize the adjacent behavior states according to the address switching sequence to obtain cross-address behavior data.
[0138] S32. Perform behavior status continuation parsing on the cross-address behavior data to obtain cross-address behavior continuation data;
[0139] In one embodiment, for adjacent addresses and The corresponding behavioral state, if satisfied , And the destination addresses are the same or belong to the same set of destination addresses, and simultaneously satisfy the following conditions: Then determine Continue Adjacent behaviors that do not meet all conditions are marked as breakpoints to be determined, and cross-address behavior continuation data is obtained.
[0140] S33. Based on cross-address behavior continuation data, perform behavior inertia extrapolation to obtain behavior inertia matching data;
[0141] In one embodiment, for the breakpoint to be determined, the average packet interval sequence of L consecutive abnormal sessions before the handover is extracted. ,calculate And obtain the extrapolated value based on the most recent change. The actual interval of the first session after the switch. satisfy Furthermore, if the destination access order and message length change direction remain continuous with those before the switch, it is marked as an inertial match, and behavioral inertial match data is obtained.
[0142] S34. Based on behavioral inertia matching data, breakpoint conflict is determined to obtain continuous data across addresses;
[0143] In one embodiment, it is checked whether adjacent actions involve a change in protocol type, a complete separation of the destination address set, or an address switching time interval exceeding [a certain threshold]. In the case of any of the above conflicts, the corresponding breakpoint is identified as a conflict breakpoint and the continuity is truncated; for records that do not have the above conflicts and have met the conditions for behavior continuation or inertial matching, their cross-address continuity is retained to obtain cross-address continuous data.
[0144] S35. Merge continuous behavior chains based on cross-address continuous data to obtain cross-address anomaly association data.
[0145] In one embodiment, consecutive records across addresses are arranged in ascending order of their occurrence time. Records that share consecutive anomaly segments and have no conflicting breakpoints in between are sequentially connected to form a sequence. The system records the continuous behavior chain, the address switching time, the corresponding abnormal segment, and the behavior status to obtain cross-address abnormal correlation data.
[0146] Optionally, S4 includes:
[0147] S41. Based on the cross-address anomaly correlation data, perform nodeization to obtain anomaly node data;
[0148] In one embodiment, each consecutive abnormal behavior segment is constructed as a node by dividing the chain according to the boundaries of the abnormal segments in the continuous behavior chain. ,in, The IPv6 address that carries this abnormal behavior, For the aforementioned behavioral state, and These are the start and end times of the abnormal segment, respectively; at the same time, the address relationship between its predecessor and successor and the corresponding abnormal segment number are retained to obtain the abnormal node data.
[0149] S42. Construct behavior-connecting edges based on abnormal node data to obtain abnormal behavior topology data;
[0150] In one embodiment, for any condition satisfying nodes and Calculate the time interval .when =30s, and when there is a behavioral continuation relationship between the two in the aforementioned cross-address continuous data, or when the destination address, destination port and protocol combination remain continuous, a directed receiving edge is established. Otherwise, no connection is established, and abnormal behavior topology data is obtained.
[0151] S43. Perform link bifurcation and convergence on the abnormal behavior topology data to obtain abnormal behavior link data;
[0152] In one embodiment, for the same node In cases where there are multiple successor nodes, the removal interval exceeds [a certain value]. Branches whose behavior or state conflicts with the preceding node; for the remaining branches, prioritize retaining the branch with the earliest start time that can continue to connect to subsequent abnormal nodes, i.e., satisfying... If multiple branches correspond to different and non-conflicting subsequent abnormal sequences, they are each retained as an independent link, resulting in abnormal behavior link data.
[0153] S44. Based on the abnormal behavior link data, perform cross-address chain sequence reconstruction to obtain abnormal behavior chain data.
[0154] In one embodiment, a node without a predecessor is selected as the head of the chain, and traversal is performed along the effective connecting edges in ascending order of node start times; combining the address sequence relationship in the address drift data, nodes with continuous behavior but carried by different IPv6 addresses are arranged sequentially as follows: When a successor node has no valid connecting edge, the current chain ends, and the chain order is re-established from the next node without a predecessor. Finally, the IPv6 address, abnormal behavior, connection relationship and occurrence time of each node are recorded to obtain abnormal behavior chain data.
[0155] Optionally, S5 includes:
[0156] S51. Threat phase anchoring is performed based on abnormal behavior chain data to obtain threat phase data;
[0157] In one embodiment, stages are mapped according to the destination access method, port changes, cross-address access relationships, and continuous communication characteristics of nodes in the abnormal behavior chain. Continuous multi-port or multi-destination address probing is anchored as the probing stage; concentrated abnormal connections targeting a fixed target are anchored as the access attempt stage; the migration or expansion of abnormal behavior from one campus network address to other addresses / target campus network nodes is anchored as the lateral propagation stage; and continuous cross-address communication with persistent abnormal behavior is anchored as the continuous activity stage. Each node is represented as... ,in, For IPv6 addresses, The threat phase after anchoring. Threat phase data is obtained at the time when the corresponding behavior occurs.
[0158] S52. Based on the threat stage data, perform cross-address propagation potential recursion to obtain threat propagation potential data;
[0159] In one embodiment, the first node of each chain of anomalous behavior is used as the propagation starting point and then... For adjacent nodes and When both belong to a pre-confirmed cross-address contiguous chain, satisfying And phase transition When, execute Otherwise, terminate the current propagation process and use the next node as the starting point of a new propagation chain. We obtained data on the spread of the threat.
