Site risk identification methods, devices, computer-readable media and electronic equipment
By combining multiple risk identification methods based on map types, website frameworks, and interface detection features, the accuracy and comprehensiveness of risk identification for cloud-based websites have been addressed, resulting in more efficient risk identification.
Patent Information
- Application Number
- CN202310253061.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2023-03-09
- Publication Date
- 2026-11-17
- Estimated Expiration
- 2043-03-09
AI Technical Summary
Existing technologies for identifying risks on cloud-based websites suffer from poorly designed features, resulting in high rates of missed and false detections. They are unable to effectively identify complex and ever-changing malicious websites and have a high false alarm rate, failing to meet the precise risk identification needs of cloud vendors.
By employing image identification features, website framework identification features, and interface detection identification features, combined with risk type tags and risk scores, risk identification is performed through multiple risk features, thereby improving the accuracy and comprehensiveness of the identification.
It improves the accuracy and comprehensiveness of risk identification for cloud-based websites, reduces false alarm rates, and enables more precise identification of risky websites.
Smart Images

Figure CN116318974B_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of cybersecurity technology, and more specifically, to a site risk identification method, apparatus, computer-readable medium, and electronic device. Background Technology
[0002] With the development of cloud technology, more and more users are using cloud resources provided by cloud providers to build websites.
[0003] However, some criminals have also begun using cloud resources to build malicious websites, and cloud providers can only deal with risky sites after the fact based on user reports. Therefore, there is an urgent need for a method that can accurately identify the risks of websites. Summary of the Invention
[0004] The embodiments of this application provide a site risk identification method, apparatus, computer-readable medium, and electronic device, which can at least improve the accuracy of risk identification to a certain extent.
[0005] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.
[0006] According to one aspect of the embodiments of this application, a site risk identification method is provided. The method includes: determining multiple risk features for risk identification, the multiple risk features including at least one of the following: image identification features, website framework identification features, and interface detection identification features; the feature information corresponding to the image identification feature is at least one fingerprint information of image resources in the site to be identified, the feature information corresponding to the website framework identification feature is fingerprint information of source code files in the site to be identified, and the feature information corresponding to the interface detection identification feature is at least a portion of the return results of the interface in the site to be identified; for the site to be identified, obtaining risk information corresponding to each of the multiple risk features; the risk information includes at least one risk type label and a risk score corresponding to the risk type label, the risk information corresponding to each type of risk feature is obtained by comparing the feature information corresponding to each type of risk feature with preset risk comparison information corresponding to each type of risk feature; and determining the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features.
[0007] According to one aspect of the embodiments of this application, a site risk identification device is provided. The device includes: a risk feature determination unit, configured to determine multiple risk features for risk identification, the multiple risk features including at least one of the following: image identification features, website framework identification features, and interface detection identification features; the feature information corresponding to the image identification feature is at least one fingerprint information of image resources in the site to be identified, the feature information corresponding to the website framework identification feature is fingerprint information of source code files in the site to be identified, and the feature information corresponding to the interface detection identification feature is at least a portion of the return results of the interface in the site to be identified; a feature risk information acquisition unit, configured to acquire risk information corresponding to each of the multiple risk features for the site to be identified; the risk information includes at least one risk type label and a risk score corresponding to the risk type label, and the risk information corresponding to each type of risk feature is obtained by comparing the feature information corresponding to each type of risk feature with preset risk comparison information corresponding to each type of risk feature; and a site risk information determination unit, configured to determine the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features.
[0008] In some embodiments of this application, based on the foregoing scheme, the multiple risk features further include keyword features, and the preset risk comparison information corresponding to the keyword features is a risk keyword and risk type label mapping table. The feature risk information acquisition unit is configured to: extract at least one keyword from the site to be identified; determine the comparison result of the at least one keyword in the risk keyword and risk type label mapping table, and obtain the risk information corresponding to the keyword features based on the comparison result.
[0009] In some embodiments of this application, based on the foregoing scheme, the device further includes a domain name information acquisition unit and a domain name information analysis unit; before determining the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features, each unit is configured to perform the following operations: the domain name information acquisition unit is configured to acquire the domain name information of the site to be identified; the domain name information analysis unit is configured to perform text risk level analysis on the domain name information through a pre-established domain name risk assessment model to obtain a risk score corresponding to the domain name risk feature; the site risk information determination unit is configured to: perform fusion processing on the risk information corresponding to each of the multiple risk features and the risk score corresponding to the domain name risk feature to obtain the risk information of the site to be identified.
[0010] In some embodiments of this application, based on the foregoing scheme, the device further includes an image resource acquisition unit, a fingerprint information generation unit, and a first construction unit; before acquiring the risk information corresponding to each risk feature among the multiple risk features, each unit is respectively configured to perform the following operations: the image resource acquisition unit is configured to acquire image resources of known risk sites; the fingerprint information generation unit is configured to generate at least one fingerprint information, namely file fingerprint and image fingerprint, for the image resources, wherein the file fingerprint is fingerprint information obtained by performing a hash operation on the image resources, and the image fingerprint is fingerprint information used for similarity comparison; the first construction unit is configured to construct a map type fingerprint and risk type label mapping table based on the at least one fingerprint information, and use the map type fingerprint and risk type label mapping table as preset risk comparison information corresponding to the map type identification feature.
[0011] In some embodiments of this application, based on the foregoing scheme, the apparatus further includes a source code file acquisition unit, a hashing unit, a removal unit, and a second construction unit; before acquiring risk information corresponding to each risk feature among multiple risk features, each unit is respectively configured to perform the following operations: the source code file acquisition unit is configured to acquire the source code files of known risk sites; the hashing unit is configured to perform hash operations on each source code file to obtain fingerprint information of each source code file; the removal unit is configured to remove whitelist information from all fingerprint information to obtain risky source code file fingerprint information; the second construction unit is configured to, based on the risky source code file... The source code file fingerprint information and risk type label mapping table are constructed using the fingerprint information of the source code file and the risk type label mapping table, and the source code file fingerprint information and risk type label mapping table are used as preset risk comparison information corresponding to the website framework identification features; the feature risk information acquisition unit is configured to: acquire the source code file to be identified of the site to be identified; perform a hash operation on each source code file to be identified to obtain the fingerprint information to be identified of each source code file; determine the comparison result of the fingerprint information to be identified in the source code file fingerprint information and risk type label mapping table, and obtain the risk information corresponding to the website framework identification features based on the comparison result.
[0012] In some embodiments of this application, based on the aforementioned scheme, the site risk information determination unit is configured to: determine the target risk type label and the risk score corresponding to the target risk type label for each type of risk feature according to the risk information corresponding to each type of risk feature in the multiple risk features; and perform fusion processing on the target risk type label and the risk score corresponding to the target risk type label for each type of risk feature to obtain the risk information of the site to be identified.
[0013] In some embodiments of this application, based on the foregoing scheme, after determining the risk information of the site to be identified, the site risk information determination unit is further configured to: determine in the cloud the risk information of sites served by other resources associated with the user on which the site to be identified depends; and integrate the risk information of the sites served by other resources and the risk information of the site to be identified to obtain the risk information of the user.
[0014] In some embodiments of this application, based on the foregoing scheme, after determining the risk information of the site to be identified, the site risk information determination unit is further configured to: compare the preset risk comparison information corresponding to at least one type of risk feature with the corresponding information of other sites, and take the site with successful comparison as the target site; compare the preset risk comparison information corresponding to other types of risk features with the corresponding information of the target site to determine the risk information of the target site.
[0015] In some embodiments of this application, based on the foregoing scheme, after determining the risk information of the site to be identified, the site risk information determining unit is further configured to: obtain the IP address registered in the cloud for creating the site to be identified; determine other IP addresses registered in the same time period as the IP address; and determine the risk information of the sites corresponding to the other IP addresses.
[0016] According to one aspect of the embodiments of this application, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the site risk identification method as described in the above embodiments.
[0017] According to one aspect of the embodiments of this application, an electronic device is provided, including: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the site risk identification method as described in the above embodiments.
[0018] According to one aspect of the embodiments of this application, a computer program product is provided, the computer program product including computer instructions stored in a computer-readable storage medium, a processor of a computer device reading the computer instructions from the computer-readable storage medium, and the processor executing the computer instructions, causing the computer device to perform the site risk identification method as described in the above embodiments.
