Equipment model identification method and equipment

Abstract

The invention discloses an equipment model identification method and equipment, and relates to the technical field of data processing, and the method comprises the steps: obtaining a plurality of underlying feature information sets of to-be-identified equipment according to a plurality of equipment management protocols, the underlying feature information contains multi-dimensional data, even if the data is incomplete, model related features can still be extracted, and the model related features can be extracted; more accurate and stable model identification is realized; by combining the feature database, after the bottom-layer feature information is matched to the device set class, the closest first device feature information set is screened out, so that the comparison range is reduced, and the data comparison efficiency is improved; according to the multiple equipment management protocols and the multiple first identification models, comprehensive model identification is realized through data complementation of the multiple equipment management protocols, and the identification accuracy is improved. By combining multi-dimensional data extraction and a set class matching mechanism, the problem of low equipment model identification accuracy is solved, and the technical effect of improving the identification accuracy is achieved.

Description

Technical Field

[0001] This application relates to the technical field of data processing, and particularly to a method and device for identifying device models. Background Art

[0002] With the rapid development of cloud computing and big data, the scale of data centers has been continuously expanding, the number of servers has been increasing, and the difficulty of device management and operation and maintenance has also increased accordingly. During the process of device management and operation and maintenance, device model identification can help operation and maintenance personnel perform correct maintenance, upgrade, etc. operations for different model devices, and avoid system failures caused by incorrect model identification. At the same time, model identification can also detect abnormal configurations and improve the security and management efficiency of the data center.

[0003] In related technologies, the model identification method mainly determines the model of the current device by extracting a specific field information (such as ProductName field information). However, fields such as Product Name are easily tampered with, resulting in inaccurate information. Moreover, the model identification in related technologies is only based on a single data source. Once the field information is abnormal or the device fails, it is easy to cause model identification failure, thereby affecting the accuracy and stability of device management. Based on this, there is an urgent need for a device model identification method to solve the technical problem of low accuracy of device model identification in related technologies. Summary of the Invention

[0004] This application provides a method and device for identifying device models to at least solve the problem of low accuracy of device model identification in related technologies.

[0005] This application provides a method for identifying device models, including:

[0006] Obtaining multiple sets of underlying feature information of the device to be identified according to multiple device management protocols of the device to be identified; wherein, the set of underlying feature information includes at least one hardware feature information and / or at least one firmware feature information;

[0007] Obtaining the current set of underlying feature information corresponding to the current device management protocol according to the multiple sets of underlying feature information, and obtaining the current feature database corresponding to the current device management protocol based on the multiple feature databases that have been constructed; wherein, the multiple device management protocols include the current device management protocol;

[0008] Determining the device set class that matches the current set of underlying feature information according to the multiple device set classes in the current feature database, screening and obtaining the first set of device feature information that matches the current set of underlying feature information from the multiple sets of device feature information corresponding to the device set class, and marking the device model corresponding to the first set of device feature information as the first identification model of the device to be identified;

[0009] Obtain the final identification model of the device to be identified according to multiple device management protocols and multiple first identification models through a preset model screening rule.

[0010] This application also provides an identification device, including:

[0011] A data acquisition module, configured to obtain multiple underlying feature information sets of the device to be identified according to multiple device management protocols of the device to be identified; wherein, the underlying feature information set includes at least one hardware feature information and / or at least one firmware feature information;

[0012] A current information acquisition module, configured to obtain a current underlying feature information set corresponding to the current device management protocol according to multiple underlying feature information sets, and obtain a current feature database corresponding to the current device management protocol based on multiple feature databases that have been constructed; wherein, the multiple device management protocols include the current device management protocol;

[0013] A screening and matching module, configured to determine a device set class that matches the current underlying feature information set according to multiple device set classes in the current feature database, and screen and obtain a first device feature information set that matches the current underlying feature information set from multiple device feature information sets corresponding to the device set class, and mark the device model corresponding to the first device feature information set as the first identification model of the device to be identified;

[0014] An identification completion module, configured to obtain the final identification model of the device to be identified according to multiple device management protocols and multiple first identification models through a preset model screening rule.

[0015] This application also provides an electronic device, including: a memory for storing a computer program; a processor for implementing the steps of any of the above identification methods when executing the computer program.

[0016] This application also provides a computer-readable storage medium, in which a computer program is stored, and wherein the computer program implements the steps of any of the above identification methods when executed by a processor.

[0017] This application also provides a computer program product, including a computer program, and the computer program implements the steps of any of the above identification methods when executed by a processor.

[0018] This application obtains the underlying feature information set of the device to be identified according to the device management protocol, which includes various types of hardware information and / or firmware information. Since the underlying feature information covers multi-dimensional data, even if a certain field is abnormal or missing, other feature information can still be used as an effective reference. Therefore, even in the case of incomplete or abnormal data, the features related to the model can still be accurately extracted to achieve more stable model identification. By combining the constructed feature database, after the underlying feature information set of the device to be identified matches the device set class, the closest first device feature information set can be quickly screened out, which effectively reduces the comparison range and improves the data comparison efficiency. Through multiple device management protocols, the respective advantages of different device management protocols can be fully utilized to analyze the feature information of the device to be identified in multiple protocol environments to the greatest extent. This not only enriches the source of feature data but also ensures that in the case of abnormal, missing, or failed acquisition of some protocol data, other protocol data can be relied on for supplementation and correction to achieve more stable and comprehensive model identification. Then, through the preset model screening rules, the multiple first identification models obtained under each protocol are further comprehensively screened to ensure that the final identification model that most conforms to the characteristics of the device to be identified is screened out from multiple candidate results. In summary, by combining the multi-dimensional data extraction of the underlying feature information and the set class matching mechanism of the feature database, this application can solve the technical problem of low accuracy in current device model identification and achieve the technical effects of improving the comprehensiveness, matching accuracy, and stability of data in the model identification process. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] To more clearly illustrate the embodiments of the present application, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0020] Figure 1 It is a schematic hardware architecture diagram of the application device identification method provided by the present application;

[0021] Figure 2 It is a schematic flowchart of the device model identification method provided by the embodiments of the present application;

[0022] Figure 3 It is a schematic flowchart of the method for device model verification provided by the embodiments of the present application;

[0023] Figure 4 It is a schematic flowchart of the method for obtaining the underlying feature information set of the device to be identified provided by the embodiments of the present application;

[0024] Figure 5 It is a schematic structural diagram of the device model identification device provided by the embodiments of the present application;

[0025] Figure 6 It is a schematic structural diagram of the electronic device provided for this application. Specific implementation manners

[0026] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the protection scope of this application.

[0027] It should be noted that in the description of this application, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusions, so that a process, method, article or device including a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such a process, method, article or device. The terms "first", "second", etc. in this application are used to distinguish similar objects, rather than to describe a specific order or sequence.

[0028] The inventive concept of this application lies in providing a device model identification method, aiming to improve the accuracy of model identification and meet the diverse and complex requirements in device management. Aiming at the problems in the related technology, such as strong dependence on a single data field and easy failure of model identification when data is abnormal, this application adopts a multi-dimensional data acquisition, feature matching and result screening mechanism to form a complete model identification process. Specifically, first, through the current device management protocol, the underlying feature information set of the device to be identified is obtained. The underlying feature information set includes the hardware information and / or firmware information of the device, ensuring the comprehensiveness and integrity of the information. Subsequently, a feature database matching mechanism is set up. By matching the underlying feature information set of the device to be identified with the device set class in the feature database, the device set class closest to the device to be identified can be quickly determined. Further, in this device set class, the first device feature information set closest to the current underlying feature information set is selected, and the model corresponding to this device feature information set is marked as the first identification model of the device to be identified. Through the classification and matching mechanism of the device set class, the low efficiency caused by matching each device one by one in the traditional method is avoided, and the accuracy and matching speed of model identification are improved. Then, to enhance the stability and adaptability of model identification, through multiple device management protocols, the underlying feature information sets of the device to be identified under different protocols are respectively obtained, and combined with the feature databases corresponding to each protocol, the acquisition of multiple first identification models is realized. Since the information obtained by different device management protocols is different, this mechanism effectively reduces the identification deviation that may be caused by abnormal data of a single protocol, and improves the stability of device model identification in a multi-protocol environment.

