A device performance data management method and device, computer device and storage medium

Abstract

The application relates to a device performance data management method and device, computer equipment and a storage medium. The method comprises the following steps: determining a device driver interface and an architecture-specific index according to a device architecture type; obtaining a platform general index, combining the architecture-specific index to obtain an index conversion relationship table; obtaining device original performance data through the device driver interface; processing the device original performance data into abstract performance data through a preset driver isolation abstraction layer; based on the index conversion relationship table, analyzing and standardizing the abstract performance data to obtain platform standard performance data, and storing the platform standard performance data in a database. The method can realize unified standard data collection and management of a heterogeneous GPU, quickly respond to customer needs, improve resource utilization, and improve the driving compatibility and stability of the system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of device performance data management, in particular to a device performance data management method and device, computer equipment and a storage medium. BACKGROUND

[0002] With the growth of artificial intelligence and high-performance computing needs, different manufacturers, such as NVIDIA, AMD, Intel, etc., have launched various different types of GPUs, which have different architectures and performance characteristics, and higher requirements for data management and performance monitoring.

[0003] In the prior art, due to the lack of unified standards and general monitoring data acquisition methods, it is difficult to achieve uniformity in data management and performance monitoring of heterogeneous GPUs, which requires directional adaptation development, not only long development cycle, but also unable to quickly respond to customer needs, and in the adaptation process, it will cause GPU idle, causing resource waste; at the same time, the directional adaptation development data management system used in the prior art usually uses monitoring scripts directly dependent on GPU driver interface, when the GPU driver interface is upgraded, it is difficult to realize timely update response, resulting in driver compatibility problems, thus easily leading to the entire data management system unusable, affecting system stability. SUMMARY

[0004] Therefore, it is necessary to provide a device performance data management method, device, computer equipment and storage medium capable of realizing unified standard data collection and management of heterogeneous GPUs, quickly responding to customer needs, improving resource utilization, improving system driver compatibility and stability.

[0005] In one aspect, a device performance data management method is provided, the method comprising:

[0006] determining a device driver interface and architecture-specific indicators according to a device architecture type;

[0007] obtaining platform general indicators, combining the architecture-specific indicators to obtain an indicator conversion relationship table;

[0008] obtaining device raw performance data through the device driver interface;

[0009] processing the device raw performance data into abstract performance data through a preset driver isolation abstraction layer;

[0010] based on the indicator conversion relationship table, analyzing and standardizing the abstract performance data to obtain platform standard performance data, and storing the platform standard performance data in a database.

[0011] In one of the embodiments, the platform general index is acquired, and the index conversion relationship table is obtained in combination with the architecture-specific index, wherein the platform general index is obtained by the following method:

[0012] The historical device usage records are acquired, and all used devices are marked;

[0013] According to the used device, the corresponding device architecture is determined, and the used architecture index is obtained;

[0014] The commonality of the used architecture index is extracted to obtain the basic general index;

[0015] The feature performance index is obtained by performing feature performance index screening and analysis on the extracted used architecture index;

[0016] According to the basic general index, the platform general index is obtained in combination with the feature performance index.

[0017] In one of the embodiments, the platform general index is acquired, and the index conversion relationship table is obtained in combination with the architecture-specific index, including:

[0018] According to the architecture-specific index, the platform general index is checked to determine whether the architecture-specific index is an already recorded index;

[0019] In response to the determination result being no, the architecture-specific index is subjected to index conversion processing to obtain a specific conversion index, and the specific conversion index is supplemented to the platform general index;

[0020] In response to the determination result being yes, the conversion relationship between the architecture-specific index and the platform general index is determined, and the index conversion relationship table is constructed based on the conversion relationship.

[0021] In one of the embodiments, the device original performance data is processed into abstract performance data through a preset driving isolation abstraction layer, wherein the driving isolation abstraction layer performs the following steps:

[0022] The device original performance data transmitted by the device driving interface is received;

[0023] The platform general index and the index conversion relationship table are acquired;

[0024] The device original performance data and the platform general index are subjected to field extraction respectively to obtain original fields and index fields;

[0025] According to the original fields, the field state of the device original performance data is judged in combination with the index fields;

[0026] According to the result of the field state judgment, the device original performance data is processed to obtain the abstract performance data, and the index conversion relationship table is updated.

[0027] In one embodiment, according to the result of the field state judgment, the device original performance data is processed to obtain the abstract performance data, and the index conversion relationship table is updated, including:

[0028] In response to the existence of the original field and the absence of the corresponding index field of the platform general index, the original field is a new field, the index field is added in the index conversion relationship table, and the corresponding conversion relationship is set to no conversion;

[0029] In response to the existence of the index field and the absence of the corresponding original field of the device original performance data, the index field is an abandoned field, and the index field is marked as disabled in the index conversion relationship table;

[0030] In response to the existence of the original field and the corresponding index field, type consistency judgment is performed based on the field types of the original field and the corresponding index field;

[0031] In response to the inconsistency of the field types of the original field and the corresponding index field, the original field is subjected to field type conversion, wherein the field type conversion at least includes one of the following: numerical value to string, string to numerical value;

[0032] In response to the success of the field type conversion, the corresponding conversion relationship in the index conversion relationship table is updated, otherwise it is marked as conversion failure;

[0033] In response to the completion of the field state judgment and the corresponding field type conversion of the device original performance data, the device original performance data is uniformly formatted and encapsulated as the abstract performance data.