[0019] In some embodiments of this application, when identifying a site for risk, risk information corresponding to each of multiple risk features is obtained. This risk information includes at least one risk type label and a risk score corresponding to the risk type label. Based on this, the risk information of the site for risk is determined according to the risk information corresponding to each of the multiple risk features. Since the determined multiple risk features include at least one of image identification features, website framework identification features, and interface detection identification features, this application proposes for the first time to use these risk features for risk identification of the site for risk. These risk features have better specificity for sites with risks, improving the accuracy of risk identification. Furthermore, this application simultaneously uses multiple risk features for risk identification, and the risk information includes both a risk type label and a risk score corresponding to the risk type label, enabling more comprehensive and accurate risk identification.
[0020] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description
[0021] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings:
[0022] Figure 1 A schematic diagram of an exemplary system architecture that can be used to implement the technical solutions of the embodiments of this application is shown;
[0023] Figure 2 A flowchart of a site risk identification method according to an embodiment of this application is shown;
[0024] Figure 3 An embodiment according to this application is shown. Figure 2 A flowchart detailing step 260 in the embodiment;
[0025] Figure 4 A flowchart of the various stages of a model according to an embodiment of this application is shown;
[0026] Figure 5 An embodiment according to this application is shown. Figure 2 A flowchart of the steps preceding step 260 and the details of step 260 in the embodiment;
[0027] Figure 6A flowchart illustrating risk identification based on website framework features according to an embodiment of this application is shown;
[0028] Figure 7 A flowchart illustrating risk identification based on map features according to an embodiment of this application is shown;
[0029] Figure 8 A flowchart illustrating risk identification based on interface detection features according to an embodiment of this application is shown;
[0030] Figure 9 An embodiment according to this application is shown. Figure 3 A flowchart of the steps preceding step 290 and the details of step 290 in the embodiment;
[0031] Figure 10 An embodiment according to this application is shown. Figure 2 A flowchart detailing step 290 in the embodiment;
[0032] Figure 11 An embodiment according to this application is shown. Figure 2 A flowchart of the steps following step 290 in the embodiment;
[0033] Figure 12 Another embodiment according to this application is shown. Figure 2 A flowchart of the steps following step 290 in the embodiment;
[0034] Figure 13 A flowchart illustrating the steps following the determination of risk information of the site to be identified, according to one embodiment of this application, is shown.
[0035] Figure 14 A block diagram of a site risk identification device according to an embodiment of this application is shown;
[0036] Figure 15 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation
[0037] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art.
[0038] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.
[0039] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.
[0040] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.
[0041] With the development of the internet, risky websites that provide harmful information or engage in inappropriate behavior frequently exist online. To create a healthy internet environment, it is necessary to manage these risky websites. Currently, many websites are built using cloud resources provided by cloud vendors; therefore, cloud vendors also need to identify the risks associated with these websites.
[0042] In related technologies, there are mainly three types of solutions for website detection:
[0043] 1. Discovery and identification of malicious web pages based on topic-based web crawling
[0044] Solution Content: This solution combines the textual characteristics of inappropriate web pages with web crawlers in search engines to proactively identify inappropriate web pages and websites on the internet. The results are then categorized and fed back to the user level for processing, thereby purifying the online environment.
[0045] The shortcomings of this scheme are: a) Based on text features for identification and matching, there is a high false positive rate. It can only be used as an auxiliary means to increase the recall rate, and cannot be used as the main means to identify malicious websites; b) When combined with search engines for malicious website identification, it cannot solve the leakage problem in scenarios such as websites not indexed by search engines, websites blocked from being indexed by search engines, and websites blocked by search engines.
[0046] 2. Identifying inappropriate image webpages using webpage features
[0047] Solution content: Use machine learning to detect the features of images on a website and identify web pages with inappropriate images.
[0048] The drawback of this approach is that image recognition-based solutions only perform well in certain scenarios, but cannot effectively identify malicious websites in other scenarios (such as counterfeit websites and websites offering cashback for fake orders).
[0049] 3. Maliciously-prone website detection based on multi-dimensional features
[0050] Solution content: By leveraging relevant features related to domain registration and resolution, and employing machine learning methods to construct a detection model, the solution aims to perform website detection.
[0051] The drawbacks of this approach are: predictions based on domain registration and resolution characteristics are linked to the behavioral habits of risky users, resulting in a high false positive rate and low recall rate. Furthermore, these behavioral characteristics may change as the cloud industry develops, and the method requires extremely high-quality training samples. Additionally, this approach can serve as a predictive method but cannot be used as a basis for accurate detection.
[0052] In summary, the inventors have found that the solutions in the related technologies mainly have the following problems:
[0053] 1. The feature design for detecting fraud risks is unreasonable: the relevant technologies select risk features in a relatively simple way, which cannot avoid missed detections caused by incomplete data sources and false detections when dealing with complex and ever-changing objects to be detected, which is not conducive to subsequent risk management.
[0054] 2. High false alarm rate of detection results: Judging the risk level of resources by detecting the output results of target websites through these risk characteristics has a high false alarm rate, which is not conducive to subsequent risk management.
[0055] Therefore, this application first provides a site risk identification method. The site risk identification method provided by the embodiments of this application can overcome the above-mentioned deficiencies. On the one hand, it proposes for the first time to use more specific risk features such as image type identification features, website frame identification features, and interface detection identification features to identify the risk of the site, which can improve the accuracy of risk identification. On the other hand, by jointly using multiple types of risk features for risk identification, and using risk type labels and the risk scores corresponding to the risk type labels as risk information for risk identification, risk identification can be performed more accurately and comprehensively.
[0056] Figure 1 A schematic diagram of an exemplary system architecture that can be used to implement the technical solutions of the embodiments of this application is shown. Figure 1As shown, the system architecture 100 may include a site administrator terminal 110, a cloud platform 120, and a user terminal 130. The cloud platform 120 specifically includes a cloud server 121 and a risk identification server 122 that are interconnected. Both the site administrator terminal 110 and the user terminal 130 establish communication connections with the cloud platform 120. The risk identification server 122 stores preset risk comparison information. The site administrator terminal 110 is the execution entity in this embodiment. When the site risk identification method provided in this application is applied... Figure 1 In the system architecture shown, a process can be as follows: First, the site administrator uses the site administrator terminal 110 to access the cloud 120, thereby using the resources of the cloud server 121 provided by the cloud 120 to establish the site to be identified; then, the risk identification server 122 accesses the site to be identified and obtains the feature information corresponding to various risk characteristics; next, the risk identification server 122 compares the feature information corresponding to various risk characteristics with the preset risk comparison information corresponding to various risk characteristics to obtain the risk information corresponding to each risk characteristic among the multiple risk characteristics; finally, the risk identification server 122 determines the risk information of the site to be identified based on the risk information corresponding to various risk characteristics.
[0057] In one embodiment of this application, the risk identification server 122 also determines the user's risk information based on the mapping relationship between the site to be identified and the user.
[0058] In one embodiment of this application, the risk identification server 122 further identifies the risks of other servers of the site to be identified based on the relationship between the domain name and the cloud server, and combines the identification results of all servers of all sites to be identified to obtain the risk information of the site to be identified.
[0059] In one embodiment of this application, after obtaining the risk information of the site to be identified, the risk information of the site to be identified is pushed to the terminal of the site manager.
[0060] It should be understood that Figure 1 The number of site administrator terminals, user terminals, cloud servers, and risk identification servers shown in the diagram is merely illustrative. Depending on implementation needs, there can be any number of site administrator terminals, user terminals, cloud servers, and risk identification servers; that is, there can be multiple user terminals and multiple site administrator terminals, and cloud servers and risk identification servers can be server clusters composed of multiple servers, etc.
[0061] It should be noted that, Figure 1 The illustration shown is merely one embodiment of this application. Although in Figure 1In the embodiment, the user terminal is a laptop computer and the site administrator terminal is a desktop computer. However, in other embodiments of this application, the user terminal and the site administrator terminal can also be various types of terminal devices such as desktop computers, workstations, tablets, vehicle terminals, and smartphones. Figure 1 In the embodiment, both the risk identification server and the cloud server are located in the cloud. However, in other embodiments of this application, the risk identification server may also be located outside the cloud, for example, it may be a separate server. Figure 1 In the embodiments described, the preset risk comparison information is stored on the risk identification server. However, in other embodiments of this application, the preset risk comparison information may also be stored in a database accessible to the risk identification server. This application does not limit the scope of protection in any way, nor should it restrict the scope of protection of this application.
