Data processing method of all-in-one machine and computer readable storage medium

Abstract

The embodiment of the invention provides a data processing method of an all-in-one machine and a computer readable storage medium, and the method comprises the steps: obtaining the whole cabinet data of the all-in-one machine, and dividing the whole cabinet data into range data and fixed value data; storing in an open source database system based on distributed file storage in a preset format; performing compliance verification on the range data and the fixed value data by using the nominal data of the whole cabinet, judging whether the range data conforms to a nominal range or not, judging whether the fixed value data conforms to a nominal value or not, performing offset correction on the range data which does not conform to the nominal range and the fixed value data which does not conform to the nominal value, and then, performing offset correction on the range data which does not conform to the nominal value; and analyzing the weights of the factors influencing the nominal data of the whole cabinet, generating compliance verification data recommendation, and sending the compliance verification data recommendation to the user. Through the data processing method and device, the problem that the data processing efficiency is low in the prior art is solved, and the effect of improving the data processing efficiency of the all-in-one machine is achieved.

Description

Technical Field

[0001] The embodiments of the present application relate to the field of computers, and more specifically, to a data processing method of an all-in-one machine and a computer-readable storage medium. Background Art

[0002] The rapid development of new generation information technologies such as cloud computing, big data, and artificial intelligence has promoted the application of intelligent all-in-one cabinets in data centers and other fields. With the acceleration of digital transformation in various industries, the demand for data centers from enterprises and institutions is increasing, and the scale of data centers is also expanding. Intelligent cabinets have become the mainstream product form of large-scale data centers with their advantages of high-density deployment, high efficiency and energy saving, and unified management. With the continuous expansion of data center scale and the increase in cabinet configuration and complexity, unified data management and compliance verification methods are particularly important.

[0003] At present, the data presentation mode of the whole cabinet is mainly through the dynamic monitoring device to obtain and store the temperature, smoke, humidity, uninterruptible power supply, node, alarm information and other data of the whole cabinet. The server node information and switch information are obtained through the device management interface respectively, and these information are stored in the storage unit of the corresponding hardware device. In addition, the commonly used compliance verification method is the rule-based verification method, which predefines a series of data compliance rules, such as data format rules, data content restriction rules (such as sensitive information processing rules), data access permission rules, etc. At each stage of data processing, the system will perform real-time inspection and verification on the data according to these preset rules to determine whether the data meets the compliance requirements.

[0004] The current technical solution has the problem of data islands, that is, the data is divided and managed, lacking uniformity. In the data retrieval process, since different devices or node devices must be accessed separately to obtain the required information, the query process is extremely cumbersome and the operation complexity is significantly increased. In addition, when faced with a large number of data query needs, this retrieval method will greatly reduce convenience and work efficiency, seriously affecting the effective use of data. In addition, the rule-based verification method has a certain degree of singleness and cannot fully meet the compliance verification requirements in complex data processing environments. When the rules need to be updated, the maintenance process is relatively complicated and often requires manual intervention to modify the rules. In addition, in the face of complex and changeable actual data scenarios, it is difficult to effectively deal with emerging compliance problems through fixed rules, resulting in insufficient flexibility and adaptability of compliance verification. In summary, the efficiency of data processing in existing technologies is low. Summary of the invention

[0005] The embodiments of the present application provide a data processing method of an all-in-one machine and a computer-readable storage medium to at least solve the problem of low efficiency of data processing in the related art.

[0006] According to an embodiment of the present application, a data processing method for an all-in-one machine is provided, comprising: obtaining whole cabinet data of the all-in-one machine, and dividing the whole cabinet data into range data and fixed value data, wherein the range data refers to data having a threshold range for parameter data compliance verification, and the fixed value data refers to data requiring a fixed value for parameter data compliance verification; storing the range data and the fixed value data in a preset format in an open source database system based on distributed file storage, and storing the whole cabinet nominal data of the all-in-one machine and the original file of the firmware version involved in the whole cabinet nominal data in the open source database system based on distributed file storage. In an open source database system for file storage; applying the whole cabinet nominal data to perform compliance check on the range data and the fixed value data, judging whether the range data conforms to the nominal range, judging whether the fixed value data conforms to the nominal value, and performing deviation correction on the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value; after performing deviation correction on the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value, analyzing the weights of factors affecting the whole cabinet nominal data, generating compliance check data recommendations, and sending the compliance check data recommendations to the user.

[0007] In some exemplary embodiments, the entire cabinet nominal data is used to perform a compliance check on the range data and the fixed value data to determine whether the range data conforms to the nominal range and whether the fixed value data conforms to the nominal value, including: using the first formula Calculate the first deviation using the second formula Calculate the second deviation, where D1 is the first deviation, the first deviation represents the deviation of the range data from the lower limit of the nominal data interval, D2 is the second deviation, the second deviation represents the deviation of the range data from the upper limit of the nominal data interval, x is the range data, a is the lower limit of the nominal data interval, and b is the upper limit of the nominal data interval; when the first deviation is 0 and the second deviation is 0, the range data meets the nominal range; when the first deviation is not 0 and the second deviation is 0, the range data does not meet the nominal range; when the first deviation is 0 and the second deviation is not 0, the range data does not meet the nominal range.

[0008] In some exemplary embodiments, the nominal data of the entire cabinet is used to perform a compliance check on the range data and the fixed value data to determine whether the range data conforms to the nominal range, and to determine whether the fixed value data conforms to the nominal value, including: using data mapping technology to establish a one-to-one correspondence between the fixed value data and the nominal data, and comparing the fixed value data with the nominal data to obtain a third deviation; when the third deviation is 0, the fixed value data conforms to the nominal value; when the third deviation is not 0, the fixed value data does not conform to the nominal value.

[0009] In some exemplary embodiments, the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value are subjected to deviation correction, including: when the first deviation is not 0 and the second deviation is 0, the range data that does not conform to the nominal range is subjected to deviation correction by the third formula x′=a+(xa)×k, wherein x′ is the value of the range data after deviation correction, k is the deviation coefficient, and 0<k≤1; when the first deviation is 0 and the second deviation is not 0, the range data is subjected to deviation correction by the fourth formula x′=b-(xb)×k, wherein x′ is the value of the range data after deviation correction, k is the deviation coefficient, and 0<k≤1.

[0010] In some exemplary embodiments, the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value are corrected, including: when the third deviation is not 0, the fixed value data that does not conform to the nominal value is corrected by means of a representational state transfer application programming interface; for the case where the third deviation is not 0 due to inconsistent firmware versions, the standard firmware in the open source database system based on distributed file storage is called to update.

[0011] In some exemplary embodiments, after the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value are adjusted and corrected, the weights of the factors affecting the nominal data of the whole cabinet are analyzed to generate compliance verification data recommendations, including: using the preprocessing module built into the automated machine learning tool to perform data cleaning, normalization and encoding operations on the factors affecting the nominal data of the whole cabinet to obtain a preprocessed data set; constructing an influencing factor matrix A = (a ij ), where a ijIndicates the importance of factor i relative to factor j, and uses a 1-9 scaling method to assign values ​​to the factors in the influencing factor matrix, where 1 represents equal importance and 9 represents absolute importance; normalizes the geometric mean of the elements in each row of the influencing factor matrix to obtain a set of weight vectors, and each element in the weight vector represents the priority ratio of the corresponding factor; adjusts the weight of the training data input into the automatic machine learning software library model based on deep learning according to the weight vector, and uses a deep learning algorithm to train and learn the training data to identify the inherent pattern of the training data, and optimizes the automatic machine learning software library model's recognition and judgment criteria for compliance features; outputs the compliance verification data recommendation according to the training result of the automatic machine learning software library model.

[0012] In some exemplary embodiments, after outputting the compliance verification data recommendation, the method further includes: collecting feedback on the compliance verification data recommendation, and transmitting the feedback data back to the training layer of the automatic machine learning software library model to optimize the automatic machine learning software library model.

[0013] In some exemplary embodiments, the geometric mean of each row of elements in the influencing factor matrix is ​​normalized to obtain a set of weight vectors, each element in the weight vector represents the priority ratio of the corresponding factor, including: using the formula Calculate the geometric mean of each row of the influencing factor matrix, where: is the geometric mean, n is the number of columns in the influencing factor matrix; the geometric mean is normalized to obtain the weight vector w=(w1, w2, ..., w n ),in, w i is the priority weight of factor i.

[0014] In some exemplary embodiments, the method also includes: when the nominal data is updated due to design requirements, storing the updated nominal data in the open source database system based on distributed file storage in the preset format; performing incremental synchronization processing on the updated nominal data to obtain new synchronization data; using the automated machine learning tool to analyze the new synchronization data, update the understanding of compliance features, and periodically generate compliance verification data recommendations to reflect the compliance standards under the latest design requirements; the method also includes: displaying the operation process, processing progress and processing results of the data processing method in the form of a page, and providing a version data information query function of the all-in-one machine.

[0015] According to another embodiment of the present application, a computer-readable storage medium is provided, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0016] Through this application, by obtaining the whole cabinet data of the integrated machine, the whole cabinet data is divided into range data and precise data, and stored in a preset format in an open source database system based on distributed file storage to achieve centralized management and efficient retrieval of data, and the original files of the nominal data of the whole cabinet of the integrated machine and the firmware version involved in the nominal data of the whole cabinet are stored in the open source database system based on distributed file storage, and then the nominal data of the whole cabinet is used to perform compliance verification on the range data and precise data, and the data that does not meet the nominal standard is adjusted and corrected. The compliance verification is combined with deviation calculation and data mapping technology to automatically correct the deviation data, ensure data quality and reduce human errors. Finally, the weights of the factors affecting the nominal data of the whole cabinet are analyzed, and the compliance verification data recommendation is output. Therefore, the problem of low efficiency of data processing in related technologies can be solved, and the effect of improving the data processing efficiency of the integrated machine can be achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a flow chart of a data processing method of an all-in-one machine according to an embodiment of the present application;

[0018] Figure 2 is a structural block diagram of a data processing device of an all-in-one machine according to an embodiment of the present application;

[0019] Figure 3 is a structural block diagram of a data acquisition module of a data processing device of an all-in-one machine according to an embodiment of the present application;

[0020] Figure 4 is a structural block diagram of a data storage module of a data processing device of an all-in-one machine according to an embodiment of the present application;

[0021] Figure 5 is a structural block diagram of a data consistency verification module of a data processing device of an all-in-one machine according to an embodiment of the present application;

[0022] Figure 6 It is a structural block diagram of a data synchronization module of a data processing device of an all-in-one machine according to an embodiment of the present application. DETAILED DESCRIPTION

[0023] The embodiments of the present application will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0024] It should be noted that the terms "first", "second", etc. in the specification, claims and the above-mentioned drawings of this application are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0025] In this embodiment, a data processing method of an all-in-one machine is provided. Figure 1 is a flow chart of a data processing method of an all-in-one machine according to an embodiment of the present application, such as Figure 1 As shown, the process includes the following steps:

[0026] Step S102, obtaining the whole cabinet data of the integrated machine, and dividing the whole cabinet data into range data and fixed value data, wherein the range data refers to data within a threshold range for parameter data compliance verification, and the fixed value data refers to data for which the parameter data compliance verification requires a fixed value;

[0027] Specifically, the data processing of the all-in-one machine in the embodiment of the present application involves a smart all-in-one micro data center (SmartRack Data Center, SRDC). SRDC deeply integrates products such as servers, storage, networks, cabinets, power distribution units (Power Distribution Unit, PDU), uninterruptible power supplies (Uninterruptible Power Supply, UPS), dynamic environmental monitoring systems, air conditioners, networking cabling, etc., and realizes all-in-one machine products through targeted development, deep optimization, and system integration. The all-in-one machine components are selected according to user needs, and the production line is transported to the user site after production, ready for use out of the box.

