AI-based intelligent calculation platform management method

Abstract

The invention relates to the technical field of artificial intelligence, in particular to an AI-based intelligent calculation platform management method. The method comprises the following steps: S1, receiving target data and carrying out segmentation processing; s2, adjusting a preset character length; step S3, storing the segmented real-time target segments on different database nodes according to the change curve; and S4, analyzing the demand of the database node, and classifying the current transmission network according to an analysis result. By flexibly adjusting the length of the preset character, segmentation and processing of the data can better adapt to actual access requirements, so that the efficiency and accuracy of data processing are improved, storage resources are reasonably distributed on different database nodes, the storage layout is optimized, the speed of data retrieval and processing is improved, and the data retrieval and processing efficiency is improved. Through demand analysis and network classification of database nodes, resources are reasonably allocated, redundancy is reduced, and it is ensured that the system is kept efficient when facing different data loads.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an AI-based intelligent computing center management method. Background Art

[0002] "Intelligent computing middle platform" refers to a computing platform that integrates artificial intelligence technology. Its main function is to provide intelligent decision-making support or services through data processing, analysis and prediction. This type of platform can integrate multiple data sources, including structured and unstructured data, and realize in-depth analysis of data through machine learning and data mining technology, thereby providing users with customized solutions or suggestions.

[0003] The patent document with announcement number CN118171743B discloses an AI intelligent computing platform management method and system, the method comprising: obtaining the business type of the target intelligent computing request, the business type is determined based on the semantic information carried by the target intelligent computing request; generating scenario-assisted confirmation questions based on the business type obtained by analysis to obtain feedback information of the scenario-assisted confirmation questions; analyzing the feedback information to match the corresponding large language model; it can be seen that the existing intelligent computing middle-end management method lacks a method that passes through efficient data classification processing, dynamic character length adjustment, real-time monitoring of access behavior, historical data analysis, and network performance evaluation, resulting in low intelligence level of data management and low network operation efficiency. Summary of the invention

[0004] To this end, the present invention provides an AI-based intelligent computing middle station management method to overcome the problem of lack of optimization of data processing and storage efficiency in the prior art, which leads to the problem of the intelligence level of the intelligent computing middle station.

[0005] To achieve the above objectives, the present invention provides an AI-based intelligent computing center management method, comprising:

[0006] Step S1, receiving target data according to business needs, classifying the target data into structured data and unstructured data, converting the target data into text data, segmenting the unstructured data with a preset character length, and obtaining real-time target segments;

[0007] Step S2, for any of the real-time target segments, extract keywords and match them in a historical database to obtain the real-time access times of the real-time target segment, determine the real-time access times according to the standard access times, and correct the preset character length to the corrected character length according to the determination result;

[0008] Step S3, obtaining the historical data volume of the received target data with a preset acquisition time, drawing a curve of the change of the historical data volume over time, and storing the segmented real-time target segments on different database nodes according to the change curve;

[0009] Step S4, analyzing the requirements of the database nodes, and classifying the current transmission network according to the analysis results into high-performance networks, standard-performance networks, and low-performance networks.

[0010] Furthermore, the unstructured data is segmented according to a preset character length, and after segmentation, a plurality of real-time target segments are obtained, including:

[0011] When the target data is determined to be unstructured data, the real-time byte length of the text data is obtained, and the real-time byte length is determined according to the standard byte length.

[0012] If the real-time byte length is less than or equal to the standard byte length, the target data is determined to be a type of data, and the text data is directly used as the real-time target segment;

[0013] If the real-time byte length is greater than the standard byte length, the target data corresponding to the text data is determined to be second-class data, and the text data is divided into a number of real-time target segments according to a preset byte segmentation length.

[0014] Further, obtaining the real-time access count of the real-time target segment includes:

[0015] For any real-time target segment, the associated keywords are obtained by extracting keywords and matching them in the historical database;

[0016] The keyword data items and the corresponding historical access times corresponding to each associated keyword are obtained in the historical database, and the maximum value of the historical access times among the associated keywords is selected as the real-time access times of the real-time target segment.

[0017] Further, correcting the preset character length to a corrected character length according to the determination result includes:

[0018] The real-time access number is determined according to the standard access number, and when the real-time access number is greater than or equal to the standard access number, the target segment obtains the real-time access frequency corresponding to the real-time target segment, and the real-time access frequency is determined according to the standard access frequency to determine whether to correct the preset byte segmentation length.