[0062] As is readily understood, the site risk identification method provided in this application is generally executed by a server, and correspondingly, the site risk identification device is generally located within the server. However, in other embodiments of this application, the terminal device may also have similar functions to the server, thereby executing the site risk identification scheme provided in this application.
[0063] Therefore, the embodiments of this application can be applied to terminals or servers. The server can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, etc., but is not limited to these. The terminal and server can be directly or indirectly connected via wired or wireless communication, and this application does not impose any restrictions.
[0064] The implementation details of the technical solutions in the embodiments of this application are described in detail below:
[0065] Figure 2 A flowchart of a site risk identification method according to an embodiment of this application is shown. The site risk identification method provided in this embodiment can be considered as a whole as a model. This site risk identification method can be executed by various computing and processing devices, such as user terminals or cloud servers. User terminals include, but are not limited to, mobile phones, computers, smart voice interaction devices, smart home appliances, vehicle terminals, and aircraft. Please refer to... Figure 2 As shown, the site risk identification method includes at least the following steps:
[0066] In step 250, multiple risk features for risk identification are determined. These multiple risk features include at least one of the following: image identification features, website framework identification features, and interface detection identification features. The feature information corresponding to the image identification feature is at least one fingerprint of the image resources in the site to be identified. The feature information corresponding to the website framework identification feature is the fingerprint of the source code file in the site to be identified. The feature information corresponding to the interface detection identification feature is at least a part of the return result of the interface in the site to be identified.
[0067] The solution in this application embodiment is based on risk characteristics for risk identification. Therefore, it is necessary to first determine which risk characteristics are used for risk identification. Pre-defined risk characteristics can be used as multiple risk characteristics for risk identification, or multiple risk characteristics for risk identification can be determined from a preset set of risk characteristics. For example, multiple risk characteristics for risk identification can be selected from a preset set of risk characteristics by sampling.
[0068] Multiple risk features can include image type identification features, website frame identification features, interface detection identification features, and keyword features simultaneously, or they can include only any two or three of these risk features, as long as the final determined multiple risk features include at least one of image type identification features, website frame identification features, and interface detection identification features.
[0069] When applying the site risk identification method provided in the embodiments of this application, risk identification can be carried out by combining multiple risk features and domain name risk features, which are special risk features.
[0070]
[0071] Table 1
[0072] Table 1 shows detailed information on various risk characteristics. It can be seen that image identification characteristics, website framework identification characteristics, and interface detection identification characteristics are specific risk characteristics, meaning they are more specific in risk identification and are therefore the primary detection methods. Keyword characteristics and domain name risk characteristics are non-specific risk characteristics, meaning they are less specific than specific risk characteristics and are therefore used as auxiliary detection methods. The "Risk Type" column in Table 1 indicates whether the corresponding risk characteristic has the ability to determine the risk type. It can be seen that, except for domain name risk characteristics, all other risk characteristics have the ability to determine the risk type.
[0073] As can be seen from Table 1, each risk characteristic has a corresponding weight. The weights will be introduced in subsequent content and will not be detailed here.
[0074] In step 260, for the site to be identified, risk information corresponding to each risk feature among multiple risk features is obtained; the risk information includes at least one risk type label and a risk score corresponding to the risk type label. The risk information corresponding to each type of risk feature is obtained by comparing the feature information corresponding to each type of risk feature with the preset risk comparison information corresponding to each type of risk feature.
[0075] If a website provides harmful information or engages in inappropriate behavior, then it is an abnormal and risky website. In other words, a risky website can be any website that does not comply with relevant laws and regulations. For example, a risky website can be a website that deceives users or a website that spreads unhealthy information.
[0076] The site to be identified can be a web site built using cloud resources provided by a cloud platform. The cloud platform provides products such as Cloud Virtual Machines (CVMs), DNSPods (domain name resolution), and Content Delivery Networks (CDNs). Users purchase these products from the cloud platform; these products are the resources needed to build and use the web site. Specifically, users build their web sites by purchasing cloud servers and other products, and register, resolve, and accelerate their web sites by purchasing DNSPods, CDNs, and other products.
[0077] Figure 3 An embodiment according to this application is shown. Figure 2 A flowchart detailing step 260 in the embodiment is provided. Please refer to [link / reference]. Figure 3 As shown, obtaining risk information corresponding to each risk feature among multiple risk features may specifically include the following steps:
[0078] In step 261, at least one keyword is extracted from the site to be identified.
[0079] The websites to be identified typically contain text information, from which keywords can be extracted. Similarly, these websites often contain images, from which keywords can be extracted. All keywords can be extracted from the website for detection. Keyword extraction can be performed using either a non-repeating or a repeating method.
[0080] In step 262, the comparison result of at least one keyword in the risk keyword and risk type label mapping table is determined, and the risk information corresponding to the keyword features is obtained based on the comparison result.
[0081] In one embodiment of this application, determining the comparison result of at least one keyword in the risk keyword and risk type label mapping table, and obtaining risk information corresponding to the keyword features based on the comparison result, includes: determining the risk type label that each keyword hits in the risk keyword and risk type label mapping table; and determining the risk score corresponding to each risk type label based on the number of times each risk type label is hit by the keyword.
[0082] For example, if a certain risk type tag is matched by keywords a certain number of times, then the risk score corresponding to that risk type tag can be:
[0083]
[0084] As is easy to understand, the number of times a risk type tag is matched by keywords can include the number of times it is matched by repeated keywords and the number of times it is matched by unique keywords, or it can only include the number of times it is matched by unique keywords. The calculation method for the risk score corresponding to different risk type tags can also be different.
[0085] Risk scores corresponding to each risk type label can also be determined using other methods. For example, each time a unique keyword hits a risk type label, 1 point can be awarded. However, cases where multiple repeated keywords hit the same risk type label are scored separately; for instance, a uniform score of 1.2 points could be awarded, or a score of 1 + ln(n), where n is the number of times the same keyword is repeated. Alternatively, a separate hit score can be assigned to each keyword, and the risk score corresponding to each risk type label can be determined by summing the hit scores for each risk type label hit.
[0086] When multiple risk features also include keyword features, the preset risk comparison information corresponding to the keyword features is a risk keyword and risk type label mapping table.
[0087] In one embodiment of this application, before obtaining the risk information corresponding to each risk feature among multiple risk features, the site risk identification method further includes: obtaining the keywords of known risk sites and the risk type label corresponding to each keyword; and constructing a risk keyword and risk type label mapping table based on the keywords and the risk type label corresponding to each keyword.
[0088] The same risk type label can correspond to multiple keywords.
[0089] For example, keywords such as "Shishicai" and "Xin Kuaishan" are labeled as "lottery risk"; while keywords such as "fund account opening" and "credit card application" are labeled as "financial risk".
[0090] The risk type label corresponding to each keyword can be determined manually or by a predetermined recognition algorithm.
[0091] Figure 4 A flowchart of the various stages of a model according to an embodiment of this application is shown. Below, in conjunction with… Figure 4 The following will further describe the solutions of the embodiments of this application. Please refer to [link / reference]. Figure 4 As shown, the model used to implement the method provided in this application mainly includes four stages: preparation, identification and assessment, user model, and governance. The process in the preparation stage is as follows: first, analyze existing risky websites; then, form various feature rules; finally, input the rules into the model. Here, feature rules refer to the preset risk comparison information corresponding to various risk features and the method for generating risk information corresponding to each type of risk feature.
[0092] The following describes the characteristics of website frame identification. Figure 5 An embodiment according to this application is shown. Figure 2 The flowchart in this embodiment details the steps preceding step 260 and step 260 itself. Please refer to [link / reference needed]. Figure 5 As shown, before obtaining the risk information corresponding to each risk feature among multiple risk features, the risk identification method for this site includes:
[0093] In step 220, the source code files of known risky sites are obtained.