[0028] First, through integrated monitoring and management tools, such as SRDC dynamic monitoring device, Redfish API (modern, open standard RESTful (Representational State Transfer) interface), Simple Network Management Protocol (SNMP) management tools, etc., the whole cabinet data of the integrated machine is automatically obtained, including the whole cabinet information, server node information, and switch information. These data fully reflect the real-time operation status of the whole cabinet and are the basis for compliance verification. Among them, the whole cabinet information includes temperature value, humidity value, smoke value, temperature value, UPS status value (UPS input voltage, output voltage) and air conditioning status value (set temperature value). The whole cabinet information is obtained through the SRDC dynamic monitoring device. The dynamic monitoring main equipment and sensors of the SRDC dynamic monitoring device include the monitoring integrated machine, front and rear temperature and humidity sensors, smoke detectors, and temperature detectors. The iView software displays the monitored information on the display screen, including temperature value, humidity value, smoke value, UPS information, and air conditioning information. Server node information includes firmware version, driver version, BMC\BIOS parameters. Use Java, Python, C and other programming languages ​​to write automatic parameter acquisition scripts, and obtain corresponding data through the Redfish interface. Redfish is a modern, open standard RESTful (Representational State Transfer) API used to manage and monitor hardware devices in data centers, such as servers, storage and network devices. It can obtain hardware device information, OS information, driver information, etc. Switch information includes version information and configuration parameters. The corresponding data is obtained through the SNMP management tool.

[0029] The acquired whole cabinet data is divided into two categories: "range data" and "fixed value data" to narrow the data retrieval scope. Range data refers to parameter data that needs to meet a certain threshold range during compliance verification, such as temperature value, humidity value, smoke value, temperature value, UPS status value (UPS input voltage, output voltage), air conditioning status value (set temperature value) and other whole cabinet information. The compliance judgment basis for these data is a preset range value. Fixed value data refers to parameter data that needs to be completely consistent with the preset fixed standard value during compliance verification, such as firmware version, driver version, BMC\BIOS parameters and other server node information, as well as version information, configuration parameters and other switch information. The compliance judgment basis is compared with the unpreset fixed value.

[0030] The purpose of classifying data into range data and fixed value data is to adopt different verification strategies according to the characteristics of different types of data in the subsequent compliance verification. For range data, verify whether it is within a reasonable range; for fixed value data, verify whether it is completely consistent with the nominal value. This classification method can more accurately identify data compliance issues and avoid the errors and inapplicability that may be caused by adopting a single verification standard for all data.

[0031] Through the above process, the embodiment of the present application can effectively obtain comprehensive cabinet data from the all-in-one machine, and through intelligent classification means, provide an accurate data classification basis for subsequent compliance verification and deviation correction, thereby improving the efficiency and accuracy of data processing.

[0032] Step S104, storing the range data and the fixed value data in a preset format in an open source database system based on distributed file storage, and storing the whole cabinet nominal data of the all-in-one machine and the original file of the firmware version involved in the whole cabinet nominal data in the open source database system based on distributed file storage;

[0033] Specifically, in order to unify the storage format and improve the data processing efficiency, the range data and the fixed value data are stored in a preset format in an open source database system based on distributed file storage. In the embodiment of the present application, the range data and the fixed value data can be stored in a MongoDB database in JSON (JavaScript Object Notation) format. MongoDB is an open source database system based on distributed file storage, which aims to provide a scalable and high-performance data storage solution for Web applications. The data structure supported by MongoDB is very flexible. It uses a BSON format similar to JSON to store data. BSON is a binary format of JSON. It not only supports all data types of JSON, but also supports more data types, such as dates, binary large objects (BLOBs), custom types, etc. Due to the flexibility of the data structure, MongoDB is very suitable for storing complex and hierarchical data, such as nested documents, arrays, etc., which makes it very effective in processing unstructured or semi-structured data. JSON is a lightweight data exchange format that is easy for people to read and write, and is also easy for machines to parse and generate. It is very suitable for storing and transmitting structured data.

[0034] In MongoDB, data is usually stored in the form of documents, each of which is a JSON object. To ensure fast retrieval and management of data, the system will design the database table structure, use the timestamp as the primary key, and create corresponding fields according to the data type (such as temperature, firmware version, etc.), and store the actual data and the obtained return value in these fields. This design makes each record unique and can maintain the time series of the data, which is convenient for tracing and analyzing historical data.

[0035] In addition to obtaining the whole cabinet data of the integrated machine, it is also necessary to collect the nominal data of the whole cabinet of the integrated machine. As an important reference standard for measuring data compliance and the normal operation status of the equipment, the nominal data covers key information such as the standard parameters, performance indicators and expected data values ​​of the equipment. In addition, these nominal data are stored in association with the corresponding firmware version original files. Each record contains complete nominal data information and corresponding firmware version original file information, where the firmware version original file information includes important attributes such as version number, release date, and file hash value. This storage method provides comprehensive and reliable data support for data compliance verification, equipment maintenance and upgrades, and troubleshooting operations, ensuring that the system can perform effective data processing and equipment management based on accurate nominal data during operation.

[0036] Similar to the whole cabinet data, the nominal data and the original firmware version files are also stored in the MongoDB database in JSON format. The difference is that the storage of these data focuses more on the integrity and relevance of the documents. Each record not only contains the detailed information of the nominal data, but also associates the corresponding original firmware version file, ensuring that accurate nominal data and related firmware version information can be quickly accessed when needed.

[0037] The above storage mechanism can effectively manage the whole cabinet data and nominal data of the integrated machine, providing a solid data foundation for subsequent data analysis, compliance verification and version management. At the same time, using a distributed file storage system such as MongoDB can cope with the massive data storage needs generated by large-scale data centers and ensure high availability and scalability of data.

[0038] Step S106, applying the above-mentioned whole cabinet nominal data to perform compliance check on the above-mentioned range data and the above-mentioned fixed value data, judging whether the above-mentioned range data conforms to the nominal range, judging whether the above-mentioned fixed value data conforms to the nominal value, and performing deviation correction on the range data that does not conform to the above-mentioned nominal range and the fixed value data that does not conform to the above-mentioned nominal value;

[0039] Specifically, nominal data refers to the performance indicators, parameter values ​​or data ranges that the SRDC cabinet or its components (such as server nodes, switches, etc.) should achieve under normal operating conditions. These data are usually defined in the equipment specifications, factory configurations or industry standards. For example, the nominal range can be the temperature threshold of the equipment, while the nominal value can be a specific firmware or software version.

[0040] Use the nominal data to perform compliance checks on the range data to check whether the range data is within the nominal range. For example, if the nominal temperature range is 15°C to 30°C, then any temperature reading outside this range will be considered non-compliant. Secondly, use the nominal value to perform compliance checks on the fixed value data. This involves exact value matching to check whether the fixed value data is completely consistent with the nominal value.

[0041] When it is found that the data does not meet the nominal range or nominal value, an offset correction is performed, that is, the current data is adjusted to meet the nominal requirements. For range data, the operating parameters of the device can be adjusted to bring the actual reading back to the nominal range; for fixed value data, the firmware or software can be updated to the nominal version.

[0042] By performing compliance verification and deviation correction, we can ensure that the data and performance of the entire SRDC cabinet always meet expectations and compliance standards, improve the reliability of equipment and the overall data center operation and maintenance efficiency. This automated compliance verification and deviation correction mechanism can significantly improve work efficiency and reduce human errors, especially when facing large-scale data collection and processing.

[0043] Step S108, after the range data that does not conform to the above nominal range and the fixed value data that does not conform to the above nominal value are adjusted and corrected, the weights of the factors affecting the above nominal data of the entire cabinet are analyzed, and compliance verification data recommendations are generated, and the above compliance verification data recommendations are sent to the user.

[0044] Specifically, the factors that affect the nominal data of the whole cabinet are: current nominal data, compliance standards, user requirements, regional differences in the SRDC whole cabinet (such as natural conditions such as ambient temperature and humidity), data fluctuations (such as instantaneous changes in system performance), etc. The factors that affect the nominal data of the whole cabinet are converted into computable data forms, usually through data preprocessing (cleaning, normalization, encoding, etc.) to facilitate subsequent analysis and machine learning model training. Next, analyze the impact of these factors on the nominal data. Weight analysis is to determine which factors have a greater impact on data compliance by training the model. According to the analyzed weights, the model will be further trained and optimized to improve its accuracy and efficiency in recommending compliance verification data.

[0045] The trained model will generate a series of compliance verification data recommendations based on the current nominal data and the weights of the influencing factors. These compliance verification data recommendations can include recommended firmware updates, environmental control parameter adjustments, device configuration optimization, etc., to help the entire cabinet data more consistent with the nominal data. Then, the generated compliance verification data recommendations are sent to the user. After the user receives and implements the recommended adjustments, the user's feedback information is collected, including the effectiveness of the adjustment results, changes in device performance, etc. This feedback information will be input into the model again to further optimize the model performance and improve the accuracy of future recommendations.

[0046] Through the above process, not only can the data that deviates from the nominal range be automatically corrected, but the key factors affecting data consistency can also be analyzed and targeted adjustment suggestions can be provided, which greatly improves the efficiency and accuracy of the data consistency verification of the entire cabinet of the all-in-one machine, while also reducing the burden of manual maintenance and improving the operational efficiency of the data center and the reliability of the equipment.

