An Industrial Internet of Things Data Center-based Data Management System and Method

Abstract

The present invention relates to the field of data management, and provides a data management system and method based on an industrial Internet of Things data center. The system includes an industrial Internet of Things user platform, an industrial Internet of Things service platform, an industrial Internet of Things management platform, an industrial Internet of Things sensing network platform, and an industrial Internet of Things perception control platform. The industrial Internet of Things management platform is configured to: determine future acquisition parameters of the sensing network sub-platform based on the remaining storage space, the first retrieval feature, and the future retrieval feature; determine the pre-increased storage amount based on the future acquisition parameters; determine the group to be deleted based on the pre-increased storage amount, the remaining storage space, and the group retrieval features and group data amounts corresponding to multiple data groups; generate a parameter update instruction based on the future acquisition parameters; and generate a data deletion instruction based on the group to be deleted. The present invention can improve the standardization and flexibility of data management and improve the data management efficiency.

Description

Technical Field

[0001] The present invention relates to the field of data management, and particularly to a data management system and method based on an industrial Internet of Things data center. Background Art

[0002] With the continuous in-depth application of industrial Internet of Things technology, enterprise-level sensor network platforms have become the core support for industrial automation and intelligence. In these platforms, the configuration of multiple sub-databases provides dedicated storage spaces for different business data, ensuring the flexibility and efficiency of data management. However, how to perform efficient data collection and storage space maintenance for each sub-platform has become the key to improving the overall performance of the system.

[0003] Therefore, it is desired to provide a data management system and method based on an industrial Internet of Things data center, which can improve the standardization and flexibility of data management and enhance data management efficiency. Summary of the Invention

[0004] The summary of the invention includes providing a data management system based on an industrial Internet of Things data center. The system includes an industrial Internet of Things user platform, an industrial Internet of Things service platform, an industrial Internet of Things management platform, an industrial Internet of Things sensor network platform, and an industrial Internet of Things perception control platform. The industrial Internet of Things management platform includes a data center. The industrial Internet of Things sensor network platform includes multiple sensor network sub-platforms and multiple sensor network sub-databases, and the multiple sensor network sub-databases correspond to the multiple sensor network sub-platforms. The industrial Internet of Things management platform is configured to: every first preset period, for a single sensor network sub-platform among the multiple sensor network sub-platforms, determine the future acquisition parameters of the sensor network sub-platform based on the remaining storage space, the first retrieval feature, and the future retrieval feature of the sensor network sub-database corresponding to the sensor network sub-platform; determine the pre-increased storage amount of the sensor network sub-database based on the future acquisition parameters; divide the data items in the sensor network sub-database to obtain multiple data groups; determine the groups to be deleted based on the pre-increased storage amount, the remaining storage space, and the group retrieval features and group data amounts corresponding to the multiple data groups; generate a parameter update instruction based on the future acquisition parameters and send it to the sensor network sub-platform to control the sensor network sub-platform to perform data acquisition based on the future acquisition parameters; and generate a data deletion instruction based on the groups to be deleted and send it to the sensor network sub-platform to control the sensor network sub-platform to delete the groups to be deleted.

[0005] The invention content includes a data management method based on an industrial Internet of Things data center. The method is executed by an industrial Internet of Things management platform in a data management system based on the industrial Internet of Things data center. The method includes, for a single sensor network sub-platform among multiple sensor network sub-platforms every first preset period: determining future acquisition parameters of a sensor network sub-database corresponding to the sensor network sub-platform based on the remaining storage space, first retrieval features, and future retrieval features of the sensor network sub-platform; determining a pre-increased storage amount of the sensor network sub-database based on the future acquisition parameters; dividing data items in the sensor network sub-database to obtain multiple data groups; determining a group to be deleted based on the pre-increased storage amount, the remaining storage space, and the group retrieval features and group data amounts corresponding to the multiple data groups; generating a parameter update instruction based on the future acquisition parameters and sending it to the sensor network sub-platform to control the sensor network sub-platform to perform data acquisition based on the future acquisition parameters; and generating a data deletion instruction based on the group to be deleted and sending it to the sensor network sub-platform to control the sensor network sub-platform to delete the group to be deleted.

[0006] The beneficial effects of the present invention include but are not limited to: (1) Through the data management system based on the industrial Internet of Things data center, an information operation closed-loop can be formed among various functional platforms, and coordinated and regular operation can be achieved under the unified management of the industrial Internet of Things management platform, realizing the informatization and intelligence of data management. (2) Adjusting the communication bandwidth based on the heartbeat parameters between the sensor network sub-database and the data center can allocate appropriate communication bandwidth for each sensor network sub-platform, avoiding low data transmission efficiency caused by too small communication bandwidth or resource waste caused by too large communication bandwidth. (3) Based on the retrieved situation of the sensor network sub-database in historical time, multiple associated platform groups and the corresponding in-group association degrees can be determined. Considering that different associated sub-platforms have different influence degrees on the sensor network sub-platform, the weights of different associated sub-platforms are adjusted, so that the future retrieval features of the sensor network sub-database corresponding to the obtained sensor network sub-platform are more accurate. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] This specification will further illustrate in the form of exemplary embodiments, and these exemplary embodiments will be described in detail through the drawings. These embodiments are not restrictive. In these embodiments, the same numbers represent the same structures, where:

[0008] Figure 1 is a schematic diagram of the platform structure of a data management system based on an industrial Internet of Things data center shown in some embodiments of this specification;

[0009] Figure 2It is an exemplary flowchart of a data management method based on an industrial Internet of Things data center as shown in some embodiments of this specification;

[0010] Figure 3 It is an exemplary flowchart of determining future retrieval features as shown in some embodiments of this specification;

[0011] Figure 4 It is an exemplary schematic diagram of a prediction model as shown in some embodiments of this specification;

[0012] Figure 5 It is an exemplary flowchart of adjusting groups to be deleted as shown in some embodiments of this specification. Detailed implementation manners

[0013] To more clearly illustrate the technical solutions of the embodiments of this specification, the accompanying drawings required for the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings in the following description are only some examples or embodiments of this specification. For those of ordinary skill in the art, without creative efforts, this specification can also be applied to other similar scenarios based on these drawings. Unless obvious from the language context or otherwise stated, the same reference numerals in the figures represent the same structure or operation.

[0014] It should be understood that the "system" used herein is a method for distinguishing different components, elements, parts, portions or assemblies at different levels. However, if other words can achieve the same purpose, the said words can be replaced by other expressions.