[0094] Accessing known risky websites can yield the corresponding source code files, which may include HTML, JS, and CSS files.
[0095] In step 230, a hash operation is performed on each source code file to obtain the fingerprint information of each source code file.
[0096] Hash algorithms such as SHA1 can be used to perform hash operations on source code files to obtain corresponding fingerprint information.
[0097] In step 240, whitelist information is removed from all fingerprint information to obtain the fingerprint information of the risky source code file.
[0098] The whitelist information can be fingerprint information of publicly available technical frameworks such as vue.min.js. By removing information from the whitelist, interference from these publicly available technical frameworks can be avoided.
[0099] In step 250, a mapping table of source code file fingerprint information and risk type label is constructed based on the risk source code file fingerprint information, and the source code file fingerprint information and risk type label mapping table are used as preset risk comparison information corresponding to the website framework identification features.
[0100] The source code file fingerprint information and risk type label mapping table contains a mapping relationship between risk source code file fingerprint information and risk type labels. The risk type labels corresponding to the risk source code file fingerprint information can be set in advance based on experience.
[0101] Please continue reading Figure 5 Obtaining risk information corresponding to each risk feature among multiple risk features can specifically include the following steps:
[0102] In step 261', the source code file of the site to be identified is obtained.
[0103] By accessing the site to be identified, the corresponding source code file to be identified can be obtained.
[0104] In step 262', a hash operation is performed on each source code file to be identified to obtain the fingerprint information of each source code file to be identified.
[0105] The hash algorithm used when generating the fingerprint information of the source code files can be used to perform hash operations on each source code file to be identified, thereby obtaining the corresponding fingerprint information.
[0106] In step 263', the comparison result of the fingerprint information to be identified in the source code file fingerprint information and risk type label mapping table is determined, and the risk information corresponding to the website framework identification features is obtained based on the comparison result.
[0107] In one embodiment of this application, determining the comparison result of the fingerprint information to be identified in the source code file fingerprint information and risk type label mapping table, and obtaining the risk information corresponding to the website framework identification features based on the comparison result, includes: determining the risk type label that the fingerprint information to be identified of each source code file is matched in the source code file fingerprint information and risk type label mapping table; and determining the risk score corresponding to each risk type label based on the number of times each risk type label is matched by the fingerprint information to be identified.
[0108] For example, if a certain risk type label is matched L times by the fingerprint information to be identified, then the risk score corresponding to that risk type label can be:
[0109]
[0110] Figure 6 A flowchart illustrating risk identification based on website framework features according to one embodiment of this application is shown. Please see [link to flowchart documentation]. Figure 6As shown, the preparation phase can include the following process: First, input the risky website into the model; then, access the risky website and obtain all its files; next, perform hash operations on the files written in JavaScript (JS files), Cascading Style Sheets (CSS files), and Hypertext Markup Language (HTML files) of the risky website to obtain corresponding fingerprint information; then, filter out public frame fingerprints from the whitelist from the obtained fingerprint information; finally, match the risk type to the fingerprint information and save the fingerprint information. During risk identification, the following process can be included: First, access the website to be tested and obtain the corresponding source code files; then, perform hash operations on all files written in JavaScript, Cascading Style Sheets, and Hypertext Markup Language (HTML) files of the website to be tested; next, match the hash results with the results in the model library; finally, output the hit rate and risk type.
[0111] In one embodiment of this application, before obtaining the risk information corresponding to each risk feature among multiple risk features, the site risk identification method further includes: obtaining image resources of known risk sites; generating at least one fingerprint information, namely file fingerprint and image fingerprint, for the image resources, wherein the file fingerprint is fingerprint information obtained by performing a hash operation on the image resources, and the image fingerprint is fingerprint information used for similarity comparison; constructing a mapping table of image type fingerprint and risk type label based on at least one fingerprint information, and using the mapping table of image type fingerprint and risk type label as preset risk comparison information corresponding to the image type identification feature.
[0112] Image resources are referred to as image types. You can acquire only image resources in specific formats such as favicon.ico from known risky websites, or you can acquire all image resources from known risky websites. For example, you can extract the favicon.ico icon from a risky website that is suspected of being a fake website, upload it to the model image type library, and associate this image type with risk type tags such as "fake website" and "phishing website." Then, you can generate at least one of file fingerprints and image fingerprints. The risk type tags corresponding to the image type can be set based on experience.
[0113] In practical applications, the online logos of companies such as banks, funds, and securities firms can be used as images to identify corresponding financial phishing websites. This feature is highly effective in identifying risks in the financial sector.
[0114] At least one fingerprint can include both file fingerprints and image fingerprints. Fingerprint information can be obtained by hashing image resources using hash algorithms such as SHA1. Image fingerprints can be generated using the mean hash algorithm, or by performing convolution processing on the image resources.
[0115] Mean hashing is a commonly used algorithm for generating image fingerprints. The generated image fingerprints can be compared and their similarity calculated. It generally involves several steps: reducing image size, simplifying colors, calculating the average pixel value, comparing pixel grayscale values, and generating an image hash. Subsequently, when comparing similarity, the Hamming distance between the two image fingerprints is used as the image similarity. The smaller the Hamming distance, the higher the image similarity.
[0116] In one embodiment of this application, obtaining risk information corresponding to each risk feature among multiple risk features may further include the following steps: obtaining image resources of the site to be identified; generating file fingerprints and image fingerprints of the image resources; determining the comparison results of the file fingerprints and image fingerprints of each image resource in the map type fingerprint and risk type label mapping table, and obtaining risk information corresponding to the map type identification features based on the comparison results.
[0117] Specifically, if the similarity between the image fingerprint of an image resource and the image fingerprint in the mapping table of image type fingerprint and risk type label is greater than a predetermined similarity threshold, or if the file fingerprint of the image resource matches the image type fingerprint in the mapping table of image type fingerprint and risk type label, then the comparison is considered successful. If the number of successful comparisons for a certain risk type label for each image resource of the site to be identified is L, then the risk score corresponding to that risk type label can be:
[0118]
[0119] The predefined similarity threshold can be set as needed, for example, it can be set to 80%.
[0120] Of course, in other embodiments of this application, a similarity of more than 80% between image fingerprints can be recorded as one successful match, a similarity of more than 60% but less than 80% can be recorded as 0.8 successful matches, and a similarity of less than 60% can be recorded as a failed match. Alternatively, the number of successful matches can be determined based on the actual similarity between image fingerprints, as long as the number of successful matches is positively correlated with the actual similarity between image fingerprints.
[0121] Figure 7 A flowchart illustrating risk identification based on map features according to an embodiment of this application is shown. See also... Figure 7 As shown, the preparation stage can include the following process: First, extract images of known risk sites with obvious risk characteristics as image types; then, form a mapping table between image types and risk types; next, input the image types and mapping table into the model; then, perform model processing, use a hash algorithm to generate file fingerprints of image types, and use an average hash algorithm to process the image types to generate image fingerprints, thus completing the preparation stage.
[0122] The risk identification process is as follows: First, screen the image resources of the site to be identified; then, compare the file fingerprint of the image resource with the file fingerprint of the image in the mapping table; next, generate the image fingerprint of the image resource and compare the image fingerprint with the image fingerprint of the image in the mapping table, and the risk identification is completed.
[0123] In one embodiment of this application, before obtaining the risk information corresponding to each risk feature among multiple risk features, the site risk identification method further includes: obtaining at least a portion of the return results of the interface of a known risk site; constructing a return result and risk type label mapping table based on the return results of the interface, and using the return results and risk type label mapping table as preset risk comparison information corresponding to the interface detection and identification features.
[0124] In one embodiment of this application, obtaining risk information corresponding to each risk feature among multiple risk features may further include the following steps: obtaining at least a portion of the return results of the interface of the site to be identified; determining the comparison results of at least a portion of the return results of each interface in the return results and risk type label mapping table, and obtaining the risk information corresponding to the interface detection and identification features based on the comparison results.
[0125] In practical applications, risk identification can be performed by selecting only the site's configuration interfaces or interfaces that return results with specific keywords.