[0047] In one embodiment of the present application, the above-mentioned whole cabinet nominal data is used to perform a compliance check on the above-mentioned range data and the above-mentioned fixed value data to determine whether the above-mentioned range data conforms to the nominal range, and to determine whether the above-mentioned fixed value data conforms to the nominal value, including: using the first formula Calculate the first deviation using the second formula Calculate the second deviation, where D1 is the first deviation, which indicates the deviation of the range data from the lower limit of the nominal data interval, D2 is the second deviation, which indicates the deviation of the range data from the upper limit of the nominal data interval, x is the range data, a is the lower limit of the nominal data interval, and b is the upper limit of the nominal data interval; when the first deviation is 0 and the second deviation is 0, the range data meets the nominal range; when the first deviation is not 0 and the second deviation is 0, the range data does not meet the nominal range; when the first deviation is 0 and the second deviation is not 0, the range data does not meet the nominal range; when the first deviation is 0 and both the second deviation are not 0, the range data does not meet the nominal range.

[0048] Specifically, for range data, the deviation is calculated to check whether it meets the nominal range. Calculate the first deviation. The first deviation D1 represents the degree of deviation between the range data x and the lower limit a of the nominal data interval. When the range data x is equal to the lower limit a of the nominal data interval, the first deviation D1 = 0, that is, there is no lower limit deviation. When the range data x is less than the lower limit a of the nominal data interval, the first deviation D1 = 0, because the deviation calculates the percentage of data exceeding the range. Here, the range data x does not exceed the lower limit a of the nominal data interval, so the deviation is 0. Through the second formula The second deviation is calculated. The second deviation D2 represents the degree of deviation between the range data x and the upper limit b of the nominal data interval. When the range data x is equal to the upper limit b of the nominal data interval, the second deviation D2 = 0, that is, there is no upper limit deviation. When the range data x is greater than the upper limit b of the nominal data interval, the second deviation D2 = 0.

[0049] If the first deviation D1 and the second deviation D2 are both 0, it means that the range data x is neither lower than the lower limit a of the nominal data interval nor higher than the upper limit b of the nominal data interval. Therefore, the range data x conforms to the nominal data interval [a, b], and the data is legal, without the need for deviation correction.

[0050] If the first deviation D1 is not 0 but the second deviation D2 is 0, it means that the range data x deviates from the lower limit a of the nominal data interval, that is, the range data x is lower than the lower limit a of the nominal data interval. At this time, the data is non-compliant and needs to be adjusted to bring the range data x back into the interval.

[0051] If the first deviation D1 is 0 but the second deviation D2 is not 0, it means that the range data x deviates from the upper limit b of the nominal data interval, that is, the range data x exceeds the upper limit b of the nominal data interval. The data is also non-compliant and also needs to be adjusted and corrected to bring the range data x back into the interval.

[0052] For range data that does not conform to the nominal range, a deflection correction operation is performed based on the direction and degree of deviation. If the deviation is due to the range data being below the lower limit a of the nominal data interval, the operating parameters of the device will be adjusted, such as increasing the power of the heater to increase the temperature reading; if the range data exceeds the upper limit b of the nominal data interval, the power consumption of the device can be reduced, such as adjusting the settings of the cooling system to reduce the temperature reading. The goal of deflection correction is to make the range data as close to the normal range of the nominal data as possible.

[0053] Through the above mechanism, the system can effectively monitor and verify the range data in the SRDC cabinet in real time, ensure that the equipment operates in the best state, and avoid performance problems or safety risks caused by data deviation from the nominal range. At the same time, this automated data verification and deviation adjustment mechanism greatly reduces the need for manual intervention and improves the operation and maintenance efficiency and automation level of the data center.

[0054] In another embodiment of the present application, the above-mentioned nominal data of the entire cabinet is used to perform a consistency check on the above-mentioned range data and the above-mentioned fixed value data to determine whether the above-mentioned range data conforms to the nominal range, and to determine whether the above-mentioned fixed value data conforms to the nominal value, including: using data mapping technology to establish a one-to-one correspondence between the above-mentioned fixed value data and the above-mentioned nominal data, and comparing the above-mentioned fixed value data with the above-mentioned nominal data to obtain a third deviation; when the above-mentioned third deviation is 0, the above-mentioned fixed value data conforms to the above-mentioned nominal value; when the above-mentioned third deviation is not 0, the above-mentioned fixed value data does not conform to the above-mentioned nominal value.

[0055] Specifically, for fixed value data, data mapping technology is used to compare and verify whether it conforms to the nominal value. First, a one-to-one correspondence is established between the fixed value data and its corresponding nominal data. This means that each fixed value data (such as the firmware version of the server node, BIOS version, etc.) has a clear corresponding nominal data, which is defined according to the design specifications, industry standards or factory configuration of the SRDC cabinet. For example, if the firmware version of the server node is nominally 2.1.0, then data mapping will ensure that the actual collected firmware version data can be directly compared with this nominal value.

[0056] Once the one-to-one correspondence is established, the next step is to compare the collected fixed value data with the nominal data. The purpose of the comparison is to check whether the actual data completely matches the nominal value. The third deviation D3 is used to quantify the difference between the fixed value data and the nominal data. If the comparison result of the fixed value data and the nominal data shows that the data is completely consistent (that is, the actual data is the same as the nominal data), then the third deviation D3 is set to 0, indicating that there is no deviation and the data meets the nominal value. Conversely, if the comparison result of the fixed value data and the nominal data shows that the data is inconsistent, the third deviation D3 is set to a non-zero value, indicating the deviation between the fixed value data and the nominal data.

[0057] If the third deviation D3 is 0, it means that the fixed value data is completely consistent with the nominal data, the data is legal, and no deviation correction is required. On the contrary, if the third deviation D3 is not 0, it means that there is a difference between the fixed value data and the nominal data, the data does not conform to the nominal value, and deviation correction is required to restore the data compliance.

[0058] For fixed value data with the third deviation D3 not equal to 0, a deviation correction operation is performed, which involves updating the device's firmware, driver, or software settings to ensure that the actual data is consistent with the nominal data. For example, if the firmware version of the server node does not match the nominal value, the firmware update process will be automatically triggered to download the correct version from the standard firmware file stored in the MongoDB database and install it on the server node, thereby eliminating the third deviation D3 and restoring the data to compliance.

[0059] The entire process is automated, greatly reducing the need for manual intervention, improving the accuracy and efficiency of data verification, and ensuring the normal operation of the entire SRDC cabinet and data consistency.

[0060] In a specific embodiment, the range data that does not meet the above-mentioned nominal range and the fixed value data that does not meet the above-mentioned nominal value are subjected to deviation correction, including: when the above-mentioned first deviation is not 0 and the above-mentioned second deviation is 0, the range data that does not meet the above-mentioned nominal range is subjected to deviation correction by the third formula x′=a+(xa)×k, wherein x′ is the numerical range data after deviation correction, k is the deviation coefficient, and 0<k≤1; when the above-mentioned first deviation is 0 and the above-mentioned second deviation is not 0, the range data is subjected to deviation correction by the fourth formula x′=b-(xb)×k, wherein x′ is the numerical value after deviation correction, k is the deviation coefficient, and 0<k≤1.

[0061] Specifically, when the first deviation is not 0 and the second deviation is 0, it means that the range data x is lower than the lower limit a of the nominal data interval, but does not exceed the upper limit b of the nominal data interval. At this time, the range data needs to be adjusted upward to reach the nominal range. The third formula x′=a+(xa)×k is used to perform deviation correction to ensure that the range data x falls within the nominal range after adjustment. The adjustment process uses the deviation coefficient k, and the value of the deviation coefficient k can be determined based on historical data, device characteristics, or a preset deviation strategy to ensure that the impact of the adjustment on device performance is minimized.

[0062] When the first deviation is 0 and the second deviation is not 0, it means that the range data x exceeds the upper limit b of the nominal data interval, but is not lower than the lower limit a of the nominal data interval. At this time, the range data needs to be adjusted downward to meet the nominal range. The fourth formula x′=b-(xb)×k is used to perform the deviation correction to ensure that the range data x meets the nominal range after adjustment.

[0063] The offset coefficient k is a key parameter used to control the adjustment range during the offset correction process. The offset coefficient k can be set through the web page of the SRDC dynamic monitoring display to achieve offset correction. The value of the offset coefficient k can be adjusted according to different situations and device characteristics. For example, when processing temperature data, the setting of the offset coefficient k may need to consider factors such as the thermal stability of the device and the rate of change of the ambient temperature to avoid excessive adjustments that may cause unstable device performance.

[0064] Through the above process, the deviation correction mechanism is used to automatically adjust the range data, which not only ensures the data consistency of the entire SRDC cabinet, but also improves the automated operation and maintenance level of the data center and enhances the stability and reliability of the system.

[0065] In a specific embodiment, the range data that does not conform to the above-mentioned nominal range and the fixed value data that does not conform to the above-mentioned nominal value are adjusted and corrected, including: when the above-mentioned third deviation is not 0, the fixed value data that does not conform to the above-mentioned nominal value is adjusted and corrected by means of a representational state transfer application programming interface; for the situation where the above-mentioned third deviation is not 0 due to inconsistent firmware versions, the standard firmware in the above-mentioned open source database system based on distributed file storage is called to update.

[0066] Specifically, for those fixed value data whose third deviation is not 0, that is, data that is inconsistent with the nominal data, these data are adjusted by calling the representational state transfer application programming interface. In this embodiment, the adjustment is performed by calling the RESTful API. The RESTful API allows the system to interact with different devices in a standard and unified way to achieve data acquisition and modification. For example, if the BIOS version of the server node does not conform to the nominal data, the system can call the RESTful API to access and update the BIOS settings.

[0067] The RESTful API call is automated, which means that it can automatically identify which fixed-value data does not meet the nominal value and automatically perform deviation correction operations without manual intervention. This process can include updating software settings, firmware versions, configuration parameters, etc. to ensure that the data is restored to the nominal range.

[0068] This is detected when the third deviation is not 0 and the data is inconsistent due to inconsistent firmware versions. Inconsistent firmware versions may affect the performance of the device and the accuracy of the data, so an update is required.

[0069] Call the standard firmware version stored in the MongoDB database for update. As an open source database system based on distributed file storage, MongoDB can store and manage firmware version files of various devices, including version number, release date, file hash value and other information. When the firmware needs to be updated, the standard firmware version of the corresponding device can be retrieved from MongoDB and the update process can be automatically executed.

[0070] The firmware update process is implemented through automated scripts or built-in firmware update tools to ensure that the update is safe, fast and will not have adverse effects on the device. After the update is completed, the system will perform a data consistency check again to confirm whether the update is successful and whether the data has met the nominal value.

[0071] Through the above steps, not only can the inconsistent state of data be identified and adjusted, but also the inconsistent firmware version can be automatically updated to ensure the data consistency of the entire SRDC cabinet and its components and the normal operation of the equipment. This mechanism fully utilizes the flexibility of RESTful API and the efficient data management capabilities of MongoDB to realize the automation of data and firmware updates, greatly improving the operation and maintenance efficiency of the data center and the reliability of the equipment.

