Network attack detection methods, devices, equipment, media, and program products

By acquiring control plane protocol feature information from multiple consecutive time windows on the backbone router network side and performing cross-window multi-protocol joint alarm determination, the problem of poor APT attack identification accuracy in existing technologies is solved, and efficient identification and detection of APT attack behavior is achieved.

CN123122989APending Publication Date: 2026-11-17CHINA UNITED NETWORK COMM GRP CO LTD +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611505139.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-24
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing technologies tend to miss low-speed, persistent, and phased attack behaviors when identifying advanced persistent threat (APT) attacks, resulting in poor identification accuracy. This is especially true in deployment scenarios where backbone router network resources are limited, making it difficult to effectively characterize multi-protocol collaborative weak anomalies and identify persistent weak anomalies across time windows.

Method used

By acquiring feature information of network-side control plane protocols in multiple consecutive time windows, anomaly labels are determined, and cross-window multi-protocol joint attack alarm judgment is performed. Multiple types of weak anomaly behaviors are aggregated, mined, and correlated to improve identification accuracy.

Benefits of technology

It achieves efficient identification of APT attack behaviors, improves the accuracy of attack behavior identification and detection sensitivity, and can capture the slow-speed and phased attack process of attackers.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123122989A_ABST
    Figure CN123122989A_ABST
Patent Text Reader

Abstract

This application provides a network attack detection method, apparatus, device, medium, and program product, relating to the field of computer network technology, for improving the accuracy of APT attack identification. The specific technical solution is as follows: Obtain feature information of N control plane protocols corresponding to the network side within M time windows. The network-side control plane protocols include at least one of the following: Border Gateway Protocol, Secure Shell Protocol, and Domain Name System (DNS). The M time windows are sequentially consecutive, and N and M are positive integers. For each time window, based on the feature information corresponding to each control plane protocol, determine the anomaly label corresponding to each control plane protocol within the M time windows. Based on the anomaly label corresponding to each control plane protocol within the M time windows, perform cross-window multi-protocol joint attack alarm judgment and output network attack alarm information. This application is applied to network anomaly detection scenarios.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This application relates to the field of computer network technology, and in particular to a network attack detection method, apparatus, device, medium, and program product. Background Technology

[0002] With the scaling up of network architectures and the increasing complexity of business systems, advanced persistent threats (APTs), which are highly covert and persistent, have become a core network security threat.

[0003] In related technologies, to address the aforementioned issues, host system logs, general network traffic data, and security device alarm information can be collected during the detection period. Based on the collected information, abnormal features existing during the detection period can be extracted. Based on these abnormal features, alarm correlation and attack chain reconstruction can be performed on the alarm information generated during the current detection period, thereby identifying APT attack behaviors in the network and achieving basic network security detection and protection.

[0004] However, the above-mentioned solution only completes the extraction of abnormal features and alarm correlation processing of the collected host system logs, general network traffic data and security device alarm information within the current detection cycle. APT attacks are characterized by low speed, continuous and phased progression. When attackers deliberately reduce the strength of a single anomaly so that the anomaly of this attack does not reach the traditional alarm threshold, it is easy to miss the APT attack process composed of multiple weak anomalies, thus causing the problem of APT attack underreporting and resulting in poor accuracy in identifying APT attacks. Summary of the Invention

[0005] This application provides a network attack detection method, apparatus, device, medium, and program product to improve the accuracy of APT attack behavior identification.

[0006] In a first aspect, embodiments of this application provide a network attack detection method, which includes: acquiring feature information of N control plane protocols corresponding to the network side in M ​​time windows, wherein the network side control plane protocols include at least one of the following: border gateway protocol, secure shell protocol, and domain name system, and the M time windows are sequentially consecutive time windows, where N and M are positive integers; for each time window, determining the abnormal label corresponding to each control plane protocol in the M time windows based on the feature information corresponding to each control plane protocol; and performing cross-window multi-protocol joint attack alarm judgment based on the abnormal label corresponding to each control plane protocol in the M time windows, and outputting network attack alarm information.

[0007] The technical solution provided in this application brings at least the following beneficial effects: by performing multi-protocol anomaly distribution judgment on multiple consecutive time windows within the detection period, all anomaly labels are obtained, and the multi-protocol anomaly labels scattered in each time window are jointly judged, thereby realizing the aggregation mining and correlation analysis of multiple weak anomaly behaviors distributed across time windows, thus improving the accuracy of APT attack behavior identification.

[0008] One possible implementation is that the feature information corresponding to the aforementioned border gateway protocol includes at least one of the following: border gateway protocol update message frequency, border gateway protocol keep-alive message frequency, and border gateway protocol route withdrawal announcement relationship data; based on the feature information corresponding to each control plane protocol information, an anomaly label corresponding to each control plane protocol in M ​​time windows is determined, including: if any feature information among the border gateway protocol update message frequency, border gateway protocol keep-alive message frequency, and border gateway protocol route withdrawal announcement relationship features in the acquired M time windows is greater than or equal to a first threshold, it is determined that the border gateway protocol has an anomaly in the M time windows, and the anomaly label corresponding to the border gateway protocol is output.

[0009] Another possible implementation is that the feature information corresponding to the aforementioned secure shell protocol includes at least one of the following: secure shell key initialization message frequency and key negotiation completion ratio; based on the feature information corresponding to each control plane protocol, determine the anomaly label corresponding to each control plane protocol in M ​​time windows, including: if the secure shell key initialization frequency in the acquired M time windows is greater than or equal to a second threshold, or the secure shell key initialization frequency is greater than or equal to a third threshold but the key negotiation completion ratio is less than a fourth threshold, determine that the secure shell protocol has an anomaly in the M time windows, and output the anomaly label corresponding to the secure shell protocol.

[0010] Another possible implementation is that the characteristic information corresponding to the above domain name system includes at least one of the following: the proportion of domain name non-response and the proportion of long-tail domain name queries; based on the characteristic information corresponding to each control plane protocol, the abnormal label corresponding to each control plane protocol in M ​​time windows is determined, including: if the proportion of domain name non-response in the obtained M time windows is greater than or equal to the fifth threshold or the proportion of long-tail domain names is greater than or equal to the sixth threshold, it is determined that the domain name system has an anomaly in the M time windows, and the abnormal label corresponding to the domain name system is output.

[0011] Another possible implementation involves outputting network attack alarm information based on the anomaly labels corresponding to each control plane protocol in M ​​time windows. This includes: summarizing the anomaly labels corresponding to each time window to obtain the time distribution results of the anomaly labels; and performing cross-window multi-protocol joint attack alarm determination based on the time distribution results.

[0012] Another possible implementation involves performing cross-window multi-protocol joint attack alarm determination based on the time distribution results. This includes: outputting attack alarm information when any two of the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol exhibit anomalies more than or equal to the seventh threshold within the observation period corresponding to the M time windows, and the number of time windows covered is greater than or equal to the preset eighth threshold; or, outputting attack alarm information when the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol all exhibit at least one anomaly within the observation period corresponding to the M time windows.