[0126] For example, if the known risky website is http: / / abc.com, analysis reveals that it is a "brushing orders and cashback" type of risky website. It has an API http: / / abc.com / api / config / get_config_default that returns JSON configuration information containing the website's configuration details. Therefore, the path to the API's return result can be extracted as: / api / config / get_config_default. The returned result from accessing this path, or keywords from the returned result, can be entered into the model, along with the corresponding risk type label "brushing orders and cashback".
[0127] When identifying a site, if the site is http: / / def.org, the corresponding return result is obtained by accessing the site's API http: / / def.org / api / config / get_config_default. The returned result or keywords within the returned result are then matched with the corresponding content in the model. If the number of successful matches for a certain risk type label is L, then the risk score corresponding to that risk type label can be:
[0128]
[0129] Although each successful match in this embodiment is counted once, in other embodiments of this application, different numbers of matches can be used to count when matching different content.
[0130] Figure 8 A flowchart illustrating risk identification based on interface detection features according to an embodiment of this application is shown. Figure 8 As shown, the preparation phase may include the following process: First, analyze the API interfaces of known websites; then, record specific interface paths and returned content; finally, match risk types and input them into the model. The risk identification phase may include the following process: First, access specific interface paths for the website to be identified; then, compare the returned results of the specific interface paths with the existing returned content in the model; finally, output the hit rate and risk type.
[0131] Please continue reading Figure 4 As shown, the identification and assessment phase involves the following process: On one hand, feature matching is performed on unknown websites to generate corresponding feature tags; on the other hand, association investigation and penetration investigation are conducted on the target website to generate corresponding feature tags. These feature tags are risk type labels. Association investigation and penetration investigation refer to investigating other resources related to the target website's resources. For example, this could be other cloud servers connected to the target website's cloud server via an API gateway, or other cloud servers under the same domain name associated with the target website's cloud server.
[0132] In step 290, the risk information of the site to be identified is determined based on the risk information corresponding to each risk feature among the multiple risk features.
[0133] By combining the risk information corresponding to various risk characteristics, the risk information of the site to be identified can be obtained. The risk information of the site to be identified can include risk type labels and risk scores corresponding to the risk type labels.
[0134] In one embodiment of this application, determining the risk information of a site to be identified based on the risk information corresponding to each of the multiple risk features includes: inputting the risk information corresponding to each of the multiple risk features into a predetermined risk information prediction model to obtain the risk information of the site to be identified output by the predetermined risk information prediction model.
[0135] Predictive risk information models can be built using various algorithms, such as machine learning algorithms or deep learning algorithms based on neural networks.
[0136] Figure 9 An embodiment according to this application is shown. Figure 3 A flowchart detailing the steps preceding step 290 and step 290 in the embodiment. Please refer to [link / reference needed]. Figure 9 As shown, the following steps may be included before step 290:
[0137] In step 270, the domain name information of the site to be identified is obtained.
[0138] The inventors discovered that risky users often purchase domain names with no real meaning and messy content due to price sensitivity and other reasons, and therefore risk identification can be carried out based on this.
[0139] Domain information can be obtained by breaking down the URL (Uniform Resource Locator) of the site to be identified.
[0140] In step 280, the risk level of the domain name information is analyzed by a pre-established domain name risk assessment model to obtain a risk score corresponding to the domain name risk characteristics.
[0141] Domain risk assessment models can be built based on various machine learning or deep learning algorithms. They can be trained by labeling known risky domains and normal domains.
[0142] The domain risk assessment model analyzes and evaluates domain information, outputting a risk score corresponding to the domain's risk characteristics. The risk score S corresponding to the domain's risk characteristics is... B Its value ranges from 0 to 10, and the larger the value, the higher the risk.
[0143] Step 290 may specifically include:
[0144] In step 290', the risk information corresponding to each risk feature among the multiple risk features and the risk score corresponding to the domain name risk feature are fused to obtain the risk information of the site to be identified.
[0145] The final risk information for the site to be identified is a fusion of the risk information corresponding to each type of risk feature and the risk score corresponding to the domain name risk feature.
[0146] Figure 10 An embodiment according to this application is shown. Figure 2 A flowchart detailing step 290 in the embodiment is provided. Please refer to [link / reference]. Figure 10 As shown, determining the risk information of a site to be identified based on the risk information corresponding to each of the multiple risk characteristics can specifically include the following steps:
[0147] In step 291, based on the risk information corresponding to each risk feature among the multiple risk features, the target risk type label for each type of risk feature and the risk score corresponding to the target risk type label are determined.
[0148] The target risk type labels for various risk characteristics may be the same or different.
[0149] For keyword features, the risk type label with the highest corresponding risk score can be used as the target risk type label. That is, the risk type label that is most frequently matched by the keyword of the site to be identified can be used as the target risk type label M of the keyword feature. A If the keywords of the site to be identified do not match any risk type label, then the target risk type label for the keyword feature can be empty.
[0150] For website frame identification features, the risk type label with the highest risk score can be used as the target risk type label. That is, the risk type label that is most frequently matched by the fingerprint information of the site to be identified can be used as the target risk type label M for website frame identification features. D If the fingerprint information of the site to be identified does not match any risk type label, then the target risk type label of the website framework identification feature can be empty.
[0151] For map type identification features, the risk type label with the highest corresponding risk score can be used as the target risk type label, that is, the risk type label with the most successful comparisons can be used as the target risk type label M for map type identification features. C If no match is found, the target risk type label for the image identification feature can be empty.
[0152] For interface detection and identification features, the risk type label with the highest corresponding risk score can be used as the target risk type label, that is, the risk type label with the most successful matches can be used as the target risk type label M of the interface detection and identification features. E If no match is found, the target risk type label for the interface detection and identification features can be empty.
[0153] In step 292, the target risk type labels of various risk characteristics and the risk scores corresponding to the target risk type labels are fused to obtain the risk information of the site to be identified.
[0154] The risk information of the site to be identified is calculated by comprehensively considering the target risk type labels of various risk characteristics and the risk scores corresponding to the target risk type labels.
[0155] In one embodiment of this application, the target risk type labels of various risk features and the risk scores corresponding to the target risk type labels are fused to obtain the risk information of the site to be identified. This includes: inputting the target risk type labels of various risk features and the risk scores corresponding to the target risk type labels into a pre-trained machine learning model to obtain the risk information of the site to be identified.
[0156] Pre-trained machine learning models can be trained based on various machine learning algorithms. These models can comprehensively analyze the target risk type labels and corresponding risk scores of various risk features to obtain accurate risk information for the sites to be identified.
[0157] In one embodiment of this application, the target risk type labels of various risk features and the risk scores corresponding to the target risk type labels are fused to obtain the risk information of the site to be identified, including: determining the risk type label of the site to be identified based on the target risk type labels of various risk features; and determining the risk score of the site to be identified based on the risk scores corresponding to the target risk type labels of various risk features.
[0158] The risk information for the site to be identified includes risk type labels and risk scores. Specifically, target risk type labels for various risk features can be input into a predefined label recognition model to obtain the risk type labels for the site, and the risk scores corresponding to the target risk type labels for various risk features can be input into a predefined score prediction model to obtain the risk scores for the site. The predefined label recognition model and the predefined score prediction model can be constructed from various models; for example, the predefined score prediction model can be constructed based on a logistic regression model.
[0159] In one embodiment of this application, each type of risk feature has a corresponding weight coefficient. The risk score of the site to be identified is determined based on the risk score corresponding to the target risk type label of each type of risk feature. This includes: determining the weighted sum of the risk scores corresponding to the target risk type label of each type of risk feature based on the weight coefficients of each type of risk feature, and using this sum as the risk score of the site to be identified.
[0160] As mentioned earlier, the model uses multiple risk features, including image type identification features, website framework identification features, interface detection identification features, and keyword features. It also includes a special type of risk feature: domain name risk features. For these five types of risk features, the risk score S of the site to be identified can be obtained using the following formula. U :
[0161] S U =*S A +*B +* C +* D +* E
[0162] Among them, S A S represents the risk score corresponding to the target risk type label of feature A. B The risk score corresponds to the target risk type label for feature B, and so on. a, b, c, d, and e are the weight coefficients for features A, B, C, D, and E, respectively. The risk characteristics corresponding to a, b, c, d, e, and A, B, C, D, and E are described in Table 1 and will not be repeated here. The weight coefficients must satisfy the following:
[0163] a + b + c + d + e = 1
[0164] The values of each weight coefficient can be set based on experience or actual conditions. For example, based on the characteristics of the features themselves and the feedback from the actual model operation, the values of each weight parameter can be set as follows:
[0165] Weight parameters a b c d e Recommended value 0.12 0.08 0.3 0.2 0.3
[0166] Table 2
[0167] As shown in Table 2, the values of each weight parameter listed therein correspond to the relative values of each weight in Table 1.
