A service deployment optimization system based on edge computing

Abstract

The application provides an edge-computing-based service deployment optimization system, comprising an edge server, a central server, a data transmission module, a service detection module and a service analysis module, the edge server is used for running a computing service model locally, the data transmission module is used for transmitting data between the edge server and the central server, the service detection module is used for detecting and counting running data of the computing service model, the service analysis module obtains an optimized deployment result according to the running data of the computing service model, the central server changes the computing service model deployed in the edge server according to the optimized deployment result, and the computing service model is used for executing a computing task. The system can continuously optimize the computing service model deployed in the edge server, meets the local task processing requirement and improves the processing efficiency.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of providing agent services, in particular to a service deployment optimization system based on edge computing. BACKGROUND

[0002] Edge computing refers to an open platform integrating network, computing, storage and application core capabilities, which provides services at the nearest end, initiates applications at the edge side, and produces faster network service responses to meet the basic needs of real-time business, application intelligence, security and privacy protection, etc. Edge computing is between physical entities and industrial connections, or at the top of physical entities.

[0003] The foregoing discussion of the background art is intended only to aid in understanding the present application. None of the foregoing discussion is to be taken as an admission that the material referred to is common general knowledge of the pertinent art.

[0004] Now many deployment systems have been developed, after our large amount of retrieval and reference, it is found that the existing deployment systems have systems such as disclosed in CN113055234B, which generally include obtaining data parameters of all task requests and performance parameters of all micro cloud servers according to edge computing network scene; then constructing a task completion time optimization problem model, and executing an approximate algorithm for each task request in order of task number from small to large to solve the task completion time minimization problem, calculating the link selection decision of all task requests and the deployment decision of VNF instances in SFC; finally, the network controller deploys VNF instances on the micro cloud server according to the deployment decision, and then allocates all VNFs in the SFC of all task requests to the VNF instances for execution. However, the system does not optimize the content after service deployment, which leads to the existence of space for improving the efficiency of task processing. SUMMARY

[0005] The present application relates to the field of providing agent services, in particular to a service deployment optimization system based on edge computing.

[0006] The present application adopts the following technical scheme:

[0007] The application discloses an edge-computing-based service deployment optimization system, which comprises an edge server, a central server, a data transmission module, a service detection module and a service analysis module, the edge server is used for running a computing service model locally, the data transmission module is used for transmitting data between the edge server and the central server, the service detection module is used for detecting and counting running data of the computing service model, the service analysis module is used for obtaining an optimized deployment result according to the running data of the computing service model, the central server is used for changing the computing service model deployed in the edge server according to the optimized deployment result, and the computing service model is used for executing a computing task.

[0008] The computing service model running in the edge server is referred to as an edge model, the computing service model running in the central server is used for executing a computing task not executed in the edge server, and the computing service model running in the central server is referred to as a central model.

[0009] The service analysis module calculates the priority P1 of the edge model according to the following formula:

[0010]

[0011] Wherein n is the processing number of the edge model, N is the processing number of all edge models and central models, T0 is the standard demand time of the computing task corresponding to the computing service model, and T1 is the time difference between the time of sending data to the edge model by the data flow unit and the time of receiving the execution result from the edge model.

[0012] The service analysis module calculates the priority P2 of the central model according to the following formula:

[0013]

[0014] Wherein m is the processing number of the central model, T2 is the time difference between the time of sending data to the central model by the data flow unit and the time of receiving the execution result from the central model, and t is the one-way transmission time of information between the edge server and the central server.

[0015] The service analysis module sorts the edge models and the central models according to the priorities from small to large, and the first Rn computing service models are used as the computing service models deployed in the edge server after optimization, and Rn is the maximum number of the computing service models capable of being run in the edge server.

[0016] Further, the central server comprises a prototype library, a copying unit, an execution unit and an interaction unit, the prototype library stores original data of all service computing models, the copying unit copies the service computing models in the prototype library, the execution unit is used for running the copied service computing models, and the interaction module is used for processing data received from the data transmission module.

[0017] Further, the data transmission module comprises a data channel and a model channel, the model channel is used for transmitting the service computing models to the edge server, the model channel is a one-way channel, and the data channel is used for transmitting ordinary data other than the service computing models, and the data channel is a two-way channel.