[0019] Further, determining the real-time access frequency according to the standard access frequency includes:

[0020] For the real-time target segment, when it is determined that the real-time access number is greater than or equal to the standard access number, the keyword data item corresponding to the maximum value of the historical access number and the corresponding historical access frequency are obtained as the real-time access frequency, and the real-time access frequency is determined according to the standard access frequency.

[0021] When the real-time access frequency is greater than the standard access frequency, the data segmentation module modifies the preset byte segmentation length corresponding to the real-time target segment according to the real-time access times and the real-time access frequency to a modified byte segmentation length.

[0022] Further, analyzing the requirements of the database node includes:

[0023] Obtaining the real-time memory utilization of any of the database nodes, comparing the real-time memory utilization with the first standard memory utilization and the second standard memory utilization, and classifying the performance type of the current transmission network according to the comparison result to obtain a classification result;

[0024] When determining that the current transmission network is a standard performance network, the bandwidth demand and the storage demand are associated for analysis, and whether the division result is accurate is verified based on the analysis result.

[0025] Further, the analysis of the associated bandwidth requirements and storage requirements includes:

[0026] The real-time flow difference is calculated according to the network input / output volume, the real-time storage difference is calculated according to the disk input / output volume, and the ratio of the real-time flow difference to the real-time storage difference is calculated to obtain the real-time storage efficiency;

[0027] Analyze real-time traffic difference, real-time storage difference, or real-time storage efficiency.

[0028] Further, analyzing the real-time flow difference, the real-time storage difference or the real-time storage efficiency includes:

[0029] Subtract the output from the input to get the real-time flow difference;

[0030] Determine the real-time flow difference according to the preset difference range.

[0031] If the real-time flow difference is within the preset difference range, the type classification is determined to be correct;

[0032] If the real-time traffic difference is not within the preset difference range, it is determined that the type classification is incorrect, and the second standard memory utilization is adjusted.

[0033] Furthermore, when the real-time flow difference is not within the preset difference range, the positive or negative value of the real-time flow difference is obtained for determination.

[0034] If the real-time traffic difference is negative, the transmission is determined to be abnormal, and priority delay control is performed;

[0035] If the real-time traffic difference is positive, subtract the disk output from the disk input to obtain the real-time storage efficiency.

[0036] Furthermore, when the real-time flow difference is positive, the real-time storage efficiency is obtained, and the standard storage efficiency is compared with the real-time storage efficiency.

[0037] If the real-time storage efficiency is less than or equal to the standard storage efficiency, the processing is determined to be abnormal, and bandwidth is allocated to increase the network bandwidth;

[0038] If the real-time storage efficiency is greater than the standard storage efficiency, storage abnormality is determined and traffic delay control is performed.

[0039] Compared with the prior art, the beneficial effects of the present invention are that by classifying the target data into structured and unstructured data, and converting the unstructured data into text data for segmentation, different types of data can be effectively organized and managed to improve the efficiency and accuracy of data access, by extracting keywords and matching them in the historical database, high-traffic data fields can be timely identified, and the preset character length can be flexibly adjusted according to the comparison between the real-time access number and the standard access number, ensuring that the data segmentation and processing can better adapt to the actual access needs, thereby improving the efficiency and accuracy of data processing, by analyzing the change of the amount of historical data over time, drawing a change curve, it is convenient to identify the trend of data traffic, so as to reasonably allocate storage resources on different database nodes, optimize the storage layout, and thus improve the speed of data retrieval and processing, by analyzing the needs of database nodes, the system can classify the performance of the current transmission network, and clearly define it as a high-performance, standard-performance or low-performance network, so as to help operation and maintenance personnel take corresponding optimization measures to ensure the efficient operation of the network, by analyzing the needs of database nodes and classifying the network, resources can be reasonably allocated, redundancy can be reduced, and the system can be ensured to maintain high efficiency when facing different data loads, which not only improves resource utilization, but also reduces resource waste. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 The present invention is a flowchart of an AI-based intelligent computing center management method according to an embodiment of the present invention. DETAILED DESCRIPTION

[0041] In order to make the objects and advantages of the present invention more clearly understood, the present invention is further described below in conjunction with embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0042] The preferred embodiments of the present invention are described below with reference to the accompanying drawings. It should be understood by those skilled in the art that these embodiments are only used to explain the technical principles of the present invention and are not intended to limit the protection scope of the present invention.