[0072] In another embodiment of the present application, after the range data that does not conform to the above-mentioned nominal range and the fixed value data that does not conform to the above-mentioned nominal value are adjusted and corrected, the weights of the factors affecting the above-mentioned nominal data of the whole cabinet are analyzed to generate compliance verification data recommendations, including: using the built-in preprocessing module of the automated machine learning tool to perform data cleaning, normalization and encoding operations on the factors affecting the above-mentioned nominal data of the whole cabinet to obtain a preprocessed data set; constructing an influencing factor matrix A=(a ij ), where a ij Indicates the importance of factor i relative to factor j, and uses a 1-9 scaling method to assign values ​​to the factors in the above-mentioned influencing factor matrix, where 1 represents equal importance and 9 represents absolute importance; the geometric mean of the elements in each row of the above-mentioned influencing factor matrix is ​​normalized to obtain a set of weight vectors, and each element in the above-mentioned weight vector represents the priority ratio of the corresponding factor; according to the above-mentioned weight vector, the weight of the training data input into the automatic machine learning software library model based on deep learning is adjusted, and the above-mentioned training data is trained and learned using a deep learning algorithm to identify the inherent pattern of the above-mentioned training data, and optimize the above-mentioned automatic machine learning software library model for identifying and judging the compliance features; according to the training results of the above-mentioned automatic machine learning software library model, the above-mentioned compliance verification data recommendation is output.

[0073] Specifically, first, the preprocessing module built into the automated machine learning tool is used to perform data cleaning, normalization, and encoding operations on various factors that affect the nominal data of the entire cabinet. In the embodiment of the present application, the preprocessing module built into AutoGluon is used to perform data cleaning, normalization, and encoding operations on various factors that affect the nominal data of the entire cabinet. AutoGluon is an open source automated machine learning software library that is easy to use and highly automated, and can automatically perform key steps such as model selection and hyperparameter adjustment. Data cleaning is intended to remove invalid or abnormal data to ensure the accuracy of analysis; normalization processing enables data of different magnitudes or dimensions to be compared on a unified scale; encoding operations convert non-numeric data into machine-readable numerical forms to facilitate subsequent analysis and modeling.

[0074] The preprocessed data set is more analytical and practical, and can be used as the basis for constructing an influencing factor matrix. Construct an influencing factor matrix A = (a ij ), each element a in the influencing factor matrix ij Indicates the importance of factor i relative to factor j. These factors can be quantified using a 1-9 scale, where 1 means that both factors are equally important and 9 means that one factor is absolutely more important than the other. The 1-9 scale is a subjective assignment method, usually derived by experts in the field or through statistical methods such as cluster analysis and factor analysis. The assignment process needs to comprehensively consider the direct and indirect effects of factors, as well as the interactions between various factors.

[0075] The geometric mean is calculated for each row of the influencing factor matrix, which can better reflect the comprehensive balance of the importance of each factor. Then, the calculated geometric mean is normalized and converted into elements in the weight vector, which represents the priority of each factor. The weight vector can more intuitively reflect the relative importance of each factor in the overall analysis. The weight vector is applied to the model training process to adjust the weight of the training data input into the deep learning-based AutoGluon model. This makes the model training pay more attention to those factors that have a greater impact on the nominal data of the entire cabinet, thereby improving the prediction accuracy and practicality of the model. The deep learning algorithm is used to learn the training data, identify the inherent patterns of the training data, and optimize the recognition and judgment criteria of the automatic machine learning software library model for the matching features. Deep learning can automatically extract features from complex data sets and achieve high-level abstraction and pattern recognition of data through multi-level neural networks.

[0076] Based on the training results of the model, a set of consistency verification data recommendations are output. These consistency verification data recommendations may include recommended adjustments to the offset coefficient k, device configuration optimization recommendations, firmware update priority ranking, etc., aiming to help users more effectively perform data consistency verification and maintenance and optimization of the entire cabinet.

[0077] The above process combines deep learning algorithms and automated machine learning tools to achieve intelligent analysis of factors that affect the nominal data of the entire cabinet and automatic recommendation of data consistency verification, thereby improving data consistency and the stability of equipment operation.

[0078] In a specific embodiment, after outputting the compliance verification data recommendation, the method further includes: collecting feedback on the compliance verification data recommendation, and transmitting the feedback data back to the training layer of the automatic machine learning software library model to optimize the automatic machine learning software library model.

[0079] Specifically, collect user feedback on the output conformity verification data recommendations. Feedback can be the result of the user's manual deflection correction, such as whether the deflection is successful, whether the corrected data is stable, whether the device performance is affected, etc. It can also be the user's evaluation of the conformity verification data recommendation itself, such as the accuracy and practicality of the recommendation. User feedback can include positive affirmation, such as the correction value recommended by the conformity verification data does solve the problem of data inconsistency; it may also include negative feedback, such as the correction value is not applicable, caused new problems or did not achieve the expected effect. Regardless of positive or negative, these feedbacks are important information for model optimization.

[0080] The collected feedback data will be integrated and preprocessed to ensure that it is compatible with the original training data and meets the requirements of model training. The preprocessed feedback data is sent back to the training layer of the automatic machine learning software library model. The training layer is the core part of model learning and optimization. It performs feature extraction, parameter adjustment and model training based on the input data to improve the model's prediction ability. The feedback data sent back will be used as additional training samples by the model for retraining the model. The model will adjust its internal parameters based on these new samples to optimize its prediction and recommendation of the compliance verification data.

[0081] By continuously collecting and learning user feedback, the model can gradually improve its generalization ability, that is, make accurate predictions and recommendations on unseen data. This is very important for dealing with new data inconsistencies that may arise in the SRDC cabinet, making the system more flexible and adaptable.

[0082] Through the above feedback mechanism, the automatic machine learning software library model can be continuously adjusted and optimized according to actual usage to provide more accurate and practical compliance verification data recommendations. This mechanism not only improves the adaptability and prediction ability of the model, but also strengthens the interaction between the system and users, ensuring that user needs are responded to in a timely manner, while ensuring the long-term stability and high performance of the SRDC whole cabinet data management, storage and compliance verification system.

[0083] In a specific embodiment, the geometric mean of each row of elements in the above-mentioned influencing factor matrix is ​​normalized to obtain a set of weight vectors, and each element in the above-mentioned weight vector represents the priority ratio of the corresponding factor, including: using the formula Calculate the geometric mean of each row of the influencing factor matrix above, where: is the geometric mean, n is the number of columns in the influencing factor matrix; the geometric mean is normalized to obtain the weight vector w = (w1, w2, ..., w n ),in, w i is the priority weight of factor i.

[0084] Specifically, each element in the weight vector represents the priority ratio of the corresponding factor, which is crucial for the intelligent analysis and recommendation of the consistency verification of the whole cabinet data. Calculate the geometric mean of each row of the influencing factor matrix. The geometric mean calculation method takes into account the influence of all factors, rather than just their simple average, which helps better reflect the relative importance of each factor. Especially when the importance of each factor varies greatly, the geometric mean can more accurately reflect the comprehensive weight of the factors.

[0085] The calculated geometric mean of each row is normalized to convert the values ​​of all elements to the range between 0 and 1, so that the elements in the weight vector can directly reflect the relative importance of each factor. After normalization, the weight vector w = (w1, w2, ..., w n ),in, w i is the priority weight of factor i.

[0086] Through normalization, the magnitude differences between different factors can be eliminated, ensuring that each element in the weight vector is compared on the same scale, making it easier for the model to understand and use this information for learning and optimization. n ) i It represents the priority ratio of the corresponding factor, and can directly reflect the relative importance of this factor in affecting the compliance of the nominal data of the entire cabinet.

[0087] The weight vector is used to adjust the weight of the training data input into the deep learning-based automatic machine learning software library model, ensuring that the model pays more attention to the factors with greater influence during the training process, thereby optimizing the model's recognition and judgment criteria for compliance features. In addition, the weight vector is not static. As the system runs, it can be dynamically adjusted based on user feedback, data changes, and model training results. This mechanism can ensure that the model always pays attention to the most relevant factors at the moment, improve the model's adaptability and prediction accuracy, and ensure that the model always maintains the optimal state.

[0088] Through the above steps, based on the geometric mean and normalization of the influencing factor matrix, a set of weight vectors reflecting the relative importance of each factor is generated. This vector not only provides more accurate training data weights for the automated machine learning software library model, but also provides key quantitative indicators for subsequent data consistency analysis and recommendations. The dynamic update and optimization mechanism of the weight vector enables the system to continuously adapt to changes in the data center environment and improve the accuracy and efficiency of its data consistency verification.

[0089] In another embodiment of the present application, the method further includes: when the nominal data is updated due to design requirements, storing the updated nominal data in the open source database system based on distributed file storage in the preset format; performing incremental synchronization processing on the updated nominal data to obtain new synchronization data; using the automated machine learning tool to analyze the new synchronization data, update the recognition of compliance features, and periodically generate compliance verification data recommendations to reflect the compliance standards under the latest design requirements;

[0090] Specifically, when nominal data is updated, it is first necessary to store the updated data in a distributed file storage-based open source database system (such as MongoDB) in a preset format (such as JSON format). The preset format ensures the consistency and compatibility of the data, which facilitates subsequent data processing and verification. Then, incremental synchronization is performed on the updated nominal data. Incremental synchronization means synchronizing only the data that has changed, rather than resynchronizing the entire database. This improves the efficiency of data synchronization and reduces unnecessary resource consumption.

[0091] Use automated machine learning tools (such as AutoGluon) to analyze the newly synchronized nominal data. Automated machine learning tools can automatically perform steps such as data preprocessing, feature selection, model training, and hyperparameter optimization, thereby significantly improving analysis efficiency and accuracy. Based on the newly synchronized data, automated machine learning tools automatically identify the inherent patterns of the data and update their understanding of compliance features. This means that automated machine learning tools can learn compliance standards under new design requirements, such as new performance indicators, parameter thresholds, etc., and adjust the model to adapt to these changes.

[0092] The above process is performed periodically, that is, when the nominal data is updated, the system automatically synchronizes the new data and analyzes the new data through automated machine learning tools to update the model's understanding of the compliance features.

[0093] Generate new compliance data recommendations based on the updated model and the latest design requirements. These compliance data recommendations can include correction suggestions for specific parameters, firmware update requirements, configuration optimization strategies, etc., all based on the latest nominal data and compliance standards.

