A cloud-edge fusion processing method and system for Internet of Things data

Abstract

The present application discloses a cloud-edge fusion processing method and system for IoT data, which belongs to the field of computer technology. Through the technical solution provided by the embodiment of the present application, the computing power of edge devices can be efficiently utilized, and the target computing power can be determined based on IoT data, data processing needs and data processing requirements. When the target computing power is less than or equal to the preset computing power of the edge device, it is processed on the edge device, avoiding the edge device from blindly trying to process data when its own computing power is insufficient, improving the processing success rate and making full use of its computing power resources without causing waste; when the target computing power is greater than the preset computing power, the data processing request and request are forwarded to the cloud for processing by the cloud, effectively utilizing the powerful computing power resources of the cloud.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a cloud-edge fusion processing method and system for Internet of Things data. Background Art

[0002] With the rapid development of the Internet of Things, a large number of IoT devices generate massive amounts of data. These IoT data need to be processed to meet the needs of different terminals. At present, when processing IoT data, we face the challenge of computing power allocation.

[0003] Traditional data processing methods often fail to fully consider the computing power characteristics of edge devices and the cloud, as well as the various requirements of terminals. For example, when processing IoT data, different data processing requirements, data processing standards, and data processing styles are often not distinguished. In existing technical solutions, there is no reasonable arrangement based on the computing power required for data processing and the computing power that the device can provide. Edge devices may still try to process data when their own computing power is insufficient, resulting in inefficient or even failed processing; or they fail to reasonably utilize the powerful computing power resources of the cloud, and cannot perform effective cloud-edge collaborative processing based on the complex requirements sent by the target terminal (such as specific data processing standards and styles).

[0004] Therefore, how to use cloud-edge fusion to process IoT data has become a hot topic of research. Summary of the invention

[0005] The embodiment of the present application provides a cloud-edge fusion processing method and system for IoT data, which can process IoT data by using cloud-edge fusion to improve the processing effect of IoT data. The technical solution is as follows:

[0006] On the one hand, a cloud-edge fusion processing method for IoT data is provided, which is applied to an edge device, and the method includes:

[0007] In response to a data processing request sent by a target terminal, obtaining the IoT data to be processed, the data processing requirements of the IoT data, and the data processing requirements of the target terminal carried in the data processing request, wherein the data processing requirements include a data processing standard and a data processing style;

[0008] Determine a target computing power required to process the IoT data based on the IoT data, the data processing demand, and the data processing requirement;

[0009] When the target computing power is less than or equal to the preset computing power, the IoT data is processed based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data, and the preset computing power is the maximum computing power that can be allocated by the edge device to the data processing request within a preset time period;

[0010] When the target computing power is greater than the preset computing power, the data processing request and the data processing requirement are forwarded to the cloud, so that the cloud processes the IoT data based on the data processing demand and the data processing requirement to obtain the data processing result of the IoT data; and the data processing result of the IoT data is obtained from the cloud.

[0011] In one aspect, a cloud-edge fusion processing system for IoT data is provided, the system comprising: an edge device and a cloud;

[0012] The edge device is used to respond to a data processing request sent by a target terminal, obtain the IoT data to be processed carried by the data processing request, the data processing requirements of the IoT data, and the data processing requirements of the target terminal, wherein the data processing requirements include a data processing standard and a data processing style;

[0013] The edge device is used to determine the target computing power required to process the IoT data based on the IoT data, the data processing demand and the data processing requirement;

[0014] The edge device is used to process the IoT data based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data when the target computing power is less than or equal to the preset computing power, and the preset computing power is the maximum computing power that can be allocated by the edge device to the data processing request within a preset time period;

[0015] The edge device is used to forward the data processing request and the data processing requirement to the cloud when the target computing power is greater than the preset computing power, so that the cloud processes the Internet of Things data based on the data processing demand and the data processing requirement to obtain the data processing result of the Internet of Things data; and obtain the data processing result of the Internet of Things data from the cloud.

[0016] On the one hand, an edge device is provided, which includes one or more processors and one or more memories, wherein at least one computer program is stored in the one or more memories, and the computer program is loaded and executed by the one or more processors to implement the cloud-edge fusion processing method for IoT data.

[0017] On the one hand, a computer-readable storage medium is provided, in which at least one computer program is stored. The computer program is loaded and executed by a processor to implement the cloud-edge fusion processing method for IoT data.

[0018] On the one hand, a computer program product or computer program is provided, which includes a program code, and the program code is stored in a computer-readable storage medium. The processor of the edge device reads the program code from the computer-readable storage medium, and the processor executes the program code, so that the edge device executes the above-mentioned cloud-edge fusion processing method for IoT data.

[0019] Through the technical solution provided by the embodiment of the present application, in terms of reasonable allocation of computing power, first of all, it is possible to efficiently utilize the computing power of edge devices, determine the target computing power based on IoT data, data processing needs and data processing requirements, and process it on the edge device when the target computing power is less than or equal to the preset computing power of the edge device, avoiding the edge device from blindly trying to process data when its own computing power is insufficient, improving the processing success rate and making full use of its computing power resources without causing waste; when the target computing power is greater than the preset computing power, it forwards data processing requests and requirements to the cloud for processing by the cloud, effectively utilizing the powerful computing power resources of the cloud. In terms of meeting terminal needs, this technical solution can meet a variety of requirements, and simultaneously obtain the data processing requirements (including standards and styles) of the target terminal when acquiring IoT data. These requirements are used when processing on edge devices or in the cloud, overcoming the problem that traditional technologies do not distinguish between different data processing standards and styles to meet the complex needs of different terminals; it can also flexibly select the processing location according to the computing power through cloud-edge fusion to achieve cloud-edge collaborative processing, and improve the overall data processing efficiency on the basis of meeting the data processing requirements of the target terminal. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0021] Figure 1 It is a schematic diagram of an implementation environment of a cloud-edge fusion processing method for IoT data provided in an embodiment of the present application;

[0022] Figure 2 It is a flow chart of a cloud-edge fusion processing method for IoT data provided by an embodiment of the present application;

[0023] Figure 3It is a flow chart of another cloud-edge fusion processing method for IoT data provided by an embodiment of the present application;

[0024] Figure 4 It is a structural diagram of a cloud-edge fusion processing system for IoT data provided by an embodiment of the present application;

[0025] Figure 5 It is a structural diagram of an edge device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0026] In order to make the objectives, technical solutions and advantages of the present application clearer, the implementation methods of the present application will be further described in detail below with reference to the accompanying drawings.

[0027] In this application, the terms "first", "second", etc. are used to distinguish identical or similar items with basically the same effects and functions. It should be understood that there is no logical or temporal dependency between "first", "second", and "nth", nor is there any limitation on quantity and execution order.