[0043] It should be noted that, in the description of the present invention, terms such as "up", "down", "left", "right", "inside" and "outside" indicating directions or positional relationships are based on the directions or positional relationships shown in the drawings. This is merely for the convenience of description and does not indicate or imply that the device or element must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation on the present invention.

[0044] In addition, it should be noted that in the description of the present invention, unless otherwise clearly specified and limited, the terms "installed", "connected", and "connected" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be indirectly connected through an intermediate medium, or it can be the internal communication of two components. For those skilled in the art, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0045] See also Figure 1 As shown, it is a flow chart of an AI-based intelligent computing middle station management method according to an embodiment of the present invention. The present invention provides an AI-based intelligent computing middle station management method, including:

[0046] Step S1, receiving target data according to business requirements, classifying the target data to obtain structured data and unstructured data, converting the target data into text data, segmenting the unstructured data with a preset character length, and obtaining a plurality of real-time target segments after segmentation;

[0047] Step S2, for any of the real-time target segments, extract keywords and match them in a historical database to obtain the real-time access times of the real-time target segment, determine the real-time access times according to the standard access times, and correct the preset character length to the corrected character length according to the determination result;

[0048] Step S3, obtaining the historical data volume of the received target data with a preset acquisition time, drawing a curve of the change of the historical data volume over time, and storing the segmented real-time target segments on different database nodes according to the change curve;

[0049] Step S4, analyzing the requirements of the database nodes, and classifying the current transmission network according to the analysis results into high-performance networks, standard-performance networks, and low-performance networks.

[0050] In step S3, when the data volume is at a peak during a certain period of time, the real-time target fragments are stored in a database node configured for high performance; during relatively stable or low-traffic periods, they are stored on nodes with lower resource usage.

[0051] In this embodiment, since effective data management is crucial in industrial cloud computing, target data is classified into structured and unstructured data, and unstructured data is converted into text data for segmentation to effectively organize and manage different types of data, so as to improve the efficiency and accuracy of data access and meet the requirements of industrial cloud computing for data processing efficiency and accuracy. Keywords are extracted and matched in the historical database to timely identify high-traffic data fields, and the preset character length is flexibly adjusted according to the comparison between the real-time access number and the standard access number, so as to ensure that the segmentation and processing of data can better meet the actual access needs, thereby improving the efficiency and accuracy of data processing. Accuracy, by analyzing the changes in historical data volume over time, drawing change curves, it is easy to identify the trend of data traffic, so as to reasonably allocate storage resources on different database nodes, optimize the storage layout, and thus improve the speed of data retrieval and processing. By analyzing the needs of database nodes, the system can classify the performance of the current transmission network and clearly define it as a high-performance, standard-performance or low-performance network, so as to help operation and maintenance personnel take corresponding optimization measures to ensure the efficient operation of the network. Through the demand analysis and network classification of database nodes, resources can be reasonably allocated, redundancy can be reduced, and the system can be ensured to remain efficient when facing different data loads, which not only improves resource utilization but also reduces resource waste.

[0052] Specifically, in step S1, the unstructured data is segmented with a preset character length to obtain real-time target segments including:

[0053] When the target data is determined to be unstructured data, the real-time byte length of the text data is obtained, and the real-time byte length is determined according to the standard byte length.

[0054] If the real-time byte length is less than or equal to the standard byte length, the target data is determined to be a type of data, and the text data is not divided into a number of real-time target segments, and the text data is directly used as the real-time target segment;

[0055] If the real-time byte length is greater than the standard byte length, the target data corresponding to the text data is determined to be second-class data, and the text data is divided into a number of real-time target segments according to a preset byte segmentation length.

[0056] When processing large-scale data, the segmented real-time target segments can be stored and retrieved more conveniently, improving the efficiency of subsequent data access.

[0057] Specifically, obtaining the real-time access count of the real-time target segment includes:

[0058] For any real-time target segment, the associated keywords are obtained by extracting keywords and matching them in the historical database;

[0059] The keyword data items and the corresponding historical access times corresponding to each associated keyword are obtained in the historical database, and the maximum value of the historical access times among the associated keywords is selected as the real-time access times of the real-time target segment.