[0029] In order to enable those skilled in the art of this technology to better understand the solution of this application, the following further elaborates on this application in combination with the accompanying drawings and specific implementation manners.

[0030] Combined with the application environment architecture or hardware architecture on which the execution of the device identification method depends, the application environment architecture or hardware architecture is described herein. Refer to Figure 1 , Figure 1 is a schematic diagram of the hardware architecture of the application device identification method provided by this application. Combined with the shown architecture, this environment includes a terminal device 101, a control device 102, a server 103 and a computer cluster 104, forming a complete device identification and data processing link.

[0031] In this architecture, the terminal device 101 serves as the user interaction entry, responsible for receiving user input instructions, query operations, and result display. The terminal device can be a computer, notebook, workstation, etc. with a graphical user interface (GUI), supporting remote access and control of the system by operation and maintenance personnel. The control device 102 serves as the control node, undertaking functions such as device management, data scheduling, and task instruction issuance. The control device establishes communication connections with the terminal device and other hardware devices, and is responsible for obtaining the characteristic data of each device according to the device management protocol, and completing model identification in combination with the characteristic database. For example, the control device generally uses a high-performance computer or server to ensure sufficient computing power and response speed when processing large-scale computer cluster data.

[0032] The server 103 serves as the device to be identified, and its device model needs to be confirmed through the model identification method. Since servers usually have complex hardware configurations and diverse firmware versions, relying solely on a single field of information (such as ProductName) may lead to misjudgment. Therefore, as the device to be identified, the server improves the accuracy of model identification through methods such as multi-protocol data fusion and feature matching. The hardware characteristic information, firmware characteristic information, and identified model information of the server 103 will be used as input parameters for model judgment to ensure the reliability of the identification result. The computer cluster 104 is the information source for constructing the characteristic database. The devices in the computer cluster include various devices with known models, and these devices cover various combinations of different manufacturers, hardware configurations, firmware versions, etc. By extracting, standardizing, and classifying the underlying characteristic information of the devices in the computer cluster, a characteristic database is formed, providing a comparison basis for model matching of the device to be identified (such as the server 103). In the figure, the computer cluster is further divided into the first type of device 105 and the second type of device 106, representing the device set classes refined according to the hardware and firmware characteristic information. This classification structure helps to quickly screen out the device set closest to the characteristics of the device to be identified, improving the model identification efficiency.

[0033] Figure 2 It is a schematic flowchart of the device model identification method provided by the embodiment of the present application. As Figure 2 shown, it includes:

[0034] S21, according to multiple device management protocols of the device to be identified, obtain multiple underlying characteristic information sets of the device to be identified; wherein, the underlying characteristic information set includes at least one hardware characteristic information and / or at least one firmware characteristic information.

[0035] In this embodiment, multiple underlying feature information sets of the device to be identified are obtained according to multiple device management protocols of the device to be identified. A device management protocol is a communication protocol for remote management and monitoring, which is responsible for providing a transmission mechanism for device status information, configuration information, and management instructions. Since the device to be identified may support multiple protocols, during the model identification process, it is necessary to combine multiple device management protocols to comprehensively obtain the underlying feature information of the device to be identified to ensure the integrity and accuracy of the information. When obtaining the underlying feature information set, the underlying feature information set includes at least one hardware feature information and / or at least one firmware feature information. Underlying feature information refers to the basic attribute information of the device, which can truly reflect the hardware configuration and firmware status of the device, and has stability and strong uniqueness. The underlying feature information can be further divided into hardware feature information and firmware feature information. Hardware feature information involves the hardware configuration features of the device, has strong stability and the characteristic of not being easily changed, and is one of the important bases for model identification. Firmware feature information refers to the information reflecting the state of the embedded software inside the device, and its data is usually related to the model, version, and upgrade status of the device. Since there are differences in the recording of hardware and firmware feature data by each device management protocol, during the data acquisition process, multiple device management protocols are fully utilized to make up for the missing or abnormal data of a single protocol. For example, different protocols may have inconsistent field names, some fields missing, or different recording methods, etc. Through multi-protocol integration, the integrity of the underlying feature information set can be effectively improved, ensuring that the feature data is comprehensively covered, and thus providing more stable and reliable data support for model identification.

[0036] S22. According to the multiple underlying feature information sets, obtain the current underlying feature information set corresponding to the current device management protocol, and based on the multiple feature databases that have been constructed, obtain the current feature database corresponding to the current device management protocol; wherein, the multiple device management protocols include the current device management protocol.

[0037] In this embodiment, the device to be identified supports multiple device management protocols, and there are differences in the underlying feature information data structures, field types, and naming rules under different protocols. Therefore, when obtaining the current underlying feature information set from multiple underlying feature information sets, subsequent matching steps are performed on the feature data corresponding to the current device management protocol. This ensures the consistency of data matching, avoids confusion caused by the mixing of multi-protocol data, and guarantees the accuracy of model identification. After obtaining the current underlying feature information set, to ensure that this information can effectively participate in model identification, it needs to be associated with the constructed feature database. The feature database is a data set aggregated based on the feature information of multiple device models and is divided according to the types of device management protocols, that is, one device management protocol corresponds to one feature database. The establishment of the feature database helps to quickly retrieve and match data during the device model identification process, improves data processing efficiency, and at the same time ensures that the content of the selected database has strong pertinence and accuracy. In this embodiment, by combining the current device management protocol to select the feature database, it is ensured that the underlying feature information of the device to be identified can be effectively utilized during the matching process, providing guarantee for the accuracy and stability of model identification.

[0038] S23. According to multiple device set classes in the current feature database, determine the device set class that the current underlying feature information set matches, and from the multiple device feature information sets corresponding to the device set class, screen and obtain the first device feature information set that matches the current underlying feature information set, and mark the device model corresponding to the first device feature information set as the first identification model of the device to be identified.

[0039] In this embodiment, the device set class is a classification structure in the feature database, which divides devices into multiple set classes according to the feature similarity of known devices. Each device set class contains device information with similar hardware configurations or firmware features, and constructs set class features by refining the common features of these devices to ensure that the feature data of each set class can have stability and representativeness within a certain range. The construction process of the device set class is based on the comprehensive analysis of device feature information to form a set class structure with strong directivity and discrimination. This device set class based on feature aggregation improves the organization efficiency of the feature database.

[0040] After determining that the current underlying feature information set of the device to be recognized matches the device set class in the current feature database, it is necessary to further screen and obtain the first device feature information set that matches the current underlying feature information set from the multiple device feature information sets included in the device set class. The device feature information set refers to a data set that reflects the hardware and / or firmware feature information of a specific device model, usually stored in the form of fields, parameters, and their corresponding values. The process of screening the first device feature information set requires comparing the underlying feature information set of the device to be recognized with the key features of each device feature information set in the device set class one by one to ensure that the screened first device feature information set has a high degree of matching in terms of parameter features, feature value ranges, and data integrity. In this way, it is possible to accurately locate the device information that is closest to the features of the device to be recognized among multiple possible device models, ensuring the accuracy of model recognition.

[0041] After completing the screening of the first device feature information set, it is also necessary to mark the device model corresponding to the first device feature information set as the first recognition model of the device to be recognized. This marking process records the preliminary recognition result of the device to be recognized under the current device management protocol. In summary, this embodiment realizes the matching of the feature information of the device to be recognized by combining the screening mechanism of the device set class, improves the accuracy of device model recognition, and reduces the instability brought by the traditional single data field recognition method.

[0042] S24. According to multiple device management protocols and multiple first recognition models, obtain the final recognition model of the device to be recognized through a preset model screening rule.

[0043] In this embodiment, a multi-protocol traversal mechanism is adopted to obtain multiple first recognition models. Regarding the settings of the multi-protocol traversal mechanism in terms of traversal order, data priority, etc., the staff can make flexible settings according to specific requirements, and this application does not limit it here. In the specific traversal process, based on the multi-protocol traversal mechanism, the underlying feature information set and the feature database under each device management protocol are retrieved one by one according to the preset protocol order. By traversing all the preset device management protocols, it is possible to ensure a comprehensive analysis of various protocol data and finally obtain multiple first recognition models corresponding to the device management protocols. Due to differences in data acquisition ranges, information integrity, field stability, etc. among different device management protocols, even in the same environment, the first recognition models obtained under different protocols may have certain differences. The core of the model screening rule lies in establishing a screening mechanism among multiple protocols and their recognition results to ensure that the final model of the device to be recognized can still be accurately judged in the case of data deviation or inconsistency. This screening rule has a certain degree of flexibility and adaptability in strategy settings and can be dynamically adjusted according to factors such as device type, data characteristics, and protocol weights to cope with data inconsistency problems in complex environments.