[0094] The goal of this mechanism is to ensure that the compliance check of the SRDC cabinet data can always be consistent with the latest design requirements. Through the continuous learning and updating of automated machine learning tools, the system can adapt to design changes in a timely manner, improve the accuracy and timeliness of data compliance checks, and thus improve the operation and maintenance efficiency of the data center and the operation stability of the equipment. When the designer changes the performance indicators or parameter thresholds, the system can respond automatically without manual intervention to adjust the data and models on a large scale. This not only saves labor costs, but also avoids potential errors caused by manual operations. In addition, by periodically updating nominal data, incremental synchronization, automated analysis, and generating recommendations, the system can continuously optimize its performance to adapt to future design changes and technological advances. This mechanism enhances the adaptability and future sustainability of the system, ensuring that it remains efficient and reliable in the ever-changing data center environment.

[0095] In summary, when the nominal data is updated, by using automated machine learning tools for analysis and model update, the present invention can achieve intelligent and efficient data compliance verification recommendations, ensuring high standards and high efficiency of data center operation and maintenance, while also demonstrating the flexibility and adaptability of the system in the face of changes in design requirements.

[0096] In another embodiment of the present application, the method further includes: displaying the operation flow, processing progress and processing result of the data processing method in the form of a page, and providing a version data information query function of the all-in-one machine.

[0097] Specifically, after the user receives the compliance verification data recommendation, the user's feedback information is collected and sent back to the model training layer to further optimize the model performance and recommendation accuracy. This part focuses on the transparency of the data processing method and user interaction. By showing the specific operation process, processing speed and final processing results of data processing in a regulated manner, the operability and user-friendliness of the system are increased. At the same time, the system also provides a version data information query function for the all-in-one machine (SRDC whole cabinet), allowing users to easily track and manage the update history of the device.

[0098] Specifically, a user interface is designed to display the entire process of data processing in the form of an HTML page. This includes the operation details of each stage, such as data collection, classification, storage, conformance verification, deflection correction, synchronization, and intelligent learning. The user interface can be in the form of charts, flow charts, timelines, etc., to clearly and intuitively present each step of data processing. The page also provides a real-time tracking function for processing progress, so that users can view the current processing steps in progress and the completion status of each step at any time. This is crucial for monitoring the efficiency and status of data processing, especially when processing large-scale data or performing complex conformance verification. The final processing results, including conformance verification results, deflected data, recommended conformance standards, etc., will be displayed to users on the page in the form of clear lists, charts, or reports. This enables users to quickly understand the results of data processing and the possible impact on the system or device.

[0099] The version data information query function of the all-in-one machine allows users to query the data information of different versions of the SRDC cabinet, including but not limited to device configuration, firmware version, performance parameters, etc. The version information query function is very important for troubleshooting, maintenance and upgrades, and historical data analysis. The user interface can also include a version comparison function, allowing users to compare the differences between different versions, which is helpful for understanding the improvement or degradation of device performance and changes in compliance standards. The query function is not limited to the version information of the entire cabinet, but can also query the detailed data parameters of specific server nodes, switches and other components, providing users with a comprehensive view of the device status.

[0100] The implementation of this series of functions improves the transparency of the system and the user interaction experience, allowing users to better understand system operations, monitor data processing, and manage device version information. For data center operation and maintenance personnel, the provision of this transparency and query function greatly simplifies daily operation and maintenance work, improves the efficiency of troubleshooting and equipment management, and also ensures the compliance and operational stability of the data center. Through intuitive page display and powerful query functions, users can more effectively perform data compliance verification and related operations.

[0101] Through the above steps, the whole cabinet data of the integrated machine is first obtained, and the whole cabinet data is divided into range data and precise data, and stored in a preset format in an open source database system based on distributed file storage to achieve centralized data management and efficient retrieval, and the original files of the nominal data of the whole cabinet of the integrated machine and the firmware version involved in the nominal data of the whole cabinet are stored in the open source database system based on distributed file storage, and then the nominal data of the whole cabinet is used to perform compliance verification on the range data and precise data, and the data that does not meet the nominal standard is adjusted and corrected. The compliance verification is combined with deviation calculation and data mapping technology to automatically correct the deviation data to ensure data quality and reduce human errors. Finally, the weights of the factors affecting the nominal data of the whole cabinet are analyzed, and the compliance verification data recommendation is output. The problem of low efficiency of data processing in related technologies is solved, and the data processing efficiency of the integrated machine is improved.

[0102] Through the description of the above implementation methods, those skilled in the art can clearly understand that the method according to the above embodiment can be implemented by software plus the necessary general hardware platform, or by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, or network device, etc.) to execute the method described in each embodiment of the present application.

[0103] In this embodiment, an all-in-one data processing device is provided, which is used to implement the above embodiments and preferred implementations, which have been described and will not be repeated. As used below, the term "module" can be a combination of software and / or hardware that implements a predetermined function. Although the device described in the following embodiments is preferably implemented in software, it is also possible and conceivable to implement it in hardware, or in a combination of software and hardware.

[0104] Figure 2 is a structural block diagram of a data processing device of an all-in-one machine according to an embodiment of the present application, such as Figure 2 As shown, the device includes a data acquisition module 1, a data storage module 2, a data compliance verification module 3, a data synchronization module 4 and a data output module 5, wherein:

[0105] The data acquisition module 1 is used to obtain the whole cabinet data of the integrated machine, and divide the whole cabinet data into range data and fixed value data, wherein the range data refers to data with a threshold range for parameter data compliance verification, and the fixed value data refers to data with a fixed value requirement for parameter data compliance verification.

[0106] Specifically, the data processing of the all-in-one machine in the embodiment of the present application involves a smart all-in-one micro data center (SmartRack Data Center, SRDC). SRDC deeply integrates products such as servers, storage, networks, cabinets, power distribution units (Power Distribution Unit, PDU), uninterruptible power supplies (Uninterruptible Power Supply, UPS), dynamic environmental monitoring systems, air conditioners, networking cabling, etc., and realizes all-in-one machine products through targeted development, deep optimization, and system integration. The all-in-one machine components are selected according to user needs, and the production line is transported to the user site after production, ready for use out of the box.

[0107] like Figure 3 As shown, the data acquisition module 1 includes a data classification acquisition module 11 and a distribution module 12. The data classification acquisition module 11 automatically obtains the whole cabinet data of the integrated machine through integrated monitoring and management tools, such as SRDC dynamic monitoring device, Redfish API (modern, open standard RESTful (Representational State Transfer) interface), Simple Network Management Protocol (SNMP) management tool, etc., including the whole cabinet information, server node information, and switch information. These data fully reflect the real-time operation status of the whole cabinet and are the basis for compliance verification. Among them, the whole cabinet information includes temperature value, humidity value, smoke value, temperature value, UPS status value (UPS input voltage, output voltage) and air conditioning status value (set temperature value). The whole cabinet information is obtained through the SRDC dynamic monitoring device. The dynamic monitoring main equipment and sensors of the SRDC dynamic monitoring device include a monitoring machine, front and rear temperature and humidity sensors, smoke detectors, and temperature detectors. The iView software displays the monitored information on the display screen, including temperature, humidity, smoke, UPS, and air conditioning information. Server node information includes firmware version, driver version, and BMC\BIOS parameters. Use Java, Python, C, and other programming languages ​​to write automatic parameter acquisition scripts and obtain corresponding data through the Redfish interface. Redfish is a modern, open standard RESTful (Representational State Transfer) API used to manage and monitor hardware devices in data centers, such as servers, storage, and network devices. It can obtain hardware device information, OS information, and driver information. Switch information includes version information and configuration parameters. The corresponding data is obtained through the SNMP management tool.

[0108] The allocation module 12 divides the acquired whole cabinet data into two categories: "range data" and "fixed value data" to narrow the data retrieval scope. Range data refers to parameter data that needs to meet a certain threshold range during compliance verification, such as temperature value, humidity value, smoke value, temperature value, UPS status value (UPS input voltage, output voltage), air conditioning status value (set temperature value) and other whole cabinet information. The compliance judgment basis for these data is a preset range value. Fixed value data refers to parameter data that needs to be completely consistent with the preset fixed standard value during compliance verification, such as firmware version, driver version, BMC\BIOS parameters and other server node information, as well as version information, configuration parameters and other switch information. The compliance judgment basis is compared with the unpreset fixed value.

[0109] The purpose of classifying data into range data and fixed value data is to adopt different verification strategies according to the characteristics of different types of data in the subsequent compliance verification. For range data, verify whether it is within a reasonable range; for fixed value data, verify whether it is completely consistent with the nominal value. This classification method can more accurately identify data compliance issues and avoid the errors and inapplicability that may be caused by adopting a single verification standard for all data.

[0110] The data storage module 2 is used to store the above-mentioned range data and the above-mentioned fixed value data in a preset format in an open source database system based on distributed file storage, and store the whole cabinet nominal data of the above-mentioned all-in-one machine and the original file of the firmware version involved in the above-mentioned whole cabinet nominal data in the above-mentioned open source database system based on distributed file storage.

[0111] Specifically, Figure 4 This is the structural diagram of the data storage module, such as Figure 4As shown, the data storage module 2 includes a collection data storage module 21 and a nominal data storage module 22. In order to unify the storage format and improve the data processing efficiency, the collection data storage module 21 stores the range data and the fixed value data in a preset format in an open source database system based on distributed file storage. In the embodiment of the present application, the range data and the fixed value data can be stored in a MongoDB database in JSON (JavaScript Object Notation) format. MongoDB is an open source database system based on distributed file storage, which aims to provide a scalable high-performance data storage solution for Web applications. The data structure supported by MongoDB is very flexible. It uses a BSON format similar to JSON to store data. BSON is a binary format of JSON. It not only supports all data types of JSON, but also supports more data types, such as dates, binary large objects (BLOBs), custom types, etc. Due to the flexibility of the data structure, MongoDB is very suitable for storing complex and hierarchical data, such as nested documents, arrays, etc., which makes it very effective in processing unstructured or semi-structured data. JSON is a lightweight data exchange format that is easy for humans to read and write, and easy for machines to parse and generate. It is very suitable for storing and transmitting structured data.

[0112] In MongoDB, data is usually stored in the form of documents, each of which is a JSON object. To ensure fast retrieval and management of data, the system will design the database table structure, use the timestamp as the primary key, and create corresponding fields according to the data type (such as temperature, firmware version, etc.), and store the actual data and the obtained return value in these fields. This design makes each record unique and can maintain the time series of the data, which is convenient for tracing and analyzing historical data.

[0113] In addition to obtaining the whole cabinet data of the integrated machine, it is also necessary to collect the nominal data of the whole cabinet of the integrated machine. As an important reference standard for measuring data compliance and normal operation of the equipment, the nominal data covers key information such as standard parameters, performance indicators and expected data values ​​of the equipment. In addition, these nominal data are stored in association with the corresponding firmware version original files. Each record contains complete nominal data information and corresponding firmware version original file information, where the firmware version original file information includes important attributes such as version number, release date, and file hash value. This storage method provides comprehensive and reliable data support for data compliance verification, equipment maintenance and upgrade, and troubleshooting, ensuring that the system can perform effective data processing and equipment management based on accurate nominal data during operation.