[0013] Secondly, embodiments of this application provide a network attack detection device, comprising: an acquisition module, a determination module, and a processing module. The acquisition module is used to acquire feature information of N control plane protocols corresponding to M time windows on the network side. The network-side control plane protocols include at least one of the following: Border Gateway Protocol, Secure Shell Protocol, and Domain Name System. The M time windows are sequentially consecutive time windows, and N and M are positive integers. The determination module is used to determine an anomaly label corresponding to each control plane protocol in the M time windows for each time window, based on the feature information corresponding to each control plane protocol. The processing module is used to perform cross-window multi-protocol joint attack alarm determination based on the anomaly labels corresponding to each control plane protocol in the M time windows, and output network attack alarm information.

[0014] One possible implementation is that the feature information corresponding to the aforementioned Border Gateway Protocol includes at least one of the following: Border Gateway Protocol update message frequency, Border Gateway Protocol keep-alive message frequency, and Border Gateway Protocol route withdrawal declaration relationship data; the aforementioned processing module is specifically used to determine that the Border Gateway Protocol is abnormal in the M time windows if any feature information among the Border Gateway Protocol update message frequency, Border Gateway Protocol keep-alive message frequency, and Border Gateway Protocol route withdrawal declaration relationship features in the acquired M time windows is greater than or equal to a first threshold, and output the abnormal label corresponding to the Border Gateway Protocol.

[0015] Another possible implementation is that the characteristic information corresponding to the aforementioned secure shell protocol includes at least one of the following: secure shell key initialization message frequency and key negotiation completion ratio; the aforementioned processing module is specifically used to determine that the secure shell protocol has an anomaly in the M time windows when the secure shell key initialization frequency is greater than or equal to the second threshold, or the secure shell key initialization frequency is greater than or equal to the third threshold but the key negotiation completion ratio is less than the fourth threshold, and output the anomaly label corresponding to the secure shell protocol.

[0016] Another possible implementation is that the characteristic information corresponding to the above domain name system includes at least one of the following: the proportion of domain name non-response and the proportion of long-tail domain name queries; the above processing module is specifically used to determine that the domain name system is abnormal in the M time windows when the proportion of domain name non-response is greater than or equal to the fifth threshold or the proportion of long-tail domain name queries is greater than or equal to the sixth threshold, and output the abnormal label corresponding to the domain name system.

[0017] Another possible implementation is that the above processing module is specifically used to summarize the abnormal labels corresponding to each time window to obtain the time distribution results of the abnormal labels; and based on the time distribution results, to perform cross-window multi-protocol joint attack alarm judgment.

[0018] Another possible implementation is that the above processing module is specifically used to output attack alarm information when any two of the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol have an anomaly count greater than or equal to the seventh threshold within the observation period corresponding to the M time windows, and the number of time windows covered is greater than or equal to the preset eighth threshold; or, when the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol all have at least one anomaly within the observation period corresponding to the M time windows, it outputs attack alarm information.

[0019] Thirdly, this application provides an electronic device comprising: a processor and a memory; the memory stores a program or instructions executable on the processor, wherein the program or instructions, when executed by the processor, implement the method of the first aspect described above.

[0020] Fourthly, this application provides a readable storage medium on which a program or instructions are stored, which, when executed by a computer, implement the method of the first aspect described above.

[0021] Fifthly, this application provides a computer program product stored in a storage medium, which, when executed by a computer, implements the method described in the first aspect.

[0022] In a sixth aspect, embodiments of this application provide a chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the method as described in the first aspect.

[0023] The beneficial effects of the second to sixth aspects mentioned above are described in the corresponding description of the first aspect and will not be repeated here. Attached Figure Description

[0024] Figure 1 A schematic diagram of the network architecture for an application of a network attack detection method provided in this application embodiment;

[0025] Figure 2 A flowchart illustrating a network attack detection method provided in an embodiment of this application;

[0026] Figure 3 A flowchart illustrating another network attack detection method provided in this application embodiment;

[0027] Figure 4 A flowchart illustrating another network attack detection method provided in this application embodiment;

[0028] Figure 5 A flowchart illustrating another network attack detection method provided in this application embodiment;

[0029] Figure 6 A flowchart illustrating another network attack detection method provided in this application embodiment;

[0030] Figure 7 A flowchart illustrating another network attack detection method provided in this application embodiment;

[0031] Figure 8 This is a schematic diagram of the structure of a network attack detection device provided in an embodiment of this application;

[0032] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0033] The network attack detection method, apparatus, equipment, media, and program products provided in this application will now be described in detail with reference to the accompanying drawings.

[0034] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0035] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0036] The terms "at least one," "at least one of," etc., used in the specification and claims of this application refer to any one, any two, or a combination of two or more of the included items. For example, at least one of a, b, and c can mean: "a," "b," "c," "a and b," "a and c," "b and c," and "a, b, and c," where a, b, and c can be single or multiple. Similarly, "at least two" refers to two or more items, and its meaning is similar to that of "at least one."

[0037] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0038] The following provides a detailed explanation of the technical terms used in the embodiments of this application.

[0039] Border Gateway Protocol (BGP) is an external gateway protocol used for exchanging routing information between autonomous systems on the Internet. It is a core routing protocol of the Internet. This protocol is mainly used to implement route reachability announcements, route updates, and route cancellations between different autonomous systems. It can adapt to large-scale, multi-topology Internet network architectures and is a core protocol for ensuring smooth routing of cross-network data transmission.

[0040] Secure Shell (SSH) is a secure network transport protocol built on the application layer. In insecure network environments, it provides encryption, authentication, and data integrity verification capabilities for remote login, data transmission, and device management between network devices. This effectively mitigates network security risks such as data eavesdropping, tampering, and forgery, and is widely used in remote network device maintenance scenarios.

[0041] The Domain Name System (DNS) is the core distributed domain name resolution system of the Internet. This system translates human-readable domain names into IP addresses that network devices can recognize, enabling the mapping and resolution of network domain names to IP addresses. It is the fundamental support system for users accessing various Internet services, ensuring accurate addressing for network access.

[0042] Border Gateway Protocol (BGP) update message frequency: This refers to the frequency at which network devices periodically or triggeredly send BGP route update messages based on the Border Gateway Protocol. BGP update messages are primarily used to announce new routes, route changes, and other routing information. This frequency directly reflects the speed of network routing topology updates and is a core indicator characterizing the dynamic changes in network routing.

[0043] Border Gateway Protocol (BGP) keepalive message frequency: This refers to the fixed frequency at which network devices send BGP keepalive messages to peer neighbor devices based on the Border Gateway Protocol (BGP). BGP keepalive messages are used to detect the connectivity and liveness status of BGP neighbor sessions, maintaining the stability of neighbor protocol sessions. This frequency directly determines the network device's sensitivity to detecting neighbor link failures and session interruptions.