[0015] Figure 1 It is a platform structure diagram of a data management system based on an industrial Internet of Things data center as shown in some embodiments of this specification.

[0016] In some embodiments, as Figure 1 shown, the data management system 100 based on the industrial Internet of Things data center includes an industrial Internet of Things user platform 110, an industrial Internet of Things service platform 120, an industrial Internet of Things management platform 130, an industrial Internet of Things sensing network platform 140, and an industrial Internet of Things perception control platform 150.

[0017] The industrial Internet of Things user platform 110 refers to a platform for interacting with users. In some embodiments, the industrial Internet of Things user platform can be configured as a terminal device. Among them, the terminal device includes a mobile device, a laptop computer, etc.

[0018] The Industrial Internet of Things (IIoT) service platform 120 refers to a platform for providing data management services based on the IIoT for users. In some embodiments, the IIoT service platform is configured as a communication network or a server, etc. The IIoT service platform can interact with the IIoT management platform and the IIoT user platform.

[0019] The IIoT management platform 130 refers to a comprehensive management platform for various data. In some embodiments, the IIoT management platform is configured as a server. In some embodiments, the IIoT management platform may further include a processor and / or a memory, etc.

[0020] In some embodiments, the IIoT management platform includes a data center 131. The data center is configured as a platform for storing information and / or data related to the data management system 100 of the IIoT data center.

[0021] The IIoT sensing network platform 140 refers to a platform for comprehensively managing sensing information. In some embodiments, the IIoT sensing network platform is configured as a communication network or a gateway, etc. The IIoT sensing network platform can interact with the IIoT management platform and the IIoT perception control platform.

[0022] In some embodiments, the IIoT sensing network platform includes multiple sensing network sub-databases and multiple sensing network sub-platforms.

[0023] As Figure 1 shown, the IIoT sensing network platform 140 includes a first sensing network sub-database 141-1, a second sensing network sub-database 141-2, etc., and a first sensing network sub-platform 142-1, a second sensing network sub-platform 142-2, etc.

[0024] The sensing network sub-database refers to a database for storing and managing the sensing information obtained by the sensing network sub-platform. The number of sensing network sub-databases is the same as that of the sensing network sub-platforms, and one sensing network sub-database corresponds to one sensing network sub-platform.

[0025] The IIoT perception control platform 150 refers to a functional platform for generating sensing information and executing control information. In some embodiments, the IIoT perception control platform includes various production and manufacturing equipment terminals and sensors. For example, the production and manufacturing terminals include automated production equipment, quality inspection equipment, etc., and the sensors include pressure sensors, image sensors, temperature sensors, etc.

[0026] The sensing network sub-platform refers to a platform that collects sensing information based on business requirements. For example, business requirements include production monitoring, energy management, etc. Correspondingly, the sensing network sub-platform includes a production monitoring platform, an energy management platform, etc. Among them, the production monitoring platform refers to a platform used to monitor the operating status of a factory production line in real time. The energy management platform refers to a platform used to collect and analyze the energy consumption data of a factory or facility.

[0027] In some embodiments, the sensing network sub-platform collects sensing information corresponding to business requirements through multiple sensing devices in the industrial Internet of Things perception and control platform, and uploads the sensing information to the industrial Internet of Things management platform. The sensing network sub-platform stores the obtained sensing information in the corresponding sensing network sub-database.

[0028] In some embodiments, the industrial Internet of Things management platform can retrieve data from the sensing network sub-database based on retrieval requirements, and upload the data to the industrial Internet of Things user platform through the industrial Internet of Things service platform. Among them, the retrieval requirement refers to the type of sensing network sub-database that the user of the industrial Internet of Things management platform wants to retrieve. The retrieval requirement corresponds to the business requirement. For example, the retrieval requirement includes retrieving data from the sensing network sub-database corresponding to the energy management platform, etc.

[0029] It can be understood that business requirements may only require the sensing information of some sensing devices. Therefore, the sensing network sub-platform can collect the sensing information of the sensing devices related to the business requirements. At the same time, if all the data collected by the sensing network sub-platform is stored in the data center, the load on the data center is relatively large. Therefore, storing the data obtained by the sensing network sub-platform through the sensing network sub-database can effectively reduce the load on the data center and clearly divide the data corresponding to different sensing network sub-platforms.

[0030] In some embodiments, the data management system 100 based on the industrial Internet of Things data center further includes a processor and a memory. The processor is configured to process information and / or data related to the data management system 100 based on the industrial Internet of Things data center. The processor includes a central processing unit (CPU), an application-specific instruction processor (ASIP), a graphics processing unit (GPU), etc. or any combination thereof.

[0031] For the foregoing detailed description, reference can be made to Figures 2 to 5 the relevant description.

[0032] Through the data management system based on the industrial Internet of Things data center, an information operation closed-loop can be formed between each functional platform, and coordinated and regular operation can be achieved under the unified management of the industrial Internet of Things management platform, realizing the informatization and intelligentization of data management.

[0033] Figure 2is an exemplary flowchart of a data management method based on an industrial Internet of Things data center as shown in some embodiments of this specification. As Figure 2 shown, the process 200 of the data management method based on the industrial Internet of Things data center includes the following steps. In some embodiments, the process 200 of the data management method based on the industrial Internet of Things data center can be executed by an industrial Internet of Things management platform (hereinafter referred to as the management platform).

[0034] In some embodiments, every first preset period, for a single sensor network sub-platform among multiple sensor network sub-platforms, the management platform can respectively execute steps 210 - 260 to generate parameter update instructions and / or data deletion instructions, etc. Among them, the first preset period is set in advance based on historical experience.

[0035] For the description of the data management system based on the industrial Internet of Things data center and each platform, see Figure 1 and its related descriptions.

[0036] Step 210, based on the remaining storage space, the first retrieval feature, and the future retrieval feature of the sensor network sub-database corresponding to the sensor network sub-platform, determine the future acquisition parameters of the sensor network sub-platform.

[0037] The remaining storage space refers to the size of the remaining storage space in the sensor network sub-database. The storage space size is represented by the amount of data that can be stored. In some embodiments, the management platform obtains the remaining storage space from the sensor network sub-database.

[0038] The first retrieval feature refers to the retrieval feature of the sensor network sub-database within a preset historical period. The retrieval feature is data used to characterize the retrieval situation of the sensor network sub-database. For example, the retrieval feature includes at least one of the number of retrieval times of the sensor network sub-database, the data items retrieved each time, and the average amount of data retrieved multiple times, etc. The preset historical period refers to a period of time in the past from the current time point. The preset historical period is set in advance based on historical experience. The data item refers to the data of the sensor network sub-platform stored in the sensor network sub-database.