[0044] Exemplarily, in the case of data inconsistency, the model screening rule can quantitatively analyze the reliability of each first identified model according to mechanisms such as weight calculation, similarity matching, and feature data cross-verification. The weight calculation can be dynamically assigned based on factors such as the stability of each protocol, the uniqueness of data features, and the coverage of device features in the identification process. For example, for some protocols with higher stability, more precise data structures, or more complete device features, the corresponding first identified model can have a higher priority during screening. Further, in the case of large deviations between multi-protocol data, the cross-verification mechanism of feature data can be further utilized to analyze key information fields, feature parameters, etc. under each protocol to ensure the accuracy of the final identified model. In this embodiment, by combining the first identified models of multiple device management protocols and using the preset model screening rule, the stability and reliability of the final identification result are ensured. The comprehensive processing of complex data in a multi-protocol environment is realized, further improving the comprehensiveness, compatibility, and anti-interference ability of the model identification method, and ensuring that accurate and stable model identification results can still be output in a diverse device environment.

[0045] In this application, by obtaining the underlying feature information set of the device to be identified according to the current device management protocol of the device, since the underlying feature information covers multi-dimensional data, even if a certain field is abnormal or missing, other feature information can still be used as an effective reference. Therefore, in the case of incomplete or abnormal data, the features related to the model can still be accurately extracted to achieve more stable model identification. By combining the constructed feature database, after the underlying feature information of the device to be identified matches the device set class, the closest first device feature information set can be quickly screened out, which effectively reduces the comparison range and improves the data comparison efficiency. Through multiple device management protocols, the respective advantages of different device management protocols can be fully utilized to analyze the feature information of the device to be identified in multiple protocol environments to the greatest extent. This not only enriches the source of feature data but also ensures that in the case of abnormal, missing, or failed acquisition of some protocol data, other protocol data can still be relied on for supplementation and correction to achieve more stable and comprehensive model identification. Then, through the preset model screening rule, the multiple first identified models obtained under each protocol are further comprehensively screened to ensure that the final identified model that best matches the features of the device to be identified is selected from multiple candidate results.

[0046] In one embodiment, Figure 3 It is a schematic flowchart of the method for verifying the device model provided by the embodiment of this application. It is the verification of the device number added after marking the device model corresponding to the first device feature information set as the first identified model of the device to be identified in step S23 above. On the basis of the above embodiment, as Figure 3As shown, the recognition method further includes:

[0047] S31. According to the current device management protocol, obtain and record the identification model information of the device to be recognized. When the first recognition model is inconsistent with the identification model information, obtain the second device feature information set corresponding to the identification model information from the current feature database.

[0048] S32. According to the second device feature information set and the current underlying feature information set, use a preset similarity algorithm to obtain the similarity between the second device feature information set and the current underlying feature information set.

[0049] S33. If the similarity between the second device feature information set and the current underlying feature information set is greater than or equal to the first similarity threshold, mark the device model corresponding to the identification model information as the second recognition model, and update the device model of the device to be recognized to include the first recognition model and the second recognition model.

[0050] S34. According to multiple device management protocols, multiple first recognition models, and at least one second recognition model, obtain the final recognition model of the device to be recognized through a preset model screening rule.

[0051] In this embodiment, to further improve the accuracy and reliability of the recognition result, it is necessary to verify and calibrate the recognition result on the basis of the preliminary recognition. First, according to the current device management protocol of the device to be recognized, obtain and record the identification model information of the device to be recognized by reading the preset field. The identification model information is the iconic information about the device model carried by the device itself, usually stored in a specific data area of the device. This field may include the model information inherent when the device leaves the factory, or the model data generated during the installation, maintenance, or configuration of the device.

[0052] After obtaining the identification model information, to further judge the reliability of this information, it is necessary to compare it with the already recognized first recognition model. The first recognition model is the preliminary recognition result obtained based on feature matching, while the identification model information is the static identification obtained by reading the device's own field, and the two may be inconsistent. When the first recognition model is inconsistent with the identification model information, it is necessary to further analyze the authenticity and credibility of the identification model information. For this purpose, obtain the second device feature information set corresponding to the identification model information from the current feature database as a reference basis for verifying the identification model information. The second device feature information set is refined from the feature data of known devices in the current feature database and has strong stability and integrity. This feature information set includes hardware feature information and / or firmware feature information, which can comprehensively reflect the device features matching the identification model information and provide a data basis for subsequent comparison and screening.

[0053] According to the second device feature information set and the current underlying feature information set of the device to be recognized, a preset similarity algorithm is used to calculate the feature similarity between the two. Exemplarily, the preset similarity algorithm can be based on various similarity calculation methods such as weighted Jaccard similarity, cosine similarity, and Euclidean distance. By comparing multi-dimensional feature data such as device hardware parameters, firmware versions, and memory specifications, the similarity degree between device models can be judged more accurately. The calculation result of the similarity will be used as a parameter for judging device model calibration to ensure reliable matching results in various situations. After the similarity calculation, it is necessary to judge the calculation result according to a preset similarity threshold. When the similarity between the second device feature information set and the current underlying feature information set is greater than or equal to the first similarity threshold, it can be determined that the identification model information has high reliability. Therefore, the device model corresponding to the identification model information is recorded as the second identification model. At the same time, to ensure the integrity of the device model information of the device to be recognized, it is necessary to update the device model of the device to be recognized to include the first identification model and the second identification model. This update mechanism aims to perform multi-dimensional data integration on the recognition results to ensure the comprehensiveness and accuracy of the model recognition results and provide more sufficient data support for subsequent screening and confirmation.

[0054] After completing the update of the first identification model and the second identification model, to ensure the stability of the final recognition result, it is necessary to further adopt a multi-protocol traversal mechanism to traverse the underlying feature information set and the feature database under multiple device management protocols to obtain multiple first identification models and second identification models corresponding to the device management protocols. The multi-protocol traversal mechanism can effectively improve the compatibility and adaptability of the device recognition method, ensure that the data characteristics of various protocols can still be fully utilized in a complex device environment, and improve the coverage of model recognition. After obtaining multiple first identification models and second identification models, it is necessary to analyze and screen these recognition results according to a preset model screening rule. The model screening rule is to comprehensively consider factors such as the credibility, data integrity, and protocol stability of each recognition result to screen out the final recognition model that most conforms to the characteristics of the device to be recognized.

[0055] This embodiment ensures that the model information of the device to be recognized can still be accurately judged when there are abnormalities in the device features, the model information is tampered with, or the device attributes change by combining the acquisition of identification model information, similarity calculation, data update, and screening mechanisms. This method has strong robustness and self-adaptability, can effectively reduce the occurrence of recognition errors caused by incomplete device information or abnormal data interference, and further improves the accuracy and stability of device model recognition.

[0056] Further, in another embodiment, this is a supplement to another situation in the above embodiment. On the basis of the above embodiment, it further includes:

[0057] S37. If the similarity between the second device feature information set and the current underlying feature information set is less than the first similarity threshold, update the device model corresponding to the first device feature information set to the first identification model.

[0058] S38. According to multiple device management protocols and multiple first identification models, obtain the final identification model of the device to be identified through a preset model screening rule.

[0059] In this embodiment, during the device model identification process, there is also a situation where the model information verification fails to be further identified. After calculating the similarity between the second device feature information set and the current underlying feature information set, it is necessary to determine the calculation result. If the similarity calculation result is less than the first similarity threshold, it indicates that there is a large difference between the second device feature information set and the current underlying feature information set of the device to be identified, and the reliability of the identification model information cannot be confirmed. At this time, the original first identification model should be used as the current optimal identification result. To ensure data consistency and stability, the device model corresponding to the first device feature information set should be updated to the first identification model. Similarly, after confirming the first identification model, to further improve the comprehensiveness and accuracy of the identification, it is necessary to traverse the underlying feature information sets and feature databases under multiple device management protocols to obtain more identification results. During the traversal process, by combining the feature information obtained from each device management protocol, screen and obtain the first identification model that matches the underlying feature information set of the device to be identified. For example, multiple first identification models obtained through the traversal mechanism. After completing the data traversal of the first identification model, to further optimize the identification result and improve its reliability, it is also necessary to screen multiple first identification models according to the preset model screening rule, and then obtain the final identification model. This process will be described in subsequent embodiments and will not be elaborated here.