[0044] Border Gateway Protocol (BGP) route withdrawal announcement relationship data refers to the associated dataset of route withdrawal information, message sending sequence, neighbor response status, and associated topology corresponding to route withdrawal messages sent by devices during BGP operation. This data comprehensively records the characteristics of the entire process of withdrawing and announcing failed and abandoned routes in the network, and can be used as the core basis for analyzing network route failures and route oscillations.

[0045] Secure Shell Key Initialization Message Frequency: This refers to the frequency with which a network device initiates SSH key initialization interactions and sends key initialization messages based on the Secure Shell protocol. Key initialization is a prerequisite for establishing a secure SSH connection. This frequency characterizes the triggering frequency of SSH secure connection initialization by the network device, reflecting the activity level of the device in initiating remote secure connections.

[0046] Secure Shell Key Negotiation Completion Rate: This refers to the ratio of the number of successful SSH key negotiations to the total number of key negotiation attempts within a statistical period. This metric quantifies the success rate of SSH secure connection key negotiations and directly reflects the stability of secure handshakes and encrypted link establishment between devices. It is a key indicator for evaluating the security and stability of SSH communication.

[0047] Domain Name Non-existent Response Percentage: This refers to the ratio of the number of responses returned by the Domain Name System (DNS) during the DNS resolution process within a statistical period, indicating that the domain name does not exist, the resolution failed, or there is no corresponding IP mapping, to the total number of DNS resolution requests. This metric reflects the proportion of invalid domain name accesses and abnormal DNS resolution requests in the network, and is used to identify abnormal network access and malicious DNS resolution probing behaviors.

[0048] Long-tail domain query ratio: This refers to the ratio of the number of DNS resolution requests for long-tail domains (domains with low query frequency, niche domains, and sparse access) to the total number of DNS resolution requests within a statistical period. Long-tail domains differ from high-frequency mainstream domains in that they are characterized by small single-access volumes, diverse domain types, and a dispersed overall proportion. This indicator can be used to analyze the long-tail distribution characteristics and abnormal access trends of network domain access.

[0049] Long-tail domains are domains whose main body is very long, not common short words, but composed of multiple words or phrases, and have a large number of characters.

[0050] The embodiments of this application provide a network attack detection method, apparatus, device, medium, and program product that can be applied to network anomaly detection scenarios.

[0051] Currently, APT attacks are characterized by high stealth, long attack cycles, and multi-stage coordinated attack methods, making them a significant security threat to network infrastructure. Backbone routers, as core forwarding and control nodes in the network architecture, are particularly vulnerable to APT attacks due to their control plane being a prime target. With the continuous expansion of network scale, conducting APT attack detection on the router network side and promptly capturing early signs of various attacks on the control plane is of great practical significance for ensuring the security of network infrastructure.

[0052] Existing APT attack detection solutions can be categorized into several implementation methods. Some solutions are based on audit logs and security device alerts, relying on attack chain models to perform semantic association and attack scenario reconstruction from scattered alerts. Other solutions are host-centric, collecting DNS logs and network traffic logs, extracting general network features, and combining them with data mining algorithms to complete anomaly detection. Still others are based on single-type logs and single-type features, completing attack identification at the single-host or single-time-window level. Still others utilize machine learning model training and expert experience mapping to achieve APT behavior identification. All of these solutions can detect some APT attack behaviors based on audit logs, network traffic logs, DNS logs, security device alerts, or attack chain models, achieving basic APT attack detection and protection in conventional scenarios.

[0053] However, the existing APT attack detection schemes mentioned above have limited ability to mine and utilize the original abnormal features of router control plane protocols. They lack dedicated detection mechanisms adapted to the characteristics of backbone router control plane protocols, making it difficult to effectively characterize weak anomalies caused by multi-protocol collaboration. They also lack the ability to identify persistent weak anomalies across time windows. Some schemes also have a high dependence on labeled samples, expert rules, and computing resources. In deployment scenarios where router network resources are limited, the operating overhead and implementation complexity are high, resulting in poor overall adaptability performance of APT attack detection in router network scenarios.

[0054] To address the aforementioned technical issues, embodiments of this application provide a network attack detection method, apparatus, device, medium, and program product. By performing multi-protocol anomaly distribution judgment on multiple consecutive time windows within a detection period, all anomaly tags are obtained. Furthermore, the multi-protocol anomaly tags scattered within each time window are jointly judged, thereby achieving aggregated mining and correlation analysis of multiple types of weak anomaly behaviors distributed across time windows, thus improving the accuracy of APT attack behavior identification.

[0055] The network attack detection method, apparatus, device, medium, and program products provided in this application will be described in detail below with reference to the accompanying drawings.

[0056] Figure 1 The diagram illustrates a network architecture for a network attack detection method provided in an embodiment of this application. For example... Figure 1 As shown, the network architecture includes a network attack detection device 101 and a terminal device 102. The network attack detection device 101 and the terminal device 102 are interconnected.

[0057] In some embodiments, the network attack detection device 101 may be a server, a computer, or a processor or processing unit within a server or computer. The server may be a single server or a server cluster consisting of multiple servers. It should be noted that the embodiments of this application do not limit the specific device form of the network attack detection device 101. Figure 1 The example shown is a network attack detection device 101 with a single server.

[0058] In some embodiments, network-side devices may be base stations, core network devices, gateway devices, routers, switches, servers, edge computing nodes, cloud processing devices, relay devices, radio frequency units, network controllers, storage service devices, or network management devices, etc., and the embodiments of this application do not specifically limit them. Figure 1 The example shown is a network-side device 102 serving as a base station.

[0059] In some embodiments, the network attack detection device 101 sends a request to the terminal device 102 for feature information corresponding to N control plane protocols on the network side within M time windows. The terminal device 102 then sends the feature information corresponding to the N control plane protocols within the M time windows to the network attack detection device 101. For each time window, the network attack detection device 101 determines an anomaly label for each control plane protocol within the M time windows based on the feature information corresponding to each control plane protocol. Based on the anomaly label for each control plane protocol within the M time windows, the network attack detection device 101 performs cross-window multi-protocol joint attack alarm determination and outputs network attack alarm information.

[0060] It should be noted that the network architecture described in the embodiments of this application is for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and does not constitute a limitation on the technical solutions provided in the embodiments of this application. As network architectures evolve, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0061] See Figure 2 This is a flowchart illustrating a network attack detection method provided in an embodiment of this application. Figure 2 As shown, the network attack detection method provided in this application embodiment can be implemented by the network attack detection device described above, specifically including the following steps 201 to 203.

[0062] Step 201: The network attack detection device acquires the feature information of the N control plane protocols on the network side in M ​​time windows.

[0063] In some embodiments, the network-side control plane protocol described above includes at least one of the following: border gateway protocol, secure shell protocol, and domain name system.

