An edge computing method for realizing heterogeneous fusion of multi-network data

Abstract

The present invention relates to the field of edge computing technology, and specifically discloses an edge computing method for realizing heterogeneous fusion of multi-network data. The method includes: S100, establishing an edge task allocation unit, which is respectively connected to an edge task processor, a cloud platform and edge devices; S200, receiving heterogeneous resource data through the edge task allocation unit, and dividing each task data into data groups with identifiers according to a preset segmentation strategy according to the encryption level corresponding to each task in the heterogeneous resource data; S300, performing allocation calculation on the data groups according to a preset allocation strategy according to the operating conditions of the edge task processor and obtaining an identification sequence of the allocation method; S400, transmitting the data in the data groups back to the edge task allocation unit after calculation. The edge task allocation unit obtains the calculation results of each task according to the identification sequence and transmits the calculation results to the cloud platform and the corresponding edge devices.

Description

Technical Field

[0001] The present invention relates to the technical field of edge computing, and specifically to an edge computing method for realizing heterogeneous fusion of multi-network data. Background Art

[0002] With the development of Internet technology and the wide application of Internet of Things technology, the requirements for computing and processing are getting higher and higher. This is mainly because the development of services and scenarios such as 5G, industrial Internet, and IoT is getting faster and faster, and there are more and more intelligent terminal devices, resulting in an increasing demand for the sinking of edge computing services. Therefore, if the traditional cloud computing method is adopted and all processing is placed in the center, it is difficult to meet the growth of a large number of edge intelligent devices. Moreover, if all data or videos are sent back to the cloud for processing, the overall cost is relatively high and there is a high latency, which cannot meet the efficiency requirements. In industrial application scenarios, bandwidth blocking, latency problems, and unpredictable network interruptions are unacceptable. Therefore, in addition to the unified control of the cloud, network nodes in the industrial field must have a certain computing ability to be able to independently judge and solve problems, and detect and predict abnormal situations in a timely and real-time manner. Therefore, the edge computing method is the trend of future computing methods.

[0003] As a way for cloud computing to extend to the edge, compared with cloud computing where software and hardware resources are all centrally deployed in large data centers far from users, edge computing is to place computing resources closer to the "edge" of users or devices, thereby reducing latency and bandwidth consumption and providing real-time processing close to the data source. On the one hand, IT services are decentralized on the edge side, ensuring computing efficiency and reducing the computing pressure on cloud computing. On the other hand, because edge services or facilities are autonomous, in the case of network disconnection between the cloud and the edge, there is a certain control ability and edge hosting ability, ensuring the stability of the computing form.

[0004] In the prior art, since edge computing resources are dynamically changing, when performing edge computing, unreasonable invocation of edge computing resources will cause congestion in the edge computing server, affecting other computing tasks. It may also be unable to effectively process computing tasks due to a small amount of resources invoked. In addition, for multi-network heterogeneous resources, there are problems with poor security during data transmission and a risk of leakage during the edge computing process. Summary of the Invention

[0005] The purpose of the present invention is to provide an edge computing method for realizing heterogeneous fusion of multi-network data, and solve the following technical problems:

[0006] How to improve the rationality and security of computing task allocation during edge computing.

[0007] The object of the present invention can be achieved by the following technical solutions:

[0008] An edge computing method for realizing heterogeneous fusion of multi-network data, the method comprising:

[0009] S100. Establish an edge task allocation unit, which is respectively connected to an edge task processor, a cloud platform and edge devices;

[0010] S200. Receive heterogeneous resource data through the edge task allocation unit, and divide each task data into data groups with identifiers according to a preset segmentation strategy according to the encryption level corresponding to each task in the heterogeneous resource data;

[0011] S300. Perform allocation calculation on the data groups according to a preset allocation strategy according to the operating conditions of the edge task processor and obtain an identifier sequence of the allocation method;

[0012] S400. Transmit the data in the data groups back to the edge task allocation unit after calculation. The edge task allocation unit obtains the calculation results of each task according to the identifier sequence, and transmits the calculation results to the cloud platform and the corresponding edge devices.

[0013] In one embodiment, the preset segmentation strategy is:

[0014] Group the task data by the data interval length L, and identify each data group in order;

[0015] The data interval length L is determined according to the encryption level.