[0060] In this embodiment, when the verification of the second device feature information set fails, the first identification model is updated as the current identification result, and further, according to multiple device management protocols and model screening rules, the comprehensive optimization of the model identification result is realized in a variable and complex device environment. While improving the identification accuracy, this method has strong stability and compatibility, and can effectively handle situations such as data anomalies, model information anomalies, and device configuration inconsistencies. And, it should be further noted here that this embodiment is directed to the situation where the similarity between the second device feature information set and the current underlying feature information set is less than the first similarity threshold under each device management protocol.

[0061] In one embodiment, Figure 4This is a schematic flowchart of the method for obtaining the underlying feature information set of the device to be recognized provided by the embodiment of the present application. It is a specific description of an implementation method for obtaining multiple underlying feature information sets of the device to be recognized in step S21 above. On the basis of the above embodiment, as Figure 4 shown, it includes:

[0062] S41. According to multiple device management protocols, select any one device management protocol as the current device management protocol, and based on the current device management protocol, establish a communication connection with the device to be recognized;

[0063] S42. According to the current device management protocol, send an instruction to the device to be recognized, and receive the original feature data returned by the device to be recognized based on the instruction;

[0064] S43. According to the original feature data, parse and extract the hardware feature information and / or firmware feature information, obtain the current underlying feature information set of the device to be recognized, and according to multiple device management protocols, obtain multiple underlying feature information sets of the device to be recognized.

[0065] In this embodiment, from multiple device management protocols, select any one device management protocol as the current device management protocol, and based on the current device management protocol, establish a communication connection with the device to be recognized. The current device management protocol refers to a protocol used for remotely obtaining and managing device information, usually including but not limited to the Intelligent Platform Management Interface (IPMI) protocol, Redfish Scalable Platforms Management (Redfish) protocol, Simple Network Management Protocol (SNMP), and Secure Shell (SSH) protocol, etc. Each protocol has different communication mechanisms and instruction specifications. Therefore, before establishing a communication connection, it is necessary to select the corresponding device management protocol according to the protocol type supported by the device to be recognized to ensure the compatibility of instruction issuance and the integrity of data return.

[0066] After successfully establishing a communication connection, according to the instruction requirements of the current device management protocol, send an instruction to the device to be identified to trigger the device to return corresponding characteristic data. The content of the instruction can be written according to the requirements of different protocol types. For example, if the current device management protocol is the IPMI protocol, the ipmitool fru print instruction can be sent to obtain the device hardware information; if the SNMP protocol is adopted, the snmpget instruction can be used to read the parameter information under a specific object identifier; and if the SSH protocol is adopted, corresponding instructions can be executed through remote login to query the device hardware configuration. During the process of sending the instruction, it is necessary to ensure the integrity and correctness of the instruction parameters to avoid device response failures caused by missing parameters or incorrect command spelling. After the instruction is sent successfully, the device to be identified will return the original characteristic data related to its hardware and firmware configuration according to the preset protocol.

[0067] After receiving the original characteristic data returned by the device to be identified, it is necessary to parse this data to extract the hardware characteristic information and / or firmware characteristic information. Since there are significant differences in the data formats returned by various device management protocols, it is necessary to analyze and identify the fields, data formats, and key content of the original data in combination with the protocol specifications. For example, the hardware information returned by the IPMI protocol may be presented in the format of Field Replaceable Unit (FRU), including information on components such as the device's processor, memory, hard disk, motherboard, etc.; the SNMP protocol may map device parameter information through object identifiers; and after the SSH protocol executes a command, it may output the specific parameters of each hardware module of the device in text form. Therefore, during the data parsing process, corresponding parsing rules need to be selected according to different protocols to ensure the correct interpretation of each piece of information in the original characteristic data.

[0068] Based on the parsing of the original data, further extract the key information related to the device's hardware and firmware to form a complete set of underlying characteristic information. The hardware characteristic information covers the core hardware configuration of the device, including but not limited to processor parameters, memory parameters, storage controller parameters, expansion slot parameters, power configuration parameters, fan parameters, and motherboard parameters. Processor parameters may include information such as processor model, number of cores, main frequency, number of threads, and cache capacity, reflecting the computing performance of the device; memory parameters may include indicators such as memory specifications, capacity, number of slots, and rate, affecting the data access speed and performance stability of the device; storage controller parameters involve the storage structure and disk management method (such as RAID mode); expansion slot parameters usually refer to the type, number, and interface specifications of the expansion slots; power configuration parameters are used to describe the power input mode and power parameters of the device; fan parameters describe the device's cooling system, including the number of fans, rotation speed, and power supply information; motherboard parameters cover information such as motherboard model and interface type, reflecting the overall hardware architecture of the device.

[0069] In addition to the hardware feature information, it also includes firmware feature information. Generally, it includes firmware manufacturer information and firmware version number information. The firmware manufacturer information is used to identify the source and compatibility of the firmware, while the firmware version number information reflects the specific version of the device firmware. Extracting the firmware feature information helps to further improve the accuracy in device model identification. Especially when the hardware parameters are similar or the device information is missing, the firmware information can be used as a judgment basis. Through this embodiment, the process of obtaining the device underlying feature information based on the current device management protocol is fully realized. This process fully considers the differences in device management protocols, the flexibility of instruction issuance, and the complexity of data parsing, ensuring that in a multi-protocol and multi-device environment, the complete hardware feature information and firmware feature information of the device to be identified can be obtained, laying a foundation for subsequent model matching and screening. In practical applications, this method can effectively handle the situation of data loss or field anomalies, improving the accuracy, stability, and reliability of the model identification process.

[0070] In an embodiment of the present application, the method for determining the device set class that matches the current underlying feature information set provided by the embodiment of the present application is a specific description of an implementation manner of the above step S23. On the basis of the above embodiment, it includes:

[0071] S51, according to multiple device set classes in the current feature database, query and obtain the common feature information sets corresponding to the device set classes to obtain multiple common feature information sets of the current feature database; wherein, the common feature information set includes multiple common feature information of multiple known devices in the device set class;

[0072] S52, from multiple common feature information sets of the current feature database, query and obtain the common feature information set that matches the current underlying feature information set to obtain the device set class that matches the current underlying feature information set.

[0073] In this embodiment, in order to ensure that the data organization form in the feature database can effectively support model identification, the feature information of known devices in the feature database is classified and managed. Specifically, by analyzing the feature information of known devices, devices with relatively high feature similarity are grouped into the same device set class. The device set class is a classification structure that aggregates known devices based on feature similarity. Its purpose is to uniformly classify devices with relatively close feature parameters, similar configuration features, and common structural features, so as to quickly match the device type similar to the device to be identified during the model identification process. The construction of the device set class is usually based on the induction and collation of the hardware information, firmware information, performance characteristics, and other basic configurations of known devices, ensuring that the devices within the set class have strong consistency in overall features, thereby improving the efficiency and accuracy of data matching.

[0074] The common feature information set of the device set class is the result of feature extraction of the device set class, which contains the common feature information of multiple known devices in the set class. The common feature information refers to the feature data with strong consistency and stability among multiple devices in the set class. These features often have strong representativeness and can effectively generalize the commonalities of the devices in the set class. The process of extracting the common feature information set involves analyzing the feature data of each device in the set class and extracting the common hardware parameters, firmware information, and other stable features among the devices. By screening the common feature information, the interference of non-critical parameters to the recognition result can be effectively reduced, and the matching efficiency and accuracy can be further improved.