[0064] In some embodiments, the above M time windows are time windows with consecutive time sequences.

[0065] In some embodiments, N and M are positive integers.

[0066] In some embodiments, the aforementioned control plane protocol feature information is multidimensional structured representation information related to the inherent attributes, traffic interaction timing characteristics, and protocol operation state changes of BGP, DNS, and SSH protocols within M time windows.

[0067] In some embodiments, the aforementioned feature information includes: Border Gateway Protocol update message frequency, Border Gateway Protocol keep-alive message frequency, Border Gateway Protocol route revocation declaration relationship data, Security Shell key initialization message frequency, key negotiation completion rate, domain name non-existent response rate, and long-tail domain name query rate, etc.

[0068] In some embodiments, the network attack detection device can acquire BGP, DNS and SSH protocol traffic on the router network side, perform time-series slicing of the protocol traffic according to a preset time window, generate protocol traffic logs for the corresponding time window, and extract and generate control plane protocol feature information corresponding to each time window based on the traffic logs corresponding to the above M time windows.

[0069] In some embodiments, the network attack detection device can perform bypass monitoring and protocol parsing of the entire network traffic on the router network side, filter and capture exclusive traffic data belonging to BGP, DNS, and SSH protocols, segment and slice the continuously captured protocol traffic data according to a preset fixed time window, aggregate and summarize the BGP, DNS, and SSH protocol traffic captured in the same time window, and generate protocol traffic logs corresponding to each time window. Each traffic log fully records the original traffic packets, interaction records, and transmission timing information of the three types of protocols within the corresponding time period.

[0070] In some embodiments, the network attack detection device can traverse the BGP, DNS, and SSH protocol traffic logs corresponding to the above M time windows, extract the inherent attribute information, temporal interaction distribution information, and state evolution information of the protocols within the corresponding time windows from the logs, and perform structured integration of the extracted multidimensional information to form feature information that can characterize the operating rules of the BGP, DNS, and SSH protocol control planes within each time window.

[0071] In some embodiments, the feature information corresponding to different time windows can objectively reflect the inherent characteristics and dynamic operational differences of the network-side control plane protocol in different time periods.

[0072] Step 202: For each time window, the network attack detection device determines the abnormal label corresponding to each control plane protocol in M ​​time windows based on the feature information corresponding to each control plane protocol information.

[0073] In some embodiments, after acquiring the feature information of the corresponding control plane protocols under each time window, the network attack detection device can standardize and organize the feature information of each control plane protocol, fuse feature dimensions and perform compliance comparison analysis, and combine the baseline feature patterns of normal protocol operation to make an overall judgment on the overall operation characteristics of BGP, DNS and SSH protocols in the current time window, and accurately determine the protocol anomaly label corresponding to the current time window based on the judgment results.

[0074] In some embodiments, the above-mentioned abnormal label can be characterized as status identification information used to identify whether there are abnormal operating behaviors, abnormal interaction states, and abnormal traffic characteristics of the corresponding control plane protocol on the router network side within a preset time window.

[0075] In some embodiments, different anomaly labels can distinguish the anomaly occurrence type and anomaly attribution dimension of different control plane protocols, and can reflect the operational deviation of each protocol within the corresponding time window.

[0076] In some embodiments, the above-mentioned anomaly labels include: anomaly labels corresponding to the Border Gateway Protocol, anomaly labels corresponding to the Secure Shell Protocol, and anomaly labels corresponding to the Domain Name System, etc.

[0077] In some embodiments, the feature information corresponding to the border gateway protocol mentioned above includes at least one of the following: border gateway protocol update message frequency, border gateway protocol keep-alive message frequency, and border gateway protocol route withdrawal announcement relationship data.

[0078] For example, combined Figure 2 ,like Figure 3 As shown, step 202 above can be implemented through step 301 as follows.

[0079] Step 301: If any of the following features in the obtained M time windows—Border Gateway Protocol update message frequency, Border Gateway Protocol keep-alive message frequency, and Border Gateway Protocol route withdrawal announcement relationship feature—is greater than or equal to the first threshold, the network attack detection device determines that the Border Gateway Protocol is abnormal in the M time windows and outputs the corresponding abnormal label for the Border Gateway Protocol.

[0080] In some embodiments, the first threshold can be preset by the network attack detection device; or it can be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0081] For example, the first threshold mentioned above can be 5, 6, or 7, etc.

[0082] In some embodiments, the network attack detection device can configure independent thresholds for the Border Gateway Protocol (BGP) update message frequency, BGP keep-alive message frequency, and BGP route withdrawal announcement relationship data, and compare the BGP update message frequency, BGP keep-alive message frequency, and BGP route withdrawal announcement relationship data with their respective thresholds. When the BGP update message frequency is greater than or equal to its corresponding threshold, or the BGP keep-alive message frequency is greater than or equal to its corresponding threshold, or the BGP route withdrawal announcement relationship data is greater than or equal to its corresponding threshold, it is determined that there is an anomaly in the BGP within that time window, and the corresponding anomaly label for the BGP is output.

[0083] In some embodiments, the above-mentioned anomaly labels are used to characterize the types of anomalies that exist in the border gateway protocol within the time window.

[0084] In some embodiments, the above-mentioned anomaly label includes at least one of the following: text identifier, identifier code, encoded value, or structured data.

[0085] In this way, network attack detection devices can identify abnormal behaviors at the routing control layer, such as route oscillation, abnormal route changes, session instability, or abnormal revocation announcements, by comparing and judging various characteristic information of the border gateway protocol. This improves the accuracy and efficiency of identifying network attack behaviors related to the border gateway protocol.

[0086] In some embodiments, the characteristic information corresponding to the above-mentioned secure shell protocol includes at least one of the following: secure shell key initialization message frequency and key negotiation completion rate.

[0087] For example, combined Figure 2 ,like Figure 4 As shown, step 202 above can be implemented through step 401 as follows.

[0088] Step 401: If the Secure Shell key initialization frequency in the M time windows is greater than or equal to the second threshold, or if the Secure Shell key initialization frequency is greater than or equal to the third threshold but the key negotiation completion rate is less than the fourth threshold, the network attack detection device determines that the Secure Shell protocol is abnormal in the M time windows and outputs the corresponding abnormal label of the Secure Shell protocol.

[0089] In some embodiments, the second threshold can be preset by the network attack detection device; or it can be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0090] For example, the second threshold mentioned above can be 15, 22, or 30, etc.

[0091] In some embodiments, the aforementioned third threshold can be preset by the network attack detection device; or, it can be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0092] For example, the third threshold mentioned above can be 2, 5, or 8, etc.

[0093] In some embodiments, the aforementioned fourth threshold can be preset by the network attack detection device; or, it can be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0094] For example, the fourth threshold mentioned above can be 0.72, 0.81, or 0.88, etc.