[0168] The risk type label for the site to be identified can be a target risk type label related to a certain type of risk characteristic, or it can be other forms of label.
[0169] In one embodiment of this application, the risk type label of the site to be identified is determined based on the target risk type label of various risk features, including: determining the target risk feature with the largest product of the corresponding weight coefficient and risk score among various risk features; and using the target risk type label of the target risk feature as the risk type label of the site to be identified.
[0170] In one embodiment of this application, determining the risk type label of the site to be identified based on the target risk type label of various risk features includes: identifying the target risk feature with the largest corresponding weight coefficient and whose corresponding target risk type label is not empty among various risk features; and using the target risk type label of the target risk feature as the risk type label of the site to be identified.
[0171] Specifically, since each type of risk feature has a corresponding weight coefficient, the target risk feature can be determined as follows: sort the various risk features in descending order of weight coefficient; starting from the risk feature at the top of the list, determine whether the target risk type label corresponding to each type of risk feature is empty, and take the first type of risk feature whose corresponding target risk type label is not empty as the target risk feature.
[0172] As mentioned earlier, since domain risk feature (B) does not have the ability to determine risk type, it does not participate in the generation of risk type labels for the site to be identified. Based on the weighting coefficients in Table 2, the judgment order of the target risk type labels corresponding to the four risk features A, C, D, and E is M. E M C M D M A That is, the risk type labels of the sites to be identified are sequentially taken as M. E M C M D M A The non-empty values in these four target risk type labels.
[0173] Figure 11 An embodiment according to this application is shown. Figure 2 A flowchart of the steps following step 290 in this embodiment. Please refer to [link / reference]. Figure 11 As shown, after step 290, the method may further include:
[0174] In step 2100, the preset risk comparison information corresponding to at least one type of risk feature is compared with the corresponding information of other sites, and the site that successfully matches the comparison is taken as the target site.
[0175] In step 2110, the preset risk comparison information corresponding to other types of risk characteristics is compared with the corresponding information of the target site to determine the risk information of the target site.
[0176] Specifically, most risky websites are built on the HTTP protocol and do not have SSL encryption at the protocol level. Therefore, we can perform large-scale feature diffusion based on the traffic data of the cloud platform.
[0177] For keyword features, keyword screening can be performed on all outgoing HTTP traffic packet data of the data center based on the keywords configured in the model. The matched keywords are then entered into the model for screening of other types of features.
[0178] For domain risk characteristics, the Host field in the header of all incoming HTTP traffic can be extracted to determine the domain risk. For high-risk input models, other types of characteristics can be screened.
[0179] For image type identification features, all outgoing HTTP traffic with content-type such as image / png can be extracted, and corresponding hash algorithms can be calculated. For the input model that matches the feature hash library, other types of features can be screened.
[0180] For website framework identification features, we can extract the packet body data of all outgoing HTTP traffic with content-type "application / javascript", "text / html", and "text / css", perform corresponding hash calculations, and screen other types of features for the input model that matches the feature hash library.
[0181] For interface detection and feature identification, all ingress HTTP traffic can be extracted to match the corresponding access path. The returned results are then matched with the feature return results. For the input model that matches the feature, other types of features are screened.
[0182] During traffic screening, image type identification and website frame feature identification may not be suitable for large-scale screening due to high computational demands. A diffusion scheme can be selected based on actual computing power. Alternatively, the other three screening methods can be used as preliminary screenings, followed by the determination of image type and website frame feature characteristics based on the screening results.
[0183] Figure 12 Another embodiment according to this application is shown. Figure 2 A flowchart of the steps following step 290 in this embodiment. Please refer to [link / reference]. Figure 12 As shown, step 290 may further include:
[0184] In step 2100', risk information of the site to be identified is determined in the cloud, which is associated with other resources served by the user on which the site to be identified depends.
[0185] The method provided in this application embodiment can be used to identify other sites in the cloud that are associated with the site to be identified, thereby obtaining the corresponding risk information.
[0186] In step 2110', the risk information of the site to be identified served by other resources and the risk information of the site to be identified are integrated to obtain the user's risk information.
[0187] The identification results for all sites to be identified can be integrated and summarized at the user level to obtain the user's risk score:
[0188]
[0189] Among them, Un For the site's identifier, For site U n Risk score;
[0190] Risk type label:
[0191]
[0192] in, For site U n Risk type label.
[0193] User risk information forms the basic user model. Furthermore, combining this with user resource usage and basic information about the user entity yields richer user model data, resulting in a more comprehensive user model. For example, user model data can be obtained across the following dimensions: "risk score," "risk type indicator," "user geographic characteristics," "resource geographic distribution characteristics," "resource product distribution characteristics," "user resource usage behavior characteristics," "user entity industry characteristics," and "user entity size characteristics."
[0194] Please continue reading Figure 4 As shown, the user modeling stage includes the following process: First, obtain resource characteristics; then, summarize user resource information based on resource characteristics; finally, combine user resource information with subject information to form a user model.
[0195] The resulting user models can be used to identify high-risk user groups, understand their behavioral characteristics, implement targeted risk control, and observe risk development trends. For example, risk scores can be used to identify top-risk users, and combined with risk type indicators, governance measures for these users can be guided. The distribution characteristics of resource ownership across products can be used to observe the distribution trends of risk development and analyze which products and technical architectures top-risk users are concentrated in. Simultaneously, this user model data can also be used to expand the user group and increase the coverage of high-risk users.
[0196] Figure 13 A flowchart illustrating the steps following the determination of risk information for the site to be identified, according to one embodiment of this application, is shown. Figure 13 As shown, the following steps may be included:
[0197] In step 1310, the IP address registered in the cloud for creating the site to be identified is obtained.
[0198] A website can only be accessed from the outside world if it uses an IP address.
[0199] In step 1320, other IP addresses that registered within the same time period as the IP address are identified.
[0200] For a period of time, a user may register multiple IP addresses on the cloud platform.
[0201] In step 1330, risk information for sites corresponding to other IP addresses is determined.
[0202] Other IP addresses are also used to build corresponding websites, so the risk information of these websites can also be identified.
[0203] In this embodiment of the application, risk screening of sites with other IP addresses is carried out by means of time period constraints, thereby realizing the diffusion of risk screening at the user level.
[0204] Of course, in other embodiments of this application, risk screening can also be performed on other websites served by resources purchased from the same IP address, from the perspective of the IP address from which the user purchased the resources.
[0205] Please continue reading Figure 4 As shown, after the user model phase ends, the governance phase begins, during which various measures can be taken to manage risks, specifically as follows: First, resource monitoring is conducted; then, if a risk is detected, a resource warning is issued; next, if the risk reaches a certain level, resources are blocked; finally, the user is removed from the user list.
[0206] By using the model proposed in this application, over 1,000 high-risk users were identified. Through the management of these users, the number of risky websites decreased by over 90%, effectively mitigating risks in the cloud. Simultaneously, the high accuracy rate ensured stable customer complaints and zero incidents.
[0207] In summary, the site risk identification method provided in this application has the following advantages: by proposing features that can more accurately identify risks, it avoids the problem of high false positive rates when facing complex and diverse websites to be detected; by designing a composite risk feature model with different weights, it avoids the problem of missed detections and false positives caused by single features, and achieves feature coverage of most currently risky websites by designing multiple risk features and weighting methods; by relying on the connection between unique resources and user identities, it establishes user models while detecting resources. It does not rely on the results of a single resource, but uses the results of the user model as the risk standard. This greatly reduces the false positive rate and also ensures recall and governance efficiency through diffusion via the user model.
[0208] The following describes an apparatus embodiment of this application, which can be used to execute the site risk identification method in the above embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the site risk identification method described above.