[0060] By extracting keywords and matching them in the historical database, we can obtain keyword data items that are highly relevant to the real-time target fields, thereby improving the accuracy of calculating the real-time access times. By associating keywords, we can accurately analyze user behavior and data request patterns, providing an in-depth understanding of data utilization. Selecting the maximum value of the historical access times as the real-time access times can reflect the most frequent usage, allowing the system to dynamically adjust data processing and storage strategies based on the access situation, optimize the storage layout of the database, and ensure that high-frequency access data blocks are preferentially stored on fast access media, thereby improving data retrieval speed.

[0061] Specifically, correcting the preset character length to a corrected character length according to the determination result includes:

[0062] Determine the real-time access times based on the standard access times.

[0063] If the real-time access number is less than the standard access number, the preset byte segmentation length corresponding to the real-time target segment is not corrected;

[0064] If the real-time access number is greater than or equal to the standard access number, the real-time access frequency corresponding to the real-time target segment is obtained, and the real-time access frequency is determined according to the standard access frequency to determine whether to modify the preset byte segmentation length.

[0065] Specifically, judging the real-time access frequency according to the standard access frequency includes:

[0066] For the real-time target segment, when it is determined that the real-time access number is greater than or equal to the standard access number, the keyword data item corresponding to the maximum value of the historical access number and the corresponding historical access frequency are obtained as the real-time access frequency, and the real-time access frequency is determined according to the standard access frequency.

[0067] If the real-time access frequency is less than or equal to the standard access frequency, the preset byte segmentation length corresponding to the real-time target segment is not corrected;

[0068] If the real-time access frequency is greater than the standard access frequency, the preset byte segmentation length corresponding to the real-time target segment is modified according to the number of real-time accesses and the real-time access frequency to a modified byte segmentation length;

[0069] Among them, Ss'=Ss×[1+(Fs-Fb) / Fs+(Ns-Nb) / Ns], Ss"=[Ss'], Ss' is the calculated simulated byte segmentation length, Ss" is the corrected byte segmentation length after correction, Ss is the preset byte segmentation length, Fs is the real-time access frequency, Fb is the standard access frequency, Ns is the real-time access number, and Nb is the standard access number.

[0070] By determining the number and frequency of real-time accesses, the data segmentation length is dynamically adjusted according to actual usage, which enhances the system's adaptability to changing traffic. When the number of real-time accesses is low, the original segmentation length is maintained to avoid unnecessary dynamic adjustments and save system resources. This option improves the system's processing efficiency and reduces the computing burden. By correcting the preset byte segmentation length, especially when the access frequency is high, data segmentation can be optimized, making data access more efficient and facilitating subsequent retrieval and analysis. The judgment and adjustment of real-time access frequency ensures that data processing can reflect the actual needs of users, reduces performance bottlenecks caused by unreasonable segmentation, and provides intelligent data segmentation decisions.

[0071] Specifically, the analysis of the requirements of the database node includes:

[0072] Obtaining the real-time memory utilization of any of the database nodes, comparing the real-time memory utilization with the first standard memory utilization and the second standard memory utilization, and classifying the performance type of the current transmission network according to the comparison result to obtain a classification result;

[0073] When determining that the current transmission network is a standard performance network, the bandwidth demand and the storage demand are associated for analysis, and whether the division result is accurate is verified based on the analysis result.

[0074] Specifically, a first standard memory utilization and a second standard memory utilization are set, and the performance types of the current transmission network are divided into the following according to the comparison results:

[0075] If the real-time memory utilization is greater than the first standard memory utilization, it is determined that the current transmission network is a low-performance network;

[0076] If the real-time memory utilization is less than the second standard memory utilization, the current transmission network is determined to be a high-performance network;

[0077] If the real-time memory utilization is between the first standard memory utilization and the second standard memory utilization, it is determined that the current transmission network is a standard performance network.

[0078] In this embodiment, the first standard memory utilization rate represents the degree of shortage of system memory resources, and the first standard memory utilization rate is set to 80%. The second standard memory utilization rate represents the degree of abundance of system memory resources, and the second standard memory utilization rate is set to 40%.

[0079] By obtaining the real-time memory utilization of database nodes, the system's memory resource status can be dynamically monitored, and potential resource shortages can be discovered in a timely manner. After determining that the network is a standard performance network, further analysis of bandwidth requirements and storage requirements can help identify specific resource requirements, optimize resource allocation, and ensure that the system can run smoothly under different conditions. The determination of real-time traffic differences and their positive and negative analysis enable the system to promptly identify network or storage anomalies, thereby taking targeted measures to ensure the reliability and efficiency of data transmission. By analyzing the priority of data packets, key data packets can be sent first, reducing delays in key tasks and improving the overall system's responsiveness.