[0095] In some embodiments, the network attack detection device can configure a corresponding threshold for the Secure Shell key initialization frequency and a corresponding threshold for the key negotiation completion ratio; and compare the Secure Shell key initialization frequency with the corresponding threshold and the key negotiation completion ratio with the corresponding threshold. When the Secure Shell key initialization frequency is greater than or equal to the corresponding threshold, or the Secure Shell key initialization frequency is greater than or equal to the corresponding threshold and the key negotiation completion ratio is less than the corresponding threshold, it is determined that there is an anomaly in the Secure Shell protocol within the time window, and the anomaly label corresponding to the Secure Shell protocol is output.

[0096] In this way, network attack detection devices can identify brute-force attacks on the management port, abnormal remote login attempts, or a large number of failed connections by comparing and judging the various characteristic information of the Secure Shell protocol, thereby improving the accuracy and efficiency of identifying network attack behaviors related to the Secure Shell protocol.

[0097] In some embodiments, the characteristic information corresponding to the above domain name system includes at least one of the following: the percentage of domain name non-existent responses and the percentage of long-tail domain name queries.

[0098] For example, combined Figure 2 ,like Figure 5 As shown, step 202 above can be implemented through step 501 as follows.

[0099] Step 501: If the percentage of domains without responses in the M time windows is greater than or equal to the fifth threshold or the percentage of long-tail domains is greater than or equal to the sixth threshold, the network attack detection device determines that the domain name system is abnormal in the M time windows and outputs the corresponding abnormal label of the domain name system.

[0100] In some embodiments, the aforementioned fifth threshold may be preset by the network attack detection device; or, it may be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0101] For example, the fifth threshold mentioned above can be 0.31, 0.35, or 0.39, etc.

[0102] In some embodiments, the aforementioned sixth threshold can be preset by the network attack detection device; or, it can be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0103] For example, the sixth threshold mentioned above can be 0.44, 0.51, or 0.57, etc.

[0104] In some embodiments, the network attack detection device can configure independent thresholds for the proportion of domain name non-response and the proportion of long-tail domain names, and compare the proportion of domain name non-response and the proportion of long-tail domain names with their respective thresholds. When the proportion of domain name non-response is greater than or equal to its corresponding threshold, or the proportion of long-tail domain names is greater than or equal to its corresponding threshold, it is determined that there is an anomaly in the domain name system within M time windows, and the anomaly label corresponding to the domain name system is output.

[0105] In this way, network attack detection devices can compare and distinguish various feature information of the Domain Name System (DNS) protocol to identify behaviors such as Domain-Generating Algorithm (DGA) domain probing, Command-and-Control (C2) domain access, abnormal resolution failures, or Domain Name System (DNS) covert communication, thereby improving the accuracy and efficiency of identifying DNS protocol-related network attack behaviors.

[0106] Step 203: The network attack detection device performs cross-window multi-protocol joint attack alarm judgment based on the abnormal labels corresponding to each control plane protocol in M ​​time windows, and outputs network attack alarm information.

[0107] In some embodiments, the above M time windows constitute a complete observation period, and the M time windows are consecutive time periods.

[0108] In some embodiments, the network attack detection device can record the detection results of the three types of protocols in each time window in chronological order to form the protocol anomaly time distribution results within an observation period, and perform joint attack alarm judgment based on the protocol anomaly time distribution results, and output network attack alarm information.

[0109] In some embodiments, the time distribution results of the above-mentioned anomaly tags can be characterized as a structured set carrying time dimension information. This structured set records each time window, the protocol anomaly tags corresponding to each time window, and the occurrence sequence relationship of the anomaly tags, which can reflect the generation and distribution of various control plane protocol anomaly tags in different time intervals.

[0110] In some embodiments, the above time distribution results are the time distribution results of abnormal labels appearing within M consecutive time windows in one observation period.

[0111] The network attack detection method provided in this application obtains feature information of N control plane protocols corresponding to the network side in M ​​time windows. The network side control plane protocols include at least one of the following: Border Gateway Protocol, Secure Shell Protocol, and Domain Name System. The M time windows are sequentially consecutive time windows, and N and M are positive integers. For each time window, based on the feature information corresponding to each control plane protocol, anomaly labels corresponding to each control plane protocol in the M time windows are determined. Based on the anomaly labels corresponding to each control plane protocol in the M time windows, cross-window multi-protocol joint attack alarm judgment is performed, and network attack alarm information is output. This application obtains all anomaly labels by performing multi-protocol anomaly distribution judgment on multiple consecutive time windows within the detection period, and jointly judges the multi-protocol anomaly labels scattered in each time window, realizing the aggregation mining and correlation analysis of multiple weak anomaly behaviors distributed across time windows, thereby improving the accuracy of APT attack behavior identification.

[0112] In some embodiments, combined with Figure 2 ,like Figure 6 As shown, step 203 above can be implemented through steps 203a and 203b.

[0113] Step 203a: The network attack detection device summarizes the abnormal labels corresponding to each time window to obtain the time distribution results of the abnormal labels.

[0114] In some embodiments, the network attack detection device can traverse the abnormal tags determined by each time window, associate and bind each time window with the corresponding generated abnormal tags, aggregate and organize the abnormal tags of all time windows, and obtain the time distribution result of the abnormal tags.

[0115] In some embodiments, the network attack detection device can establish a mapping association between the time identifier information of each time window and the abnormal label generated under that time window, thereby completing the association and binding between each time window and the corresponding generated abnormal label.

[0116] In some embodiments, the network attack detection device can collect, sort, and structure all time windows that have been associated and bound and their corresponding abnormal tags according to the chronological order, thereby realizing the aggregation and organization of abnormal tags for all time windows.

[0117] Step 203b: The network attack detection device performs a cross-window multi-protocol joint attack alarm judgment based on the time distribution results.

[0118] In some embodiments, the network attack detection device can match the corresponding thresholds to the abnormal labels obtained from the sum of each window based on the above time distribution results, and compare the thresholds of the abnormal labels corresponding to the Border Gateway Protocol, the Domain Name System, and the Secure Shell Protocol to perform attack alarm determination.

[0119] In some embodiments, the network attack detection device can output attack alarm information when it determines that there is a multi-protocol coordination anomaly within the observation period based on preset multi-protocol anomaly determination conditions.

[0120] In some embodiments, the aforementioned attack alarm information is alarm characterization data output by the network attack detection device based on the multi-protocol anomaly determination conditions within the observation period, used to characterize attack events where there are multi-protocol coordination anomalies on the router network side within the observation period.

[0121] In some embodiments, the aforementioned attack alarm information is alarm characterization information generated based on the time distribution results of abnormal labels, which is used to feed back network attack-related detection results to the service processing side.

[0122] In some embodiments, the attack alarm information includes time range information corresponding to the observation period, protocol type identifier of the anomaly, anomaly label corresponding to each protocol, time window record corresponding to the anomaly, and alarm event type identifier, etc.