[0075] After obtaining multiple common feature information sets of the feature database, a matching operation needs to be performed in the feature database based on the current underlying feature information set of the device to be recognized. The matching process is to use the underlying feature information of the device to be recognized as the matching basis and compare each of the multiple common feature information sets in the feature database one by one. During the matching process, the common feature information set that most conforms to the features of the device to be recognized needs to be screened according to factors such as the feature type, the integrity of the feature field, the parameter value range, and the stability of the feature data. Since the common feature information set is derived from the common features of the device set class, it is more stable and valuable in the matching process and can effectively avoid matching deviations caused by abnormal individual fields or data missing. After completing the matching of the common feature information set, it is necessary to further determine the device set class corresponding to the common feature information set. Since each common feature information set is associated with a specific device set class, the device set class that matches the underlying feature information of the device to be recognized can be quickly located through the matching result. This matching mechanism not only improves the efficiency of model recognition but also effectively reduces the data interference caused by problems such as complex device feature information and large parameter differences.

[0076] In another specific embodiment, in the process of determining the device set class that matches the current underlying feature information set, a clustering algorithm can also be introduced to further improve the accuracy of device set class division, including:

[0077] S501, based on the feature database, use the clustering algorithm to divide and obtain multiple device set classes of the feature database;

[0078] S502, according to the multiple device set classes, obtain the device set class that matches the underlying feature information.

[0079] In this embodiment, the clustering algorithm is an unsupervised learning method that aims to divide data into several independent clusters based on the similarity between data samples. Samples within each cluster have high similarity, while samples between different clusters have significant differences. The clustering algorithm can achieve more accurate feature classification in the case of complex, diverse, and high-dimensional feature information.

[0080] In a specific implementation, first, obtain the underlying feature information sets of multiple known devices from the set of known devices. Since these feature information may have different data types (such as strings, numerical types, enumeration types, etc.), data preprocessing is performed before applying the clustering algorithm. Specifically, data standardization, feature normalization, feature encoding, etc. can be used to ensure the unity and consistency of the feature data. After completing the data preprocessing, select an appropriate clustering algorithm for clustering analysis of the feature data. Common clustering algorithms include, but are not limited to, the K-means algorithm, hierarchical clustering algorithm, etc. In the specific clustering process, similarity measurement methods (such as Euclidean distance, cosine similarity, Manhattan distance, etc.) can be used to calculate the similarity between device feature data. Taking the K-means algorithm as an example, first, preset the number of clusters k according to the data scale; then, randomly initialize k cluster centers and assign each device feature data point to the closest cluster center; then, recalculate the new cluster centers based on the sample data in each cluster and repeat this process continuously until the cluster centers no longer change significantly or reach the preset number of iterations, thus obtaining a stable clustering result.

[0081] After completing the clustering analysis, analyze each clustering result, obtain the feature information with a high frequency in each cluster, and use this feature information as the common feature information set of the cluster. If there is feature information in the underlying feature information set of the device to be recognized that highly matches a certain common feature information set, then classify the device to be recognized into the corresponding device set class. The clustering algorithm can fully explore the internal structure of the feature data. Especially in the case of high data dimension and complex device feature information, it can more accurately discover the potential patterns and rules in the data, thereby improving the accuracy of the division of the device set class.

[0082] Furthermore, in a specific embodiment, the method for obtaining the first device feature information set provided in the embodiment of the present application is a further description of the implementation manner of obtaining the first device feature information set in step S23 above. On the basis of the above embodiment, it includes:

[0083] S61, based on the device set class that matches the current underlying feature information set, obtain the device feature information sets of multiple known devices in the device set class;

[0084] S62. Obtain the first unique feature information sets of multiple known devices based on the device feature information sets of the known devices and the common feature information set of the device set class.

[0085] S63. Obtain the second unique feature information set of the device to be recognized based on the current underlying feature information set and the common feature information set of the device set class.

[0086] S64. Query and obtain the first unique feature information set that matches the second unique feature information set according to multiple first unique feature information sets, and mark the device feature information set corresponding to the first unique feature information set as the first device feature information set.

[0087] In this embodiment, after the current underlying feature information set is matched with the device set class in the current feature database, obtain the device feature information sets of multiple known devices in the device set class as the basic data source for subsequent feature matching. After obtaining the device feature information sets of multiple known devices in the device set class, it is necessary to further extract the first unique feature information set of each known device in combination with the common feature information set of the device set class. The common feature information set refers to the common characteristics among the devices in the device set class, including relatively general feature information such as hardware structure and basic parameters. The unique feature information set refers to the unique and distinguishable feature information of each device outside the common features. By removing the common feature information from the device feature information set of the known device, the obtained first unique feature information set has stronger recognition ability, which helps to accurately judge the device model. While obtaining the first unique feature information set, it is also necessary to extract the second unique feature information set of the device to be recognized based on the current underlying feature information set of the device to be recognized and in combination with the common feature information set of the matching device set class. The second unique feature information set represents the uniqueness of the device to be recognized within the device set class.

[0088] Next, perform a feature matching operation based on the first unique feature information sets of the known devices. During this process, compare the second unique feature information set of the device to be recognized with the first unique feature information sets one by one to screen out the most matching first unique feature information set. This comparison process can use string similarity, numerical difference degree or other similarity algorithms to ensure the accuracy and stability of the matching result. After the matching is completed, the device feature information set corresponding to the screened first unique feature information set can be marked as the first device feature information set. The marking result of the first device feature information set will be used as the first recognized model of the device to be recognized, providing data support for subsequent model recognition, screening and confirmation. Through this embodiment, according to the common feature information and unique feature information of the device set class, it not only avoids the interference of common features on model recognition, but also strengthens the role of unique feature information in recognition, thus effectively improving the accuracy and stability of the model recognition of the device to be recognized.

[0089] In a specific embodiment, the method for obtaining the final identification model of the device to be identified in the first case provided by the embodiments of the present application is an illustration of an implementation manner for obtaining the final identification model of the device to be identified in step S24 above. Based on the above embodiments, it includes:

[0090] S71, based on multiple underlying feature information sets of the device to be identified, according to multiple first device feature information sets, obtain the similarity score between the underlying feature information set and the first device feature information set that matches the underlying feature information set, so as to obtain the similarity score of the first identification model;

[0091] S72, according to the similarity score of the first identification model and the preset weight threshold of the device management protocol corresponding to the first identification model, calculate the weighted scores of multiple first identification models; wherein, the weighted score is the product of the similarity score and the preset weight threshold;

[0092] S73, perform normalization processing on the weighted scores of multiple first identification models to obtain the probability values of multiple first identification models;

[0093] S74, screen and obtain the first identification model with the largest probability value as the final identification model of the device to be identified.

[0094] In this embodiment, based on multiple underlying feature information sets of the device to be identified, according to multiple first device feature information sets, calculate the similarity score between each underlying feature information set and the first device feature information set that matches the underlying feature information set. Each device management protocol corresponds to an underlying feature information. After each underlying feature information is matched, there will be a corresponding first device feature information. The similarity score is an index to measure the matching degree of two data sets at the feature level, reflecting the similarity degree of the underlying feature information set of the device to be identified and each first device feature information set in terms of feature parameters, data structure, and device configuration. The choice of calculating the similarity score depends on the type of feature information, the number of feature fields, and the dimensional characteristics of the data. Exemplarily, a weighted similarity algorithm is adopted to assign different weight coefficients to different types of feature data to highlight the matching degree of key parameters and reduce the interference of non-key parameters or volatile parameters on the matching result. Parameters with stronger stability such as the processor model and motherboard information can be assigned higher weights, while some volatile or less influential fields for model identification (such as fan speed, firmware update status, etc.) can be assigned lower weights. Through the weighting mechanism, further improve the accuracy and anti-interference ability of the similarity score in a complex environment with large data differences.

[0095] After calculating the similarity score for each pair of data, it is necessary to further calculate the weighted similarity score in combination with the device management protocol corresponding to each first recognition model. Each device management protocol has different stability, data integrity, and feature specificity during the device model recognition process. Therefore, corresponding preset weight thresholds are assigned to different protocols to reflect the reliability of each protocol for model recognition and the credibility of the data. The determination of the preset weight threshold is based on the analysis of the characteristics of each protocol, including the data stability of the protocol, the data coverage, the integrity of the field information, and the application breadth of each protocol in actual device operation and maintenance. Protocols with stronger stability and more complete field information are usually assigned higher weights to enhance the credibility of the data under that protocol.