[0028] Cloud-edge fusion: refers to the combination of cloud computing and edge computing, and the establishment of a collaborative working mechanism between cloud and edge devices to achieve efficient data processing and analysis. This fusion model can give full play to the powerful computing power of cloud computing and the low latency and high bandwidth advantages of edge computing, providing enterprises with more flexible and efficient digital solutions.

[0029] Edge devices: devices that process and store data at the edge of the network, usually near the data source, such as sensors, smartphones, IoT devices, etc. Compared with traditional cloud computing models, edge devices can process data in real time at the location where it is generated, thereby reducing latency and improving response speed.

[0030] Internet of Things: refers to connecting billions of physical devices around the world to the Internet through various information sensors, radio frequency identification technology, global positioning system and other devices and technologies to achieve real-time data collection and sharing. The core of the Internet of Things is to connect objects in the real world to the Internet so that they can communicate and interact with each other.

[0031] IoT data: refers to data collected through IoT devices (such as sensors, actuators, etc.). These data may include device status, environmental parameters (such as temperature, humidity, light, etc.), location information, etc. IoT data are characterized by diversity, real-time and massive data types, including not only traditional structured data, but also unstructured data (such as images, videos, audio, etc.).

[0032] It should be noted that the information (including but not limited to user device information, user personal information, etc.), data (including but not limited to data used for analysis, stored data, displayed data, etc.) and signals involved in this application are all authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant laws, regulations and standards of relevant countries and regions. In particular, the IoT data involved in the embodiments of this application is fully processed only after being fully authorized by the user.

[0033] In the related technology, there is a technical solution that uses cloud-edge fusion to process data. Generally speaking, manufacturers will configure which data is processed by edge devices and which data is processed by the cloud, so as to achieve cooperation between edge devices and the cloud and improve the efficiency and effectiveness of data processing.

[0034] However, since the computing power of edge devices is much smaller than that of the cloud, relying on configuration to decide whether data is processed by edge devices or the cloud may result in the edge devices being unable to process the assigned data. For example, in the IoT scenario, the amount and data types of IoT data generated are diverse, and the needs and requirements for processing IoT data are also different. Therefore, some IoT data can be processed smoothly by edge devices, while other IoT data cannot be processed smoothly by edge devices due to computing power limitations, resulting in errors in IoT data processing or slow processing speed, and poor user experience.

[0035] The technical solution provided by the embodiment of the present application can improve the above problems, increase the probability of smooth processing of IoT data, and thus improve the user experience. In other words, the technical solution provided by the embodiment of the present application requires edge devices to have strong computing power, which is suitable for scenarios where the amount of IoT data is huge and the processing is difficult. The focus of the solution is to improve the processing effect of IoT data rather than efficiency.

[0036] Figure 1 This is a schematic diagram of an implementation environment of a cloud-edge fusion processing method for IoT data provided by an embodiment of the present application, see Figure 1 , the implementation environment may include an edge device 110 and a cloud 140 .

[0037] The edge device 110 is connected to the cloud 140 via a wireless network or a wired network. Optionally, the edge device 110 is a smart phone, a tablet computer, a laptop computer, a desktop computer, etc., but is not limited thereto. The edge device 110 is installed and runs an application program that supports IoT data processing.

[0038] The cloud 140 is an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, distribution networks (Content Delivery Network, CDN), and big data and artificial intelligence platforms. The cloud 140 can provide background services for applications running on the edge device 110.

[0039] The following is a description of the cloud-edge fusion processing method for IoT data provided in an embodiment of the present application. Figure 2 This is a flow chart of a cloud-edge fusion processing method for IoT data provided by an embodiment of the present application, see Figure 2 , taking the execution of edge device as an example, the method includes the following steps.

[0040] 201. In response to a data processing request sent by a target terminal, the edge device obtains the IoT data to be processed carried in the data processing request, the data processing requirements of the IoT data, and the data processing requirements of the target terminal, wherein the data processing requirements include a data processing standard and a data processing style.

[0041] Among them, the target terminal is a terminal with IoT data processing needs, and the data processing request is used to request processing of the IoT data to be processed. In an embodiment of the present application, the target terminal is an IoT device, and the edge device is a management device connected to multiple IoT devices. The IoT device and the edge device are usually located in the same space. For example, the IoT device includes air conditioners, cameras, washing machines, televisions, etc., and the edge device is a management device connected to air conditioners, cameras, washing machines, televisions, etc. IoT data is data related to IoT devices. IoT data includes multiple types of data. For example, IoT data includes device information, device attributes, and collected data of IoT devices. Data processing needs refer to the specific needs and expectations of users or organizations for data processing in specific business scenarios. Data processing requirements refer to the rules, standards, or constraints that must be followed when processing data. Accordingly, data processing requirements include data processing standards and data processing styles.

[0042] 202. The edge device determines the target computing power required to process the IoT data based on the IoT data, the data processing demand, and the data processing requirement.

[0043] Among them, the target computing power refers to the computing power required to process the IoT data according to the data processing needs and the data processing requirements.

[0044] 203. When the target computing power is less than or equal to the preset computing power, the edge device processes the IoT data based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data. The preset computing power is the maximum computing power that can be allocated to the data processing request by the edge device within a preset time period.

[0045] Among them, if the target computing power is less than or equal to the preset computing power, it means that the edge device has the ability to process the IoT data according to the data processing needs and the data processing requirements. There are two meanings here. First, the computing power of the edge device must be greater than the target computing power. Second, the computing power that the edge device can allocate to the task of processing the IoT data is sufficient. Therefore, the IoT data can be directly processed by the edge device.

[0046] 204. When the target computing power is greater than the preset computing power, the edge device forwards the data processing request and the data processing requirement to the cloud, so that the cloud processes the IoT data based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data. The data processing result of the IoT data is obtained from the cloud.

[0047] Among them, if the target computing power is greater than the preset computing power, it means that the edge device does not have the ability to process the IoT data according to the data processing needs and the data processing requirements. Therefore, it is handed over to the cloud with sufficient computing power for processing, thereby improving the processing effect of the IoT data.

[0048] Through the technical solution provided by the embodiment of the present application, in terms of reasonable allocation of computing power, first of all, it is possible to efficiently utilize the computing power of edge devices, determine the target computing power based on IoT data, data processing needs and data processing requirements, and process it on the edge device when the target computing power is less than or equal to the preset computing power of the edge device, avoiding the edge device from blindly trying to process data when its own computing power is insufficient, improving the processing success rate and making full use of its computing power resources without causing waste; when the target computing power is greater than the preset computing power, it forwards data processing requests and requirements to the cloud for processing by the cloud, effectively utilizing the powerful computing power resources of the cloud. In terms of meeting terminal needs, this technical solution can meet a variety of requirements, and simultaneously obtain the data processing requirements (including standards and styles) of the target terminal when acquiring IoT data. These requirements are used when processing on edge devices or in the cloud, overcoming the problem that traditional technologies do not distinguish between different data processing standards and styles to meet the complex needs of different terminals; it can also flexibly select the processing location according to the computing power through cloud-edge fusion to achieve cloud-edge collaborative processing, and improve the overall data processing efficiency on the basis of meeting the data processing requirements of the target terminal.