[0123] In some embodiments, if the network attack detection device determines that there is no multi-protocol coordination anomaly within the observation period based on preset multi-protocol anomaly determination conditions, it may not output attack alarm information.

[0124] In this way, the network attack detection device makes judgments based on the abnormal distribution of protocols within multiple consecutive time windows. It can identify APT attack processes that are slowly advanced, implemented in stages, and transferred between different protocols, thereby improving the ability to capture and detect advanced persistent threat attacks.

[0125] In some embodiments, combined with Figure 6 ,like Figure 7 As shown, step 203a above can be implemented through step 601 or step 602.

[0126] Step 601: If any two of the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol exhibit anomalies greater than or equal to the seventh threshold within the observation period corresponding to M time windows, and the number of time windows covered is greater than or equal to the preset eighth threshold, the network attack detection device outputs an attack alarm message.

[0127] In some embodiments, the aforementioned seventh threshold can be preset by the network attack detection device; or, it can be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0128] For example, the seventh threshold mentioned above can be 2, 3, or 4, etc.

[0129] In some embodiments, the aforementioned eighth threshold can be preset by the network attack detection device; or, it can be user-defined. The specific threshold can be determined according to actual usage requirements, and this application embodiment does not impose any limitations.

[0130] For example, the eighth threshold mentioned above can be 1, 2, or 3, etc.

[0131] In some embodiments, the network attack detection device can configure an independent threshold for the number of protocol anomalies and an independent threshold for the number of covered time windows. It can also compare the number of anomalies and the number of covered time windows of any two protocols among the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol within an observation period consisting of M time windows with their respective thresholds. When the number of anomalies of any two protocols among the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol within the observation period is greater than or equal to their corresponding thresholds, and the number of covered time windows is greater than or equal to their corresponding thresholds, an attack alarm message is output.

[0132] Step 602: If the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol all exhibit at least one anomaly within the observation period corresponding to the M time windows, the network attack detection device outputs an attack alarm message.

[0133] In some embodiments, the network attack detection device can acquire the anomaly labels of the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol within an observation period consisting of M time windows, and determine whether there are any anomaly labels for the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol within the observation period. If the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol all show at least one anomaly within the observation period corresponding to the M time windows, an attack alarm message is output.

[0134] In this way, the network attack detection device can transform various abnormal behaviors in the router's network-side control plane protocol into protocol anomaly labels, and further describe the combination relationship of abnormal behaviors in the protocol and time dimensions through the protocol anomaly time distribution results within a continuous time window. This enables lightweight detection of low-frequency, continuous, cross-protocol linked APT attack behaviors, thereby improving the capture capability and detection sensitivity of advanced persistent threat attacks.

[0135] The network attack detection method of this application will be described below through specific embodiments.

[0136] The implementation process of the network attack detection method provided in this application includes the following steps S1 to S5.

[0137] S1. The network attack detection device collects the control plane traffic on the router's network side.

[0138] For example, a traffic acquisition probe, bypass detection device, or log acquisition module is deployed on the router network side to continuously collect traffic from three control plane protocols: BGP, DNS, and SSH. The collected content can include packet timestamps, protocol types, source addresses, destination addresses, packet types, response codes, route update fields, route revocation fields, SSH key negotiation fields, DNS query domain names and resolution results, etc. The acquisition method can be mirror port acquisition, bypass acquisition, router log interface acquisition, or other methods capable of acquiring control plane protocol behavior. The acquisition module divides continuous traffic into multiple time windows, denoted as t1, t2, ..., tN, according to a preset period. Each time window serves as the basic processing unit for a protocol anomaly detection. This approach allows for the analysis of control information from different protocols at a unified time granularity, facilitating subsequent determination of whether abnormal behavior persists across multiple time windows or whether there is a linkage between multiple protocols.

[0139] S2. The network attack detection device generates BGP anomaly tags based on BGP control information.

[0140] For example, within any given time window, the number of BGP Update messages, BGP Keepalive messages, Update messages carrying route cancellation information, and Update messages carrying route advertisement information are counted. Based on the statistical results, the frequency of BGP Update messages and BGP Keepalive messages per unit time, as well as the ratio of route cancellations to route advertisements, are determined. Frequent occurrences of BGP Update messages may indicate abnormal route updates, route oscillations, or abnormal route changes; abnormal frequency of BGP Keepalive messages may indicate unstable BGP session states; and an abnormally high number of route cancellations relative to route advertisements may indicate large-scale route cancellations, route leaks, or abnormal path changes. These BGP anomaly characteristics characterize abnormal situations in the BGP control plane from three perspectives: route updates, session keepalives, and route cancellations. The BGP anomaly characteristics are compared with preset BGP anomaly thresholds. These preset BGP anomaly thresholds can be determined based on RFC specifications, engineering experience, or local network operating baselines. If any of the following characteristics—BGPUpdate frequency, BGP Keepalive frequency, or BGP revocation declaration relationship—meets the abnormal condition, the BGP protocol is determined to be abnormal within the current time window and recorded as a BGP abnormality; otherwise, the BGP protocol is determined to be normal within the current time window. The BGP abnormality detection result is used for subsequent cross-window multi-protocol joint determination.

[0141] S3, the network attack detection device generates SSH anomaly tags based on SSH control information.

[0142] Within any given time window, the number of SSH key exchange initialization messages and the number of SSH key negotiation completion messages are counted. Based on the statistical results, the frequency of SSH KeyInit messages and SSHNewKey messages per unit time is determined, along with the ratio of key negotiation completions to key exchange initializations. Specifically, frequent SSH KeyInit messages indicate a large number of SSH connections or key exchange initialization requests; conversely, a low number of SSH NewKey messages indicates that many connection attempts failed to complete key negotiation. This phenomenon typically corresponds to brute-force attacks on the management port, abnormal remote login attempts, automated scanning, or unauthorized access detection. By observing both KeyInit and NewKey messages simultaneously, anomaly detection can be avoided by relying solely on connection counts, thus providing a more accurate characterization of abnormal access behavior in the SSH management channel. When the frequency of SSH KeyInit messages exceeds a preset threshold, or when the frequency of SSH KeyInit messages is high while the key negotiation completion rate is below a preset threshold, the SSH protocol is considered abnormal within the current time window and recorded as an SSH anomaly; otherwise, the SSH protocol is considered normal within the current time window. The SSH anomaly detection results are used for subsequent cross-window multi-protocol joint determination. Furthermore, DNS anomaly labels are generated based on DNS control information. Within any given time window, the number of DNS resolution failure messages, the total number of DNS messages, the number of long-tail domain name queries, and the total number of DNS queries are counted. Based on the statistical results, the proportion of NXDOMAIN responses in DNS messages and the proportion of long-tail domain name queries in DNS queries are determined. Specifically, a high proportion of NXDOMAIN responses may indicate a large number of invalid domain name probes, DGA domain name accesses, or C2 domain name probes; a high proportion of long-tail domain name queries may indicate abnormal long domain names, multi-level subdomains, encoded domain names, or DNS covert communication behavior. Long-tail domain names can be determined based on the domain name string length, the number of subdomain levels, or the local network operating baseline. When the proportion of NXDOMAIN responses or the proportion of long-tail domain name queries exceeds a preset threshold, the DNS protocol is determined to be abnormal within the current time window and recorded as a DNS anomaly; otherwise, the DNS protocol is determined to be normal within the current time window. The DNS anomaly detection results are used for subsequent cross-window multi-protocol joint determination.