[0016] In one embodiment, the preset allocation strategy is:

[0017] Obtain the estimated calculation amount C of all tasks for , obtain the operating conditions of each edge computing device, and the operating state includes the maximum calculation amount C max in the first specific duration and the current calculation amount C to be processed await ;

[0018] According to the formula calculate the remaining calculation amount C of the edge task processor edge , α i is the available calculation amount coefficient, and α i is less than 1;

[0019] Compare the estimated calculation amount C for with the remaining calculation amount C edge :

[0020] If C for ≤C edge , then allocate the data group to the edge task processor;

[0021] If C for > C edge , then allocate the non-overflowing part of the data group to the edge task processor and the overflowing part of the data group to the cloud platform.

[0022] In one embodiment, the method for allocating the data group to the edge task processor is as follows:

[0023] SS100. Obtain the edge task processors that are idle at the t i time point, where the t i time point is the time point of the second specific duration from the current time point t i-1 ;

[0024] SS200. Randomly allocate the data group according to the calculation amount per unit duration of the edge task processor and the start calculation point from the t i duration;

[0025] where i = 1, 2,..., m, t m - t0 = the first specific duration, t i - t i-1 = the second specific duration;

[0026] SS300. Repeat steps S1 and S2 by incrementing i until i = m.

[0027] In one embodiment, the generation process of the identification sequence is as follows:

[0028] When randomly allocating the data group, record the corresponding processing platform number of the data group and its order on the platform, and link the recorded data with the identification corresponding to the data group;

[0029] For the data group of each task, form a sequence of the recorded data according to the identification order, and use this as the identification sequence.

[0030] In one embodiment, the method for obtaining the calculation result is as follows:

[0031] According to the corresponding processing platform number and its order on the platform in each task identification sequence, recombine the calculation data of the data group to obtain the calculation result.

[0032] In one embodiment, step S300 further includes:

[0033] Encrypt the allocated data group and send the encrypted data group to the edge task processor and the cloud platform.

[0034] In one embodiment, in steps S1 and S3, the idle edge task processors are sorted according to the priority coefficient of each edge task processor, and are allocated in sequence according to the sorting order.

[0035] In one embodiment, the calculation method of the priority coefficient of the edge task processor is as follows:

[0036] Through the formula Calculate the priority coefficient P of the edge task processor;

[0037] Among them, δ1 and δ2 are the transmission weight coefficient and the calculation weight coefficient respectively, L is the path distance between the edge task processor and the edge task allocation unit, and v t Is the transmission speed from the edge task processor to the edge task allocation unit, C th Is the preset task volume, and v c Is the calculation speed of the edge task processor.

[0038] Advantages of the present invention:

[0039] (1) Through the segmentation of the data group, the present invention can make better use of the calculation resource allocation process, improve the allocation rationality of edge computing, and the segmentation and tracing back processes of the data group can implement the encryption mechanism of the data fusion process, thereby improving the security of the data transmission process.

[0040] (2) By analyzing the calculation amount to be processed, the present invention reasonably allocates the calculation tasks to the edge task processor and the cloud platform according to the size of the calculation amount, ensuring the calculation efficiency and the rationality of the calculation amount allocation.

[0041] (3) The present invention can continuously adjust the task allocation according to the specific state of the edge task processor, thereby realizing the timely adjustment of data allocation during the dynamic task change of the edge task manager, and further ensuring the rationality of data allocation.

[0042] (4) By sorting the priority of the edge task processors related to the edge devices and allocating them in sequence, this method can enable the edge task processors with higher priorities to be preferentially allocated, thereby more reasonably allocating the calculation tasks and ensuring the rationality of edge computing allocation and the calculation efficiency. Description of the Drawings

[0043] The following further describes the present invention with reference to the drawings.

[0044] Figure 1 Is the step flow chart of the edge computing method for realizing multi-network data heterogeneous fusion of the present invention;

[0045] Figure 2 Is the step flow chart of the method for allocating the data group to the edge task processor in the present invention. Detailed Embodiments

[0046] Next, the technical solutions in the embodiments of the present invention will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention.