[0049] The above steps 201-204 are a brief introduction to the cloud-edge fusion processing method for IoT data provided by the embodiment of the present application. The following will combine some examples to more clearly explain the cloud-edge fusion processing method for IoT data provided by the embodiment of the present application. Figure 3 , taking the execution subject as an edge device as an example, the method includes the following steps.

[0050] 301. In response to a data processing request sent by a target terminal, the edge device obtains the IoT data to be processed carried in the data processing request, the data processing requirements of the IoT data, and the data processing requirements of the target terminal, wherein the data processing requirements include a data processing standard and a data processing style.

[0051] Among them, the target terminal is a terminal with IoT data processing needs, and the data processing request is used to request processing of the IoT data to be processed. In an embodiment of the present application, the target terminal is an IoT device, and the edge device is a management device connected to multiple IoT devices. The IoT device and the edge device are usually located in the same space. For example, the IoT device includes air conditioners, cameras, washing machines, televisions, etc., and the edge device is a management device connected to air conditioners, cameras, washing machines, televisions, etc. IoT data is data related to IoT devices. IoT data includes multiple types of data. For example, IoT data includes device information, device attributes, and collected data of IoT devices. Data processing needs refer to the specific needs and expectations of users or organizations for data processing in specific business scenarios. Data processing requirements refer to the rules, standards, or constraints that must be followed when processing data. Accordingly, data processing requirements include data processing standards and data processing styles.

[0052] In a possible implementation, in response to a data processing request sent by a target terminal, the edge device parses the data processing request to obtain the IoT data to be processed, the data processing requirements of the IoT data, and the data processing requirements of the target terminal.

[0053] Among them, the method of parsing the data processing request is set by the technical personnel according to the actual situation, and the embodiment of the present application does not limit this.

[0054] 302. The edge device determines the basic computing power required to process the IoT data based on the IoT data and the data processing requirements.

[0055] In one possible implementation, the edge device determines the data type and data volume of the IoT data. Based on the data type and the data volume, the edge device determines a first initial computing power for processing the IoT data. The edge device determines the requirement description information of the data processing requirement. Based on the requirement description information of the data processing requirement, the edge device determines a second initial computing power that meets the data processing requirement. The edge device determines a data summary corresponding to the IoT data. Based on the data summary corresponding to the IoT data and the data processing requirement, the edge device determines a third initial computing power for processing the IoT data. The edge device merges the first initial computing power, the second initial computing power, and the third initial computing power to obtain the basic computing power required to process the IoT data.

[0056] The first initial computing power, the second initial computing power and the third initial computing power are initial computing powers of different aspects, and the basic computing power can be determined by using the first initial computing power, the second initial computing power and the third initial computing power. The demand description information is a natural language description of the data processing requirements. The data summary corresponding to the IoT data refers to the simplified form of the IoT data, and the data volume of the data summary is much smaller than the data volume of the IoT data.

[0057] In order to explain the above implementation more clearly, the above implementation will be explained in several parts below.

[0058] In the first part, the edge device determines the data type and amount of the IoT data.

[0059] In a possible implementation, the edge device performs data analysis on the IoT data to obtain the data type and data volume of the IoT data.

[0060] In the second part, the edge device determines a first initial computing power for processing the IoT data based on the data type and the data volume.

[0061] In a possible implementation, the edge device generates a first computing power acquisition request based on the data type and the data volume and determines a first reference computing power for processing the IoT data. The first computing power acquisition request is used to request the corresponding second reference computing power, and the second reference computing power is used to assist in determining the first initial computing power. The edge device sends the first computing power acquisition request to the cloud, so that the cloud returns multiple second reference computing powers and the credibility of each second reference computing power based on the first computing power acquisition request. The multiple second reference computing powers include multiple third reference computing powers and a fourth reference computing power. The third reference computing power is sent to the cloud by other edge devices, and the fourth reference computing power is determined by the cloud. The edge device determines the first initial computing power for processing the IoT data based on the first reference computing power, the multiple second reference computing powers, and the credibility of each second reference computing power.

[0062] Among them, the first reference computing power request is sent to the cloud, and is used to request the cloud to determine multiple second reference computing powers corresponding to the data type and data volume and the credibility of each second reference computing power. Other edge devices refer to edge devices connected to the cloud, and the number of other edge devices is multiple, and other edge devices do not include this edge device. The cloud forwards the first reference computing power request to other edge devices, and the other edge devices determine the third reference computing power based on the first reference computing power request. In some implementations, the credibility of each second reference computing power is determined by the cloud. For the credibility of the third reference computing power, the cloud determines it based on the historical information of the corresponding other edge devices. The historical information includes the difference between the reference computing power determined by the other edge devices and the actual computing power. The actual computing power is obtained after the cloud processes the IoT data, or the edge device uploads the IoT data to the cloud after processing.

[0063] In order to explain the above implementation more clearly, the following describes a method in which the edge device in the above implementation determines the first reference computing power for processing the IoT data based on the data type and the data volume.

[0064] In some embodiments, the edge device inputs the data type and the data amount into a first computing power determination model, extracts features of the data type and the data amount through the first computing power determination model, and obtains a first computing power determination feature. The edge device fully connects and normalizes the first computing power determination feature through the first computing power determination model to obtain the first reference computing power.

[0065] Among them, the first computing power determination model is a regression model, which can map the data type and data amount to the first reference computing power. The first computing power determination model can adopt any type of regression model, and the embodiment of the present application is not limited to this.

[0066] In order to explain the above implementation more clearly, the following describes a method in which the edge device in the above implementation determines the first initial computing power for processing the IoT data based on the first reference computing power, the multiple second reference computing powers, and the credibility of each second reference computing power.

[0067] In some embodiments, the edge device determines a first weight of the first reference computing power. The edge device determines a second weight of each second reference computing power based on the credibility of each second reference computing power. The edge device performs weighted fusion of the first reference computing power and multiple second reference computing powers based on the first weight and multiple second weights to obtain the first initial computing power.