[0096] The weighted score is calculated by multiplying the similarity score by the preset weight threshold to combine the matching degree of each first recognition model with the reliability of its corresponding protocol. This calculation method can reasonably highlight the matching results of more reliable protocols in the case of certain differences in multi-protocol data, and avoid misjudgments caused by abnormal data or deviations in the results of individual protocols. The introduction of the weighted score further improves the scientific nature of data fusion, ensuring that the most valuable recognition results can still be selected in a complex environment. After calculating the weighted score, it is necessary to normalize the weighted scores of multiple first recognition models. The purpose of normalization is to convert the weighted scores of each first recognition model into standardized probability values to ensure the comparability of different protocols and different recognition results on the same scale. The normalization process maps all weighted scores to the [0, 1] interval according to their total ratio, representing the confidence of each first recognition model in the form of probability values. The normalized probability value reflects the credibility of each first recognition model. The higher the probability value, the higher the matching degree of the recognition model, and the more likely it is to be the true model of the device to be recognized. Normalization not only helps the fusion of data from different protocols but also effectively avoids the result imbalance caused by numerical deviations, further improving the stability and consistency of the recognition results. After completing the normalization process, it is necessary to select the first recognition model with the largest probability value as the final recognition model of the device to be recognized based on the probability screening mechanism.

[0097] In another embodiment, the application embodiment provides another method for obtaining the final recognition model of the device to be recognized. It is a specific description of the method for obtaining the final recognition model of the device to be recognized in step S34 above. Based on the above embodiment, it includes:

[0098] S81, based on multiple underlying feature information sets of the device to be recognized, and according to multiple first device feature information sets, obtain the similarity score between the current underlying feature information set and the first device feature information set that matches the current underlying feature information set, so as to obtain the similarity score corresponding to the first recognition model;

[0099] S82. Obtain the similarity score between the second device feature information set and the current underlying feature information set to obtain the similarity score of the second recognition model.

[0100] S83. Obtain the same device management protocol corresponding to the first recognition model and the second recognition model, and obtain the preset weight threshold of the device management protocol.

[0101] S84. Calculate and obtain the weighted score of the first recognition model and the weighted score of the second recognition model according to the similarity score corresponding to the first recognition model, the similarity score corresponding to the second recognition model, and the preset weight threshold.

[0102] S85. Normalize the weighted scores of multiple first recognition models and second recognition models to obtain the probability values of the first recognition model and the second recognition model.

[0103] S86. Based on the probability values, screen out the recognition model with the largest probability value as the final recognition model of the device to be recognized.

[0104] In this embodiment, based on multiple underlying feature information sets of the device to be recognized, compare the obtained multiple first device feature information sets, and calculate the similarity score between each underlying feature information set and the corresponding first device feature information set. When calculating the similarity, methods such as character matching, numerical comparison, or vector distance calculation can be combined to improve the accuracy of the similarity score. For example, for numerical feature data, methods such as cosine similarity, Euclidean distance, and Mahalanobis distance can be used; for string-like feature data, methods such as Levenshtein distance and Jaccard similarity can be used. By comprehensively considering different types of data features, more accurate similarity calculation can be achieved. Then, obtain the second recognition model corresponding to the second device feature information set, and calculate the similarity score between the second device feature information set and the current underlying feature information set based on the current underlying feature information set. The calculation method of the similarity score with the first recognition model is similar and will not be elaborated here.

[0105] To further ensure the rationality of the scores of each model, it is also necessary to obtain the device management protocol corresponding to each identified model, and determine its preset weight threshold based on the device management protocol. Since different device management protocols have different data integrity and feature coverage in model identification, different weight thresholds are assigned to different protocols. For example, the IPMI protocol is strong in data integrity, stability and field coverage, so it is given a higher weight; while the SSH protocol is usually given a lower weight because its data is greatly affected by the operating system environment and has poor stability. Using the similarity score and preset weight threshold obtained in the above steps, the weighted score of the first identified model and the weighted score of the second identified model are calculated respectively. Specifically, the weighted score of the first identified model can be obtained by multiplying the similarity score of the model by the preset weight threshold of the device management protocol to which it belongs, and the weighted score of the second identified model is the same. The weighting process takes into account the feature matching degree and protocol reliability of each model, so that the final score is more scientific and fair. Then, the weighted scores of the first identified model and the second identified model are normalized so that different scores can be compared and screened under the same dimension. The purpose of normalization is to map the weighted scores of each identification model to a fixed interval (such as between 0 and 1) to ensure the comparability of different score results. Finally, based on the results of normalization, the identification model with the highest probability value is selected as the final identification model of the device to be identified. If there are two identification models with the same and highest probability, the two identification models can be output at the same time for the staff to choose.

[0106] In a specific embodiment, based on the above embodiment, further description is given here on building a feature database, which specifically includes:

[0107] S2101, based on multiple known devices in the known device cluster, according to multiple device management protocols, multiple underlying information sets of the multiple known devices are obtained, and multiple underlying feature information sets of the multiple known devices are obtained through feature extraction;

[0108] S2102, performing data standardization processing on the underlying feature information sets according to the multiple underlying feature information sets to obtain multiple standard feature information sets;

[0109] S2103, reading the system event log of the known device, obtaining the number of device model modifications of the known device, and querying to obtain the modification number weight that matches the number of device model modifications;

[0110] S2104, obtaining a protocol weight corresponding to the current device management protocol, and obtaining device feature information sets of multiple known devices according to the protocol weight, the modification number weight and the standard feature information set;

[0111] S2105. Obtain the weighted weights of two device feature information sets according to the string similarity of any two device feature information sets, and use the weighted similarity algorithm to calculate the weighted similarity of the two device feature information sets. When the weighted similarity is greater than or equal to the second similarity threshold, classify the two device feature information sets into the same device set class;

[0112] S2106. Traverse multiple device feature information sets in the known device cluster to obtain multiple device set classes corresponding to the known device cluster;

[0113] S2107. Extract the common feature information of multiple device feature information sets belonging to the same device set class to form the common feature set of the device set class, and obtain the unique feature information sets of multiple known devices according to the common feature set and multiple device feature information sets;

[0114] S2108. Traverse multiple device set classes, and obtain and store multiple common feature sets and the unique feature information sets of multiple known devices into the feature database.

[0115] In this embodiment, constructing the feature database is the basis for ensuring the reliability and accuracy of the model identification result. First, based on the current device management protocol, establish a communication connection with the known device cluster, and on this basis, obtain the underlying information sets of multiple known devices in the known device cluster. The known device cluster refers to a set of devices whose models, configurations, and feature information have been clearly defined, usually including device groups of different manufacturers, models, and specifications, with strong diversity and representativeness. By combining multiple protocols, fully obtain the basic information such as the hardware configuration, firmware version, interface information, and topology structure of the known devices to form a complete underlying information set. After obtaining the underlying information sets of multiple known devices, it is necessary to extract the feature data related to the device hardware and firmware from the original data through feature extraction methods. For example, the processor model, memory specification, motherboard model, firmware version, etc. The feature extraction method can combine technical means such as string matching, regular expression parsing, and keyword field extraction to ensure that the data fields extracted from different protocols can be uniformly formatted.

[0116] After feature extraction is completed, to further improve the comparability and consistency of data, it is necessary to perform data standardization on multiple underlying feature information sets. The purpose of data standardization is to eliminate the format differences in data returned by different devices and ensure high stability of feature information during storage and matching. The data standardization processing rules include but are not limited to field renaming, unit conversion, and redundant field elimination. For example, the pen models returned by different protocols may have different field names (such as "Processor Model", "CPU Type", etc.), so these fields need to be uniformly renamed to "processor model"; unit conversion is for numerical data such as storage capacity, bandwidth, power, etc., and convert them to a unified unit (such as converting MB to GB, MHz to GHz, etc.); in addition, redundant, irrelevant or redundant fields need to be eliminated to reduce the interference of invalid data on feature matching. The standardized data set has stronger readability and comparability, laying a foundation for subsequent model matching and device classification.