[0034] In one embodiment, based on the index conversion relationship table, the abstract performance data is parsed and standardized to obtain platform standard performance data, including:

[0035] All fields in the abstract performance data are traversed to obtain abstract data fields;

[0036] Based on the name of the abstract data field, the corresponding conversion rule is determined by searching in the index conversion relationship table;

[0037] According to the conversion rule, the abstract performance data is subjected to data standardization processing and storage adaptability processing respectively to obtain platform standard performance data, wherein the selection of the storage adaptability processing is based on the type of the database.

[0038] According to the conversion rule, the abstract performance data is subjected to data standardization processing and storage adaptability processing respectively to obtain platform standard performance data, wherein the selection of the storage adaptability processing is based on the type of the database.

[0039] In one of the embodiments, after the platform standard performance data is stored in the database, the following steps are included:

[0040] Based on the general index, the platform standard performance data is obtained from the database for visual output to obtain a basic dashboard.

[0041] The user-specified business type is obtained, and the index analysis is performed according to the business type to obtain a key performance indicator.

[0042] According to the key performance indicator, the corresponding general index in the basic dashboard is filtered to generate a recommended configuration dashboard.

[0043] On the other hand, a device performance data management device is provided, and the device includes:

[0044] A device identification module is configured to determine a device driver interface and an architecture-specific index according to a device architecture type.

[0045] An index conversion relationship generation module is configured to obtain a platform general index and combine the architecture-specific index to obtain an index conversion relationship table.

[0046] A driving data acquisition module is configured to obtain device original performance data through the device driver interface.

[0047] An isolated abstract processing module is configured to process the device original performance data into abstract performance data based on a preset driving isolated abstract layer.

[0048] A data analysis and storage module is configured to analyze and standardize the abstract performance data based on the index conversion relationship table to obtain platform standard performance data, and store the platform standard performance data in a database.

[0049] In another aspect, a computer device is provided, which includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor implements the following steps when executing the computer program:

[0050] determine a device driving interface and an architecture-specific index according to a device architecture type;

[0051] obtain a platform general index, and obtain an index conversion relationship table in combination with the architecture-specific index;

[0052] obtain device original performance data through the device driving interface;

[0053] process the device original performance data into abstract performance data through a preset driving isolation abstraction layer;

[0054] perform analysis and standardization on the abstract performance data based on the index conversion relationship table to obtain platform standard performance data, and store the platform standard performance data into a database.

[0055] In another aspect, a computer readable storage medium is provided, and a computer program is stored on the computer readable storage medium, and the computer program is executed by a processor to implement the following steps:

[0056] determine a device driving interface and an architecture-specific index according to a device architecture type;

[0057] obtain a platform general index, and obtain an index conversion relationship table in combination with the architecture-specific index;

[0058] obtain device original performance data through the device driving interface;

[0059] process the device original performance data into abstract performance data through a preset driving isolation abstraction layer;

[0060] perform analysis and standardization on the abstract performance data based on the index conversion relationship table to obtain platform standard performance data, and store the platform standard performance data into a database.

[0061] The device performance data management method and device, computer device and storage medium described above, by combining the architecture-specific index and the platform general index, constructing the index conversion relationship table and standardizing the device performance data, can help to solve the problem of standardizing the performance data of devices of different architectures, thereby improving data compatibility and usability, reducing data management complexity caused by differences in hardware architectures, and avoiding time and resource waste caused by directional adaptation development. Meanwhile, by using the preset driving isolation abstraction layer, the data acquisition and the driving interface can be decoupled, the strong dependence of data acquisition on the driving interface in the traditional technology is avoided, the risk that the interface becomes unusable due to driving upgrade, thereby affecting normal data acquisition and system usability, is reduced, and the stability and reliability of the system are improved. BRIEF DESCRIPTION OF DRAWINGS

[0062] Figure 1Figure 1 is a diagram of an application environment of a device performance data management method according to an embodiment;

[0063] Figure 2 Figure 2 is a flowchart of a device performance data management method according to an embodiment;

[0064] Figure 3 Figure 3 is a block diagram of a device performance data management apparatus according to an embodiment;

[0065] Figure 4 Figure 4 is a diagram of an internal structure of a computer device according to an embodiment. DETAILED DESCRIPTION

[0066] In order to make the purposes, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and should not be used to limit the present application.

[0067] The device performance data management method provided by the present application can be applied in an application environment as shown in Figure 1 . In the application environment, a terminal 102 communicates with a server 104 through a network. The terminal 102 sends a device performance data calling request to the server 104. After receiving the device performance data calling request, the server 104 returns platform standard performance data stored in a database to the terminal 102 according to the device performance data management method. The terminal 102 can be, but is not limited to, various personal computers, notebook computers, smart phones, tablet computers and portable wearable devices. The server 104 can be implemented by an independent server or a server cluster composed of multiple servers.

[0068] In an embodiment, as shown in Figure 2 , a device performance data management method is provided. Taking the server 104 in Figure 1 as an example, the device performance data management method includes the following steps:

[0069] Step 201: Determine the device driver interface and architecture-specific indicators according to the device architecture type.

[0070] The device architecture type refers to the manufacturer or supplier of the device (such as NVIDIA, AMD, Intel, etc.). These architectures provide different types of hardware devices, such as GPUs, CPUs, memories, etc. The device driver interface is the interface between the hardware and the operating system or management software, which is responsible for the management, monitoring and control of the device. The system or software can obtain the information of the device through the driver interface. The architecture-specific indicators are defined by the device architecture and are used to measure and monitor the specific data or parameters of the device performance. Different architectures may use different names, data formats or measurement standards.

[0071] Step 202, obtaining platform general indicators, combining architecture-specific indicators, and obtaining an index conversion relationship table.