[0068] Among them, the second weight is positively correlated with the credibility of the corresponding second reference computing power, that is, the higher the credibility of the second reference computing power, the higher the corresponding second weight; the lower the credibility of the second reference computing power, the lower the corresponding second weight. In some embodiments, substituting the credibility into the first relationship data can obtain the corresponding second weight. The first relationship data is in the form of a function. The first relationship data is set by the technician according to the actual situation, and the embodiment of the present application does not limit this. The first weight of the first reference computing power is generated by the edge device, and the first weight can reflect the credibility of the first reference computing power.

[0069] The third part is the requirement description information that the edge device determines the data processing requirement.

[0070] In a possible implementation, the edge device extracts features of the data processing requirement to obtain a processing requirement feature of the data processing requirement. Based on the processing requirement feature and multiple first prompt texts, the edge device determines multiple sub-processing requirements corresponding to the data processing requirement, where different sub-processing requirements correspond to different aspects, and different first prompt texts are used to indicate the generation of sub-processing requirements in different aspects. The edge device determines sub-requirement description information of each sub-processing requirement. Based on the sub-requirement description information of each sub-processing requirement, the edge device generates requirement description information of the data processing requirement.

[0071] Among them, the first prompt text is in the form of a latent vector, and the first prompt text is set by the technician according to the actual situation, and the embodiment of the present application does not limit this.

[0072] For example, the edge device performs multiple full connections on the data processing requirement to obtain the processing requirement characteristics of the data processing requirement. The edge device combines the processing requirement characteristics with each first prompt text to obtain multiple first combination information, and one first combination information includes the processing requirement characteristics and a first prompt text. The edge device inputs the multiple first combination information into the first information generation model, decodes the multiple first combination information through the first information generation model, and obtains the multiple sub-processing requirements, and one sub-processing requirement corresponds to one first combination information. The edge device inputs the multiple sub-processing requirements into the first language model, encodes and decodes each sub-processing requirement through the first language model, and obtains the sub-requirement description information of each sub-processing requirement. The edge device merges the sub-requirement description information of each sub-processing requirement to obtain the requirement description information of the data processing requirement.

[0073] Among them, in the embodiment of the present application, the difference between the data processing requirements and the requirement description information is that the data processing requirements are in the form of vectors and codes, and the requirement description information is in the form of natural language. The first information generation model is an encoding and decoding model based on the attention mechanism, and the first language model is a large language model. The embodiment of the present application does not limit the structure of the first information generation model and the first language model.

[0074] Part 4: The edge device determines the second initial computing power that meets the data processing requirements based on the requirement description information of the data processing requirements.

[0075] In a possible implementation, the edge device extracts features from the demand description information to obtain demand description features of the demand description information. Based on the demand description features, the edge device generates a second computing power acquisition request and determines a fifth reference computing power that meets the data processing requirements. The second computing power acquisition request is used to request the acquisition of the corresponding sixth reference computing power, and the sixth reference computing power is used to assist in determining the second initial computing power. The edge device sends the second computing power acquisition request to the cloud, so that the cloud returns multiple sixth reference computing powers and the credibility of each sixth reference computing power based on the second computing power acquisition request. The multiple sixth reference computing powers include a seventh reference computing power and an eighth reference computing power. The seventh reference computing power is sent to the cloud by the target edge device, and the eighth reference computing power is determined by the cloud. The target edge device is another edge device whose credibility is greater than or equal to the credibility threshold. The edge device determines the second initial computing power for processing the IoT data based on the fourth reference computing power, the sixth reference computing power, and the credibility of the sixth reference computing power.

[0076] The second reference computing power request is sent to the cloud to request the cloud to determine multiple sixth reference computing powers corresponding to the demand description characteristics and the credibility of each sixth reference computing power. The cloud forwards the second reference computing power request to the target edge device, and the target edge device determines the seventh reference computing power based on the second reference computing power request.

[0077] In order to explain the above implementation more clearly, the following describes a method in which the edge device in the above implementation performs feature extraction on the demand description information to obtain the demand description features of the demand description information.

[0078] In some embodiments, the edge device encodes the requirement description information based on an attention mechanism to obtain requirement description features of the requirement description information.

[0079] In order to explain the above implementation more clearly, the following describes a method in which the edge device in the above implementation determines the fourth reference computing power that meets the data processing requirement based on the requirement description characteristics.

[0080] In some embodiments, the edge device inputs the demand description feature into the second computing power determination model, and fully connects the demand description feature through the second computing power determination model to obtain the second computing power determination feature. The edge device fully connects and normalizes the second computing power determination feature through the second computing power determination model to obtain the fifth reference computing power.

[0081] Among them, the second computing power determination model is a regression model, which can map the demand description characteristics into the fifth reference computing power. The second computing power determination model can adopt any type of regression model, and the embodiment of the present application is not limited to this.

[0082] In order to explain the above implementation more clearly, the following describes a method in which the edge device in the above implementation determines the second initial computing power for processing the IoT data based on the fifth reference computing power, the sixth reference computing power, and the credibility of the sixth reference computing power.

[0083] In some embodiments, the edge device determines a third weight of the fifth reference computing power. The edge device determines a fifth weight of each sixth reference computing power based on the credibility of each sixth reference computing power. The edge device performs weighted fusion of the fifth reference computing power and multiple sixth reference computing powers based on the third weight and multiple fifth weights to obtain the second initial computing power.

[0084] Part 5: The edge device determines the data summary corresponding to the IoT data.

[0085] Among them, the data summary can be regarded as a simplified form of the IoT data, that is, the data volume of the data summary is smaller than that of the IoT data. Since the technical solution provided in the embodiment of the present application is applied in scenarios where the IoT data volume is large and the processing difficulty is high, converting the IoT data into a data summary can improve the efficiency of data processing.

[0086] In a possible implementation, the edge device extracts a summary of the IoT data to obtain a data summary corresponding to the IoT data.

[0087] For example, the edge device inputs the IoT data into the summary extraction model, and extracts features of the IoT data through the summary extraction model to obtain data features of the IoT data. The edge device performs multiple rounds of iterative decoding of the data features of the IoT data through the summary extraction model to obtain the data summary corresponding to the IoT data.

[0088] Among them, the summary extraction model is a regression model, which can reduce the mapping of IoT data with large data volume into data summaries with small data volume.

[0089] Part 6: The edge device determines the third initial computing power for processing the IoT data based on the data summary corresponding to the IoT data and the data processing requirements.