[0047] Please refer to Figure 1 As shown, in one embodiment, an edge computing method for realizing heterogeneous fusion of multi-network data is provided. The method includes:

[0048] S100. Establish an edge task allocation unit, and the edge task allocation unit is respectively connected to an edge task processor, a cloud platform, and edge devices;

[0049] S200. Receive heterogeneous resource data through the edge task allocation unit, and divide each task data into data groups with identifiers according to a preset segmentation strategy according to the encryption level corresponding to each task in the heterogeneous resource data;

[0050] S300. Perform allocation calculation on the data groups according to a preset allocation strategy according to the operating conditions of the edge task processor and obtain an identifier sequence of the allocation method;

[0051] S400. Transmit the data in the data groups back to the edge task allocation unit after calculation. The edge task allocation unit obtains the calculation results of each task according to the identifier sequence, and transmits the calculation results to the cloud platform and the corresponding edge devices.

[0052] Through the above technical solutions, by establishing an edge task allocation unit and connecting it to an edge task processor, a cloud platform, and edge devices, the functions of receiving, sending, and allocating computing tasks are realized. Then, heterogeneous computing resources are received through the edge task allocation unit to achieve unified allocation of data in different network forms and different protocols. For each computing task, it is divided into multiple data groups and each data group is identified. Then, according to the operating conditions of the edge task processor, the data groups are allocated and an identifier sequence of the allocation method is obtained. Therefore, after the calculation is completed by the edge task processor, the calculation results can be traced back and recombined according to the identifier sequence, and then the calculation results of each task can be obtained. The calculation results are transmitted to the cloud platform and the corresponding edge devices, thereby completing the process of edge computing. In this process, the segmentation of data groups can make better use of the process of allocating computing resources, improving the allocation rationality of edge computing. And the segmentation and trace-back processes of data groups can implement the encryption mechanism of the data fusion process, thereby improving the security of the data transmission process.

[0053] In step S100, the edge task allocation unit is implemented by an intelligent device with DTS in the prior art, which will not be elaborated here; in step S200, the specific process of splitting and packing the data group is implemented by the prior art, which will not be elaborated here; in step S300, the form of data group transmission depends on the specific application scenario and is not limited here; in step S400, the process of receiving and backtracking and reorganizing the data group is the prior art, which will not be elaborated here.

[0054] As an implementation manner of the present invention, the preset splitting strategy is:

[0055] Group the task data according to the data interval length L, and sequentially identify each data group after grouping;

[0056] The data interval length L is determined according to the encryption level.

[0057] Through the above technical solution, the data group is split by using the preset splitting strategy. Specifically, the task data is grouped according to the data interval length L, and each data group after grouping is sequentially identified. This splitting method can make the amount of calculation to be processed by each data group relatively similar. In addition, different encryption levels result in different selections of the data interval length L. When the encryption level is high, L is selected to be shorter, thereby increasing the complexity of data fusion and improving the confidentiality level. The specific selection of the L length is not limited here.

[0058] As an implementation manner of the present invention, the preset allocation strategy is:

[0059] Obtain the estimated calculation amount C of all tasks for , obtain the operating conditions of each edge computing device, and the operating state includes the maximum calculation amount C under the first specific duration max and the current calculation amount C to be processed await ;

[0060] According to the formula calculate the remaining calculation amount C of the edge task processor edge , α i is the available calculation amount coefficient, and α i is less than 1;

[0061] Compare the estimated calculation amount C for with the remaining calculation amount C edge :

[0062] If C for ≤C edge , then allocate the data group to the edge task processor;

[0063] If C for >C edge, the data group of the non-overflow part is allocated to the edge task processor, and the data group of the overflow part is allocated to the cloud platform.

[0064] Through the above technical solution, by analyzing the amount of computation to be processed and reasonably allocating the computing tasks to the edge task processor and the cloud platform according to the size of the computation amount. Specifically, the maximum computation amount of all edge task processors is predicted according to the first specific duration, where the first specific duration can be selected relatively according to the cloud computing transmission and processing time, ensuring the computing efficiency and the rationality of the computation amount allocation.

[0065] In the above technical solution, the remaining computation amount C edge in the formula of i α is the available computation amount coefficient, which is determined according to the specific state of each edge task processor. By multiplying this coefficient α i with the maximum computation amount C max , the available computation amount of each edge task processor can be obtained. This setting can reserve the basic computing power for each edge task processor and ensure the operation of the basic functions of the edge task processor.