[0072] Specifically, historical device usage records are obtained to obtain used architecture indicators; based on the used architecture indicators, basic general indicators are extracted; the basic general indicators are further supplemented according to the used architecture indicators and the architecture-specific indicators, respectively, to obtain platform general indicators; the conversion relationship between the architecture-specific indicators and the platform general indicators is determined according to the platform general indicators, and the index conversion relationship table is constructed.

[0073] Step 203, obtaining device original performance data through a device driver interface.

[0074] Among them, the device original performance data refers to the performance data collected directly from the hardware through the device driver interface without processing, which is usually the original number related to the running state of the device hardware, and the data format is specific to the architecture.

[0075] Specifically, a data acquisition request is sent to the hardware device through the device driver interface to obtain real-time performance data of the device, i.e., device original performance data.

[0076] Step 204, processing the device original performance data into abstract performance data through a preset driver isolation abstraction layer.

[0077] Specifically, the device original performance data transmitted by the device driver interface is received through the driver isolation abstraction layer; the platform general indicators and the index conversion relationship table are obtained; the field state of the device original performance data is judged according to the platform general indicators; the device original performance data is processed according to the results of the field state judgment to obtain abstract performance data, and the index conversion relationship table is updated.

[0078] Step 205, based on the index conversion relationship table, the abstract performance data is parsed and standardized to obtain platform standard performance data, and the platform standard performance data is stored in the database.

[0079] Specifically, based on the platform general indicators, the field type of the abstract data field is verified, and in response to the result of the verification being inconsistent, the type of the abstract data field is converted based on the index conversion relationship table, otherwise, no conversion is needed; the abstract performance data with the consistent verification result is subjected to data standardization processing and storage adaptability processing to obtain platform standard performance data.

[0080] In the device performance data management method, the index conversion relationship table is constructed by combining the architecture-specific indicators and the platform universal indicators, and the device performance data is standardized, which helps to solve the problem of standardizing the performance data of different architecture devices, thereby improving the data compatibility and usability, reducing the data management complexity caused by the difference in hardware architecture, and avoiding the waste of time and resources caused by directional adaptation development. At the same time, by pre-setting the driving isolation abstraction layer, the data acquisition and driving interface implementation can be decoupled, avoiding the strong dependence of data acquisition on driving interface in the traditional technology, reducing the risk of interface unavailability caused by driving upgrade, thereby affecting the normal acquisition of data and causing the system to be unavailable, improving the stability and reliability of the system.

[0081] In one embodiment, the platform universal indicators are obtained by combining the architecture-specific indicators, and an index conversion relationship table is obtained, wherein the platform universal indicators are obtained by the following method:

[0082] Obtain historical device usage records, and mark all used devices;

[0083] According to the used device, the corresponding device architecture is determined, and the used architecture indicators are obtained;

[0084] Commonality extraction is performed on the used architecture indicators to obtain basic universal indicators;

[0085] The extracted used architecture indicators are subjected to feature performance indicator screening and analysis to obtain feature performance indicators;

[0086] The platform universal indicators are obtained according to the basic universal indicators combined with the feature performance indicators.

[0087] Specifically, by obtaining historical device usage records, marking all used devices, and extracting universal indicators according to architecture indicators, the platform performance data management system has higher universality and adaptability, which helps to build a highly available platform universal indicator, thereby providing a basis for subsequent data standardization and performance analysis, reducing the dependence on single architecture devices, and improving the expansibility of the system.

[0088] It is worth noting that the used architecture indicators are obtained by commonality extraction to obtain basic universal indicators, which specifically includes the following steps:

[0089] Collect and record the usage of historical devices, including device model, architecture information and used indicator data, etc.

[0090] According to the device information in the historical records, the device architecture that has been used is determined, and the used architecture indicators are extracted from the indicator data provided by each device architecture.

[0091] By comparing the metrics data provided by different architectures, we can identify the common metrics among them, which are the basic general metrics.

[0092] Define the basic, general indicators and their data formats to ensure that they can be uniformly identified and processed by the system.

[0093] It is worth noting that the basic general metrics are obtained through commonality extraction, which can identify common standards across different device architectures and efficiently determine the most critical metrics for a device. However, due to differences in some metrics between different device architectures, some key metrics may be omitted. Therefore, characteristic performance metrics are needed to supplement the basic general metrics, thereby obtaining high-availability platform general metrics. The specific steps for obtaining characteristic performance metrics include:

[0094] Obtain the screening criteria, which are based on at least one of the following dimensions: relevance, importance, and availability;

[0095] The previously used architecture metrics after extracting the basic general metrics are recorded as the remaining architecture metrics.

[0096] Based on the basic general indicators, correlation analysis is performed on the remaining architectural indicators in turn to obtain the correlation scores between the remaining architectural indicators and the basic general indicators.

[0097] Obtain historical device performance data, combine it with historical device usage records, assess the impact of remaining architecture metrics on device performance, and assign importance scores.

[0098] A reliability score is obtained based on the difficulty of data collection corresponding to the remaining architecture indicators. If the data can be collected directly, the reliability score is 1. For each additional calculation required to obtain the data corresponding to the indicator, the reliability score decreases by 10-30%.

[0099] Based on the remaining architecture metrics, the corresponding relevance score, importance score, and usability score are assigned corresponding preset weights, and a weighted calculation is performed to obtain the feature score.

[0100] The remaining architecture metrics that are in the top 10-20% of feature scores and meet the preset threshold are selected as feature performance metrics.

[0101] In one embodiment, platform-wide metrics are obtained, and combined with architecture-specific metrics, a metric transformation relationship table is obtained, including:

[0102] Based on the architecture-specific metrics, check among the platform's general metrics to determine whether the architecture-specific metrics are already entered.

[0103] In response to the determination result being no, the architecture-specific indicators are subjected to indicator conversion processing to obtain specific converted indicators, and the specific converted indicators are supplemented into the platform general indicators.