[0090] In a possible implementation, the edge device generates a third computing power acquisition request and determines a ninth reference computing power for processing the IoT data based on the data summary corresponding to the IoT data and the data processing requirements. The third computing power acquisition request is used to obtain the corresponding ninth reference computing power after the request, and the ninth reference computing power is used to assist in determining the third initial computing power. The edge device sends the third computing power acquisition request to the cloud, so that the cloud returns multiple tenth reference computing powers and the credibility of each tenth reference computing power based on the third computing power acquisition request. The multiple tenth reference computing powers include multiple eleventh reference computing powers and a twelfth reference computing power. The eleventh reference computing power is sent to the cloud by other edge devices, and the twelfth reference computing power is determined by the cloud. The edge device determines the third initial computing power for processing the IoT data based on the first reference computing power, the multiple tenth reference computing powers, and the credibility of each tenth reference computing power.

[0091] In order to explain the above implementation more clearly, the following describes a method in which the edge device in the above implementation determines the ninth reference computing power for processing the IoT data based on the data summary corresponding to the IoT data and the data processing requirements.

[0092] In some embodiments, the edge device inputs the data summary and the data processing requirement into a third computing power determination model, extracts features from the data summary and the data processing requirement through the third computing power determination model, and obtains a third computing power determination feature. The edge device fully connects and normalizes the third computing power determination feature through the third computing power determination model to obtain the ninth reference computing power.

[0093] Among them, the third computing power determination model is a regression model, which can map the data summary and data processing requirements into the ninth reference computing power. The third computing power determination model can adopt any type of regression model, and the embodiment of the present application is not limited to this.

[0094] In order to explain the above implementation more clearly, the following describes a method in which the edge device in the above implementation determines the third initial computing power for processing the IoT data based on the first reference computing power, the multiple tenth reference computing powers, and the credibility of each tenth reference computing power.

[0095] In some embodiments, the edge device determines a first weight of the first reference computing power. The edge device determines a fifth weight of each tenth reference computing power based on the credibility of each tenth reference computing power. The edge device performs weighted fusion of the first reference computing power and multiple tenth reference computing powers based on the first weight and multiple fifth weights to obtain the third initial computing power.

[0096] Part 7: The edge device integrates the first initial computing power, the second initial computing power and the third initial computing power to obtain the basic computing power required to process the IoT data.

[0097] In a possible implementation, the edge device performs weighted fusion on the first initial computing power, the second initial computing power, and the third initial computing power to obtain the basic computing power required to process the IoT data.

[0098] Among them, the weights of weighted fusion are set by technical personnel according to actual conditions, and the embodiments of the present application do not limit this.

[0099] 303. The edge device determines a first computing power correction coefficient for processing the IoT data based on the data processing requirement and the data processing standard.

[0100] In one possible implementation, the edge device determines the requirement description information of the data processing requirement. The edge device determines a first initial correction coefficient corresponding to the data processing requirement based on the requirement description information of the data processing requirement. The edge device determines the standard description information of the data processing standard. The edge device determines a second initial correction coefficient corresponding to the data processing standard based on the standard description information of the data processing standard. The edge device merges the first initial correction coefficient and the second initial correction coefficient to obtain the first computing power correction coefficient.

[0101] In order to explain the above implementation more clearly, the above implementation will be explained in several parts below.

[0102] Part 1: The edge device determines the requirement description information for the data processing requirement.

[0103] In a possible implementation, the edge device extracts features of the data processing requirement to obtain a processing requirement feature of the data processing requirement. Based on the processing requirement feature and multiple second prompt texts, the edge device determines multiple sub-processing requirements corresponding to the data processing requirement, where different sub-processing requirements correspond to different aspects, and different second prompt texts are used to indicate the generation of sub-processing requirements of different aspects. The edge device determines sub-requirement description information of each sub-processing requirement. Based on the sub-requirement description information of each sub-processing requirement, the edge device generates requirement description information of the data processing requirement.

[0104] Among them, the second prompt text is in the form of a latent vector, and the second prompt text is set by the technician according to the actual situation, and the embodiment of the present application does not limit this.

[0105] For example, the edge device performs multiple full connections on the data processing requirement to obtain the processing requirement feature of the data processing requirement. The edge device combines the processing requirement feature with each second prompt text to obtain multiple second combination information, and one second combination information includes the processing requirement feature and a second prompt text. The edge device inputs the multiple second combination information into the second information generation model, and decodes the multiple second combination information through the second information generation model to obtain the multiple sub-processing requirements, and one sub-processing requirement corresponds to one second combination information. The edge device inputs the multiple sub-processing requirements into the second language model, encodes and decodes each sub-processing requirement through the second language model, and obtains the sub-requirement description information of each sub-processing requirement. The edge device merges the sub-requirement description information of each sub-processing requirement to obtain the requirement description information of the data processing requirement.

[0106] Among them, in the embodiment of the present application, the difference between the data processing requirements and the requirement description information is that the data processing requirements are in the form of vectors and codes, and the requirement description information is in the form of natural language. The second information generation model is an encoding and decoding model based on the attention mechanism, and the second language model is a large language model. The embodiment of the present application does not limit the structure of the second information generation model and the second language model.

[0107] In the second part, the edge device determines a first initial correction coefficient corresponding to the data processing requirement based on the requirement description information of the data processing requirement.

[0108] In a possible implementation, the edge device extracts features from the requirement description information to obtain a requirement description feature of the requirement description information. The edge device inputs the requirement description feature into a first coefficient determination model, and performs full connection on the requirement description feature through the first coefficient determination model to obtain a first coefficient determination feature. The edge device performs full connection and normalization on the first coefficient determination feature through the first coefficient determination model to obtain the first initial correction coefficient.

[0109] Part 3: The edge device determines the standard description information of the data processing standard.

[0110] In a possible implementation, the edge device extracts features of the data processing standard to obtain processing standard features of the data processing standard. Based on the processing standard features and multiple third prompt texts, the edge device determines multiple sub-processing standards corresponding to the data processing standard, where different sub-processing standards correspond to different aspects, and different third prompt texts are used to indicate the generation of sub-processing standards of different aspects. The edge device determines sub-standard description information of each sub-processing standard. Based on the sub-standard description information of each sub-processing standard, the edge device generates standard description information of the data processing standard.

[0111] Among them, the third prompt text is in the form of a latent vector. The third prompt text is set by the technician according to the actual situation, and the embodiment of the present application does not limit this.

[0112] For example, the edge device performs multiple full connections on the data processing standard to obtain the processing standard feature of the data processing standard. The edge device combines the processing standard feature with each third prompt text to obtain multiple third combination information, and one third combination information includes the processing standard feature and a third prompt text. The edge device inputs the multiple third combination information into the third information generation model, and decodes the multiple third combination information through the third information generation model to obtain the multiple sub-processing standards, and one sub-processing standard corresponds to one third combination information. The edge device inputs the multiple sub-processing standards into the third language model, encodes and decodes each sub-processing standard through the third language model, and obtains the sub-standard description information of each sub-processing standard. The edge device merges the sub-standard description information of each sub-processing standard to obtain the standard description information of the data processing standard.