[0039] In some embodiments, the management platform obtains the first retrieval feature by counting the historical retrieval process of the sensor network sub-database within the preset historical period.

[0040] The future retrieval feature refers to the retrieval feature of the sensor network sub-database within a preset future period. The preset future period is set in advance based on historical experience.

[0041] In some embodiments, the management platform determines the future retrieval feature through various methods. For example, the management platform obtains the future retrieval feature by manual input.

[0042] In some embodiments, the management platform may also obtain multiple associated sub - platforms of the sensing network sub - platform; and determine future retrieval features based on the first stored data corresponding to the multiple associated sub - platforms.

[0043] An associated sub - platform refers to other sensing network sub - platforms that are associated with the current sensing network sub - platform.

[0044] In some embodiments, the management platform obtains associated sub - platforms through various methods. For example, the management platform obtains associated sub - platforms by manual input.

[0045] In some embodiments, the management platform may also determine multiple associated sub - platforms based on multiple associated platform groups. For details on this part, see Figure 3 and its related descriptions.

[0046] The first stored data refers to data related to the data storage of the sensing network sub - database corresponding to the associated sub - platform. In some embodiments, the first stored data includes the historical storage sequence of the associated sub - platform, etc. The historical storage sequence is a sequence composed of the data storage rates of multiple sampling points in a preset historical period by the sensing network sub - database corresponding to the associated sub - platform. Among them, the multiple sampling points are preset based on historical experience.

[0047] In some embodiments, the data storage rate is represented by the amount of data stored per unit time.

[0048] In some embodiments, for a single associated sub - platform among the multiple associated sub - platforms, the management platform constructs a first target vector based on the first stored data of the sensing network sub - database corresponding to the associated sub - platform, matches it in the first vector database, obtains multiple first feature vectors that meet the first similarity condition, and determines the mean of the labels of the multiple first feature vectors as the candidate retrieval feature of the sensing network sub - database corresponding to the associated sub - platform. The first similarity condition includes that the similarity to the first target vector is greater than the first similarity threshold, and the first similarity threshold is set based on experience. The similarity between vectors is negatively correlated with the vector distance. The vector distance includes the Euclidean distance, etc.

[0049] In some embodiments, the first vector database is constructed based on historical data. For example, the management platform constructs multiple first feature vectors based on multiple historical first stored data corresponding to the sensing network sub - database of the associated sub - platform in the first historical time in the historical data, and determines the actual retrieval feature of the sensing network sub - database corresponding to the associated sub - platform at the second historical time as the label corresponding to the first feature vector. The first historical time is earlier than the second historical time.

[0050] In some embodiments, the management platform determines the candidate retrieval features of the sensor network sub-databases corresponding to multiple associated sub-platforms through the above method, performs a weighted sum of the retrieval times among the multiple candidate retrieval features, and uses the sum result as the retrieval times in the future retrieval features of the sensor network sub-database corresponding to the sensor network sub-platform. The management platform may also use the data items retrieved among the multiple candidate retrieval features as the data items to be retrieved in the future retrieval features, and count the total amount of data of the retrieved data items as the data volume in the future retrieval features. The weights of different associated sub-platforms are set based on experience.

[0051] In some embodiments, the management platform determines the future retrieval features based on the intra-group association degree corresponding to multiple associated sub-platforms and the first stored data. For more details, see Figure 3 and the relevant descriptions.

[0052] By referring to the retrieval features of other sensor network sub-platforms with relatively high relevance, the future retrieval features of the current sensor network sub-platform can be determined more accurately.

[0053] The future acquisition parameters refer to the acquisition parameters of the sensor network sub-platform in a preset future time period. The acquisition parameters include the acquisition frequency, acquisition accuracy, etc. Among them, the acquisition accuracy refers to the accuracy of acquiring sensor information. For example, the pixel value of image sensor information, etc.

[0054] In some embodiments, the management platform determines the future acquisition parameters through various methods. For example, the management platform constructs a vector to be clustered based on the remaining storage space, the first retrieval feature, and the future retrieval feature, and constructs clustering vectors based on the historical remaining storage space, the historical first retrieval feature, and the historical future retrieval feature in the historical data. The management platform performs clustering on the vector to be clustered and the multiple clustering vectors to obtain multiple clustering clusters, determines the clustering cluster containing the vector to be clustered as the target cluster, and uses the future acquisition parameter with the best acquisition effect among the future acquisition parameters corresponding to all the clustering vectors in the target cluster as the future acquisition parameter of the sensor network sub-platform. The acquisition effect is positively correlated with the retrieval success rate of the data items in the sensor network sub-database by the management platform after acquiring data based on the future acquisition parameters. The higher the retrieval success rate, the better the acquisition effect. The retrieval success rate refers to the ratio of the number of data items actually retrieved by the management platform from the sensor network sub-database to the number of data items that need to be retrieved. If the sensor network sub-database includes all the data items that the management platform needs to retrieve, the retrieval success rate is 100%.

[0055] The future acquisition parameters corresponding to the clustering vectors can be determined in advance according to the historical data. For example, the actual acquisition parameters of the sensor network sub-platform corresponding to the clustering vector at the third historical time are determined as the future acquisition parameters of the fourth historical time corresponding to the clustering vector. The third historical time is earlier than the fourth historical time.

[0056] In some embodiments, the management platform may adjust the future acquisition parameters of the target sub-platform based on the future acquisition parameters of multiple associated sub-platforms corresponding to the target sub-platform. For more details, see Figure 4 and the related descriptions.

[0057] Step 220: Determine the pre-increased storage capacity of the sensor network sub-database based on the future acquisition parameters.

[0058] The pre-increased storage capacity refers to the increased data storage capacity of the sensor network sub-database within a preset future time period.

[0059] In some embodiments, the management platform determines the pre-increased storage capacity based on the future acquisition parameters of the sensor network sub-platform. For example, the management platform takes the result of multiplying the acquisition frequency within a preset future time period by the amount of data collected by the sensor device each time as the pre-increased storage capacity. Here, the amount of data collected by the sensor device each time is a fixed value.