[0104] In response to the determination result being yes, a conversion relationship between the architecture-specific indicators and the platform general indicators is determined, and an indicator conversion relationship table is constructed based on the conversion relationship.

[0105] Specifically, the embodiment can ensure continuous updating and integrity of the platform general indicator library, improve adaptability to new devices and architecture indicators, reduce resource waste caused by device idling, avoid the defects of hard coding or customized development in the traditional method, and enhance the flexibility and long-term applicability of the system by judging whether the architecture-specific indicators are already recorded indicators and performing necessary conversion and supplementation.

[0106] In one embodiment, the device raw performance data is processed into abstract performance data by a preset driver isolation abstraction layer, wherein the driver isolation abstraction layer performs the following steps:

[0107] The device raw performance data transmitted by the device driver interface is received;

[0108] The platform general indicators and the indicator conversion relationship table are obtained;

[0109] The device raw performance data and the platform general indicators are subjected to field extraction respectively to obtain raw fields and indicator fields;

[0110] According to the raw fields, the device raw performance data is subjected to field state judgment in combination with the indicator fields;

[0111] According to the results of the field state judgment, the device raw performance data is processed to obtain abstract performance data, and the indicator conversion relationship table is updated.

[0112] Specifically, the embodiment processes and standardizes the raw performance data by the driver isolation abstraction layer, improves the processing efficiency of the performance data, and ensures the compatibility of cross-platform devices; at the same time, by setting the driver isolation abstraction layer, the system can maintain the stability of the data processing logic when the device driver changes, and effectively reduces the system instability problem caused by the upgrade of the driver.

[0113] It is worth noting that the driver isolation abstraction layer and the device driver communicate through the device driver interface to receive the device raw performance data, and the driver isolation abstraction layer processes the device raw performance data into abstract performance data, and then sends the abstract performance data to the monitoring script, which in turn sends the abstract performance data for subsequent processing.

[0114] Specifically, without the driving isolation abstraction layer, when the device driver is upgraded or switched to a driver of a different architecture, the monitoring script must be adjusted accordingly for the new driver version or architecture API. The update of the device driver can change the data format, API interface, or other behaviors, which means that the monitoring script may need to be modified to adapt to these changes, resulting in increased maintenance and development costs. However, the introduction of the driving isolation abstraction layer can encapsulate the driving differences into an intermediary layer for unified interface and data format and data conversion.

[0115] Specifically, the driving isolation abstraction layer accesses the data of different device drivers by defining a unified standard interface (such as a unified API or data structure). Regardless of the changes in the driver version or architecture, the monitoring script only needs to interact with the driving isolation abstraction layer and does not need to concern the specific implementation of the driver. At the same time, the driving isolation abstraction layer converts the data of different drivers, so that the monitoring script can obtain unified format data regardless of which device driver it obtains data from. The driving isolation abstraction layer is responsible for mapping and format conversion of the original data fields, ensuring that the script does not depend on the implementation details of a specific driver and avoiding frequent modifications of the script. In addition, when processing the original performance data, the driving isolation abstraction layer updates the index conversion relationship table based on the comparison results of the original performance data and the platform general indicators, which can better adapt to changes in the driving interface and indicators.

[0116] In one embodiment, the original performance data of the device is processed to obtain abstract performance data and update the index conversion relationship table according to the results of the field state judgment, including:

[0117] In response to the existence of the original field and the non-existence of the corresponding index field of the platform general indicator, the original field is a new field, and the index field is added in the index conversion relationship table, and the corresponding conversion relationship is set as no conversion;

[0118] In response to the existence of the index field and the non-existence of the corresponding original field of the device original performance data, the index field is an abandoned field, and the index field is marked as disabled in the index conversion relationship table;

[0119] In response to the existence of the original field and the corresponding index field, type consistency judgment is performed based on the field types of the original field and the corresponding index field;

[0120] In response to the inconsistency of the field types of the original field and the corresponding index field, the original field is subjected to field type conversion, wherein the field type conversion at least includes one of the following: numerical value to string, string to numerical value;

[0121] In response to the field type conversion being successful, the corresponding conversion relationship in the index conversion relationship table is updated, otherwise, it is marked as conversion failure;

[0122] In response to the field status judgment and the corresponding field type conversion of the device original performance data being completed, the device original performance data is uniformly formatted and encapsulated as abstract performance data.

[0123] It is worth noting that in response to the field types of the original field and the corresponding index field being inconsistent, the original field is subjected to field type conversion, which at least includes one of the following: numerical value to string, string to numerical value, and specifically includes the following steps:

[0124] The field type of the original field is obtained, which is at least one of the following: numerical value, string;

[0125] In response to the field type of the original field being numerical value, it is directly converted, and the field type after conversion is string, and the corresponding conversion relationship in the index conversion relationship table is updated;

[0126] In response to the character type of the original field being string, the original field is preferentially converted to double type numerical value; it is worth noting that numerical value to string generally does not fail, but string to numerical value may fail, therefore, in response to double type numerical value conversion failure, the corresponding conversion relationship in the index conversion relationship table is marked as conversion failure, triggering the conversion failure processing mechanism.