[0113] Among them, in the embodiment of the present application, the difference between the data processing standard and the standard description information is that the data processing standard is in the form of vectors and codes, and the standard description information is in the form of natural language. The third information generation model is an encoding and decoding model based on the attention mechanism, and the third language model is a large language model. The embodiment of the present application does not limit the structure of the third information generation model and the third language model.

[0114] Part 4: The edge device determines a second initial correction coefficient corresponding to the data processing standard based on the standard description information of the data processing standard.

[0115] In a possible implementation, the edge device extracts features from the standard description information to obtain standard description features of the standard description information. The edge device inputs the standard description features into a second coefficient determination model, and fully connects the standard description features through the second coefficient determination model to obtain a second coefficient determination feature. The edge device fully connects and normalizes the second coefficient determination feature through the second coefficient determination model to obtain the second initial correction coefficient.

[0116] The fifth part: the edge device fuses the first initial correction coefficient and the second initial correction coefficient to obtain the first computing power correction coefficient.

[0117] In a possible implementation, the edge device performs weighted fusion on the first initial correction coefficient and the second initial correction coefficient to obtain the first computing power correction coefficient.

[0118] Among them, the weights of weighted fusion are set by technical personnel according to actual conditions, and the embodiments of the present application do not limit this.

[0119] 304. The edge device determines a second computing power correction coefficient for processing the IoT data based on the data processing requirement and the data processing style.

[0120] In a possible implementation, the edge device determines the style description information of the data processing style. Based on the style description information of the data processing style, the edge device determines a third initial correction coefficient corresponding to the data processing style. The edge device merges the first initial correction coefficient and the third initial correction coefficient to obtain the second computing power correction coefficient.

[0121] In order to explain the above implementation more clearly, the above implementation will be explained in several parts below.

[0122] In the first part, the edge device determines the style description information of the data processing style.

[0123] In a possible implementation, the edge device extracts features of the data processing style to obtain processing style features of the data processing style. Based on the processing style features and multiple fourth prompt texts, the edge device determines multiple sub-processing styles corresponding to the data processing style, where different sub-processing styles correspond to different aspects, and different fourth prompt texts are used to indicate the generation of sub-processing styles of different aspects. The edge device determines sub-style description information of each sub-processing style. Based on the sub-style description information of each sub-processing style, the edge device generates style description information of the data processing style.

[0124] Among them, the fourth prompt text is in the form of a latent vector, and the fourth prompt text is set by the technician according to the actual situation, and the embodiment of the present application does not limit this.

[0125] For example, the edge device performs multiple full connections on the data processing style to obtain the processing style features of the data processing style. The edge device combines the processing style features with each fourth prompt text to obtain multiple fourth combination information, where one fourth combination information includes the processing style features and a fourth prompt text. The edge device inputs the multiple fourth combination information into the fourth information generation model, and decodes the multiple fourth combination information through the fourth information generation model to obtain the multiple sub-processing styles, where one sub-processing style corresponds to one fourth combination information. The edge device inputs the multiple sub-processing styles into the fourth language model, and encodes and decodes each sub-processing style through the fourth language model to obtain sub-style description information of each sub-processing style. The edge device merges the sub-style description information of each sub-processing style to obtain the style description information of the data processing style.

[0126] Among them, in the embodiment of the present application, the difference between the data processing style and the style description information is that the data processing style is in the form of vectors and codes, and the style description information is in the form of natural language. The fourth information generation model is an encoding and decoding model based on the attention mechanism, and the fourth language model is a large language model. The embodiment of the present application does not limit the structure of the fourth information generation model and the fourth language model.

[0127] In the second part, the edge device determines a third initial correction coefficient corresponding to the data processing style based on the style description information of the data processing style.

[0128] In a possible implementation, the edge device extracts features from the style description information to obtain style description features of the style description information. The edge device inputs the style description features into a third coefficient determination model, and performs full connection on the style description features through the third coefficient determination model to obtain a third coefficient determination feature. The edge device performs full connection and normalization on the third coefficient determination feature through the third coefficient determination model to obtain the third initial correction coefficient.

[0129] In the third part, the edge device fuses the first initial correction coefficient and the third initial correction coefficient to obtain the second computing power correction coefficient.

[0130] In a possible implementation, the edge device performs weighted fusion on the first initial correction coefficient and the third initial correction coefficient to obtain the second computing power correction coefficient.

[0131] Among them, the weights of weighted fusion are set by technical personnel according to actual conditions, and the embodiments of the present application do not limit this.

[0132] 305. The edge device determines the target computing power required to process the IoT data based on the basic computing power, the first computing power correction coefficient, and the second computing power correction coefficient.

[0133] Among them, the target computing power refers to the computing power required to process the IoT data according to the data processing needs and the data processing requirements.

[0134] In a possible implementation, the edge device multiplies the basic computing power with the first computing power correction coefficient and the second computing power correction coefficient to obtain the target computing power required to process the IoT data.

[0135] Among them, the weights of weighted fusion are set by technical personnel according to actual conditions, and the embodiments of the present application do not limit this.

[0136] Optionally, after step 305, the edge device can execute the following step 306 or 307, which is not limited in the embodiment of the present application.

[0137] 306. When the target computing power is less than or equal to the preset computing power, the edge device processes the IoT data based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data. The preset computing power is the maximum computing power that can be allocated by the edge device for the data processing request within a preset time period.

[0138] Among them, if the target computing power is less than or equal to the preset computing power, it means that the edge device has the ability to process the IoT data according to the data processing needs and the data processing requirements. There are two meanings here. First, the computing power of the edge device must be greater than the target computing power. Second, the computing power that the edge device can allocate to the task of processing the IoT data is sufficient. Therefore, the IoT data can be directly processed by the edge device.

[0139] In a possible implementation, the edge device extracts features of the data processing demand and the data processing requirement to obtain a processing demand feature of the data processing demand and a processing requirement feature of the data processing requirement. The edge device fuses the processing demand feature and the processing requirement feature to target processing data processing features. The edge device processes the IoT data based on the target data processing feature to obtain a data processing result of the IoT data.

[0140] 307. When the target computing power is greater than the preset computing power, the edge device forwards the data processing request and the data processing requirement to the cloud, so that the cloud processes the IoT data based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data. The data processing result of the IoT data is obtained from the cloud.

[0141] Among them, if the target computing power is greater than the preset computing power, it means that the edge device does not have the ability to process the IoT data according to the data processing needs and the data processing requirements. Therefore, it is handed over to the cloud with sufficient computing power for processing, thereby improving the processing effect of the IoT data.

[0142] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present application, which will not be described one by one here.