[0117] After completing the data standardization process, further read the system event logs of each known device to obtain the number of times the device model has been modified for each known device. The system event log (System Event Log, abbreviated as: SEL) usually records historical data such as the operating status, hardware changes, firmware upgrades, and fault information of the device. By analyzing the number of times the device model has been modified in the log, it is possible to determine whether the model information of the device has been manually changed, the hardware has been replaced, or there has been accidental interference. According to the number of times the device model has been modified, further query the matching modification weight. The modification weight is used to analyze the stability of the device model information. For devices with a larger number of modification times, the model information may have greater uncertainty, so a lower weight should be assigned; while devices with fewer modification times or stable operation have a higher weight, further improving the credibility of their feature information.

[0118] Next, obtain the protocol weight corresponding to each device management protocol. The protocol weight is an important parameter that measures the reliability and accuracy of different device management protocols during the data acquisition process. Since different device management protocols have differences in the integrity, stability, and accuracy of data extraction, when integrating feature data, it is necessary to introduce protocol weights to reasonably allocate the data influence degree of each protocol. The acquisition of protocol weights can be comprehensively analyzed based on multiple indicators. For example, by analyzing historical data to statistically calculate parameters such as the recognition success rate, false alarm rate, and data missing rate of each protocol. Protocols with higher stability, stronger data integrity, and higher recognition accuracy are usually assigned higher weights, while protocols with higher data missing rates and larger recognition error rates should be assigned lower weights, thereby obtaining the protocol weight ω modelAccording to the standard feature information sets of multiple known devices, multiply the modification times weight and the protocol weight by the standard feature information in the standard feature information sets to obtain the device feature information sets of the multiple known devices; that is

[0119]

[0120]

[0121] where ω model is the protocol weight, and ω change is the modification times weight.

[0122] Furthermore, it is necessary to arbitrarily select two device feature information sets from the known device set, calculate their string similarity to obtain the weighted weights of the two device feature information sets. The calculation method of the string similarity can adopt weighted Jaccard similarity algorithm, cosine similarity algorithm, etc., and calculate the matching degree of the two by comparing the proportion of the common features and the overall features in the device feature strings. Subsequently, according to the weighted weights, use the weighted similarity algorithm to further calculate the weighted similarity of the two device feature information sets. If the weighted similarity of the two device feature information sets is greater than or equal to the second similarity threshold, classify the two device feature information sets into the same device set class.

[0123] Exemplarily, arbitrarily select two device feature information sets from the known device set, denoted as T1 and T2 respectively. Each device feature information set consists of multiple features, including but not limited to hardware information, firmware information, device model, device manufacturer, etc., and these features are mostly represented in the form of strings. Since different features have different importance in device recognition, a weighted mechanism needs to be introduced during similarity calculation to ensure that key features obtain higher matching weights.

[0124] In the specific process of calculating the string similarity, the weighted Jaccard similarity algorithm can be adopted. The core idea of this algorithm is to calculate the similarity of the two by comparing the proportion of the common features and all features in the two feature information sets. Its calculation formula is as follows:

[0125]

[0126] Among them, ω represents the feature weight coefficient; T1(x) and T2(x) respectively represent the feature values in the device feature information sets T1 and T2. By introducing a weighting mechanism, it is ensured that features with higher weights are given priority in the matching process, improving the accuracy of feature matching. For example, weights can be assigned according to the matching similarity between strings. That is, by using a vectorization method, after converting the device feature strings into vectors, the similarity is calculated through the ratio of the dot product of the vectors to the vector norms. After calculating the string similarity, it is necessary to calculate the weighted similarity value based on the similarity result. Specifically, according to the similarity result obtained from the above weighted Jaccard similarity formula, combined with the weighted weights of each field in the device feature information set, the weighted similarity between the two is calculated. If the weighted similarity of two device feature information sets is greater than or equal to the second similarity threshold, it is considered that the matching degree between them is high, and the two device feature information sets are classified into the same device set class. Through this method, the similarity between device feature information sets can be effectively distinguished, ensuring a more accurate division of device set classes, thereby improving the construction quality of the feature database and providing more reliable data support for subsequent device model identification.

[0127] In each device set class, further extract the common feature information of multiple device feature information under this set class to form the common feature set of the device set class. The common feature set represents the common features of this type of device, facilitating the rapid identification of this device group in subsequent model predictions. At the same time, it is necessary to combine the feature information of each known device and extract the unique feature information of each device as the key feature to distinguish different models of devices within the same set class. Through the combination of the common feature set and the unique feature set, the common and individual features of the device can be more comprehensively described, thereby improving the accuracy of model identification. After refining the common feature set and the unique feature set, further establish the mapping relationship between each common feature set, each unique feature set and the known device model based on the corresponding relationship between the device set class and the known device. This mapping relationship is the core data structure in the feature database, which can achieve the rapid matching of the features of the device to be identified and the known device features during the device model identification process. Finally, store the above multiple common feature sets, multiple unique feature sets, the device models of the known devices and their mapping relationships in the feature database together to complete the construction of the feature database. It can be seen that the constructed feature database has the advantages of wide data sources, complete feature information, accurate feature matching, and strong model stability.

[0128] Through the description of the above embodiments, those skilled in the art can clearly understand that the method according to the above embodiments can be implemented by means of software plus a necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases, the former is a better implementation method.

[0129] Figure 5 This is a schematic structural diagram of the device model identification device provided by the embodiment of the present application. As Figure 5 shown, the embodiment of the present application also provides a device model identification device, including:

[0130] A data acquisition module 501, configured to acquire multiple underlying feature information sets of the device to be identified according to multiple device management protocols of the device to be identified; wherein, the underlying feature information set includes at least one hardware feature information and / or at least one firmware feature information;

[0131] A current information acquisition module 502, configured to acquire a current underlying feature information set corresponding to the current device management protocol according to multiple underlying feature information sets, and acquire a current feature database corresponding to the current device management protocol based on multiple pre-constructed feature databases; wherein, the multiple device management protocols include the current device management protocol;

[0132] A screening and matching module 503, configured to determine a device set class matching the current underlying feature information set according to multiple device set classes in the current feature database, screen and obtain a first device feature information set matching the current underlying feature information set from multiple device feature information sets corresponding to the device set class, and mark the device model corresponding to the first device feature information set as the first identification model of the device to be identified;

[0133] An identification completion module 504, configured to obtain the final identification model of the device to be identified through a preset model screening rule according to multiple device management protocols and multiple first identification models.

[0134] For the description of the features in the embodiment corresponding to the device model identification device, reference can be made to the relevant description of the corresponding embodiment of the identification method, which will not be elaborated here one by one.

[0135] Figure 6 This is a schematic structural diagram of the electronic device provided by the present application. As Figure 6 shown, the electronic device 6 provided in this embodiment includes: at least one processor 61 and a memory 62. Optionally, the device 6 further includes a communication component 63. Among them, the processor 61, the memory 62, and the communication component 63 are connected through a bus.

[0136] In a specific implementation process, at least one processor 61 executes computer execution instructions stored in the memory 62, so that at least one processor 61 executes the above-mentioned identification method embodiment.

[0137] For the specific implementation process of the processor 61, reference can be made to the above method embodiment, and its implementation principle and technical effect are similar, which will not be elaborated here in this embodiment.

[0138] In the above embodiments, it should be understood that the processor may be a central processing unit (CPU), or may also be other general-purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), etc. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the application can be directly implemented by the execution of the hardware processor, or can be implemented by the combination of hardware and software modules in the processor.

[0139] The memory may include a random access memory (RAM), and may also include a non-volatile memory (NVM), such as at least one disk memory.

[0140] The bus may be an industry standard architecture (ISA) bus, a peripheral component interconnect (PCI) bus, an extended industry standard architecture (EISA) bus, etc. The bus can be divided into an address bus, a data bus, a control bus, etc. For the sake of convenience of representation, the buses in the drawings of the present application are not limited to only one bus or one type of bus.

[0141] The embodiments of the present application also provide a computer-readable storage medium, in which a computer program is stored, and the computer program is configured to execute the steps in the above-described identification method embodiments when running.

[0142] In an exemplary embodiment, the above computer-readable storage medium may include, but is not limited to: USB flash drives, read-only memories (ROMs), random access memories (RAMs), mobile hard disks, magnetic disks, or optical discs and other media that can store computer programs.