[0143] S4. The network attack detection device summarizes protocol anomaly results within a continuous time window.

[0144] For example, after detecting anomalies in three protocol types—BGP, DNS, and SSH—the detection results for each protocol type within each time window are recorded chronologically, forming a temporal distribution of protocol anomalies within an observation period. This temporal distribution includes at least the BGP, DNS, and SSH anomaly results for each time window. Using this temporal distribution, it can be determined whether a particular protocol anomaly persists across multiple time windows, or whether different protocol anomalies occur sequentially or concurrently within the same observation period. Compared to a single alarm, anomaly results within consecutive time windows better reflect the persistence, stealth, and phased progression characteristics of APT attacks.

[0145] S5. The network attack detection device performs joint APT determination based on protocol anomaly results within a continuous time window.

[0146] For example, the number of time windows in which BGP, DNS, and SSH protocols exhibit anomalies within a single observation period is counted, and the time window range covered by anomalies for any two protocols is also counted. An APT attack alert is output when either of the following conditions is met: Condition 1: Anomalies occur multiple times in both protocols within the observation period, and the number of time windows covered by anomalies for both protocols reaches a preset span threshold; Condition 2: Anomalies occur at least once in each of the three protocols (BGP, DNS, and SSH) within the same observation period. If neither of the above conditions is met, no APT attack alert is output, and the detection continues into the next observation period.

[0147] In this way, by performing multi-protocol anomaly distribution judgment on multiple consecutive time windows within the detection period, all anomaly labels are obtained, and the multi-protocol anomaly labels scattered in each time window are jointly judged, realizing the aggregation mining and correlation analysis of multiple weak anomaly behaviors distributed across time windows, thereby improving the accuracy of APT attack behavior identification.

[0148] It should be noted that the descriptions of each step S1 to S5 in this embodiment can be found in the descriptions in the above embodiments, and will not be repeated here.

[0149] It should be noted that the above-described method embodiments, or the various possible implementations of the method embodiments, can be executed individually, or, provided there is no conflict, they can be combined with each other. The specific implementation can be determined according to actual usage requirements, and this application embodiment does not impose any restrictions on this.

[0150] As can be seen, the above mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, the embodiments of this application provide corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the modules and algorithm steps of the various examples described in the embodiments disclosed herein, the embodiments of this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0151] This application embodiment can divide the network attack detection device into functional modules according to the above method example. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. Optionally, the module division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0152] In some embodiments, this application also provides a network attack detection device. This network attack detection device may include one or more functional modules for implementing the network attack detection method described in the above embodiments.

[0153] For example, Figure 8 This is a schematic diagram of a network attack detection device provided in an embodiment of this application. Figure 8 As shown, the network attack detection device 900 includes: an acquisition module 901, a determination module 902, and a processing module 903.

[0154] The acquisition module 901 is used to acquire feature information of N control plane protocols on the network side across M time windows. The network-side control plane protocols include at least one of the following: Border Gateway Protocol, Secure Shell Protocol, and Domain Name System. The M time windows are sequentially consecutive, and N and M are positive integers. The determination module 902 is used to determine the anomaly label corresponding to each control plane protocol across the M time windows for each time window, based on the feature information corresponding to each control plane protocol. The processing module 903 is used to perform cross-window multi-protocol joint attack alarm determination based on the anomaly label corresponding to each control plane protocol across the M time windows, and output network attack alarm information.

[0155] The network attack detection device provided in this application obtains all the abnormal labels by performing multi-protocol abnormal distribution judgment on multiple consecutive time windows within the detection period, and jointly judges the multi-protocol abnormal labels scattered in each time window, thereby realizing the aggregation mining and correlation analysis of multiple weak abnormal behaviors distributed across time windows, thus improving the accuracy of APT attack behavior identification.

[0156] In some embodiments, the feature information corresponding to the border gateway protocol includes at least one of the following: border gateway protocol update message frequency, border gateway protocol keep-alive message frequency, and border gateway protocol route withdrawal announcement relationship data; the processing module 903 is specifically used to determine that the border gateway protocol has an anomaly in the M time windows when any feature information among the border gateway protocol update message frequency, border gateway protocol keep-alive message frequency, and border gateway protocol route withdrawal announcement relationship features in the acquired M time windows is greater than or equal to a first threshold, and output the anomaly label corresponding to the border gateway protocol.

[0157] In other embodiments, the feature information corresponding to the aforementioned Secure Shell Protocol includes at least one of the following: Secure Shell key initialization message frequency and key negotiation completion ratio; the aforementioned processing module 903 is specifically used to determine that the Secure Shell Protocol has an anomaly in the M time windows when the Secure Shell key initialization frequency in the acquired M time windows is greater than or equal to a second threshold, or the Secure Shell key initialization frequency is greater than or equal to a third threshold but the key negotiation completion ratio is less than a fourth threshold, and output the anomaly label corresponding to the Secure Shell Protocol.

[0158] In some other embodiments, the characteristic information corresponding to the domain name system includes at least one of the following: the proportion of domain name non-response and the proportion of long-tail domain name queries; the processing module 903 is specifically used to determine that the domain name system is abnormal in the M time windows when the proportion of domain name non-response is greater than or equal to the fifth threshold or the proportion of long-tail domain names is greater than or equal to the sixth threshold, and output the abnormal label corresponding to the domain name system.

[0159] In some other embodiments, the processing module 903 is specifically used to summarize the abnormal labels corresponding to each time window to obtain the time distribution results of the abnormal labels; and based on the time distribution results, to perform cross-window multi-protocol joint attack alarm judgment.

[0160] In some other embodiments, the processing module 903 is specifically used to output attack alarm information when any two of the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol exhibit anomalies more than or equal to a seventh threshold within the observation period corresponding to M time windows, and the number of time windows covered is greater than or equal to a preset eighth threshold; or, when the Border Gateway Protocol, Domain Name System, and Secure Shell Protocol all exhibit anomalies at least once within the observation period corresponding to M time windows, the attack alarm information is output.

[0161] It should be noted that the network attack detection device can implement all the processes implemented in the above method embodiments and achieve the same beneficial effects. To avoid repetition, it will not be described again here.