[0127] Specifically, the conversion failure processing mechanism includes:

[0128] Recording conversion failure information, which includes: failure timestamp, original performance data, failed field, and failure reason;

[0129] The conversion failure information is notified of an alarm, and the data collection and processing of the corresponding original performance data is stopped;

[0130] A preset automatic repair rule is obtained, and according to the failure reason of the conversion failure information, it is queried whether there is a repair processing mode corresponding to the situation in the preset automatic repair rule, wherein the preset automatic repair rule includes: writing a script to automatically clean and format data for a specific data format error;

[0131] In response to the existence of a repair processing mode corresponding to the situation, the original performance data is adjusted according to the corresponding repair processing mode, and re-conversion is attempted;

[0132] In response to the absence of a repair processing method for the corresponding condition or re-conversion failure, the original performance data is rolled back according to the failure timestamp, the original performance data is restored to the state before the failure, and it is marked as priority processing data;

[0133] The priority processing data is processed through manual correction processing or artificial intelligence analysis processing until the priority processing data is successfully converted, the corresponding original performance data is restored for data collection and processing, and the corresponding processing method is recorded in the preset repair rule.

[0134] Specifically, by using the above automatic processing method, the problems caused by conversion failure can be significantly reduced, and the response ability and stability of the system can be improved, thereby more effectively managing and processing abnormal situations in the data conversion process.

[0135] Specifically, the embodiment processes and updates the index conversion relationship table according to the field state judgment of the device original performance data, which helps to process the inconsistent data format problem and ensures the formatting and uniformity of the data, effectively improving the quality and consistency of the data, and ensuring that the data can smoothly enter the subsequent standardization and storage link.

[0136] In one embodiment, based on the index conversion relationship table, the abstract performance data is parsed and standardized to obtain platform standard performance data, including:

[0137] All fields in the abstract performance data are traversed to obtain abstract data fields;

[0138] Based on the name of the abstract data field, the corresponding conversion rule is determined in the index conversion relationship table;

[0139] Based on the platform general index, the field type of the abstract data field is verified, and in response to the inconsistent result of the verification, the abstract data field is converted according to the conversion rule, otherwise, no conversion is required;

[0140] According to the conversion rule, the abstract performance data is respectively processed by data standardization and storage adaptability, to obtain platform standard performance data, wherein the selection of storage adaptability is based on the type of the database.

[0141] Specifically, in the embodiment, the data field is converted and processed according to the index conversion relationship table in the standardization process, which improves the compatibility of the data, helps to optimize data storage and subsequent analysis process, ensures the data quality and improves the efficiency of subsequent query and operation.

[0142] In one embodiment, after the platform standard performance data is stored in the database, it includes:

[0143] Based on the general index, the platform standard performance data is obtained from the database for visual output, and the basic dashboard is obtained;

[0144] The user-specified business type is obtained, and the key performance indicators are obtained according to the business type;

[0145] According to the key performance indicators, the corresponding general indicators in the basic dashboard are screened, and the recommended configuration dashboard is generated.

[0146] Specifically, in the embodiment, after the platform standard performance data is stored in the database, the visual output and the key performance indicator analysis of the business type can provide customized performance analysis and display results for the user, which is beneficial to improve the performance information acquisition efficiency of the user in diversified use scenarios, and generate recommended configurations according to business requirements, improve the decision efficiency of the user and the practicability of the system.

[0147] It should be understood that, although Figure 2 The steps in the flowchart are displayed in sequence according to the arrows, but these steps are not necessarily executed in sequence according to the arrows. Unless otherwise specified in this article, the execution of these steps has no strict order limitation, and these steps can be executed in other order. Moreover, Figure 2 At least part of the steps in the flowchart can include multiple sub-steps or multiple stages, which are not necessarily executed at the same time, but can be executed at different times, and the execution order of these sub-steps or stages is not necessarily sequential, but can be executed in rotation or alternation with other steps or sub-steps or stages of other steps.

[0148] In one embodiment, as shown in Figure 3 A device performance data management apparatus is provided, comprising: a device identification module, an index conversion relationship generation module, a driving data acquisition module, an isolation abstraction processing module and a data analysis and storage module, wherein:

[0149] The device identification module is configured to determine the device driver interface and the architecture-specific index according to the device architecture type;

[0150] The index conversion relationship generation module is configured to obtain platform general indexes, and combine the architecture-specific indexes to obtain an index conversion relationship table;

[0151] The driving data acquisition module is configured to obtain device original performance data through the device driver interface;

[0152] The isolation abstraction processing module is configured to process the device original performance data into abstract performance data based on a preset driving isolation abstraction layer;

[0153] The data analysis and storage module is configured to analyze and standardize the abstract performance data based on the index conversion relationship table to obtain platform standard performance data, and store the platform standard performance data in a database.

[0154] The index conversion relationship generation module is configured to obtain historical device usage records, mark all used devices, determine corresponding device architectures according to the used devices, and obtain used architecture indexes; perform commonality extraction on the used architecture indexes to obtain basic common indexes; perform feature performance index screening and analysis on the extracted used architecture indexes to obtain feature performance indexes; obtain platform common indexes based on the basic common indexes and the feature performance indexes.

[0155] The index conversion relationship generation module is further configured to check the platform common indexes according to the architecture-specific indexes, determine whether the architecture-specific indexes are already recorded indexes, perform index conversion processing on the architecture-specific indexes to obtain specific conversion indexes and record the specific conversion indexes in the platform common indexes in response to a determination result of no, and determine a conversion relationship between the architecture-specific indexes and the platform common indexes and construct an index conversion relationship table based on the conversion relationship in response to a determination result of yes.

[0156] The driving data collection module is configured to send a data acquisition request to a hardware device through a device driver interface to obtain real-time performance data of the device, i.e., device original performance data.

[0157] The isolation and abstraction processing module is configured to obtain abstract performance data processed by the driving isolation and abstraction layer, and the driving data collection module is configured to call the driving isolation and abstraction layer to perform the following steps: receive device original performance data transmitted by the device driver interface; obtain the platform common indexes and the index conversion relationship table; perform field extraction on the device original performance data and the platform common indexes to obtain original fields and index fields; perform field state judgment on the device original performance data according to the original fields and the index fields; and perform processing on the device original performance data according to a result of the field state judgment to obtain the abstract performance data and update the index conversion relationship table.