[0209] Figure 14A block diagram of a site risk identification device according to an embodiment of this application is shown. (Refer to...) Figure 14 As shown, a site risk identification device 1400 according to an embodiment of this application includes: a risk feature determination unit 1410, a feature risk information acquisition unit 1420, and a site risk information determination unit 1430. The risk feature determination unit 1410 is used to determine multiple risk features for risk identification, the multiple risk features including at least one of the following: image identification features, website framework identification features, and interface detection identification features; the feature information corresponding to the image identification feature is at least one fingerprint information of the image resources in the site to be identified, the feature information corresponding to the website framework identification feature is the fingerprint information of the source code files in the site to be identified, and the feature information corresponding to the interface detection identification feature is at least a part of the return result of the interface in the site to be identified; the feature risk information acquisition unit 1420 is used to acquire risk information corresponding to each of the multiple risk features for the site to be identified; the risk information includes at least one risk type label and a risk score corresponding to the risk type label, and the risk information corresponding to each type of risk feature is obtained by comparing the feature information corresponding to each type of risk feature with the preset risk comparison information corresponding to each type of risk feature; the site risk information determination unit 1430 is used to determine the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features.
[0210] In some embodiments of this application, based on the foregoing scheme, the multiple risk features further include keyword features, and the preset risk comparison information corresponding to the keyword features is a risk keyword and risk type label mapping table. The feature risk information acquisition unit 1420 is configured to: extract at least one keyword from the site to be identified; determine the comparison result of the at least one keyword in the risk keyword and risk type label mapping table, and obtain the risk information corresponding to the keyword features based on the comparison result.
[0211] In some embodiments of this application, based on the foregoing scheme, the device further includes a domain name information acquisition unit and a domain name information analysis unit; before determining the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features, each unit is configured to perform the following operations: the domain name information acquisition unit is configured to acquire the domain name information of the site to be identified; the domain name information analysis unit is configured to perform text risk level analysis on the domain name information through a pre-established domain name risk assessment model to obtain a risk score corresponding to the domain name risk feature; the site risk information determination unit is configured to: perform fusion processing on the risk information corresponding to each of the multiple risk features and the risk score corresponding to the domain name risk feature to obtain the risk information of the site to be identified.
[0212] In some embodiments of this application, based on the foregoing scheme, the device further includes an image resource acquisition unit, a fingerprint information generation unit, and a first construction unit; before acquiring the risk information corresponding to each risk feature among the multiple risk features, each unit is respectively configured to perform the following operations: the image resource acquisition unit is configured to acquire image resources of known risk sites; the fingerprint information generation unit is configured to generate at least one fingerprint information, namely file fingerprint and image fingerprint, for the image resources, wherein the file fingerprint is fingerprint information obtained by performing a hash operation on the image resources, and the image fingerprint is fingerprint information used for similarity comparison; the first construction unit is configured to construct a map type fingerprint and risk type label mapping table based on the at least one fingerprint information, and use the map type fingerprint and risk type label mapping table as preset risk comparison information corresponding to the map type identification feature.
[0213] In some embodiments of this application, based on the foregoing scheme, the apparatus further includes a source code file acquisition unit, a hashing unit, a removal unit, and a second construction unit; before acquiring risk information corresponding to each risk feature among multiple risk features, each unit is respectively configured to perform the following operations: the source code file acquisition unit is configured to acquire the source code files of known risk sites; the hashing unit is configured to perform hash operations on each source code file to obtain fingerprint information of each source code file; the removal unit is configured to remove whitelist information from all fingerprint information to obtain risky source code file fingerprint information; the second construction unit is configured to, based on the risky source code file... The fingerprint information constructs a mapping table between source code file fingerprint information and risk type label, and uses the source code file fingerprint information and risk type label mapping table as preset risk comparison information corresponding to the website framework identification features; the feature risk information acquisition unit 1420 is configured to: acquire the source code files to be identified of the site to be identified; perform hash operations on each source code file to be identified to obtain the fingerprint information to be identified of each source code file; determine the comparison result of the fingerprint information to be identified in the source code file fingerprint information and risk type label mapping table, and obtain the risk information corresponding to the website framework identification features based on the comparison result.
[0214] In some embodiments of this application, based on the aforementioned scheme, the site risk information determination unit 1420 is configured to: determine the target risk type label and the risk score corresponding to the target risk type label for each type of risk feature according to the risk information corresponding to each type of risk feature in the multiple risk features; and perform fusion processing on the target risk type label and the risk score corresponding to the target risk type label for each type of risk feature to obtain the risk information of the site to be identified.
[0215] In some embodiments of this application, based on the foregoing scheme, after determining the risk information of the site to be identified, the site risk information determination unit 1420 is further configured to: determine in the cloud the risk information of sites served by other resources associated with the user on which the site to be identified depends; and integrate the risk information of the sites served by other resources and the risk information of the site to be identified to obtain the risk information of the user.
[0216] In some embodiments of this application, based on the foregoing scheme, after determining the risk information of the site to be identified, the site risk information determination unit 1420 is further configured to: compare the preset risk comparison information corresponding to at least one type of risk feature with the corresponding information of other sites, and take the site that successfully matches as the target site; compare the preset risk comparison information corresponding to other types of risk features with the corresponding information of the target site to determine the risk information of the target site.
[0217] In some embodiments of this application, based on the foregoing scheme, after determining the risk information of the site to be identified, the site risk information determination unit 1420 is further configured to: obtain the IP address registered in the cloud for creating the site to be identified; determine other IP addresses registered in the same time period as the IP address; and determine the risk information of the sites corresponding to the other IP addresses.
[0218] Figure 15 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown.
[0219] It should be noted that, Figure 15 The computer system 1500 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.
[0220] like Figure 15 As shown, the computer system 1500 includes a Central Processing Unit (CPU) 1501, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 1502 or programs loaded from storage portion 1508 into Random Access Memory (RAM) 1503, such as performing the methods described in the above embodiments. Various programs and data required for system operation are also stored in RAM 1503. The CPU 1501, ROM 1502, and RAM 1503 are interconnected via bus 1504. An Input / Output (I / O) interface 1505 is also connected to bus 1504.
[0221] The following components are connected to I / O interface 1505: an input section 1506 including a keyboard, mouse, etc.; an output section 1507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 1508 including a hard disk, etc.; and a communication section 1509 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 1509 performs communication processing via a network such as the Internet. A drive 1510 is also connected to I / O interface 1505 as needed. Removable media 1511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 1510 as needed so that computer programs read from them can be installed into storage section 1508 as needed.
[0222] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 1509, and / or installed from removable medium 1511. When the computer program is executed by central processing unit (CPU) 1501, it performs various functions defined in the system of this application.
[0223] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.
[0224] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0225] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.
[0226] In one aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to perform the methods described in the above embodiments.
[0227] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.
[0228] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the method according to the embodiments of this application.
[0229] It is understood that in the specific implementation of this application, data related to site risk identification is involved. When the above embodiments of this application are applied to specific products or technologies, user permission or consent is required, and the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions.
[0230] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.
[0231] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.