[0114] Similar to the whole cabinet data, the nominal data storage module 22 stores the nominal data and the original firmware version files in the MongoDB database in JSON format. The difference is that the storage of these data focuses more on the integrity and relevance of the documents. Each record not only contains the detailed information of the nominal data, but also associates the corresponding original firmware version file, ensuring that accurate nominal data and related firmware version information can be quickly accessed when needed.

[0115] The data consistency check module 3 is used to apply the above-mentioned whole cabinet nominal data to perform consistency check on the above-mentioned range data and the above-mentioned fixed value data, to determine whether the above-mentioned range data conforms to the nominal range, and to determine whether the above-mentioned fixed value data conforms to the nominal value, and to perform deviation correction on the range data that does not conform to the above-mentioned nominal range and the fixed value data that does not conform to the above-mentioned nominal value.

[0116] Specifically, nominal data refers to the performance indicators, parameter values ​​or data ranges that the SRDC cabinet or its components (such as server nodes, switches, etc.) should achieve under normal operating conditions. These data are usually defined in the equipment specifications, factory configurations or industry standards. For example, the nominal range can be the temperature threshold of the equipment, while the nominal value can be a specific firmware or software version.

[0117] Figure 5 The structure diagram of the data consistency verification module is shown in Figure 1. Figure 5 As shown, the data conformance check module 3 includes a data check module 31 and a data deviation adjustment module 32. The data check module 31 uses the nominal data to perform conformance check on the range data. This usually involves numerical comparison to check whether the range data is within the nominal range. For example, if the nominal temperature range is 15°C to 30°C, any temperature reading outside this range will be considered non-conforming. Secondly, the fixed value data is checked using the nominal value, which involves precise value matching to check whether the data is completely consistent with the nominal value.

[0118] When it is found that the data does not meet the nominal range or nominal value, the data deviation adjustment module 32 will perform deviation correction, that is, adjust the current data to meet the nominal requirements. For range data, the operating parameters of the device can be adjusted to bring the actual reading back to the nominal range; for fixed value data, the firmware or software can be updated to the nominal version.

[0119] The data synchronization module 4 is used to analyze the weights of factors affecting the nominal data of the entire cabinet after adjusting the range data that does not conform to the above nominal range and the fixed value data that does not conform to the above nominal value, generate compliance verification data recommendations, and send the compliance verification data recommendations to the user.

[0120] Specifically, Figure 6The structure diagram of the data synchronization module is shown in Figure 1. Figure 6 As shown, the data synchronization module 4 includes a data recording module 41 and an intelligent module 42, wherein:

[0121] The data recording module 41 obtains the binary log data written in real time to maintain the continuity of the writing business. After the whole cabinet data adjustment task is completed, the whole cabinet data recovery point is created by creating a version snapshot, and the incremental synchronization operation of the database is performed. The adjustment data record in the data adjustment module is stored in the MongoDB database with the timestamp as the table primary key, and the historical data record is retained for subsequent reference.

[0122] Based on the automated machine learning tool AutoGluon, the intelligent module 42 generates compliance verification data recommendations for user selection. First, multiple factors that may affect the nominal data of the whole cabinet are identified. The factors that affect the nominal data of the whole cabinet mainly include: current nominal data, compliance standards, user requirements, regional differences in the SRDC whole cabinet (such as natural conditions such as ambient temperature and humidity), data fluctuations (such as instantaneous changes in system performance), etc. The identified factors are converted into a computable data form, usually through data preprocessing (cleaning, normalization, encoding, etc.) to facilitate subsequent analysis and machine learning model training. Next, the impact of these factors on the nominal data is analyzed. Weight analysis is to determine which factors have a greater impact on data compliance by training the model, which usually involves feature selection and importance ranking of the model. According to the analyzed factor weights, the model will be further trained and optimized to improve its accuracy and efficiency in recommending compliance verification data.

[0123] The trained model will generate a series of compliance verification data recommendations based on the current nominal data and the weights of the influencing factors. These recommendations can include recommended firmware updates, adjustments to environmental control parameters, and optimized equipment configurations to help the data of the entire cabinet and its components closer to the nominal data. The generated compliance verification data recommendations are then sent to the user. After the user receives and implements the recommended adjustments, user feedback is collected, including the effectiveness of the adjustment results, changes in equipment performance, etc. This feedback information will be input into the model again to further optimize model performance and improve the accuracy of future recommendations.

[0124] The data output module 5 is used to display the operation flow, processing progress and processing results of the above data processing method in the form of a page, and provide a version data information query function of the above all-in-one machine.

[0125] Specifically, after the user receives the compliance verification data recommendation, the user's feedback information is collected and sent back to the model training layer to further optimize the model performance and recommendation accuracy. This part focuses on the transparency of the data processing method and user interaction. By showing the specific operation process, processing speed and final processing results of data processing in a regulated manner, the operability and user-friendliness of the system are increased. At the same time, the system also provides a version data information query function for the all-in-one machine (SRDC whole cabinet), allowing users to easily track and manage the update history of the device.

[0126] Specifically, a user interface is designed to display the entire process of data processing in the form of an HTML page. This includes the operation details of each stage, such as data collection, classification, storage, conformance verification, deflection correction, synchronization, and intelligent learning. The user interface can be in the form of charts, flow charts, timelines, etc., to clearly and intuitively present each step of data processing. The page also provides a real-time tracking function for processing progress, so that users can view the current processing steps in progress and the completion status of each step at any time. This is crucial for monitoring the efficiency and status of data processing, especially when processing large-scale data or performing complex conformance verification. The final processing results, including conformance verification results, deflected data, recommended conformance standards, etc., will be displayed to users on the page in the form of clear lists, charts, or reports. This enables users to quickly understand the results of data processing and the possible impact on the system or device.

[0127] The version data information query function of the all-in-one machine allows users to query the data information of different versions of the SRDC cabinet, including but not limited to device configuration, firmware version, performance parameters, etc. The version information query function is very important for troubleshooting, maintenance and upgrades, and historical data analysis. The user interface can also include a version comparison function, allowing users to compare the differences between different versions, which is helpful for understanding the improvement or degradation of device performance and changes in compliance standards. The query function is not limited to the version information of the entire cabinet, but can also query the detailed data parameters of specific server nodes, switches and other components, providing users with a comprehensive view of the device status.

[0128] In one embodiment of the present application, the data verification module includes a calculation submodule, a first judgment submodule, a second judgment submodule and a third judgment submodule, wherein:

[0129] The calculation submodule is used to adopt the first formula Calculate the first deviation using the second formula Calculate the second deviation, where D1 is the first deviation, which indicates the deviation between the range data and the lower limit of the nominal data interval, D2 is the second deviation, which indicates the deviation between the range data and the upper limit of the nominal data interval, x is the range data, a is the lower limit of the nominal data interval, and b is the upper limit of the nominal data interval;

[0130] The first judgment submodule is used for determining that when the first deviation is 0 and the second deviation is 0, the range data meets the nominal range;

[0131] The second judgment submodule is used for determining that when the first deviation is not 0 and the second deviation is 0, the range data does not conform to the nominal range;

[0132] The third judgment submodule is used for judging that when the first deviation is 0 and the second deviation is not 0, the range data does not conform to the nominal range.

[0133] Specifically, for range data, the deviation is calculated to check whether it meets the nominal range. Calculate the first deviation. The first deviation D1 represents the degree of deviation between the range data x and the lower limit a of the nominal data interval. When the range data x is equal to the lower limit a of the nominal data interval, the first deviation D1 = 0, that is, there is no lower limit deviation. When the range data x is less than the lower limit a of the nominal data interval, the first deviation D1 = 0, because the deviation calculates the percentage of data exceeding the range. Here, the range data x does not exceed the lower limit a of the nominal data interval, so the deviation is 0. Through the second formula The second deviation is calculated. The second deviation D2 represents the degree of deviation between the range data x and the upper limit b of the nominal data interval. When the range data x is equal to the upper limit b of the nominal data interval, the second deviation D2 = 0, that is, there is no upper limit deviation. When the range data x is greater than the upper limit b of the nominal data interval, the second deviation D2 = 0.

[0134] If the first deviation D1 and the second deviation D2 are both 0, it means that the range data x is neither lower than the lower limit a of the nominal data interval nor higher than the upper limit b of the nominal data interval. Therefore, the range data x conforms to the nominal data interval [a, b], and the data is legal, without the need for deviation correction.

[0135] If the first deviation D1 is not 0 but the second deviation D2 is 0, it means that the range data x deviates only from the lower limit a of the nominal data interval, that is, the range data x is lower than the lower limit a of the nominal data interval. At this time, the data is non-compliant and needs to be adjusted to bring the range data x back into the interval.

[0136] If the first deviation D1 is 0 but the second deviation D2 is not 0, it means that the range data x deviates from the upper limit b of the nominal data interval, that is, the range data x exceeds the upper limit b of the nominal data interval. The data is also non-compliant and also needs to be adjusted and corrected to bring the range data x back into the interval.

[0137] For range data that does not conform to the nominal range, a deflection correction operation is performed based on the direction and degree of deviation. If the deviation is due to the range data being below the lower limit a of the nominal data interval, the operating parameters of the device will be adjusted, such as increasing the power of the heater to increase the temperature reading; if the range data exceeds the upper limit b of the nominal data interval, the power consumption of the device can be reduced, such as adjusting the settings of the cooling system to reduce the temperature reading. The goal of deflection correction is to make the range data as close to the normal range of the nominal data as possible.

[0138] In another embodiment of the present application, the data verification module includes a comparison submodule, a fourth judgment submodule and a third judgment submodule, wherein:

[0139] The comparison submodule is used to establish a one-to-one correspondence between the fixed value data and the nominal data by using data mapping technology, and compare the fixed value data with the nominal data to obtain a third deviation;

[0140] The fourth judgment submodule is used for determining that when the third deviation is 0, the fixed value data meets the nominal value;

[0141] The seventh judgment submodule is used for determining that when the third deviation is not 0, the fixed value data does not conform to the nominal value.

[0142] Specifically, for fixed value data, data mapping technology is used to compare and verify whether it conforms to the nominal value. First, a one-to-one correspondence is established between the fixed value data and its corresponding nominal data. This means that each fixed value data (such as the firmware version of the server node, BIOS version, etc.) has a clear corresponding nominal data, which is defined according to the design specifications, industry standards or factory configuration of the SRDC cabinet. For example, if the firmware version of the server node is nominally 2.1.0, then data mapping will ensure that the actual collected firmware version data can be directly compared with this nominal value.