[0158] The isolation abstract processing module is further configured to call the driving isolation abstract layer to perform the following steps: in response to the existence of the original field and the non-existence of the corresponding index field of the platform general index, the original field is a new field, an index field is added in the index conversion relationship table, and the corresponding conversion relationship is set as no conversion; in response to the existence of the index field and the non-existence of the corresponding original field of the device original performance data, the index field is an obsolete field, the index field is marked as disabled in the index conversion relationship table; in response to the existence of the original field and the corresponding index field, type consistency is judged based on the field types of the original field and the corresponding index field; in response to the inconsistency of the field types of the original field and the corresponding index field, the original field is subjected to field type conversion, wherein the field type conversion at least includes one of the following: numerical value to string, string to numerical value; in response to the success of the field type conversion, the corresponding conversion relationship in the index conversion relationship table is updated, otherwise, it is marked as conversion failure; in response to the completion of the field state judgment and the corresponding field type conversion of the device original performance data, the device original performance data is uniformly formatted and encapsulated as abstract performance data.

[0159] The data analysis and storage module is configured to traverse all fields in the abstract performance data to obtain abstract data fields; based on the names of the abstract data fields, the corresponding conversion rules are determined by searching in the index conversion relationship table; based on the platform general index, the field types of the abstract data fields are verified, and in response to the inconsistent verification result, the abstract data fields are subjected to type conversion according to the conversion rules, otherwise, no conversion is needed; the abstract performance data is subjected to data standardization processing and storage adaptability processing according to the conversion rules to obtain platform standard performance data, wherein the selection of the storage adaptability processing is based on the type of the database.

[0160] In one embodiment, the database is connected with a data display module configured to obtain platform standard performance data from the database based on the general index for visual output to obtain a basic dashboard; obtain a user-specified business type, perform index analysis according to the business type to obtain key performance indicators; and perform filtering on the corresponding general index in the basic dashboard according to the key performance indicators to generate a recommended configuration dashboard.

[0161] The specific limitations of the device performance data management device can be referred to the limitations of the device performance data management method in the foregoing, which will not be repeated here. Each module in the above device performance data management device can be realized by software, hardware and their combinations in whole or in part. The above modules can be embedded in or independent of the processor in the computer device in hardware form, or can be stored in the memory in the computer device in software form, so as to be called and executed by the processor to perform the operations corresponding to each module.

[0162] In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 4 As shown, the computer device includes a processor, memory, network interface, and database connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and database. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores device performance data management data. The network interface communicates with external terminals via a network connection. When the computer program is executed by the processor, it implements a device performance data management method.

[0163] Those skilled in the art will understand that Figure 4 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0164] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to perform the following steps:

[0165] Based on the device architecture type, determine the device driver interface and architecture-specific metrics;

[0166] Obtain platform-wide metrics and combine them with architecture-specific metrics to obtain a metric conversion relationship table;

[0167] Obtain raw device performance data through the device driver interface;

[0168] The device's raw performance data is processed into abstract performance data through a preset driver isolation abstraction layer.

[0169] Based on the indicator transformation relationship table, the abstract performance data is parsed and standardized to obtain the platform standard performance data, which is then stored in the database.

[0170] In one embodiment, the processor, when executing a computer program, also performs the following steps:

[0171] Retrieve historical device usage records and mark all previously used devices;

[0172] Based on the previously used equipment, determine the corresponding equipment architecture and obtain the previously used architecture indicators;

[0173] Commonality extraction is performed on the used architecture indicators to obtain basic common indicators;

[0174] Feature performance indicators are screened and analyzed from the extracted used architecture indicators to obtain feature performance indicators;

[0175] Platform common indicators are obtained according to the basic common indicators in combination with the feature performance indicators.

[0176] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0177] According to the architecture-specific indicators, a check is performed in the platform common indicators to determine whether the architecture-specific indicators are already recorded indicators;

[0178] In response to a determination result of no, the architecture-specific indicators are subjected to indicator conversion processing to obtain specific conversion indicators, and the specific conversion indicators are recorded in the platform common indicators;

[0179] In response to a determination result of yes, a conversion relationship between the architecture-specific indicators and the platform common indicators is determined, and a conversion relationship table is constructed based on the conversion relationship.

[0180] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0181] The device raw performance data transmitted by the device driver interface is received;

[0182] The platform common indicators and the conversion relationship table are obtained;

[0183] Field extraction is performed on the device raw performance data and the platform common indicators to obtain raw fields and indicator fields;

[0184] Field state judgment is performed on the device raw performance data according to the raw fields in combination with the indicator fields;

[0185] According to the results of the field state judgment, the device raw performance data is processed to obtain abstract performance data, and the conversion relationship table is updated.

[0186] In one embodiment, the processor, when executing the computer program, also implements the following steps:

[0187] In response to the existence of the raw fields and the nonexistence of corresponding indicator fields in the platform common indicators, the raw fields are new fields, and the indicator fields are added in the conversion relationship table, and the corresponding conversion relationship is set to no conversion;

[0188] In response to the existence of the indicator fields and the nonexistence of corresponding raw fields in the device raw performance data, the indicator fields are abandoned fields, and the indicator fields are marked as disabled in the conversion relationship table;

[0189] in response to the original field and the corresponding index field both existing, type consistency judgment is performed based on the field types of the original field and the corresponding index field;

[0190] in response to the field types of the original field and the corresponding index field being inconsistent, field type conversion is performed on the original field, wherein the field type conversion at least includes one of the following: numerical value to string, string to numerical value;

[0191] in response to the field type conversion being successful, the corresponding conversion relationship in the index conversion relationship table is updated, otherwise, it is marked as conversion failure;

[0192] in response to the device original performance data completing the field state judgment and the corresponding field type conversion, the device original performance data is uniformly formatted and encapsulated as abstract performance data.