[0018] Further, the edge server comprises a loading execution unit and a data flow unit, the loading execution unit is used for receiving, loading and executing the service computing models, and the data flow unit is used for data interaction with the loading execution unit and the data channel.

[0019] Further, the system comprises two stages when processing the computing task, a simple deployment stage and an optimized deployment stage, the edge server submits a model application to the central server in the simple deployment stage, the central server sends corresponding computing service models to the edge server, and after the computing service models running on the edge server reach an upper limit, the central server optimizes and deploys the computing service models deployed on the edge server at a fixed time interval.

[0020] The present application has the following beneficial effects:

[0021] The computing service models for processing tasks of the system run on the edge client and the central client respectively, the number and time of the running computing service models are monitored, the monitoring data are analyzed and processed, more suitable computing service model data running on the edge client are obtained, the computing service models deployed on the edge client are optimized and adjusted according to the analysis result, and the processing efficiency of the tasks is improved.

[0022] For further understanding of the features and technical contents of the present application, please refer to the following detailed description and drawings of the present application, however, the drawings provided are only used for reference and illustration, and are not used to limit the present application. DETAILED DESCRIPTION

[0023] Figure 1 It is a schematic diagram of the overall structure framework of the present application;

[0024] Figure 2 It is a schematic diagram of the central server of the present application;

[0025] Figure 3 A schematic diagram of a data transmission module of the present application is shown in FIG. 1.

[0026] Figure 4 A schematic diagram of an edge server of the present application is shown in FIG. 2.

[0027] Figure 5 A schematic diagram of a simple deployment stage operation flow of the present application is shown in FIG. 3. DETAILED DESCRIPTION

[0028] The following is a detailed description of the embodiments of the present application, which is provided to give a person skilled in the art a better understanding of the advantages and effects of the present application. The present application can be implemented or applied by other different embodiments, and the details in the present specification can be modified and changed in various ways based on different views and applications without departing from the spirit of the present application. In addition, the drawings of the present application are only simple schematic illustrations and are not drawn according to the actual size, and it is declared in advance. The following embodiments will further illustrate the related technical content of the present application in detail, but the disclosed content is not used to limit the protection scope of the present application.

[0029] Embodiment 1.

[0030] The present embodiment provides an edge computing-based service deployment optimization system, which combines Figure 1 , including an edge server, a central server, a data transmission module, a service detection module and a service analysis module, the edge server is used to run a computing service model locally for a user, the data transmission module is used to transmit data between the edge server and the central server, the service detection module is used to detect and count the running data of the computing service model, the service analysis module obtains an optimized deployment result according to the running data of the computing service model, the central server changes the computing service model deployed in the edge server according to the optimized deployment result, and the computing service model is used to execute a computing task;

[0031] The computing service model running in the edge server is called an edge model, and the computing service model running in the central server is used to execute a computing task that is not executed in the edge server, and the computing service model running in the central server is called a central model;

[0032] The service analysis module calculates the priority P1 of the edge model according to the following formula:

[0033]

[0034] Wherein, n is the processing number of the edge model, N is the processing number of all edge models and central models, T0 is the standard demand time of the computing task corresponding to the computing service model, T1 is the time difference between the data flow unit sending data to the edge model and receiving the execution result from the edge model;

[0035] The service analysis module calculates the priority P2 of the central model according to the following formula:

[0036]

[0037] Wherein, m is the processing number of the central model, T2 is the time difference between the data flow unit sending data to the central model and receiving the execution result from the central model, t is the one-way transmission time of information between the edge server and the central server;

[0038] The service analysis module sorts the edge model and the central model according to the priority from small to large, and the first Rn computing service models are deployed on the edge server after optimization as the computing service models, and Rn is the maximum number of computing service models that the edge server can run;

[0039] The central server includes a prototype library, a copying unit, an execution unit and an interaction unit, the prototype library saves the original data of all service computing models, the copying unit copies the service computing models in the prototype library, the execution unit is used to run the copied service computing models, and the interaction module is used to process the data received from the data transmission module;

[0040] The data transmission module includes a data channel and a model channel, the model channel is used to transmit the service computing model to the edge server, the model channel is a one-way channel, and the data channel is used to transmit ordinary data which is not a service computing model, the data channel is a two-way channel;

[0041] The edge server includes a loading execution unit and a data flow unit, the loading execution unit is used to receive, load and execute the service computing module, and the data flow unit is used to interact with the loading execution unit and the data channel;

[0042] The system includes two stages when processing computing tasks, the first stage is a simple deployment stage, and the second stage is an optimized deployment stage, the edge server submits a model application to the central server in the simple deployment stage, the central server sends the corresponding computing service model to the edge server, and after the computing service model running on the edge server reaches the upper limit, the central server optimizes and deploys the computing service model deployed in the edge server at a fixed time interval.