[0060] In some embodiments, the management platform adjusts the pre-increased storage capacity based on the first compression amount average value and the second compression amount average value of the sensor network sub-database in multiple historical time intervals. For example, the management platform may take the sum of the first compression amount average value and the second compression amount average value of multiple historical time intervals as the normal compression amount of the sensor network sub-database, and take the difference between the pre-increased storage capacity and the normal compression amount as the final pre-increased storage capacity. Here, the first compression amount refers to the compression amount when the sensor network sub-platform performs duplicate data compression on the data collected by the sensor device. The second compression amount refers to the compression amount when the sensor network sub-platform performs data precision compression on the data collected by the sensor device. The multiple historical time intervals are set based on experience.

[0061] Step 230: Divide the data items in the sensor network sub-database to obtain multiple data groups.

[0062] A data group refers to a group composed of multiple data items. In some embodiments, the management platform divides the data items in the sensor network sub-database based on the retrieval time series to obtain multiple data groups. For example, the management platform clusters the multiple data items in the sensor network sub-database based on the retrieval time series corresponding to the data items, and takes the multiple clustering clusters obtained after clustering as the multiple data groups. Here, the retrieval time series refers to the sequence composed of the times when the data items are retrieved multiple times in the historical data. The management platform counts the data items retrieved each time in the first retrieval feature, obtains the time when each data item is retrieved based on the historical data, and forms the retrieval time series with the times when each data item is retrieved multiple times.

[0063] Step 240: Determine the group to be deleted based on the pre-increased storage capacity, the remaining storage space, and the group retrieval features and group data amounts corresponding to the multiple data groups.

[0064] The grouped retrieval feature refers to the retrieval feature of data groups in a preset future time period. In some embodiments, the grouped retrieval feature includes at least one of the total retrieval times and the total retrieval volume of multiple data items within a data group in a preset future time period, etc.

[0065] The grouped data volume refers to the total data volume of multiple data items within a data group. The management platform takes the sum value of the data volumes of multiple data items calculated within the data group as the grouped data volume.

[0066] In some embodiments, the management platform determines the grouped retrieval feature corresponding to each data group based on the future retrieval feature. For example, the management platform counts the number of occurrences of data items within a data group in the future retrieval feature, takes the product of the number and the retrieval times in the future retrieval feature as the total retrieval times in the grouped retrieval feature, and takes the product of the data volume of the data items that appear in the future retrieval feature and the retrieval times in the future retrieval feature as the total retrieval volume.

[0067] The group to be deleted refers to the data group that needs to be deleted. The group to be deleted includes one or more data groups.

[0068] In some embodiments, when the remaining storage space meets the preset deletion condition, the management platform iteratively includes the data group with the least total retrieval times into the group to be deleted based on multiple data groups until the storage space of the sensor network sub-database after deleting the group to be deleted does not meet the preset deletion condition, and then the determination of the group to be deleted is completed. Among them, the preset deletion condition includes that the storage space is less than the pre-increased storage volume or the difference between the storage space and the pre-increased storage volume is less than the storage volume threshold. The storage volume threshold is preset based on historical experience. The storage space of the sensor network sub-database after deleting the group to be deleted is represented by the sum value of the remaining storage space before iteration and the grouped data volume of the group to be deleted.

[0069] In some embodiments, the management platform can adjust the group to be deleted based on the estimated retrieval feature. For more content, see Figure 5 and related descriptions.

[0070] Step 250: Generate a parameter update instruction based on future acquisition parameters and send it to the sensor network sub-platform to control the sensor network sub-platform to perform data acquisition based on the future acquisition parameters.

[0071] The parameter update instruction refers to the instruction for controlling the sensor network sub-platform to update the acquisition parameters. In some embodiments, the management platform converts the future acquisition parameters into machine instructions and sends the machine instructions as the parameter update instruction to the sensor network sub-platform to control the sensor network sub-platform to perform data acquisition based on the future acquisition parameters.

[0072] Step 260: Generate a data deletion instruction based on the grouping to be deleted and send it to the sub-platform of the sensing network to control the sub-platform of the sensing network to delete the grouping to be deleted.

[0073] The data deletion instruction refers to an instruction for controlling the sub-platform of the sensing network to delete the grouping to be deleted. In some embodiments, the management platform converts the grouping to be deleted into a machine instruction and sends the machine instruction as the data deletion instruction to the sub-platform of the sensing network to control the sub-platform of the sensing network to delete the grouping to be deleted.

[0074] Generating a parameter update instruction through future acquisition parameters can control the sub-platform of the sensing network to perform more comprehensive data acquisition, avoiding the situation where corresponding data items cannot be retrieved. By determining the grouping to be deleted and generating a data deletion instruction, unused data can be deleted in a timely manner, improving the standardization and flexibility of data management and enhancing data management efficiency.

[0075] In some embodiments, every first preset period, the management platform can also determine the heartbeat parameters between the sub-database of the sensing network and the data center based on the future acquisition parameters, determine the communication bandwidth between the sub-platform of the sensing network and the data center based on the heartbeat parameters, and adjust the communication parameters between the sub-platform of the sensing network and the data center based on the communication bandwidth.

[0076] In some embodiments, the heartbeat parameters include parameters related to data transmission between the sub-database of the sensing network and the data center. For example, the heartbeat parameters include at least one of a heartbeat period and the type of heartbeat transmission information, etc.

[0077] The heartbeat period refers to the time interval for the sub-database of the sensing network to transmit heartbeat transmission information to the data center.

[0078] The heartbeat transmission information refers to information related to data transmission between the sub-database of the sensing network and the data center. The type of heartbeat transmission information includes at least one of status information, storage information, and the number of connections, etc.

[0079] The status information is used to reflect the current status of the sub-database of the sensing network. For example, the sub-database of the sensing network is available or faulty, etc. The storage information includes the total amount of data stored in the sub-database of the sensing network and the remaining storage space, etc. The number of connections refers to the number of data centers or sub-platforms of the sensing network connected to the sub-database of the sensing network.

[0080] In some embodiments, the management platform queries the heartbeat parameters corresponding to the future acquisition parameters in the first preset table as the current heartbeat parameters based on the future acquisition parameters. The first preset table is constructed by technicians based on historical experience and includes multiple pairs of corresponding relationships between future acquisition parameters and heartbeat parameters.

[0081] In some embodiments, every second preset period, in response to the load data of the communication link between multiple sub-platforms of the sensing network and the data center not meeting the preset load condition, the management platform adjusts the heartbeat parameter based on the current stored data of multiple associated sub-platforms corresponding to the sub-platform of the sensing network.