[0193] In one embodiment, the processor executing the computer program also implements the following steps:

[0194] traversing all fields in the abstract performance data to obtain abstract data fields;

[0195] based on the name of the abstract data field, searching in the index conversion relationship table to determine the corresponding conversion rule;

[0196] based on the platform general index, the field type of the abstract data field is verified, in response to the verification result being inconsistent, the abstract data field is type converted according to the conversion rule, otherwise, no conversion is needed;

[0197] according to the conversion rule, the abstract performance data is respectively subjected to data standardization processing and storage adaptability processing to obtain platform standard performance data, wherein the selection of storage adaptability processing is based on the type of the database.

[0198] In one embodiment, the processor executing the computer program also implements the following steps:

[0199] based on the general index, the platform standard performance data is obtained from the database for visual output to obtain a basic dashboard;

[0200] obtaining a user-specified business type, performing index analysis according to the business type to obtain key performance indicators;

[0201] according to the key performance indicators, the corresponding general index in the basic dashboard is filtered to generate a recommended configuration dashboard.

[0202] In one embodiment, a computer readable storage medium is provided, which stores a computer program, and the computer program is executed by a processor to implement the following steps:

[0203] determine the device driver interface and the architecture-specific index according to the device architecture type;

[0204] obtain the platform general index, combine the architecture-specific index, and obtain an index conversion relationship table;

[0205] obtain the device original performance data through the device driver interface;

[0206] process the device original performance data into abstract performance data through the preset driver isolation abstraction layer;

[0207] based on the index conversion relationship table, analyze and standardize the abstract performance data to obtain platform standard performance data, and store the platform standard performance data in a database.

[0208] In one embodiment, the computer program is executed by the processor to further implement the following steps:

[0209] According to the architecture-specific index, check in the platform general index to determine whether the architecture-specific index is an already recorded index;

[0210] In response to a negative determination result, perform index conversion processing on the architecture-specific index to obtain a specific conversion index, and supplement the specific conversion index to the platform general index;

[0211] In response to a positive determination result, determine the conversion relationship between the architecture-specific index and the platform general index, and based on the conversion relationship, construct an index conversion relationship table.

[0212] In one embodiment, the computer program is executed by the processor to further implement the following steps:

[0213] receive the device original performance data transmitted by the device driver interface;

[0214] obtain the platform general index and the index conversion relationship table;

[0215] extract fields from the device original performance data and the platform general index to obtain original fields and index fields;

[0216] According to the original fields, combine the index fields, and perform field state judgment on the device original performance data;

[0217] According to the results of the field state judgment, process the device original performance data to obtain abstract performance data, and update the index conversion relationship table.

[0218] In one embodiment, the computer program is executed by the processor to further implement the following steps:

[0219] In response to the original field existing and the platform general index not existing a corresponding index field, the original field is a new field, and a new index field is added in the index conversion relationship table, and a corresponding conversion relationship is set as no conversion;

[0220] In response to the index field existing and the device original performance data not existing a corresponding original field, the index field is an abandoned field, and the index field is marked as disabled in the index conversion relationship table.

[0221] In response to the original field and the corresponding index field both existing, type consistency is judged based on the field types of the original field and the corresponding index field.

[0222] In response to the field types of the original field and the corresponding index field being inconsistent, the original field is subjected to field type conversion, wherein the field type conversion at least includes one of the following: numerical value to string, string to numerical value.

[0223] In response to the field type conversion being successful, the corresponding conversion relationship in the index conversion relationship table is updated, otherwise, it is marked as conversion failure.

[0224] In response to the device original performance data completing the field state judgment and the corresponding field type conversion, the device original performance data is uniformly formatted and encapsulated as abstract performance data.

[0225] In one embodiment, the computer program is executed by the processor to further implement the following steps:

[0226] All fields in the abstract performance data are traversed to obtain abstract data fields.

[0227] Based on the name of the abstract data field, a corresponding conversion rule is determined by searching in the index conversion relationship table.

[0228] Based on the platform general index, the field type of the abstract data field is verified, and in response to the verification result being inconsistent, the abstract data field is subjected to type conversion according to the conversion rule, otherwise, no conversion is needed.

[0229] According to the conversion rule, the abstract performance data is subjected to data standardization processing and storage adaptability processing respectively to obtain platform standard performance data, wherein the selection of storage adaptability processing is based on the type of the database.

[0230] In one embodiment, the computer program is executed by the processor to further implement the following steps:

[0231] Based on the general index, the platform standard performance data is obtained from the database for visual output to obtain a basic dashboard.

[0232] Obtaining a user-specified service type, performing index analysis according to the service type, and obtaining a key performance indicator;

[0233] According to the key performance indicator, filtering corresponding general indicators in the basic dashboard, and generating a recommended configuration dashboard.

[0234] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment methods can be completed by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer readable storage medium. When the computer program is executed, the processes of the above-mentioned embodiments of the methods can be included. Any reference to memory, storage, database or other medium used in the embodiments provided in the present application can include non-volatile and / or volatile memory. Non-volatile memory can include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory can include random access memory (RAM) or external cache memory. As an illustration but not limitation, RAM is available in various forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous link (Synchlink) DRAM (SLDRAM), memory bus (Rambus) direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0235] The technical features of the above embodiments can be combined in any way. In order to make the description concise, not all possible combinations of the technical features in the above embodiments are described, but as long as the combinations of the technical features do not exist, they should be considered as the scope of the present application.