[0160] S53. Based on the threat propagation potential data, perform link transition simulation to obtain threat chain simulation data;
[0161] In one embodiment, the propagation potential and threat stage of adjacent nodes are compared according to the sequence of the anomalous behavior chain. When ,and At that time, establish a threat jump edge If stage reversal occurs or time intervals exceed [a certain threshold], [the situation will be different]. If the preceding abnormal behavior chain has no node connection, the corresponding transition is truncated. Connect the edges along the continuous valid transitions to obtain the threat chain projection data.
[0162] S54. Based on the threat chain simulation data, perform situation boundary mapping to obtain security situation data.
[0163] In one embodiment, the start and end times of each valid threat chain, the set of IPv6 addresses involved, the highest threat stage, and the maximum propagation potential are extracted. For threat chains with overlapping time intervals and shared anomalous nodes, the time boundaries are merged, and the impact range is determined by the union of the earliest start and latest end times of the merged chain. The current highest threat stage and propagation depth are then output. By analyzing the affected address range and duration, security situation data can be obtained.
[0164] Therefore, the embodiments should be regarded as exemplary and non-limiting in all respects, and the scope of the invention is defined by the appended application documents rather than the foregoing description. Thus, it is intended that all variations falling within the meaning and scope of the equivalents of the application documents be incorporated into the invention.
[0165] The above description is merely a specific embodiment of the present invention, enabling those skilled in the art to understand or implement the invention. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the invention. Therefore, the present invention is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features of the invention herein.
Claims
1. A campus network security situation assessment method based on abnormal traffic characteristics, characterized in that, The method includes: S1. Obtain abnormal traffic segment data; S2. Perform temporary address drift association based on abnormal traffic fragment data to obtain address drift data; S3. Based on address drift data, determine the continuity of cross-address behavior to obtain cross-address anomaly correlation data; S4. Reconstruct the abnormal behavior chain based on cross-address abnormal correlation data to obtain abnormal behavior chain data; S5. Based on the abnormal behavior chain data, perform cross-address threat chain inference to obtain security situation data.
2. The method according to claim 1, characterized in that, S1 includes: Obtain raw IPv6 traffic data from the campus network; Traffic session aggregation is performed on raw IPv6 traffic data to obtain IPv6 traffic session data. Local anomaly deviations are extracted from IPv6 traffic session data to obtain abnormal traffic data. The abnormal traffic data is segmented into time-series segments to obtain abnormal traffic segment data.
3. The method according to claim 1, characterized in that, S2 include: Based on abnormal traffic fragment data, roaming coupling drift and multi-address ambiguity resolution are performed to obtain first address drift data and second address drift data, respectively; The drift data at the first address and the drift data at the second address are overlapped to obtain the drift overlapped data; Behavioral address disambiguation is performed on the drift-overlapping data to obtain drift-related data; Drift timing reconstruction is performed based on drift correlation data to obtain address drift data.
4. The method according to claim 3, characterized in that, The roaming coupling drift includes: Access handover data is obtained by extracting abnormal traffic fragment data; Address replacement adjacency resolution is performed based on roaming handover data to obtain address replacement data; Migration path determination is performed on the address replacement data to obtain roaming address association data; Based on the roaming address association data, coupled drift reconstruction is performed to obtain the first address drift data.
5. The method according to claim 3, characterized in that, The multi-address home ambiguity resolution includes: The abnormal traffic fragment data is identified by address coexistence intervals to obtain multi-address coexistence data; Based on the multiple access coexistence data, abnormal fragment attribution aggregation is performed to obtain multiple access attribution data; Parallel address inheritance determination is performed on the multi-address data to obtain multi-address association data; The second address drift data is obtained by resolving the attribution ambiguity based on the multi-address attribution association data.
6. The method according to claim 3, characterized in that, The behavior address disambiguation includes: Based on drift overlap data, address behavior competition map is constructed to obtain behavior attribution competition map data; Inheritance residual fingerprints are extracted based on behavior attribution competition graph data to obtain behavior inheritance fingerprint data; Based on behavioral inheritance fingerprint data, counterfactual attribution exclusion is performed to obtain address attribution exclusion data; Based on address-based exclusion data, unique inheritance chain convergence is performed to obtain drift-related data.
7. The method according to claim 1, characterized in that, S3 includes: Cross-address behavior data is obtained by extracting cross-address behavior data based on address drift data. The behavior status continuation parsing of cross-address behavior data is performed to obtain cross-address behavior continuation data; Behavioral inertia extrapolation is performed based on cross-address behavior continuation data to obtain behavior inertia matching data; Breakpoint conflict detection is performed based on behavioral inertia matching data to obtain continuous data across addresses; By merging continuous behavior chains based on continuous data across addresses, cross-address anomaly correlation data is obtained.
8. The method according to claim 1, characterized in that, S4 includes: Based on cross-address anomaly correlation data, node-based anomaly node data is obtained; Based on the abnormal node data, behavior-connecting edges are constructed to obtain abnormal behavior topology data; Link bifurcation and convergence are performed on the abnormal behavior topology data to obtain abnormal behavior link data; Cross-address chain sequence reconstruction is performed based on abnormal behavior chain data to obtain abnormal behavior chain data.
9. The method according to claim 1, characterized in that, S5 include: Threat phase anchoring is performed based on abnormal behavior chain data to obtain threat phase data; Based on the threat phase data, the cross-address propagation potential is recursively extrapolated to obtain the threat propagation potential data; Threat chain simulation data is obtained by performing link transition simulation based on threat propagation potential data. Security situation data is obtained by mapping the situation boundary based on threat chain simulation data.