[0143] Once the one-to-one correspondence is established, the next step is to compare the collected fixed value data with the nominal data. The purpose of the comparison is to check whether the actual data completely matches the nominal value. The third deviation D3 is used to quantify the difference between the fixed value data and the nominal data. If the comparison result of the fixed value data and the nominal data shows that the data is completely consistent (that is, the actual data is the same as the nominal data), then the third deviation D3 is set to 0, indicating that there is no deviation and the data meets the nominal value. Conversely, if the comparison result of the fixed value data and the nominal data shows that the data is inconsistent, the third deviation D3 is set to a non-zero value, indicating the deviation between the fixed value data and the nominal data.

[0144] If the third deviation D3 is 0, it means that the fixed value data is completely consistent with the nominal data, the data is legal, and no deviation correction is required. On the contrary, if the third deviation D3 is not 0, it means that there is a difference between the fixed value data and the nominal data, the data does not conform to the nominal value, and deviation correction is required to restore the data compliance.

[0145] For fixed value data with the third deviation D3 not equal to 0, a deviation correction operation is performed, which involves updating the device's firmware, driver, or software settings to ensure that the actual data is consistent with the nominal data. For example, if the firmware version of the server node does not match the nominal value, the firmware update process will be automatically triggered to download the correct version from the standard firmware file stored in the MongoDB database and install it on the server node, thereby eliminating the third deviation D3 and restoring the data to compliance.

[0146] In a specific embodiment, the data deflection adjustment module includes a first deflection adjustment correction submodule and a second deflection adjustment correction submodule, wherein:

[0147] The first deviation correction submodule is used for, when the first deviation is not 0 and the second deviation is 0, using a third formula x′=a+(xa)×k to perform deviation correction on the range data that does not meet the above nominal range, wherein x′ is the range data after deviation correction, k is the deviation coefficient, and 0<k≤1;

[0148] The second deviation correction submodule is used to perform deviation correction on the above range data using the fourth formula x′=b-(xb)×k when the above first deviation is 0 and the above second deviation is not 0, wherein x′ is the value after deviation correction, k is the deviation coefficient, and 0<k≤1.

[0149] Specifically, when the first deviation is not 0 and the second deviation is 0, it means that the range data x is lower than the lower limit a of the nominal data interval, but does not exceed the upper limit b of the nominal data interval. At this time, the range data needs to be adjusted upward to reach the nominal range. The third formula x′=a+(xa)×k is used to perform deviation correction to ensure that the range data x falls within the nominal range after adjustment. The adjustment process uses the deviation coefficient k, and the value of the deviation coefficient k can be determined based on historical data, device characteristics, or a preset deviation strategy to ensure that the impact of the adjustment on device performance is minimized.

[0150] When the first deviation is 0 and the second deviation is not 0, it means that the range data x exceeds the upper limit b of the nominal data interval, but is not lower than the lower limit a of the nominal data interval. At this time, the range data needs to be adjusted downward to meet the nominal range. The fourth formula x′=b+(xb)×k is used to perform the deviation correction to ensure that the range data x meets the nominal range after adjustment.

[0151] The offset coefficient k is a key parameter used to control the adjustment range during the offset correction process. The offset coefficient k can be set through the web page of the SRDC dynamic monitoring display to achieve offset correction. The value of the offset coefficient k can be adjusted according to different situations and device characteristics. For example, when processing temperature data, the setting of the offset coefficient k may need to consider factors such as the thermal stability of the device and the rate of change of the ambient temperature to avoid excessive adjustments that may cause unstable device performance.

[0152] In a specific embodiment, the data deflection module includes a third deflection correction submodule and an update submodule, wherein:

[0153] The third deviation correction submodule is used for, when the third deviation is not 0, performing deviation correction on the fixed value data that does not meet the nominal value by means of a representational state transfer application programming interface;

[0154] The update submodule is used to call the standard firmware in the open source database system based on distributed file storage to update when the third deviation is not 0 due to inconsistent firmware versions.

[0155] Specifically, for those fixed value data whose third deviation is not 0, that is, data that is inconsistent with the nominal data, these data are adjusted by calling the representational state transfer application programming interface. In this embodiment, the adjustment is performed by calling the RESTful API. The RESTful API allows the system to interact with different devices in a standard and unified way to achieve data acquisition and modification. For example, if the BIOS version of the server node does not conform to the nominal data, the system can call the RESTful API to access and update the BIOS settings.

[0156] The RESTful API call is automated, which means that it can automatically identify which fixed-value data does not meet the nominal value and automatically perform deviation correction operations without manual intervention. This process can include updating software settings, firmware versions, configuration parameters, etc. to ensure that the data is restored to the nominal range.

[0157] This is detected when the third deviation is not 0 and the data is inconsistent due to inconsistent firmware versions. Inconsistent firmware versions may affect the performance of the device and the accuracy of the data, so an update is required.

[0158] Call the standard firmware version stored in the MongoDB database for update. As an open source database system based on distributed file storage, MongoDB can store and manage firmware version files of various devices, including version number, release date, file hash value and other information. When the firmware needs to be updated, the standard firmware version of the corresponding device can be retrieved from MongoDB and the update process can be automatically executed.

[0159] The firmware update process is implemented through automated scripts or built-in firmware update tools to ensure that the update is safe, fast and will not have adverse effects on the device. After the update is completed, the system will perform a data consistency check again to confirm whether the update is successful and whether the data has met the nominal value.

[0160] In another embodiment of the present application, the intelligent module includes a preprocessing submodule, a construction submodule, a normalization processing submodule, a first optimization submodule and an output submodule, wherein:

[0161] The preprocessing submodule is used to use the preprocessing module built into the above-mentioned automatic machine learning tool to perform data cleaning, normalization and encoding operations on the factors affecting the above-mentioned nominal data of the entire cabinet to obtain a preprocessed data set;

[0162] The construction submodule is used to construct the influencing factor matrix A = (a ij ), where a ij Indicates the importance of factor i relative to factor j. Use a 1-9 scale to assign values ​​to the factors in the above influencing factor matrix, where 1 means equally important and 9 means absolutely important.

[0163] The normalization processing submodule is used to normalize the geometric mean of each row of elements in the above-mentioned influencing factor matrix to obtain a set of weight vectors. Each element in the above-mentioned weight vector represents the priority ratio of the corresponding factor.

[0164] The first optimization submodule is used to adjust the weight of the training data input into the deep learning-based automatic machine learning software library model according to the weight vector, and use the deep learning algorithm to train the training data to identify the inherent pattern of the training data, and optimize the recognition and judgment criteria of the automatic machine learning software library model for the compliance features;

[0165] The output submodule is used to output the above-mentioned compliance verification data recommendation according to the training results of the above-mentioned automatic machine learning software library model.

[0166] Specifically, first, the preprocessing module built into the automated machine learning tool is used to perform data cleaning, normalization, and encoding operations on various factors that affect the nominal data of the entire cabinet. In the embodiment of the present application, the preprocessing module built into AutoGluon is used to perform data cleaning, normalization, and encoding operations on various factors that affect the nominal data of the entire cabinet. AutoGluon is an open source automated machine learning software library that is easy to use and highly automated, and can automatically perform key steps such as model selection and hyperparameter adjustment. Data cleaning is intended to remove invalid or abnormal data to ensure the accuracy of analysis; normalization processing enables data of different magnitudes or dimensions to be compared on a unified scale; encoding operations convert non-numeric data into machine-readable numerical forms to facilitate subsequent analysis and modeling.

[0167] The preprocessed data set is more analytical and practical, and can be used as the basis for constructing an influencing factor matrix. Construct an influencing factor matrix A = (a ij ), each element a in the influencing factor matrix ij Indicates the importance of factor i relative to factor j. These factors can be quantified using a 1-9 scale, where 1 means that both factors are equally important and 9 means that one factor is absolutely more important than the other. The 1-9 scale is a subjective assignment method, usually derived by experts in the field or through statistical methods such as cluster analysis and factor analysis. The assignment process needs to comprehensively consider the direct and indirect effects of factors, as well as the interactions between various factors.

[0168] The geometric mean is calculated for each row of the influencing factor matrix, which can better reflect the comprehensive balance of the importance of each factor. Then, the calculated geometric mean is normalized and converted into elements in the weight vector, which represents the priority of each factor. The weight vector can more intuitively reflect the relative importance of each factor in the overall analysis. The weight vector is applied to the model training process to adjust the weight of the training data input into the deep learning-based AutoGluon model. This makes the model training pay more attention to those factors that have a greater impact on the nominal data of the entire cabinet, thereby improving the prediction accuracy and practicality of the model. The deep learning algorithm is used to learn the training data, identify the inherent patterns of the training data, and optimize the recognition and judgment criteria of the automatic machine learning software library model for the matching features. Deep learning can automatically extract features from complex data sets and achieve high-level abstraction and pattern recognition of data through multi-level neural networks.

[0169] Based on the training results of the model, a set of consistency verification data recommendations are output. These recommendations may include the recommended adjustment of the offset coefficient k, the optimization suggestions for device configuration, the priority sorting of firmware updates, etc., aiming to help users more effectively perform data consistency verification and maintenance and optimization of the entire cabinet.

[0170] In a specific embodiment, the above-mentioned intelligent module includes a second optimization sub-module, and the above-mentioned second optimization sub-module is used to collect feedback on the above-mentioned compliance verification data recommendation, and transmit the feedback data back to the training layer of the above-mentioned automatic machine learning software library model to optimize the above-mentioned automatic machine learning software library model.

[0171] Specifically, collect user feedback on the output conformity verification data recommendations. Feedback can be the result of the user's manual deflection correction, such as whether the deflection is successful, whether the corrected data is stable, whether the device performance is affected, etc. It can also be the user's evaluation of the conformity verification data recommendation itself, such as the accuracy and practicality of the recommendation. User feedback can include positive affirmation, such as the correction value recommended by the conformity verification data does solve the problem of data inconsistency; it may also include negative feedback, such as the correction value is not applicable, caused new problems or did not achieve the expected effect. Regardless of positive or negative, these feedbacks are important information for model optimization.

[0172] The collected feedback data will be integrated and preprocessed to ensure that it is compatible with the original training data and meets the requirements of model training. The preprocessed feedback data is sent back to the training layer of the automatic machine learning software library model. The training layer is the core part of model learning and optimization. It performs feature extraction, parameter adjustment and model training based on the input data to improve the model's prediction ability. The feedback data sent back will be used as additional training samples by the model for retraining the model. The model will adjust its internal parameters based on these new samples to optimize its prediction and recommendation of the compliance verification data.

[0173] By continuously collecting and learning user feedback, the model can gradually improve its generalization ability, that is, make accurate predictions and recommendations on unseen data. This is very important for dealing with new data inconsistencies that may arise in the SRDC cabinet, making the system more flexible and adaptable.