[0143] The embodiments of the present application also provide a computer program product, the above computer program product includes a computer program, and when the computer program is executed by a processor, it implements the steps in any of the above-described device model identification method embodiments.

[0144] Embodiments of the present application further provide another computer program product, including a non-volatile computer-readable storage medium storing a computer program, and the computer program, when executed by a processor, implements the steps in any of the above-described method embodiments for identifying device models.

[0145] Those skilled in the art can further realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of both. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0146] The above has introduced the identification method provided by the present application in detail. Specific examples have been used herein to illustrate the principle and implementation manner of the present application. The description of the above embodiments is only used to help understand the method and its core idea of the present application. It should be noted that for those of ordinary skill in the art in the technical field, without departing from the principle of the present application, several improvements and modifications can be made to the present application, and these improvements and modifications also fall within the protection scope of the claims of the present application.

Claims

1. A device model identification method, characterized in that: include: According to multiple device management protocols of the device to be identified, multiple underlying feature information sets of the device to be identified are obtained; wherein the underlying feature information set includes at least one hardware feature information and / or at least one firmware feature information; According to the multiple underlying feature information sets, a current underlying feature information set corresponding to the current device management protocol is obtained, and based on the multiple feature databases that have been constructed, a current feature database corresponding to the current device management protocol is obtained; wherein the multiple device management protocols include the current device management protocol; Determine, according to the multiple device set classes in the current feature database, the device set class that matches the current underlying feature information set, filter and obtain a first device feature information set that matches the current underlying feature information set from the multiple device feature information sets corresponding to the device set class, and mark the device model corresponding to the first device feature information set as the first identification model of the device to be identified; According to the plurality of device management protocols and the plurality of first identification models, a final identification model of the device to be identified is obtained through a preset model screening rule.

2. The identification method according to claim 1, characterized in that: After marking the device model corresponding to the first device feature information set as the first identification model of the device to be identified, the identification method further includes: According to the current device management protocol, obtaining and recording the identification model information of the device to be identified, and when the first identification model and the identification model information are inconsistent, obtaining a second device feature information set corresponding to the identification model information from the current feature database; According to the second device feature information set and the current underlying feature information set, a preset similarity algorithm is used to obtain the similarity between the second device feature information set and the current underlying feature information set; If the similarity between the second device feature information set and the current underlying feature information set is greater than or equal to a first similarity threshold, marking the device model corresponding to the identification model information as a second identification model, and updating the device model of the device to be identified to include the first identification model and the second identification model; According to the plurality of device management protocols, the plurality of first identification models and at least one of the second identification models, a final identification model of the device to be identified is obtained through a preset model screening rule.

3. The identification method according to claim 2, characterized in that: The method of obtaining the final identification model of the device to be identified according to the plurality of device management protocols, the plurality of first identification models and the at least one second identification model by using a preset model screening rule includes: Based on the multiple underlying feature information sets of the device to be identified, and according to the multiple first device feature information sets, obtaining a similarity score between the current underlying feature information set and a first device feature information set matching the current underlying feature information set, so as to obtain a similarity score corresponding to the first identification model; Obtaining a similarity score between the second device feature information set and the current underlying feature information set to obtain a similarity score for the second identification model; Acquire the same device management protocol corresponding to the first identification model and the second identification model, and acquire a preset weight threshold of the device management protocol; Calculate and obtain a weighted score of the first identification model and a weighted score of the second identification model according to the similarity score corresponding to the first identification model, the similarity score corresponding to the second identification model, and the preset weight threshold; Normalizing the weighted scores of the plurality of first identification models and the second identification models to obtain probability values ​​of the first identification model and the second identification model; Based on the probability values, an identification model with the largest probability value is screened out as the final identification model of the device to be identified.

4. The identification method according to claim 2, characterized in that: Also includes: If the similarity between the second device feature information set and the current underlying feature information set is less than a first similarity threshold, updating the device model corresponding to the first device feature information set to a first identification model; According to the plurality of device management protocols and the plurality of first identification models, a final identification model of the device to be identified is obtained through a preset model screening rule.

5. The identification method according to claim 1, characterized in that: Determining, according to the multiple device set classes in the current feature database, a device set class that matches the current underlying feature information set, comprising: According to the multiple device collection classes in the current feature database, query and obtain the public feature information set corresponding to the device collection class, so as to obtain multiple public feature information sets of the current feature database; wherein the public feature information set includes multiple public feature information of multiple known devices in the device collection class; A common feature information set matching the current underlying feature information set is queried and acquired from a plurality of common feature information sets in the current feature database to acquire a device set class matching the current underlying feature information set.

6. The identification method according to claim 5, characterized in that: The step of screening and acquiring a first device feature information set matching the current underlying feature information set from a plurality of device feature information sets corresponding to the device set class includes: Based on a device set class matching the current underlying feature information set, obtaining device feature information sets of a plurality of known devices in the device set class; Acquire a first unique feature information set of a plurality of the known devices according to the device feature information set of the known devices and the common feature information set of the device set class; Acquire a second unique feature information set of the device to be identified according to the current underlying feature information set and the common feature information set of the device set class; According to the plurality of first unique feature information sets, a first unique feature information set matching the second unique feature information set is queried and acquired, and a device feature information set corresponding to the first unique feature information set is marked as a first device feature information set.

7. The identification method according to claim 1 or 4, characterized in that: The method of obtaining the final identification model of the device to be identified according to the plurality of device management protocols and the plurality of first identification models by using a preset model screening rule includes: Based on the multiple underlying feature information sets of the device to be identified, and according to the multiple first device feature information sets, obtaining similarity scores of the underlying feature information set and the first device feature information set matching the underlying feature information set, so as to obtain the similarity score of the first identification model; Calculating weighted scores of the plurality of first identification models according to the similarity scores of the first identification models and the preset weight threshold of the device management protocol corresponding to the first identification model; wherein the weighted score is the product of the similarity score and the preset weight threshold; Normalizing the weighted scores of the plurality of first identification models to obtain probability values ​​of the plurality of first identification models; The first identification model with the largest probability value is selected as the final identification model of the device to be identified.

8. The identification method according to claim 1 or 2, characterized in that: The step of acquiring multiple sets of underlying feature information of the device to be identified according to multiple device management protocols of the device to be identified includes: According to the multiple device management protocols, any device management protocol is selected as the current device management protocol, and based on the current device management protocol, a communication connection with the device to be identified is established; According to the current device management protocol, send an instruction to the device to be identified, and receive original feature data returned by the device to be identified based on the instruction; According to the original feature data, hardware feature information and / or firmware feature information is parsed and extracted to obtain the current underlying feature information set of the device to be identified, and according to the multiple device management protocols, multiple underlying feature information sets of the device to be identified are obtained.

9. The identification method according to claim 1, characterized in that: The construction of the feature database includes: Based on multiple known devices in the known device cluster, according to the multiple device management protocols, multiple underlying information sets of the multiple known devices are obtained, and multiple underlying feature information sets of the multiple known devices are obtained through feature extraction; According to the plurality of underlying feature information sets, performing data standardization processing on the underlying feature information sets to obtain a plurality of standard feature information sets; Read the system event log of the known device, obtain the number of device model modifications of the known device, and query to obtain the modification number weight that matches the number of device model modifications; Acquire a protocol weight corresponding to the current device management protocol, and acquire device feature information sets of the plurality of known devices according to the protocol weight, the modification number weight and the standard feature information set; According to the string similarity of any two device feature information sets, the weighted weights of the two device feature information sets are obtained, and the weighted similarity of the two device feature information sets is calculated using a weighted similarity algorithm, and when the weighted similarity is greater than or equal to a second similarity threshold, the two device feature information sets are classified into the same device set class; Traversing multiple device feature information sets in the known device cluster to obtain multiple device set classes corresponding to the known device cluster; Extracting common feature information of multiple device feature information sets belonging to the same device set class to form a common feature set of the device set class, and acquiring unique feature information sets of multiple known devices based on the common feature set and the multiple device feature information sets; Traverse the plurality of device set classes, obtain and store a plurality of common feature sets and a plurality of unique feature information sets of known devices into the feature database.

10. An electronic device, characterized in that: include: Memory for storing computer programs; A processor, configured to implement the steps of the identification method according to any one of claims 1 to 9 when executing the computer program.

Images