[0082] In some embodiments, the management platform can adjust the heartbeat parameter between the sensing network sub-database corresponding to each sensing network sub-platform and the data center based on the current stored data of multiple associated sub-platforms corresponding to each sensing network sub-platform. The following takes a single sensing network sub-platform as an example for illustration.

[0083] The second preset period refers to the period during which the management platform adjusts the heartbeat parameter. The second preset period is set based on experience. The second preset period is shorter than the first preset period.

[0084] The load data refers to the data traffic on the communication link between the data center and multiple sensing network sub-platforms. Among them, multiple sensing network sub-platforms share a communication link to transmit information to the data center.

[0085] In some embodiments, the management platform obtains the load data by real-time monitoring of the data traffic on the communication link.

[0086] The preset load condition refers to the condition for judging whether the heartbeat parameter needs to be adjusted. In some embodiments, the preset load condition includes that the load data is less than the preset traffic threshold, etc. The preset traffic threshold is set based on experience.

[0087] The current stored data refers to the data related to the data storage of the sensing network sub-database corresponding to the associated sub-platform corresponding to the sensing network sub-platform at the current moment. In some embodiments, the current stored data includes the data storage rate of the sensing network sub-database at the current moment, etc. For more content about the data storage rate, see step 210 and its related descriptions.

[0088] In some embodiments, the management platform calculates the weighted sum of the current stored data of multiple associated sub-platforms corresponding to the sensing network sub-platform, queries the heartbeat period adjustment amount and the type of heartbeat transmission information corresponding to the weighted sum in the second preset table based on the weighted sum, adjusts the heartbeat period based on the heartbeat period adjustment amount, and forms a new adjustment parameter by combining the adjusted heartbeat period and the type of heartbeat transmission information. The weights of different associated sub-platforms are set based on experience.

[0089] The second preset table is constructed by technicians based on historical experience and includes multiple pairs of corresponding relationships between the weighted sum and the heartbeat period adjustment amount and the type of heartbeat transmission information.

[0090] When the load data exceeds the preset traffic threshold, the heartbeat parameter is adjusted based on the data storage rate of the sub-database of the sensing network at the current moment, which can timely adjust the heartbeat parameter to reduce the data traffic on the communication link when the load of the communication link is large, and ensure the normal transmission of data in the communication link.

[0091] In some embodiments, the management platform allocates the communication bandwidth of each sensing network sub-platform based on the ratio of the heartbeat parameters between the sub-databases of the sensing networks corresponding to the multiple sensing network sub-platforms and the data center. For example, the management platform allocates the same proportion of communication bandwidth for each sensing network sub-platform based on the ratio of the heartbeat parameters. Among them, the ratio of the heartbeat parameters is represented by the ratio of the weighted sum of the heartbeat period and the amount of data transmitted in the heartbeat. Among them, the weights of the heartbeat period and the data transmitted in the heartbeat are set based on experience. The weight of the heartbeat period can be negative.

[0092] Communication parameters refer to the parameters related to the communication between the sensing network sub-platform and the data center. In some embodiments, the communication parameters include the routing parameters of the routers on the communication link between the sensing network sub-platform and the data center, etc. The routing parameters include the IP addresses of the sensing network sub-platform and the data center and the corresponding bandwidth limits, etc.

[0093] In some embodiments, the management platform takes the communication bandwidth corresponding to the sensing network sub-platform as the bandwidth limit in the routing parameters.

[0094] Adjusting the communication bandwidth based on the heartbeat parameters between the sub-database of the sensing network and the data center can allocate appropriate communication bandwidth for each sensing network sub-platform, avoiding low data transmission efficiency caused by too small communication bandwidth or resource waste caused by too large communication bandwidth.

[0095] It should be noted that the above description of the process 200 of the data management method based on the industrial Internet of Things data center is only for illustration and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process 200 of the data management method based on the industrial Internet of Things data center under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0096] Figure 3 is an exemplary flowchart for determining future retrieval features shown in some embodiments of this specification. As Figure 3 shown, the process 300 for determining future retrieval features includes the following steps.

[0097] Step 310, based on the second retrieval features corresponding to multiple sensing network sub-platforms, determine multiple associated platform groups and the degree of intra-group association corresponding to the multiple associated platform groups.

[0098] The second retrieval feature refers to data related to the retrieval time of the sub-database of the sensor network corresponding to the sub-platform of the sensor network. In some embodiments, the second retrieval feature includes the time when the sub-database of the sensor network is retrieved each time in the historical data. The second retrieval feature is obtained based on the historical data.

[0099] An associated platform group refers to a combination of sub-platforms of a sensor network that are associated with each other. Each associated platform group includes two sub-platforms of a sensor network that are associated with each other.

[0100] In some embodiments, the management platform calculates the explicit association value and the implicit association value corresponding to any two sub-platforms of the sensor network based on the second retrieval features corresponding to multiple sub-platforms of the sensor network. In response to the sum of the explicit association value and the implicit association value being greater than the association value threshold, the management platform determines these two sub-platforms of the sensor network as a group of associated platform groups. The association value threshold is set based on experience.

[0101] The explicit association value is positively correlated with the number of occurrences of positive events in multiple historical time intervals. A positive event refers to the number of times that the sub-databases of two sub-platforms of the sensor network are retrieved simultaneously in a single historical time interval being greater than a preset explicit threshold. The preset explicit threshold is set based on experience. Among them, being retrieved simultaneously means that in a single retrieval instruction issued by the data center, it is required to retrieve the data of the sub-databases of two sub-platforms of the sensor network simultaneously. For the description of multiple historical time intervals, see step 220 and its related descriptions.

[0102] In some embodiments, the preset explicit threshold can be positively correlated with the mean value of the number of sub-databases of the sensor network required to be retrieved in multiple retrieval instructions.

[0103] The greater the mean value of the number of sub-databases of the sensor network required to be retrieved in multiple retrieval instructions, the lower the correlation between the two sub-platforms of the sensor network that are retrieved simultaneously. At this time, it is necessary to appropriately increase the preset explicit threshold to ensure the accuracy of the explicit association value.

[0104] The implicit association value is positively correlated with the number of occurrences of negative events in multiple historical time intervals. A negative event refers to the number of times that the sub-databases of two sub-platforms of the sensor network are retrieved separately being less than a preset implicit threshold in two or more consecutive historical time intervals. The preset implicit threshold is set based on experience.

[0105] In some embodiments, the preset implicit threshold can be positively correlated with the weighted sum of the retrieval frequencies of the sub-databases of the sensor network corresponding to two sub-platforms of the sensor network. Among them, the retrieval frequency is the ratio of the mean value of the number of times the sub-database of the sensor network is retrieved in multiple historical time intervals to the duration of a single historical time interval. The weights corresponding to the two sub-platforms of the sensor network are set based on experience.