[0080] Specifically, the analysis of associated bandwidth requirements and storage requirements includes:

[0081] The real-time flow difference is calculated according to the network input / output volume, the real-time storage difference is calculated according to the disk input / output volume, and the ratio of the real-time flow difference to the real-time storage difference is calculated to obtain the real-time storage efficiency;

[0082] Analyze real-time traffic difference, real-time storage difference, or real-time storage efficiency.

[0083] Specifically, the analysis of the real-time traffic difference, the real-time storage difference or the real-time storage efficiency includes:

[0084] Subtract the output from the input to get the real-time flow difference;

[0085] Determine the real-time flow difference according to the preset difference range.

[0086] If the real-time flow difference is within the preset difference range, the type classification is determined to be correct;

[0087] If the real-time traffic difference is not within the preset difference range, it is determined that the type classification is incorrect, and the second standard memory utilization is adjusted to the second corrected memory utilization;

[0088] Wherein, Wb2'=Wb2×[1+|ΔQs-ΔQb| / |ΔQs+ΔQb|], Wb2 is the second standard memory utilization, Wb2' is the second corrected memory utilization, ΔQs is the real-time flow difference, and ΔQb is the preset flow difference;

[0089] The preset difference range indicates an acceptable flow difference range, and the preset flow difference indicates an acceptable flow difference. The preset flow difference range is [-200MB, 200MB], and the preset flow difference is 0MB.

[0090] Based on the feedback of real-time traffic difference, the second standard memory utilization is adaptively increased to ensure the flexibility and stability of the system under dynamic load conditions.

[0091] Specifically, when the real-time flow difference is not within the preset difference range, the positive or negative value of the real-time flow difference is obtained for determination.

[0092] If the real-time traffic difference is negative, the transmission is determined to be abnormal, and priority delay control is performed;

[0093] If the real-time traffic difference is positive, subtract the disk output from the disk input to obtain the real-time storage efficiency.

[0094] In this embodiment, if the real-time traffic difference is positive, it means that the network receives more data than it sends, which may lead to increased storage pressure; if the real-time traffic difference is negative, it means that the data sent is more than the data received, which may indicate that the network transmission is stable and storage is available.

[0095] Specifically, priority delay control is to analyze data packets and set high priority for key data packets so as to give priority to sending key data.

[0096] Specifically, when the real-time flow difference is positive, the real-time storage efficiency is obtained, and the standard storage efficiency is compared with the real-time storage efficiency.

[0097] If the real-time storage efficiency is less than or equal to the standard storage efficiency, the processing is determined to be abnormal, and bandwidth is allocated to increase the network bandwidth;

[0098] If the real-time storage efficiency is greater than the standard storage efficiency, storage abnormality is determined and traffic delay control is performed.

[0099] When storage anomalies are detected, the problem can be solved by increasing the network bandwidth to ensure that data transmission is not affected by insufficient bandwidth.

[0100] Specifically, traffic delay control is to limit the data flow in the network to prevent the writing speed of the storage unit from failing to keep up with the receiving speed of the network data.

[0101] So far, the technical solutions of the present invention have been described in conjunction with the preferred embodiments shown in the accompanying drawings. However, it is easy for those skilled in the art to understand that the protection scope of the present invention is obviously not limited to these specific embodiments. Without departing from the principle of the present invention, those skilled in the art can make equivalent changes or substitutions to the relevant technical features, and the technical solutions after these changes or substitutions will fall within the protection scope of the present invention.

[0102] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. An AI-based intelligent computing platform management method, characterized in that: include, Step S1, receiving target data according to business requirements, classifying the target data to obtain structured data and unstructured data, converting the target data into text data, segmenting the unstructured data with a preset character length, and obtaining a plurality of real-time target segments after segmentation; Step S2, for any of the real-time target segments, extract keywords and match them in a historical database to obtain the real-time access times of the real-time target segment, determine the real-time access times according to the standard access times, and determine whether the preset character length meets the access requirements according to the determination result; Step S3, obtaining the historical data volume of the received target data with a preset acquisition time, drawing a curve of the change of the historical data volume over time, and storing the segmented real-time target segments on different database nodes according to the change curve; Step S4, analyzing the requirements of the database nodes, and classifying the current transmission network according to the analysis results into high-performance networks, standard-performance networks, and low-performance networks.