[0143] Through the technical solution provided by the embodiment of the present application, in terms of reasonable allocation of computing power, first of all, it is possible to efficiently utilize the computing power of edge devices, determine the target computing power based on IoT data, data processing needs and data processing requirements, and process it on the edge device when the target computing power is less than or equal to the preset computing power of the edge device, avoiding the edge device from blindly trying to process data when its own computing power is insufficient, improving the processing success rate and making full use of its computing power resources without causing waste; when the target computing power is greater than the preset computing power, it forwards data processing requests and requirements to the cloud for processing by the cloud, effectively utilizing the powerful computing power resources of the cloud. In terms of meeting terminal needs, this technical solution can meet a variety of requirements, and simultaneously obtain the data processing requirements (including standards and styles) of the target terminal when acquiring IoT data. These requirements are used when processing on edge devices or in the cloud, overcoming the problem that traditional technologies do not distinguish between different data processing standards and styles to meet the complex needs of different terminals; it can also flexibly select the processing location according to the computing power through cloud-edge fusion to achieve cloud-edge collaborative processing, and improve the overall data processing efficiency on the basis of meeting the data processing requirements of the target terminal.

[0144] Figure 4 This is a schematic diagram of the structure of a cloud-edge fusion processing system for IoT data provided by an embodiment of the present application, see Figure 4 , the system includes: an edge device 401 and a cloud 402.

[0145] The edge device is used to respond to a data processing request sent by a target terminal, obtain the IoT data to be processed carried by the data processing request, the data processing requirements of the IoT data, and the data processing requirements of the target terminal, which include data processing standards and data processing styles.

[0146] The edge device is used to determine the target computing power required to process the IoT data based on the IoT data, the data processing needs and the data processing requirements.

[0147] The edge device is used to process the IoT data based on the data processing demand and the data processing requirement to obtain the data processing result of the IoT data when the target computing power is less than or equal to the preset computing power. The preset computing power is the maximum computing power that the edge device can allocate to the data processing request within a preset time period.

[0148] The edge device is used to forward the data processing request and the data processing requirement to the cloud when the target computing power is greater than the preset computing power, so that the cloud processes the IoT data based on the data processing demand and the data processing requirement to obtain the data processing result of the IoT data. The data processing result of the IoT data is obtained from the cloud.

[0149] It should be noted that: the cloud-edge fusion processing system for IoT data provided in the above embodiment only uses the division of the above functional modules as an example when processing IoT data. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the edge device is divided into different functional modules to complete all or part of the functions described above. In addition, the cloud-edge fusion processing system for IoT data provided in the above embodiment and the cloud-edge fusion processing method embodiment for IoT data belong to the same concept. The specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0150] Through the technical solution provided by the embodiment of the present application, in terms of reasonable allocation of computing power, first of all, it is possible to efficiently utilize the computing power of edge devices, determine the target computing power based on IoT data, data processing needs and data processing requirements, and process it on the edge device when the target computing power is less than or equal to the preset computing power of the edge device, avoiding the edge device from blindly trying to process data when its own computing power is insufficient, improving the processing success rate and making full use of its computing power resources without causing waste; when the target computing power is greater than the preset computing power, it forwards data processing requests and requirements to the cloud for processing by the cloud, effectively utilizing the powerful computing power resources of the cloud. In terms of meeting terminal needs, this technical solution can meet a variety of requirements, and simultaneously obtain the data processing requirements (including standards and styles) of the target terminal when acquiring IoT data. These requirements are used when processing on edge devices or in the cloud, overcoming the problem that traditional technologies do not distinguish between different data processing standards and styles to meet the complex needs of different terminals; it can also flexibly select the processing location according to the computing power through cloud-edge fusion to achieve cloud-edge collaborative processing, and improve the overall data processing efficiency on the basis of meeting the data processing requirements of the target terminal.

[0151] Figure 5It is a structural diagram of an edge device provided in an embodiment of the present application. The edge device 500 may have relatively large differences due to different configurations or performances, and may include one or more processors (Central Processing Units, CPU) 501 and one or more memories 502, wherein the one or more memories 502 store at least one computer program, and the at least one computer program is loaded and executed by the one or more processors 501 to implement the methods provided in the above-mentioned various method embodiments. Of course, the edge device 500 may also have components such as a wired or wireless network interface, a keyboard, and an input and output interface for input and output. The edge device 500 may also include other components for implementing device functions, which will not be repeated here.

[0152] In an exemplary embodiment, a computer-readable storage medium is also provided, such as a memory including a computer program, and the computer program can be executed by a processor to complete the cloud-edge fusion processing method for IoT data in the above embodiment. For example, the computer-readable storage medium can be a read-only memory (ROM), a random access memory (RAM), a compact disc read-only memory (CD-ROM), a magnetic tape, a floppy disk, an optical data storage device, etc.

[0153] In an exemplary embodiment, a computer program product or a computer program is also provided, which includes a program code, and the program code is stored in a computer-readable storage medium. The processor of the edge device reads the program code from the computer-readable storage medium, and the processor executes the program code, so that the edge device executes the above-mentioned cloud-edge fusion processing method for IoT data.

[0154] In some embodiments, the computer program involved in the embodiments of the present application may be deployed and executed on an edge device, or on multiple edge devices located at one location, or on multiple edge devices distributed at multiple locations and interconnected by a communication network. Multiple edge devices distributed at multiple locations and interconnected by a communication network may constitute a blockchain system.

[0155] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0156] The above are only optional embodiments of the present application and are not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A cloud-edge fusion processing method for IoT data, characterized in that: Applied to an edge device, the method comprises: In response to a data processing request sent by a target terminal, obtaining the IoT data to be processed, the data processing requirements of the IoT data, and the data processing requirements of the target terminal carried in the data processing request, wherein the data processing requirements include a data processing standard and a data processing style; Determine a target computing power required to process the IoT data based on the IoT data, the data processing demand, and the data processing requirement; The determining, based on the IoT data, the data processing demand, and the data processing requirement, a target computing power required for processing the IoT data includes: Based on the IoT data and the data processing requirements, determine the basic computing power required to process the IoT data; based on the data processing requirements and the data processing standards, determine a first computing power correction coefficient for processing the IoT data; based on the data processing requirements and the data processing style, determine a second computing power correction coefficient for processing the IoT data; based on the basic computing power, the first computing power correction coefficient, and the second computing power correction coefficient, determine the target computing power required to process the IoT data; Based on the IoT data and the data processing requirements, determining the basic computing power required to process the IoT data includes: Determine the data type and data volume of the IoT data; determine a first initial computing power for processing the IoT data based on the data type and data volume; determine requirement description information of the data processing requirement, where the requirement description information is a natural language description of the data processing requirement; determine a second initial computing power that meets the data processing requirement based on the requirement description information of the data processing requirement; determine a data summary corresponding to the IoT data; determine a third initial computing power for processing the IoT data based on the data summary corresponding to the IoT data and the data processing requirement; merge the first initial computing power, the second initial computing power and the third initial computing power to obtain the basic computing power required to process the IoT data; When the target computing power is less than or equal to the preset computing power, the IoT data is processed based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data, and the preset computing power is the maximum computing power that can be allocated by the edge device to the data processing request within a preset time period; When the target computing power is greater than the preset computing power, the data processing request and the data processing requirement are forwarded to the cloud, so that the cloud processes the IoT data based on the data processing demand and the data processing requirement to obtain the data processing result of the IoT data; and the data processing result of the IoT data is obtained from the cloud.