Claims
1. A site risk identification method characterized by, The method includes: Multiple risk features are identified for risk identification, including interface detection and identification features and at least one of the following: image identification features and website framework identification features; the feature information corresponding to the image identification features is at least one fingerprint information of image resources in the site to be identified, the feature information corresponding to the website framework identification features is the fingerprint information of source code files in the site to be identified, and the feature information corresponding to the interface detection and identification features is at least a portion of the return results of the interface in the site to be identified. Obtain at least a portion of the returned results from the API of a known risky site; Based on the return results of the interface, a mapping table between the return results and risk type labels is constructed, and the return results and risk type label mapping table are used as the preset risk comparison information corresponding to the detection and identification features of the interface; For the site to be identified, risk information corresponding to each of the multiple risk features is obtained. The risk information includes at least one risk type label and a risk score corresponding to the risk type label. The risk information corresponding to each type of risk feature is obtained by comparing the feature information corresponding to each type of risk feature with preset risk comparison information corresponding to each type of risk feature. Obtaining the risk information corresponding to each of the multiple risk features includes: obtaining at least a portion of the return results of the interface of the site to be identified; determining the comparison results of at least a portion of the return results of each interface in the return results and risk type label mapping table, and obtaining the risk information corresponding to the interface detection and identification features based on the comparison results. The site to be identified is a network site built using cloud resources provided by a cloud platform in the cloud. Determining the comparison results of at least a portion of the return results of each interface in the return results and risk type label mapping table, and obtaining the risk information corresponding to the interface detection and identification features based on the comparison results, includes: matching at least a portion of the return results with the content of the risk type label mapping table, determining the risk score corresponding to the risk type label based on the number of successful matches corresponding to the risk type label; and outputting the matching results and risk type. Based on the risk information corresponding to each of the multiple risk features, the risk information of the site to be identified is determined. The risk information is used to identify the risky website, which is a website that provides harmful information or engages in harmful behavior. Obtain the IP address registered in the cloud for creating the site to be identified; Identify other IP addresses that registered within the same time period as the stated IP address; Determine the risk information of the sites corresponding to the other IP addresses; In the cloud, identify risk information about sites served by other resources associated with users dependent on the resources relied upon by the site to be identified; The risk information of the sites served by other resources and the risk information of the sites to be identified are integrated to obtain the risk information of the user, and the risk information of the user is used as the basic user model; Based on the user resource usage, the basic information of the subject, and the basic user model, a complete user model is generated. The complete user model includes risk score, risk type index, and user model data in at least one of the following dimensions: user geographic characteristics, resource geographic distribution characteristics, resource product distribution characteristics, user resource usage behavior characteristics, user subject industry characteristics, and user subject size characteristics. Top-risk users are identified based on risk scores from various well-developed user models. Based on the risk type indicators in each complete user model, determine the governance measures for risky users, and govern the corresponding risky users in accordance with the governance measures.
2. The site risk identification method of claim 1, wherein, The multiple risk features also include keyword features, and the preset risk comparison information corresponding to the keyword features is a risk keyword and risk type label mapping table. Obtaining the risk information corresponding to each type of risk feature in the multiple risk features includes: Extract at least one keyword from the site to be identified; Determine the comparison result of the at least one keyword in the risk keyword and risk type label mapping table, and obtain the risk information corresponding to the keyword feature based on the comparison result.
3. The method of claim 2, wherein, Before determining the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features, the method further includes: Obtain the domain name information of the site to be identified; By using a pre-established domain name risk assessment model, the risk level of the domain name information is analyzed to obtain a risk score corresponding to the domain name risk characteristics. The step of determining the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features includes: The risk information corresponding to each of the multiple risk features and the risk score corresponding to the domain name risk feature are fused together to obtain the risk information of the site to be identified.
4. The site risk identification method according to claim 1, characterized in that, Before obtaining the risk information corresponding to each of the multiple risk characteristics, the method further includes: Obtain image resources from known risky websites; Generate at least one fingerprint information, namely file fingerprint and image fingerprint, for the image resource. The file fingerprint is fingerprint information obtained by performing a hash operation on the image resource, and the image fingerprint is fingerprint information used for similarity comparison. A mapping table of map fingerprints and risk type labels is constructed based on the at least one fingerprint information, and the mapping table of map fingerprints and risk type labels is used as the preset risk comparison information corresponding to the map identification features.
5. The site risk identification method according to claim 1, characterized in that, Before obtaining the risk information corresponding to each of the multiple risk characteristics, the method further includes: Obtain the source code files of known risky websites; Perform a hash operation on each of the aforementioned source code files to obtain the fingerprint information of each source code file; Remove whitelist information from all fingerprint information to obtain the fingerprint information of the risky source code file; Based on the risk source code file fingerprint information, a source code file fingerprint information and risk type label mapping table are constructed, and the source code file fingerprint information and risk type label mapping table are used as preset risk comparison information corresponding to the website framework identification features; The step of obtaining risk information corresponding to each of the multiple risk features includes: Obtain the source code file to be identified for the site to be identified; Perform a hash operation on each of the source code files to be identified to obtain the fingerprint information of each source code file to be identified. The comparison result of the fingerprint information to be identified in the source code file fingerprint information and risk type label mapping table is determined, and the risk information corresponding to the website framework identification feature is obtained based on the comparison result.
6. The site risk identification method according to claim 1, characterized in that, The step of determining the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features includes: Based on the risk information corresponding to each of the multiple risk features, determine the target risk type label for each type of risk feature and the risk score corresponding to the target risk type label; The risk information of the site to be identified is obtained by fusing the target risk type labels of various risk characteristics and the risk scores corresponding to the target risk type labels.
7. The site risk identification method according to any one of claims 1-6, characterized in that, After determining the risk information of the site to be identified, the method further includes: The preset risk comparison information corresponding to at least one type of risk characteristic is compared with the corresponding information of other sites, and the sites that are successfully compared are taken as target sites. The preset risk comparison information corresponding to other types of risk characteristics is compared with the corresponding information of the target site to determine the risk information of the target site.
8. A site risk identification device, characterized in that, The device includes: A risk feature determination unit is used to determine multiple risk features for risk identification. The multiple risk features include interface detection and identification features and at least one of the following: image type identification features and website framework identification features. The feature information corresponding to the image type identification feature is at least one fingerprint information of image resources in the site to be identified. The feature information corresponding to the website framework identification feature is the fingerprint information of source code files in the site to be identified. The feature information corresponding to the interface detection and identification feature is at least a portion of the return results of the interface in the site to be identified. The device is also used to: obtain at least a portion of the return results of the interface of a known risk site; construct a mapping table of return results and risk type labels based on the return results of the interface, and use the return results and risk type label mapping table as preset risk comparison information corresponding to the interface detection and identification features; A feature risk information acquisition unit is used to acquire risk information corresponding to each of the multiple risk features for a site to be identified. The risk information includes at least one risk type label and a risk score corresponding to the risk type label. The risk information corresponding to each type of risk feature is obtained by comparing the feature information corresponding to each type of risk feature with preset risk comparison information corresponding to each type of risk feature. Acquiring the risk information corresponding to each of the multiple risk features includes: acquiring at least a portion of the return results of the interface of the site to be identified; determining the comparison results of at least a portion of the return results of each interface in the return results and risk type label mapping table, and obtaining the risk information corresponding to the interface detection and identification features based on the comparison results; the site to be identified is a network site built on the cloud using cloud resources provided by a cloud platform; the feature risk information acquisition unit is configured to: match at least a portion of the return results with the content of the risk type label mapping table, determine the risk score corresponding to the risk type label based on the number of successful matches corresponding to the risk type label; and output the matching results and risk type. The site risk information determination unit is used to determine the risk information of the site to be identified based on the risk information corresponding to each of the multiple risk features. The risk information is used to determine the risky website, which is a website that provides harmful information or engages in harmful behavior. After determining the risk information of the site to be identified, the site risk information determination unit is further configured to: obtain the IP address registered in the cloud for creating the site to be identified; determine other IP addresses registered within the same time period as the IP address; determine the risk information of the sites corresponding to the other IP addresses; determine the risk information of sites served by other resources associated with the user on which the site to be identified depends; and integrate the risk information of the sites served by other resources and the risk information of the site to be identified to obtain the risk information of the user. The device is further configured to: use the user's risk information as a basic user model; generate a complete user model based on the user's resource usage, the subject's basic information, and the basic user model, wherein the complete user model includes risk scores, risk type indicators, and user model data in at least one of the following dimensions: user geographic characteristics, resource geographic distribution characteristics, resource product distribution characteristics, user resource usage behavior characteristics, user subject industry characteristics, and user subject size characteristics; identify top-risk users based on the risk scores in each complete user model; determine governance measures for risky users based on the risk type indicators in each complete user model, and govern the corresponding risky users according to the governance measures.
9. A computer-readable medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the site risk identification method as described in any one of claims 1 to 7.
10. An electronic device, characterized in that, include: One or more processors; A storage device for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the site risk identification method as described in any one of claims 1 to 7.
11. A computer program product, characterized in that, The computer program product includes computer instructions stored in a computer-readable storage medium, a processor of a computer device reading the computer instructions from the computer-readable storage medium, and the processor executing the computer instructions to cause the computer device to perform the site risk identification method as described in any one of claims 1 to 7.
Citation Information
Patent Citations
Phishing website recognizing method and phishing website recognizing system
CN104615760A
Webpage recognition method and device, and server
CN107204960A
Interface data exposure detection method based on dynamic tag
CN115618085A