[0106] When the retrieval frequency is relatively small, it indicates that the two sub-databases of the sensing network are retrieved less frequently. At this time, the preset implicit threshold can be appropriately reduced to avoid the situation where the number of times the two sub-databases of the sensing network are retrieved alone is always less than the preset implicit threshold.

[0107] The degree of intra-group association refers to the degree of association between two sub-platforms of the sensing network in the associated platform group. Each sub-platform of the sensing network in the associated platform group corresponds to a degree of intra-group association respectively.

[0108] In some embodiments, based on the second retrieval characteristics corresponding to multiple sub-platforms of the sensing network, the management platform counts the number of times the sub-databases of the sensing network corresponding to two sub-platforms of the associated platform group are retrieved alone and the number of times they are retrieved simultaneously in multiple historical time intervals, and uses the ratio of the number of times each sub-database of the sensing network is retrieved simultaneously to the number of times it is retrieved alone as the degree of intra-group association of the sub-platform of the sensing network corresponding to the sub-database of the sensing network.

[0109] Step 320: Determine multiple associated sub-platforms based on multiple associated platform groups.

[0110] In some embodiments, the two sub-platforms of the sensing network in each associated platform group are associated sub-platforms with each other. The management platform can count all the associated platform groups containing the current sub-platform of the sensing network, and determine the other sub-platforms of the sensing network in these associated platform groups as the associated sub-platforms of the current sub-platform of the sensing network. For more content about the associated sub-platforms, see Figure 2 and related descriptions. The management platform can determine the associated sub-platforms of all sub-platforms of the sensing network in the above manner.

[0111] Step 330: Determine future retrieval characteristics based on the degree of intra-group association corresponding to multiple associated sub-platforms and the first stored data.

[0112] For more content about the first stored data and future retrieval characteristics, see Figure 2 and related descriptions.

[0113] In some embodiments, the management platform allocates weights to multiple associated sub-platforms when determining future retrieval characteristics based on the proportion of the degree of intra-group association corresponding to the multiple associated sub-platforms of the current sub-platform of the sensing network. For example, the management platform allocates the same proportion of weights to each associated sub-platform based on the proportion of the degree of intra-group association.

[0114] The management platform obtains the future retrieval characteristics of the sub-database of the sensing network corresponding to the current sub-platform of the sensing network by weighted summation based on multiple candidate retrieval characteristics of the sub-databases of the sensing network corresponding to multiple associated sub-platforms. For the description of obtaining future retrieval characteristics by weighted summation based on multiple candidate retrieval characteristics, see Figure 2and related descriptions.

[0115] In some embodiments, the management platform constructs a data retrieval map based on multiple associated platform groups, the degree of association within the group, and the second stored data corresponding to multiple sensor network sub-platforms; based on the data retrieval map, through a prediction model, it determines future retrieval characteristics. For more details, see Figure 4 and related descriptions.

[0116] Based on the retrieval situation of the sensor network sub-database in historical time, multiple associated platform groups and the corresponding degree of association within the group can be determined. Considering that different associated sub-platforms have different influence degrees on the sensor network sub-platforms, the weights of different associated sub-platforms are adjusted accordingly, so that the future retrieval characteristics of the sensor network sub-database corresponding to the sensor network sub-platforms are more accurate.

[0117] It should be noted that the above description of process 300 for determining future retrieval characteristics is only for illustration and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to process 300 for determining future retrieval characteristics under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0118] Figure 4 is an exemplary schematic diagram of a prediction model shown in some embodiments of this specification.

[0119] In some embodiments, as Figure 4 shown, the management platform constructs a data retrieval map 440 based on multiple associated platform groups 410, the degree of association within the group 420 corresponding to multiple associated sub-platforms, and the second stored data 430 corresponding to multiple sensor network sub-platforms. Based on the data retrieval map 440, it determines future retrieval characteristics 460 through a prediction model 450.

[0120] For more details about associated platform groups, future retrieval characteristics, and the degree of association within the group, see Figure 2 and Figure 3 and related descriptions.

[0121] The second stored data refers to the data related to the data storage of the sensor network sub-database corresponding to the sensor network sub-platform. In some embodiments, the second stored data includes the historical storage sequence of the sensor network sub-database corresponding to the sensor network sub-platform. For the description of the historical storage sequence, see step 210 and its related descriptions.

[0122] The data retrieval map refers to a graph structure that can reflect the data retrieval situation of the sensor network sub-databases corresponding to multiple sensor network sub-platforms. The graph structure is a data structure composed of nodes and edges. The edges connect the nodes, and the nodes and edges can have characteristics.

[0123] In some embodiments, the management platform constructs a data retrieval map based on multiple associated platform groups, the degree of intra-group association corresponding to multiple associated sub-platforms, and the second stored data corresponding to multiple sensor network sub-platforms. For example, the nodes of the data retrieval map (such as node 441, etc.) include sensor network sub-platforms, and the node features include the second stored data of the sensor network sub-database corresponding to the sensor network sub-platform.

[0124] The edges in the data retrieval map can represent the association relationships between nodes. For example, if two sensor network sub-platforms belong to an associated platform group, there are two directed edges with opposite directions between the two sensor network sub-platforms (such as edge 442, etc.), and the edge features of the directed edge include the degree of group association of the sensor network sub-platform corresponding to the starting node of the directed edge in the associated platform group.

[0125] The prediction model is a model used to determine future retrieval features. In some embodiments, the prediction model is a machine learning model. For example, the prediction model is any one or combination of a Graph Neural Network (GNN) model, etc. or other custom model structures.

[0126] In some embodiments, the input of the prediction model includes the data retrieval map. The output of the prediction model includes the future retrieval features of the sensor network sub-databases corresponding to each node in the data retrieval map.

[0127] In some embodiments, the management platform trains a prediction model based on multiple labeled training samples. For example, the management platform can input multiple training samples into an initial prediction model, construct a loss function based on the output of the initial prediction model and the labels of the training samples, iteratively update the parameters of the initial prediction model based on the loss function, and end the iteration when the iteration completion condition is met to obtain a trained prediction model. Among them, the method of iterative update includes but is not limited to the gradient descent method, etc. The iteration completion conditions include the convergence of the loss function or the number of iterations reaching a threshold, etc.