[0066] As an implementation manner of the present invention, please refer to Figure 2 shown in the figure, the method for allocating the data group to the edge task processor is as follows:

[0067] SS100. Obtain the edge task processors that are idle at the time point t i in the edge task processors, where the time point t i is the time point of the second specific duration away from the current time point t i-1 ;

[0068] SS200. Randomly allocate the data group according to the computation amount per unit duration of the edge task processor and the distance from the start computing point to t i duration;

[0069] where i = 1, 2,..., m, t m - t0 = the first specific duration, t i - t i-1 = the second specific duration;

[0070] SS300. Repeat steps S1 and S2 by incrementing i until i = m.

[0071] Through the above technical solution, the first specific duration is evenly divided into several second specific durations, and based on the current time point, the edge task processors that are idle at the time point t1 are obtained. According to the computation amount C unit per unit duration of the edge task processor and the start computing point t b from t iDuration randomly allocated data groups, where the computational amount per unit duration is C unit Varies according to different edge task processors. The start calculation point is the time point when each edge task processor finishes the previous calculation task. Therefore, through the formula C unit *(t1 - t b )), the computational amount of a single edge task processor during the period from t0 to t1 can be calculated. Therefore, by allocating data groups according to this computational amount, the allocation of data groups can be adjusted more appropriately according to the current computational status of each edge task processor. Additionally, in this embodiment, by repeating steps S1 and S2 with an increment of i at intervals of the second specific duration, the allocation of tasks can be continuously adjusted according to the specific status of the edge task processors, and thus the timely adjustment of data allocation during the dynamic task change process of the edge task manager can be achieved, further ensuring the rationality of data allocation.

[0072] In the above technical solution, the random allocation of data is generated through a random algorithm, and the random process can ensure the disordered state of the task group allocation, thereby playing a role in encrypting the data fusion process.

[0073] As an implementation manner of the present invention, the generation process of the identification sequence is as follows:

[0074] When randomly allocating data groups, record the serial number of the processing platform corresponding to the data group and the order on this platform, and link the recorded data with the identifier corresponding to the data group;

[0075] For the data groups of each task, form a sequence of the recorded data according to the identifier order, and use this as the identification sequence.

[0076] Through the above technical solution, by recording the serial number of the processing platform corresponding to the data group and the order on this platform when allocating the data group, and linking the recorded data with the identifier corresponding to the data group, the recording of the random allocation method is realized. Then, for the data groups of each task, a sequence of the recorded data is formed according to the identifier order, and thus the identification sequence of each task can be obtained.

[0077] As an implementation manner of the present invention, the method for obtaining the calculation result is as follows:

[0078] According to the serial number of the processing platform corresponding to and the order on this platform in each task identification sequence, recombine the calculation data of the data group to obtain the calculation result.

[0079] Through the above technical solution, according to the relevant information corresponding to the identification sequence of each task, when the calculation data of the data group is transmitted back, the calculation data is recombined according to the corresponding relationship of the relevant information, and thus the calculation result can be obtained.

[0080] As an implementation manner of the present invention, step S300 further includes:

[0081] Encrypt the allocated data group and send the encrypted data group to the edge task processor and the cloud platform.

[0082] Through the above technical solution, by encrypting the allocated data group, the double encryption effect of the data group can be achieved, ensuring the security of data during the edge computing process.

[0083] In the above embodiment, the encryption method selects an asymmetric encryption algorithm, and its specific implementation process is prior art and will not be elaborated here.

[0084] As an implementation manner of the present invention, in steps S1 and S3, the idle edge task processors are sorted according to the priority coefficient of each edge task processor, and are allocated in sequence according to the sorting order.

[0085] Through the above technical solution, by performing priority sorting on the edge task processors related to the edge device and allocating them in sequence, this method can enable the edge task processors with higher priorities to be preferentially allocated, and thus can more reasonably allocate computing tasks, ensuring the rationality and computing efficiency of edge computing allocation.

[0086] As an implementation manner of the present invention, the calculation method of the priority coefficient of the edge task processor is:

[0087] Through the formula Calculate the priority coefficient P of the edge task processor;

[0088] Wherein, δ1 and δ2 are the transmission weight coefficient and the calculation weight coefficient respectively, L is the path distance between the edge task processor and the edge task allocation unit, v t is the transmission speed from the edge task processor to the edge task allocation unit, C th is the preset task volume, v c is the calculation speed of the edge task processor.