2. The AI-based intelligent computing center management method according to claim 1 is characterized in that: The unstructured data is segmented according to a preset character length, and a number of real-time target segments are obtained after segmentation, including: When the target data is determined to be unstructured data, the real-time byte length of the text data is obtained, and the real-time byte length is determined according to the standard byte length. If the real-time byte length is less than or equal to the standard byte length, the target data is determined to be a type of data, and the text data is not directly used as a real-time target segment; If the real-time byte length is greater than the standard byte length, the target data corresponding to the text data is determined to be second-class data, and the text data is divided into a number of real-time target segments according to a preset byte segmentation length.

3. The AI-based intelligent computing middle station management method according to claim 1 is characterized in that: Obtaining the real-time access count of the real-time target segment includes: For any real-time target segment, the associated keywords are obtained by extracting keywords and matching them in the historical database; The keyword data items and the corresponding historical access times corresponding to each associated keyword are obtained in the historical database, and the maximum value of the historical access times among the associated keywords is selected as the real-time access times of the real-time target segment.

4. The AI-based intelligent computing middle station management method according to claim 1 is characterized in that: Correcting the preset character length to a corrected character length according to the determination result includes: The real-time access number is determined according to the standard access number, and when the real-time access number is greater than or equal to the standard access number, the target segment obtains the real-time access frequency corresponding to the real-time target segment, and the real-time access frequency is determined according to the standard access frequency to determine whether to correct the preset byte segmentation length.

5. The AI-based intelligent computing center management method according to claim 1 is characterized in that: Determining the real-time access frequency based on the standard access frequency includes: For the real-time target segment, when it is determined that the real-time access number is greater than or equal to the standard access number, the keyword data item corresponding to the maximum value of the historical access number and the corresponding historical access frequency are obtained as the real-time access frequency, and the real-time access frequency is determined according to the standard access frequency. When the real-time access frequency is greater than the standard access frequency, the preset byte segmentation length corresponding to the real-time target segment is modified according to the real-time access times and the real-time access frequency to be modified to the modified byte segmentation length.

6. The AI-based intelligent computing middle station management method according to claim 1 is characterized in that: The requirements for the database node are analyzed including: Obtaining the real-time memory utilization of any of the database nodes, comparing the real-time memory utilization with the first standard memory utilization and the second standard memory utilization, and classifying the performance type of the current transmission network according to the comparison result to obtain a classification result; When determining that the current transmission network is a standard performance network, the bandwidth demand and the storage demand are associated for analysis, and whether the division result is accurate is verified based on the analysis result.

7. The AI-based intelligent computing middle station management method according to claim 6 is characterized in that: The analysis of bandwidth requirements and storage requirements includes: The real-time flow difference is calculated according to the network input / output volume, the real-time storage difference is calculated according to the disk input / output volume, and the ratio of the real-time flow difference to the real-time storage difference is calculated to obtain the real-time storage efficiency; Analyze real-time traffic difference, real-time storage difference, or real-time storage efficiency.

8. The AI-based intelligent computing middle station management method according to claim 7 is characterized in that: Analysis of real-time flow difference, real-time storage difference or real-time storage efficiency includes: Subtract the output from the input to get the real-time flow difference; Determine the real-time flow difference according to the preset difference range. If the real-time flow difference is within the preset difference range, the type classification is determined to be correct; If the real-time traffic difference is not within the preset difference range, it is determined that the type classification is incorrect, and the second standard memory utilization is adjusted.

9. The AI-based intelligent computing middle station management method according to claim 8 is characterized in that: When the real-time flow difference is not within the preset difference range, the positive and negative values ​​of the real-time flow difference are obtained for judgment. If the real-time traffic difference is negative, the transmission is determined to be abnormal, and priority delay control is performed; If the real-time traffic difference is positive, subtract the disk output from the disk input to obtain the real-time storage efficiency.

10. The AI-based intelligent computing middle station management method according to claim 9 is characterized in that: When the real-time flow difference is positive, the real-time storage efficiency is obtained, and the standard storage efficiency is compared with the real-time storage efficiency. If the real-time storage efficiency is less than or equal to the standard storage efficiency, the processing is determined to be abnormal, and bandwidth is allocated to increase the network bandwidth; If the real-time storage efficiency is greater than the standard storage efficiency, storage abnormality is determined and traffic delay control is performed.

Images