[0128] The training samples include sample data retrieval maps, and the sample data retrieval maps include historical data retrieval maps determined based on historical data at the fifth historical time. The nodes and their features, and the edges and their features of the historical data retrieval maps are similar to the above descriptions.

[0129] The label of the training sample refers to the actual retrieval feature of each sample node at the sixth historical time in the sample data retrieval map. The sixth historical time is later than the fifth historical time.

[0130] Based on the data retrieval map, the future retrieval features are determined through a prediction model, and the future retrieval features of the sensor network sub-databases corresponding to multiple sensor network sub-platforms can be determined simultaneously, so as to realize the unified management of multiple sensor network sub-platforms. By using a machine learning model to determine the future retrieval features, the efficiency and accuracy of determining the future retrieval features can be improved.

[0131] In some embodiments, the management platform determines the update priorities corresponding to multiple sensor network sub-platforms through the data retrieval map; based on the update priorities, multiple future acquisition parameters corresponding to the multiple sensor network sub-platforms are determined. Among them, for the target sub-platform whose update priority meets the preset adjustment condition, the future acquisition parameters of the target sub-platform are adjusted based on the future acquisition parameters of multiple associated sub-platforms corresponding to the target sub-platform.

[0132] The update priority is the order of adjusting the future acquisition parameters. In some embodiments, the management platform sorts the out-degrees of each node in the data retrieval map from large to small, and uses the sorting result as the update priority of the sensor network sub-platform corresponding to each node. The smaller the value corresponding to the sorting result, the higher the ranking, and the higher the corresponding update priority. Among them, the out-degree of each node refers to the number of directed edges pointing from this node to other nodes.

[0133] In some embodiments, the management platform determines the sensor network sub-platform whose update priority meets the preset adjustment condition as the target sub-platform, and adjusts the future acquisition parameters of the target sub-platform based on the future acquisition parameters of multiple associated sub-platforms corresponding to the target sub-platform. Among them, the preset adjustment condition is that the update priority is lower than the priority threshold. The priority threshold is set based on experience.

[0134] In some embodiments, the management platform screens multiple associated platform groups including the target sub-platform, and uses the multiple associated platform groups with the intra-group association degree greater than the association threshold as multiple target associated platform groups. In response to the future acquisition parameters of other sensor network sub-platforms in the multiple target associated platform groups meeting the second preset condition, the management platform adjusts (such as reducing the preset adjustment amount, etc.) the future acquisition parameters of the target sub-platform based on the preset adjustment amount. The association threshold and the preset adjustment amount are set based on experience.

[0135] The second preset condition includes that the weighted sum of the future acquisition parameters of multiple other sensor network sub-platforms is greater than the preset acquisition threshold. The preset acquisition threshold and the weight of each sensor network sub-platform are set based on experience.

[0136] In some embodiments, the weight of the sensor network sub-platform is positively correlated with the intra-group association degree corresponding to the sensor network sub-platform in the associated platform group including the target sub-platform.

[0137] Adjust the future acquisition parameters of the sensor network sub-platform based on the update priority, which can preferentially adjust the future acquisition parameters of the sensor network sub-platform with a higher importance level, reduce the acquisition parameters of the sensor network sub-platform with a lower priority by a preset adjustment amount, and thus preferentially ensure the normal acquisition and transmission of data from the sensor network sub-platform with a higher importance level.

[0138] Figure 5 It is an exemplary flowchart for adjusting the group to be deleted shown in some embodiments of this specification. As Figure 5 shown, the process 500 for adjusting the group to be deleted includes the following steps.

[0139] Step 510, obtain multiple associated data groups of the data group.

[0140] An associated data group refers to a data group that is associated with the current data group. In some embodiments, the management platform clusters multiple data groups based on the retrieval time and retrieved data items in the historical data to obtain multiple clustering clusters, and determines the data groups in the same clustering cluster as the current data group as the associated data groups of the current data group. For more information about data groups, see Figure 2 and related descriptions.

[0141] Step 520, based on the multiple associated data groups and the grouping retrieval characteristics of the data group, determine the estimated retrieval characteristics of the data group.

[0142] For more information about the grouping retrieval characteristics, see Figure 2 and related descriptions.

[0143] The estimated retrieval characteristic refers to the grouping retrieval characteristic of the current data group within a preset future period.

[0144] In some embodiments, the management platform constructs a feature vector based on the multiple associated data groups and the grouping retrieval characteristics of the data group, performs vector matching, determines the mean of the labels of the matched feature vectors as the candidate grouping retrieval characteristic corresponding to the associated data group, and based on the candidate retrieval characteristics corresponding to the multiple associated data groups, determines the weighted sum of the multiple candidate retrieval characteristics as the estimated retrieval characteristic. The weights of different associated data groups are set based on experience. Among them, the method for determining the candidate grouping retrieval characteristic corresponding to the associated data group is similar to the method for determining the candidate retrieval characteristic corresponding to the associated sub-platform in step 210, and its implementation method can be seen in step 210 and its related description.

[0145] Step 530, adjust the group to be deleted based on the estimated retrieval characteristic.

[0146] In some embodiments, in response to the difference between the storage space of the sub-database of the sensing network after deleting the packet to be deleted and the pre-increased storage amount being higher than the storage space threshold after determining the packet to be deleted, the management platform determines the packet to be deleted with the largest weighted sum of each parameter in the estimated retrieval features among the multiple packets to be deleted as the data packet that does not need to be deleted. The management platform iteratively executes the above operation until the difference between the storage space of the sub-database of the sensing network after deleting the packet to be deleted and the pre-increased storage amount is not higher than the storage space threshold. The storage space threshold is set based on experience. For more information about the packet to be deleted and the pre-increased storage amount, see Figure 2 and related descriptions.

[0147] Based on the estimated retrieval features, the packet to be deleted is adjusted in real time. When there is still a certain amount of storage space in the storage space of the sub-database of the sensing network after deleting the packet to be deleted excluding the pre-increased storage amount, the data packets that are more easily retrievable are retained as much as possible to ensure the integrity of the data to the greatest extent.

[0148] It should be noted that the above description of the process 500 for adjusting the packet to be deleted is only for illustration and explanation, and does not limit the scope of application of this specification. For those skilled in the art, various modifications and changes can be made to the process 500 for adjusting the packet to be deleted under the guidance of this specification. However, these modifications and changes are still within the scope of this specification.

[0149] Some embodiments of this specification also provide a computer-readable storage medium, and the storage medium stores computer instructions. When the computer reads the computer instructions in the storage medium, the computer executes the data management method based on the industrial Internet of Things data center described in any one of the above embodiments.