[0089] Through the above technical solution, a calculation method of the priority coefficient P is provided. The influence of transmission time on the priority is judged according to the path distance and transmission speed between the edge task processor and the edge task allocation unit. The influence of calculation time on the priority is judged through the preset task volume and calculation speed, and weighted calculation is performed through the transmission weight coefficient δ1 and the calculation weight coefficient δ2. Among them, C this the preset task volume, and its value is a fixed quantity. The larger the P value, the higher its priority. Therefore, the priority coefficient calculated in this way can sort the edge task processors by considering the factors of data transmission speed and the computing speed of the edge task processors, and achieve the priority allocation of the edge task processors with higher priorities.

[0090] The above has described in detail an embodiment of the present invention. However, the content described above is only a preferred embodiment of the present invention and cannot be considered as limiting the scope of implementation of the present invention. All equivalent changes and improvements made within the scope of the application of the present invention shall still fall within the scope covered by the patent of the present invention.

Claims

1. An edge computing method for realizing heterogeneous fusion of multi-network data, characterized in that, The method includes: S100. Establish an edge task allocation unit, which is respectively connected to an edge task processor, a cloud platform, and edge devices; S200. Receive heterogeneous resource data through the edge task allocation unit, and divide each task data into data groups with identifiers according to a preset segmentation strategy based on the encryption level corresponding to each task in the heterogeneous resource data; S300. Perform allocation calculations on the data groups according to a preset allocation strategy based on the operating conditions of the edge task processor and obtain an identifier sequence of the allocation method; S400. Transmit the calculated data in the data groups back to the edge task allocation unit. The edge task allocation unit obtains the calculation results of each task according to the identifier sequence and transmits the calculation results to the cloud platform and the corresponding edge devices; The preset segmentation strategy is: Group the task data by the data interval length L, and identify each data group in sequence after grouping; The data interval length L is determined according to the encryption level; Obtain the estimated computational volume C of all tasks for , obtain the operating conditions of each edge task processor, where the operating conditions include the maximum computational volume C under a first specific duration max and the current computational volume C to be processed await ; According to the formula calculate the remaining computing power C of the edge task processor edge , α i is the available computing power coefficient, and α i is less than 1; Compare the estimated computation amount C for with the remaining computation amount C edge as follows: If C for ≤ C edge , the data group is assigned to the edge task processor; If C for > C edge , then allocate the non-overflowing part of the data group to the edge task processor and allocate the overflowing part of the data group to the cloud platform; The method for allocating the data groups to the edge task processor is: SS100, obtain the edge task processor that is idle at time t i The edge task processor that is idle at time point t i The time point is the time point at a second specific duration from the current time point t i-1 The time point of the second specific duration; SS200, randomly allocate data groups according to the computing capacity per unit time of the edge task processor and the start computing point distance t i Duration where i = 1, 2, …, m, t m - t0 = the first specific duration, t i - t i-1 = the second specific duration; SS300. Repeat steps S1 and S2 by incrementing i until i = m.

2. The edge computing method for realizing multi-network data heterogeneous fusion according to claim 1, wherein The generation process of the identifier sequence is: When randomly allocating data groups, record the serial number of the edge task processor corresponding to the data group and its order in the edge task processor, and link the recorded data with the identifier corresponding to the data group; For the data groups of each task, form a sequence of the recorded data according to the identifier order, and use this as the identifier sequence.

3. The edge computing method for realizing multi-network data heterogeneous fusion according to claim 2, wherein The method for obtaining the calculation result is: According to the serial number of the edge task processor corresponding to each task in the identifier sequence and its order in the edge task processor, recombine the calculation data of the data group to obtain the calculation result.

4. An edge computing method for realizing multi-network data heterogeneous fusion according to claim 1, characterized in that, Step S300 also includes: Encrypt the allocated data groups and send the encrypted data groups to the edge task processor and the cloud platform.

5. The edge computing method for realizing multi-network data heterogeneous fusion according to claim 1, wherein, Idle edge task processors are sorted according to the priority coefficient of each edge task processor, and are allocated in sequence according to the sorting order.

6. The edge computing method for realizing multi-network data heterogeneous fusion according to claim 5, characterized in that The calculation method of the priority coefficient of the edge task processor is: Through the formula Calculate the priority coefficient P of the edge task processor; Among them, δ1 and δ2 are the transmission weight coefficient and the calculation weight coefficient respectively, L is the path distance between the edge task processor and the edge task allocation unit, and v t is the transmission speed from the edge task processor to the edge task allocation unit, and C th is the preset task volume, and v c is the calculation speed of the edge task processor.

Images