[0174] In a specific embodiment, the normalization processing submodule is used to adopt the formula Calculate the geometric mean of each row of the influencing factor matrix above, where: is the geometric mean, n is the number of columns in the influencing factor matrix; the geometric mean is normalized to obtain the weight vector w = (w1, w2, ..., w n ),in, w i is the priority weight of factor i.

[0175] Specifically, each element in the weight vector represents the priority ratio of the corresponding factor, which is crucial for the intelligent analysis and recommendation of the consistency verification of the whole cabinet data. Calculate the geometric mean of each row of the influencing factor matrix. The geometric mean calculation method takes into account the influence of all factors, rather than just their simple average, which helps better reflect the relative importance of each factor. Especially when the importance of each factor varies greatly, the geometric mean can more accurately reflect the comprehensive weight of the factors.

[0176] The calculated geometric mean of each row is normalized to convert the values ​​of all elements to the range between 0 and 1, so that the elements in the weight vector can directly reflect the relative importance of each factor. After normalization, the weight vector w = (w1, w2, ..., w n ),in, w i is the priority weight of factor i.

[0177] Through normalization, the magnitude differences between different factors can be eliminated, ensuring that each element in the weight vector is compared on the same scale, making it easier for the model to understand and use this information for learning and optimization. n ) i It represents the priority ratio of the corresponding factor, and can directly reflect the relative importance of this factor in affecting the compliance of the nominal data of the entire cabinet.

[0178] The weight vector is used to adjust the weight of the training data input into the deep learning-based automatic machine learning software library model, ensuring that the model pays more attention to the factors with greater influence during the training process, thereby optimizing the model's recognition and judgment criteria for compliance features. In addition, the weight vector is not static. As the system runs, it can be dynamically adjusted based on user feedback, data changes, and model training results. This mechanism can ensure that the model always pays attention to the most relevant factors at the moment, improve the model's adaptability and prediction accuracy, and ensure that the model always maintains the optimal state.

[0179] In another embodiment of the present application, the data output module is used to display the operation flow, processing progress and processing results of the data processing method in the form of a page, and provide a version data information query function of the all-in-one machine.

[0180] Specifically, after the user receives the compliance verification data recommendation, the user's feedback information is collected and sent back to the model training layer to further optimize the model performance and recommendation accuracy. This part focuses on the transparency of the data processing method and user interaction. By showing the specific operation process, processing speed and final processing results of data processing in a regulated manner, the operability and user-friendliness of the system are increased. At the same time, the system also provides a version data information query function for the all-in-one machine (SRDC whole cabinet), allowing users to easily track and manage the update history of the device.

[0181] Specifically, a user interface is designed to display the entire process of data processing in the form of an HTML page. This includes the operation details of each stage, such as data collection, classification, storage, conformance verification, deflection correction, synchronization, and intelligent learning. The user interface can be in the form of charts, flow charts, timelines, etc., to clearly and intuitively present each step of data processing. The page also provides a real-time tracking function for processing progress, so that users can view the current processing steps in progress and the completion status of each step at any time. This is crucial for monitoring the efficiency and status of data processing, especially when processing large-scale data or performing complex conformance verification. The final processing results, including conformance verification results, deflected data, recommended conformance standards, etc., will be displayed to users on the page in the form of clear lists, charts, or reports. This enables users to quickly understand the results of data processing and the possible impact on the system or device.

[0182] The version data information query function of the all-in-one machine allows users to query the data information of different versions of the SRDC cabinet, including but not limited to device configuration, firmware version, performance parameters, etc. The version information query function is very important for troubleshooting, maintenance and upgrades, and historical data analysis. The user interface can also include a version comparison function, allowing users to compare the differences between different versions, which is helpful for understanding the improvement or degradation of device performance and changes in compliance standards. The query function is not limited to the version information of the entire cabinet, but can also query the detailed data parameters of specific server nodes, switches and other components, providing users with a comprehensive view of the device status.

[0183] It should be noted that the above modules can be implemented by software or hardware. For the latter, it can be implemented in the following ways, but not limited to: the above modules are all located in the same processor; or the above modules are located in different processors in any combination.

[0184] An embodiment of the present application also provides a computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to execute the steps of any of the above method embodiments when running.

[0185] In some exemplary embodiments, the computer-readable storage medium may include, but is not limited to, various media that can store computer programs, such as a USB flash drive, a read-only memory (ROM), a random access memory (RAM), a removable hard disk, a magnetic disk, or an optical disk.

[0186] Obviously, those skilled in the art should understand that the above modules or steps of the present application can be implemented by a general computing device, they can be concentrated on a single computing device, or distributed on a network composed of multiple computing devices, they can be implemented by a program code executable by a computing device, so that they can be stored in a storage device and executed by the computing device, and in some cases, the steps shown or described can be executed in a different order than here, or they can be made into individual integrated circuit modules, or multiple modules or steps in them can be made into a single integrated circuit module for implementation. In this way, the present application is not limited to any specific combination of hardware and software.

[0187] The above is only a preferred embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the present application may have various modifications and changes. Any modification, equivalent replacement, improvement, etc. made within the principles of the present application shall be included in the protection scope of the present application.

Claims

1. A data processing method for an all-in-one machine, characterized in that: include: Acquire the whole cabinet data of the integrated machine, and divide the whole cabinet data into range data and fixed value data, wherein the range data refers to data with a threshold range for parameter data compliance verification, and the fixed value data refers to data with a fixed value requirement for parameter data compliance verification; The range data and the fixed value data are stored in a preset format in an open source database system based on distributed file storage, and the whole cabinet nominal data of the all-in-one machine and the original file of the firmware version involved in the whole cabinet nominal data are stored in the open source database system based on distributed file storage; Apply the whole cabinet nominal data to perform compliance check on the range data and the fixed value data, determine whether the range data conforms to the nominal range, and determine whether the fixed value data conforms to the nominal value, and perform deviation correction on the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value; After the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value are adjusted and corrected, the weights of the factors affecting the nominal data of the entire cabinet are analyzed, compliance verification data recommendations are generated, and the compliance verification data recommendations are sent to the user.

2. The method according to claim 1, characterized in that Applying the whole cabinet nominal data to perform compliance check on the range data and the fixed value data, determining whether the range data complies with the nominal range, and determining whether the fixed value data complies with the nominal value, includes: Using the first formula Calculate the first deviation using the second formula Calculate the second deviation, where D1 is the first deviation, The first deviation represents the deviation between the range data and the lower limit of the nominal data interval, D2 is the second deviation, the second deviation is the deviation between the range data and the upper limit of the nominal data interval, × is the range data, a is the lower limit of the nominal data interval, and b is the upper limit of the nominal data interval; When the first deviation is 0 and the second deviation is 0, the range data conforms to the nominal range; when the first deviation is not 0 and the second deviation is 0, the range data does not conform to the nominal range; When the first deviation is 0 and the second deviation is not 0, the range data does not conform to the nominal range.

3. The method according to claim 1, characterized in that Applying the whole cabinet nominal data to perform compliance check on the range data and the fixed value data, determining whether the range data complies with the nominal range, and determining whether the fixed value data complies with the nominal value, includes: Using data mapping technology to establish a one-to-one correspondence between the fixed value data and the nominal data, and comparing the fixed value data with the nominal data to obtain a third deviation; When the third deviation is 0, the fixed value data conforms to the nominal value; When the third deviation is not 0, the fixed value data does not conform to the nominal value.

4. The method according to claim 2, characterized in that: Performing deviation correction on the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value, including: When the first deviation is not 0 and the second deviation is 0, the third formula x′=a+(xa)×k is used to perform deviation correction on the range data that does not meet the nominal range, wherein x′ is the range data after deviation correction, k is the deviation coefficient, and 0<k≤1; When the first deviation is 0 and the second deviation is not 0, the fourth formula x′=b-(xb)×k is used to perform deviation correction on the range data, wherein x′ is the value after deviation correction, k is the deviation coefficient, and 0<k≤1.

5. The method according to claim 3, characterized in that: Performing deviation correction on the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value, including: When the third deviation is not 0, adjusting and correcting the fixed value data that does not conform to the nominal value by means of a representational state transfer application programming interface; In the case where the third deviation is not 0 due to inconsistent firmware versions, the standard firmware in the open source database system based on distributed file storage is called to update.

6. The method according to claim 1, characterized in that After the range data that does not conform to the nominal range and the fixed value data that does not conform to the nominal value are adjusted and corrected, the weights of the factors affecting the nominal data of the entire cabinet are analyzed to generate compliance verification data recommendations, including: Using a preprocessing module built into an automated machine learning tool to perform data cleaning, normalization, and encoding operations on factors affecting the nominal data of the entire cabinet, to obtain a preprocessed data set; Construct the influencing factor matrix A = (a ij ), where a ij Indicates the importance of factor i relative to factor j, using a 1-9 scale to assign values ​​to the factors in the influencing factor matrix, 1 indicates equal importance, and 9 indicates absolute importance; normalizes the geometric mean of the elements in each row of the influencing factor matrix to obtain a set of weight vectors, each element in the weight vector indicates the priority ratio of the corresponding factor; The weight of the training data input into the automatic machine learning software library model based on deep learning is adjusted according to the weight vector, and the training data is trained and learned using a deep learning algorithm to identify the inherent patterns of the training data and optimize the automatic machine learning software library model's recognition and judgment criteria for compliance features; based on the training results of the automatic machine learning software library model, the compliance verification data recommendation is output.

7. The method according to claim 6, characterized in that After outputting the compliance verification data recommendation, the method further includes: Collect feedback on the compliance verification data recommendations, and transmit the feedback data back to the training layer of the automatic machine learning software library model to optimize the automatic machine learning software library model.

8. The method according to claim 6, characterized in that A set of weight vectors is obtained by normalizing the geometric mean of the elements in each row of the influencing factor matrix, and each element in the weight vector represents the priority ratio of the corresponding factor, including: Using the fifth formula Calculate the geometric mean of each row of elements in the influencing factor matrix, where: is the geometric mean, and n is the number of columns of the influencing factor matrix; The geometric mean is normalized to obtain the weight vector W=(W1, W2, ..., W n ),in, W i is the priority weight of factor i.

9. The method according to claim 1, characterized in that: The method further comprises: When the nominal data is updated due to design requirements, the updated nominal data is stored in the open source database system based on distributed file storage in the preset format; the updated nominal data is incremented Synchronize and process to obtain new synchronized data; use an automated machine learning tool to analyze the new synchronized data, update the knowledge of compliance features, and periodically generate compliance verification data recommendations to reflect the compliance standards under the latest design requirements; the method also includes: The operation flow, processing progress and processing results of the data processing method are displayed in the form of a page, and the The version data information query function of the all-in-one machine is described.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program, when executed by a processor, implements the steps of the data processing method of the all-in-one machine described in any one of claims 1 to 9.

Images