[0236] The above embodiments only express several implementation manners of the present application, and the description is specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be noted that for those of ordinary skill in the art, without departing from the concept of the present application, some modifications and improvements can be made, which are all within the scope of the present application. Therefore, the scope of protection of the patent of the present application should be subject to the appended claims.

Claims

1. A method for managing equipment performance data, characterized in that, include: Based on the device architecture type, determine the device driver interface and architecture-specific metrics; Obtain platform-wide metrics and combine them with architecture-specific metrics to obtain a metric conversion relationship table; The device's raw performance data is obtained through the device driver interface. The device's raw performance data is processed into abstract performance data through a preset driver isolation abstraction layer. The driver isolation abstraction layer performs the following steps: receiving the device's raw performance data transmitted from the device driver interface; obtaining the platform's general indicators and the indicator conversion relationship table; extracting fields from both the device's raw performance data and the platform's general indicators to obtain raw fields and indicator fields; determining the field status of the device's raw performance data based on the raw fields and the indicator fields; and processing the device's raw performance data according to the results of the field status determination to obtain the abstract performance data and updating the indicator conversion relationship table. Based on the indicator conversion relationship table, the abstract performance data is parsed and standardized to obtain platform standard performance data, which is then stored in the database.

2. The equipment performance data management method according to claim 1, characterized in that, Obtain platform-wide metrics and combine them with architecture-specific metrics to obtain a metric transformation relationship table. The platform-wide metrics are obtained through the following methods: Retrieve historical device usage records and mark all previously used devices; Based on the previously used equipment, the corresponding equipment architecture is determined, and the previously used architecture indicators are obtained; Commonalities were extracted from the previously used architectural metrics to obtain basic general metrics; The extracted previously used architecture metrics are then subjected to feature performance metric screening and parsing to obtain the feature performance metrics; Based on the aforementioned basic general indicators and combined with the aforementioned characteristic performance indicators, the platform's general indicators are obtained.

3. The equipment performance data management method according to claim 1, characterized in that, Obtain platform-wide metrics and combine them with architecture-specific metrics to obtain a metric transformation relationship table, including: Based on the architecture-specific metrics, check the platform's general metrics to determine whether the architecture-specific metrics are already recorded. If the judgment result is negative, the architecture-specific indicators are converted to obtain specific converted indicators, and the specific converted indicators are added to the platform general indicators. If the judgment result is yes, then the conversion relationship between the architecture-specific indicators and the platform-general indicators is determined, and based on the conversion relationship, the indicator conversion relationship table is constructed.

4. The equipment performance data management method according to claim 1, characterized in that, The step of processing the original device performance data based on the result of the field status judgment to obtain the abstract performance data and updating the indicator conversion relationship table includes: If the original field exists, but the platform general indicator does not have a corresponding indicator field, then the original field is a new field. In this case, the indicator field is added to the indicator conversion relationship table, and the corresponding conversion relationship is set to no conversion. If the indicator field exists, but the original performance data of the device does not have a corresponding original field, then the indicator field is a deprecated field, and the indicator field is marked as disabled in the indicator conversion relationship table; If both the original field and the corresponding indicator field exist, a type consistency judgment is performed based on the field types of the original field and the corresponding indicator field. In response to the inconsistency between the field types of the original field and the corresponding indicator field, the field type of the original field is converted, wherein the field type conversion includes at least one of the following: numeric to string, string to numeric; If the field type conversion is successful, the corresponding conversion relationship in the indicator conversion relationship table will be updated; otherwise, it will be marked as a conversion failure. In response to the completion of the field status judgment and the corresponding field type conversion of the original device performance data, the original device performance data is uniformly formatted and encapsulated into the abstract performance data.

5. The equipment performance data management method according to claim 1, characterized in that, Based on the aforementioned indicator conversion relationship table, the abstract performance data is parsed and standardized to obtain platform standard performance data, including: Iterate through all fields in the abstract performance data to obtain the abstract data fields; Based on the name of the abstract data field, the corresponding conversion rule is determined by searching the indicator conversion relationship table. Based on the platform's general metrics, the field type of the abstract data field is validated. If the validation result is inconsistent, the abstract data field is converted according to the conversion rules; otherwise, no conversion is required. According to the transformation rules, the abstract performance data is subjected to data standardization and database adaptation processing to obtain platform standard performance data. The selection of the database adaptation processing is based on the type of the database.

6. The equipment performance data management method according to claim 1, characterized in that, After storing the platform's standard performance data in the database, the following steps are included: Based on the general metrics, the platform's standard performance data is retrieved from the database and visualized to obtain a basic dashboard. Obtain the service type specified by the user, perform indicator analysis based on the service type, and obtain key performance indicators; Based on the key performance indicators, the corresponding general indicators in the basic dashboard are filtered to generate a recommended configuration dashboard.

7. A device for managing equipment performance data, used to perform the method as described in any one of claims 1-6, characterized in that, The device includes: The device identification module is used to determine the device driver interface and architecture-specific indicators based on the device architecture type. The indicator conversion relationship generation module is used to obtain platform-wide indicators and combine them with architecture-specific indicators to obtain an indicator conversion relationship table. The drive data acquisition module is used to acquire the original performance data of the device through the device drive interface; The isolation and abstraction processing module is used to process the original performance data of the device into abstract performance data based on a preset driver isolation and abstraction layer. The data parsing and storage module is used to parse and standardize the abstract performance data based on the indicator conversion relationship table to obtain platform standard performance data, and store the platform standard performance data in the database.

8. A computer device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images