[0162] In the case where the functions of the integrated modules described above are implemented in hardware, this application provides a possible structural schematic diagram of the electronic device involved in the above embodiments. For example... Figure 9 As shown, the electronic device 90 includes: a processor 92, a communication interface 93, and a bus 94. Optionally, the electronic device 90 may also include a memory 91.

[0163] Processor 92 may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 92 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 92 may also be a combination that implements computational functions, such as including one or more microprocessor combinations, a combination of a DSP and a microprocessor, etc.

[0164] Communication interface 93 is used to connect with other devices via a communication network. This communication network can be Ethernet, wireless access network, wireless local area network (WLAN), etc.

[0165] The memory 91 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0166] In one possible implementation, the memory 91 can exist independently of the processor 92. The memory 91 can be connected to the processor 92 via a bus 94 and is used to store instructions or program code. When the processor 92 calls and executes the instructions or program code stored in the memory 91, it can implement the network attack detection method provided in the embodiments of this application.

[0167] In another possible implementation, memory 91 can also be integrated with processor 92.

[0168] Bus 94 can be an Extended Industry Standard Architecture (EISA) bus, etc. Bus 94 can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0169] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the service calling device can be divided into different functional modules to complete all or part of the functions described above.

[0170] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the various processes of the above-described network attack detection method embodiments and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0171] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0172] This application also provides a readable storage medium storing a program or instructions that, when executed by a computer, implement the network attack detection method provided in the above embodiments. It is understood that all or part of the processes in the above method embodiments can be executed by computer instructions instructing related hardware; the readable storage medium can be any of the foregoing embodiments or memory; the readable storage medium can also be an external storage device of the service invocation device, such as a pluggable hard drive, SmartMedia Card (SMC), Secure Digital (SD) card, flash card, etc., equipped on the service invocation device. Further, the readable storage medium can include both internal storage units of the service invocation device and external storage devices. The readable storage medium is used to store the computer program and other programs and data required by the service invocation device. The readable storage medium can also be used to temporarily store data that has been output or will be output.

[0173] This application also provides a computer program product, which is stored in a storage medium and, when executed by a computer, implements the network attack detection method provided in the above embodiments.

[0174] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0175] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0176] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for detecting network attacks, characterized in that, include: Obtain the feature information of N control plane protocols on the network side in M ​​time windows. The network side control plane protocols include at least one of the following: border gateway protocol, secure shell protocol, and domain name system. The M time windows are time-series continuous time windows, and N and M are positive integers. For each time window, based on the feature information corresponding to each control plane protocol information, an anomaly label corresponding to each control plane protocol in the M time windows is determined; Based on the anomaly labels corresponding to each control plane protocol in the M time windows, cross-window multi-protocol joint attack alarm determination is performed, and network attack alarm information is output.

2. The network attack detection method according to claim 1, characterized in that, The characteristic information corresponding to the Border Gateway Protocol includes at least one of the following: Border Gateway Protocol update message frequency, Border Gateway Protocol keep-alive message frequency, and Border Gateway Protocol route withdrawal announcement relationship data; The step of determining the anomaly label for each control plane protocol within the M time windows based on the feature information corresponding to each control plane protocol information includes: If any of the following features in the obtained M time windows—Border Gateway Protocol update message frequency, Border Gateway Protocol keep-alive message frequency, and Border Gateway Protocol route withdrawal declaration relationship feature—is greater than or equal to a first threshold, it is determined that the Border Gateway Protocol is abnormal in the M time windows, and the corresponding abnormal label of the Border Gateway Protocol is output.

3. The network attack detection method according to claim 1, characterized in that, The characteristic information corresponding to the secure enclosure protocol includes at least one of the following: secure enclosure key initialization message frequency and key negotiation completion rate; The step of determining the anomaly label for each control plane protocol within the M time windows based on the feature information corresponding to each control plane protocol information includes: If the secure shell key initialization frequency in the M time windows is greater than or equal to the second threshold, or if the secure shell key initialization frequency is greater than or equal to the third threshold but the key negotiation completion rate is less than the fourth threshold, it is determined that the secure shell protocol is abnormal in the M time windows, and the abnormal label corresponding to the secure shell protocol is output.

4. The network attack detection method according to claim 1, characterized in that, The characteristic information corresponding to the domain name system includes at least one of the following: the percentage of domain name non-responders and the percentage of long-tail domain name queries; The step of determining the anomaly label for each control plane protocol within the M time windows based on the feature information corresponding to each control plane protocol information includes: If the percentage of no responses to the domain name in the M time windows is greater than or equal to the fifth threshold or the percentage of long-tail domain names is greater than or equal to the sixth threshold, it is determined that the domain name system is abnormal in the M time windows, and the corresponding abnormal label of the domain name system is output.

5. The network attack detection method according to claim 1, characterized in that, The method involves determining cross-window multi-protocol joint attack alarms based on the anomaly labels corresponding to each control plane protocol within the M time windows, and outputting network attack alarm information, including: The anomaly labels corresponding to each time window are summarized to obtain the time distribution results of the anomaly labels; Based on the time distribution results, a cross-window multi-protocol joint attack alarm determination is performed, and network attack alarm information is output.

6. The network attack detection method according to claim 5, characterized in that, Based on the time distribution results, a cross-window multi-protocol joint attack alarm determination is performed, and network attack alarm information is output, including: If any two of the Border Gateway Protocol, the Domain Name System, and the Secure Shell Protocol exhibit anomalies greater than or equal to the seventh threshold within the observation period corresponding to the M time windows, and the number of time windows covered is greater than or equal to a preset eighth threshold, an attack alarm message will be output; or... If at least one anomaly occurs in the Border Gateway Protocol, the Domain Name System, and the Secure Shell Protocol within the observation period corresponding to the M time windows, the attack alarm information will be output.

7. A device for detecting network attacks, characterized in that, The network attack detection device includes an acquisition module, a determination module, and a processing module; The acquisition module is used to acquire feature information of N control plane protocols on the network side in M ​​time windows. The network side control plane protocols include at least one of the following: border gateway protocol, secure shell protocol, and domain name system. The M time windows are time-series continuous time windows, and N and M are positive integers. The determining module is used to determine the anomaly label corresponding to each control plane protocol in the M time windows based on the feature information corresponding to each control plane protocol information for each time window. The processing module is used to determine cross-window multi-protocol joint attack alarms based on the abnormal labels corresponding to each control plane protocol in the M time windows, and output network attack alarm information.

8. An electronic device, characterized in that, It includes a processor and a memory, the memory storing programs or instructions that can run on the processor, the programs or instructions being executed by the processor to implement the network attack detection method as described in any one of claims 1-6.

9. A readable storage medium, characterized in that, The readable storage medium stores a program or instructions that, when executed by a computer, implement the network attack detection method as described in any one of claims 1-6.

10. A computer program product, characterized in that, The computer program product is stored in a storage medium, and when executed by a computer, the computer program product implements the network attack detection method as described in any one of claims 1-6.