2. The method according to claim 1, characterized in that The determining, based on the data type and the data volume, a first initial computing power for processing the IoT data includes: Based on the data type and the data volume, generate a first computing power acquisition request and determine a first reference computing power for processing the IoT data, wherein the first computing power acquisition request is used to request acquisition of a corresponding second reference computing power, and the second reference computing power is used to assist in determining the first initial computing power; Sending the first computing power acquisition request to the cloud, so that the cloud returns a plurality of second reference computing powers and the credibility of each of the second reference computing powers based on the first computing power acquisition request, wherein the plurality of second reference computing powers include a plurality of third reference computing powers and a fourth reference computing power, wherein the third reference computing power is sent to the cloud by other edge devices, and the fourth reference computing power is determined by the cloud; Based on the first reference computing power, the multiple second reference computing powers, and the credibility of each of the second reference computing powers, a first initial computing power for processing the Internet of Things data is determined.

3. The method according to claim 1, characterized in that The requirement description information for determining the data processing requirement includes: Extracting features of the data processing requirements to obtain processing requirement features of the data processing requirements; Based on the processing requirement characteristics and the plurality of first prompt texts, determining a plurality of sub-processing requirements corresponding to the data processing requirement, wherein different sub-processing requirements correspond to different aspects, and different first prompt texts are used to indicate the generation of sub-processing requirements of different aspects; Determining sub-requirement description information of each of the sub-processing requirements; Based on the sub-requirement description information of each of the sub-processing requirements, the requirement description information of the data processing requirement is generated.

4. The method according to claim 1, characterized in that: The determining, based on the requirement description information of the data processing requirement, a second initial computing power that meets the data processing requirement includes: Extracting features from the requirement description information to obtain requirement description features of the requirement description information; Based on the requirement description feature, generate a second computing power acquisition request and determine a fifth reference computing power that meets the data processing requirement, wherein the second computing power acquisition request is used to request to obtain a corresponding sixth reference computing power, and the sixth reference computing power is used to assist in determining the second initial computing power; Sending the second computing power acquisition request to the cloud, so that the cloud returns a plurality of sixth reference computing powers and the credibility of each of the sixth reference computing powers based on the second computing power acquisition request, wherein the plurality of sixth reference computing powers include a seventh reference computing power and an eighth reference computing power, the seventh reference computing power is sent to the cloud by the target edge device, the eighth reference computing power is determined by the cloud, and the target edge device is another edge device whose credibility is greater than or equal to the credibility threshold; Based on the fifth reference computing power, the sixth reference computing power, and the credibility of the sixth reference computing power, a second initial computing power for processing the Internet of Things data is determined.

5. The method according to claim 1, characterized in that The determining, based on the data processing requirement and the data processing standard, a first computing power correction coefficient for processing the IoT data includes: Determine requirement description information of the data processing requirement; Determining a first initial correction coefficient corresponding to the data processing requirement based on the requirement description information of the data processing requirement; Determine standard descriptive information of the data processing standard; Determining a second initial correction coefficient corresponding to the data processing standard based on the standard description information of the data processing standard; The first initial correction coefficient and the second initial correction coefficient are combined to obtain the first computing power correction coefficient.

6. The method according to claim 5, characterized in that The determining, based on the data processing requirement and the data processing style, a second computing power correction coefficient for processing the IoT data includes: Determining style description information of the data processing style; Determining a third initial correction coefficient corresponding to the data processing style based on the style description information of the data processing style; The first initial correction coefficient and the third initial correction coefficient are combined to obtain the second computing power correction coefficient.

7. The method according to claim 1, characterized in that The processing of the IoT data based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data includes: Extracting features from the data processing demand and the data processing requirement to obtain processing requirement features of the data processing demand and processing requirement features of the data processing requirement; The processing demand feature and the processing requirement feature are merged to obtain a target data processing feature; The Internet of Things data is processed based on the target data processing characteristics to obtain a data processing result of the Internet of Things data.

8. A cloud-edge fusion processing system for IoT data, characterized in that: The system includes: an edge device and a cloud; The edge device is used to respond to a data processing request sent by a target terminal, obtain the IoT data to be processed carried by the data processing request, the data processing requirements of the IoT data, and the data processing requirements of the target terminal, wherein the data processing requirements include a data processing standard and a data processing style; The edge device is used to determine the target computing power required to process the IoT data based on the IoT data, the data processing demand and the data processing requirement; The edge device is used to determine the basic computing power required to process the IoT data based on the IoT data and the data processing requirements; determine a first computing power correction coefficient for processing the IoT data based on the data processing requirements and the data processing standards; determine a second computing power correction coefficient for processing the IoT data based on the data processing requirements and the data processing style; determine the target computing power required to process the IoT data based on the basic computing power, the first computing power correction coefficient, and the second computing power correction coefficient; The edge device is used to determine the data type and data volume of the IoT data; based on the data type and the data volume, determine the first initial computing power for processing the IoT data; determine the demand description information of the data processing requirement, where the demand description information is a natural language description of the data processing requirement; based on the demand description information of the data processing requirement, determine the second initial computing power that meets the data processing requirement; determine the data summary corresponding to the IoT data; based on the data summary corresponding to the IoT data and the data processing requirement, determine the third initial computing power for processing the IoT data; merge the first initial computing power, the second initial computing power and the third initial computing power to obtain the basic computing power required to process the IoT data; The edge device is used to process the IoT data based on the data processing demand and the data processing requirement to obtain a data processing result of the IoT data when the target computing power is less than or equal to the preset computing power, and the preset computing power is the maximum computing power that can be allocated by the edge device to the data processing request within a preset time period; The edge device is used to forward the data processing request and the data processing requirement to the cloud when the target computing power is greater than the preset computing power, so that the cloud processes the Internet of Things data based on the data processing demand and the data processing requirement to obtain the data processing result of the Internet of Things data; and obtain the data processing result of the Internet of Things data from the cloud.

Images