[0150] In addition, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0151] In some embodiments, the numerical parameters used in the specification and claims are all approximate values, and these approximate values can be changed according to the characteristics required by individual embodiments.

[0152] If there are inconsistencies or conflicts between the descriptions, definitions, and / or uses of terms in the materials cited in this specification and the content described in this specification, the descriptions, definitions, and / or uses of terms in this specification shall prevail.

Claims

1. A data management system based on an industrial Internet of Things data center, characterized in that: The system includes an industrial Internet of Things user platform, an industrial Internet of Things service platform, an industrial Internet of Things management platform, an industrial Internet of Things sensor network platform, and an industrial Internet of Things perception control platform; The industrial Internet of Things management platform includes a data center, the industrial Internet of Things sensor network platform includes a plurality of sensor network sub-platforms and a plurality of sensor network sub-databases, and the plurality of sensor network sub-databases correspond to the plurality of sensor network sub-platforms; The industrial Internet of Things management platform is configured to: At every first preset period, for a single sensor network sub-platform among the plurality of sensor network sub-platforms: Determine the future acquisition parameters of the sensor network sub-platform based on the remaining storage space of the sensor network sub-database corresponding to the sensor network sub-platform, the first retrieval feature and the future retrieval feature; wherein the first retrieval feature refers to the retrieval feature of the sensor network sub-database within a preset historical period, the future retrieval feature refers to the retrieval feature of the sensor network sub-database within a preset future period, and the retrieval feature refers to data used to characterize the retrieval situation of the sensor network sub-database; Based on the future acquisition parameters, determining the pre-increment storage capacity of the sensor network sub-database; Dividing the data items in the sensor network sub-database to obtain a plurality of data groups; Determining a group to be deleted based on the pre-increment storage amount, the remaining storage space, and group retrieval features and group data amounts corresponding to the plurality of data groups; Based on the future acquisition parameters, a parameter update instruction is generated and sent to the sensor network sub-platform to control the sensor network sub-platform to perform data acquisition based on the future acquisition parameters; Based on the group to be deleted, a data deletion instruction is generated and sent to the sensor network sub-platform to control the sensor network sub-platform to delete the group to be deleted.

2. The system according to claim 1, characterized in that The industrial Internet of Things management platform is further configured to: Acquire multiple associated sub-platforms of the sensor network sub-platform; The future retrieval feature is determined based on the first stored data corresponding to the multiple associated sub-platforms.

3. The system according to claim 2, characterized in that The industrial Internet of Things management platform is further configured to: Determining a plurality of associated platform groups and intra-group association degrees corresponding to the plurality of associated platform groups based on the second retrieved features corresponding to the plurality of sensor network sub-platforms; Based on the multiple associated platform groups, determining the multiple associated sub-platforms; The future retrieval feature is determined based on the intra-group association levels corresponding to the multiple associated sub-platforms and the first stored data.

4. The system according to claim 3, characterized in that The industrial Internet of Things management platform is further configured to: constructing a data retrieval graph based on the multiple associated platform groups, the intra-group association degrees corresponding to the multiple associated sub-platforms, and the second stored data corresponding to the multiple sensor network sub-platforms, wherein the second stored data includes the first stored data; Based on the data retrieval map, the future retrieval features are determined by a prediction model, and the prediction model is a machine learning model.

5. The system according to claim 1, wherein: The industrial Internet of Things management platform is further configured to: For a single data packet among the plurality of data packets: Acquire multiple associated data groups of the data group; Determining an estimated call feature of the data group based on the plurality of associated data groups and the group call features of the data group; Based on the estimated retrieved features, the to-be-deleted group is adjusted.

6. A data management method based on an industrial Internet of Things data center, characterized in that: The method is performed by an industrial Internet of Things management platform in a data management system based on an industrial Internet of Things data center, and the method includes: Every first preset period, for a single sensor network sub-platform among the multiple sensor network sub-platforms: Determine the future acquisition parameters of the sensor network sub-platform based on the remaining storage space, the first retrieval feature and the future retrieval feature of the sensor network sub-database corresponding to the sensor network sub-platform; wherein the first retrieval feature refers to the retrieval feature of the sensor network sub-database within a preset historical period, the future retrieval feature refers to the retrieval feature of the sensor network sub-database within a preset future period, and the retrieval feature refers to data used to characterize the retrieval situation of the sensor network sub-database; Based on the future acquisition parameters, determining the pre-increment storage capacity of the sensor network sub-database; Dividing the data items in the sensor network sub-database to obtain a plurality of data groups; Determining a group to be deleted based on the pre-increment storage amount, the remaining storage space, and group retrieval features and group data amounts corresponding to the plurality of data groups; Based on the future acquisition parameters, a parameter update instruction is generated and sent to the sensor network sub-platform to control the sensor network sub-platform to perform data acquisition based on the future acquisition parameters; Based on the group to be deleted, a data deletion instruction is generated and sent to the sensor network sub-platform to control the sensor network sub-platform to delete the group to be deleted.

7. The method according to claim 6, characterized in that The method further comprises: Acquire multiple associated sub-platforms of the sensor network sub-platform; The future retrieval feature is determined based on the first stored data corresponding to the multiple associated sub-platforms.

8. The method according to claim 7, characterized in that The determining the future retrieval feature based on the first stored data corresponding to the multiple associated sub-platforms includes: Determining a plurality of associated platform groups and intra-group association degrees corresponding to the plurality of associated platform groups based on the second retrieved features corresponding to the plurality of sensor network sub-platforms; Based on the multiple associated platform groups, determining the multiple associated sub-platforms; The future retrieval feature is determined based on the intra-group association levels corresponding to the multiple associated sub-platforms and the first stored data.

9. The method according to claim 8, characterized in that The method further comprises: constructing a data retrieval graph based on the multiple associated platform groups, the intra-group association degrees corresponding to the multiple associated sub-platforms, and the second stored data corresponding to the multiple sensor network sub-platforms, wherein the second stored data includes the first stored data; Based on the data retrieval map, the future retrieval features are determined by a prediction model, and the prediction model is a machine learning model.

10. The method according to claim 6, characterized in that The method further comprises: For a single data packet among the plurality of data packets: Acquire multiple associated data groups of the data group; Determining an estimated call feature of the data group based on the plurality of associated data groups and the group call features of the data group; Based on the estimated retrieved features, the to-be-deleted group is adjusted.

Images