[0043] Embodiment two.

[0044] The embodiment includes all the contents of embodiment one, and provides an edge computing-based service deployment optimization system, which comprises an edge server, a central server, a data transmission module, a service detection module and a service analysis module. The central server contains a service computing model, the edge server can collect running data of a local terminal of a client and load the service computing model, the data transmission module is used for transmitting data between the edge server and the central server, the service detection module is used for detecting running data of the service computing model in the edge server, and the service analysis module optimizes deployment of the service computing model in the edge server according to the running data of the service computing model.

[0045] In combination Figure 2 , the central server comprises a prototype library, a copying unit, an execution unit and an interaction unit. The prototype library stores original data of all service computing models. The copying unit copies the service computing models in the prototype library. The execution unit is used for running the copied service computing models. The interaction module is used for processing data received from the data transmission module.

[0046] In combination Figure 3 , the data transmission module comprises a data channel and a model channel. The model channel is used for transmitting the service computing model to the edge server. The model channel is a one-way channel. The data channel is used for transmitting ordinary data other than the service computing model. The data channel is a two-way channel.

[0047] In combination Figure 4 , the edge server comprises a loading execution unit and a data flow communication unit. The loading execution unit is used for receiving, loading and executing the service computing module. The data flow communication unit is used for interacting with the loading execution unit and the data channel to execute data required for a task and execution results.

[0048] The system comprises two stages when processing a computing task. The first stage is a simple deployment stage, and the second stage is an optimized deployment stage. In combination Figure 5 , the running process of the simple deployment stage comprises the following steps:

[0049] S1, the data flow communication unit in the edge server receives running data collected in a local terminal;

[0050] S2, the data flow communication unit sends a model application to the central server according to the type of the running data;

[0051] S3, the interaction unit receives the model application and sends a copying list to the copying unit;

[0052] S4, the copying unit copies the corresponding service computing model in the prototype library according to the model information in the copy list;

[0053] S5, the interaction unit sends the service computing model data copied to the edge server through the model channel;

[0054] S6, the loading execution unit in the edge server receives the service computing model from the model channel, and after loading is completed, the loading execution unit sends the loading completion information to the data flow unit;

[0055] S7, the data flow unit sends the corresponding collection data to the loading execution unit, and the service computing model in the loading execution unit processes the collection data to obtain an execution result, which is sent to the data flow unit;

[0056] S8, the data flow unit feeds back the execution result to the local terminal;

[0057] The data flow unit creates a data identification table according to the loading completion information, when the collection data is in the data identification table, the collection data is sent to the loading execution unit, when the collection data is not in the data identification table, the data flow unit sends a model application to the central server, if the service computing model that the loading execution unit can run reaches the upper limit, the service computing model copied by the copying unit is saved to the execution unit, and the data flow unit sends the collection data to the execution unit of the central server for operation processing;

[0058] The computing service model running in the loading execution unit is called an edge model, the computing service model running in the execution unit is called a central model, when the service computing model that the loading execution unit can run reaches the upper limit, the service detection module counts the processing times and processing time of the edge model and the central model, when the counted time reaches the basic duration, the system formally enters the optimization deployment stage, the service analysis module determines whether the central model needs to replace the existing edge model to realize the deployment optimization effect according to the model running data obtained by counting;

[0059] The service analysis module calculates the priority P1 of the edge model according to the following formula:

[0060]

[0061] Wherein, n is the processing number of the edge model, N is the processing number of all edge models and central models, T0 is the standard demand time of the computing task corresponding to the computing service model, T1 is the time difference between the data flow unit sending data to the edge model and receiving the execution result from the edge model;

[0062] The service analysis module calculates the priority P2 of the central model according to the following formula:

[0063]

[0064] Wherein, m is the processing number of the central model, T2 is the time difference between the data flow unit sending data to the central model and receiving the execution result from the central model, and t is the one-way transmission time of information between the edge server and the central server;

[0065] The relationship between N and m and n satisfies:

[0066] N = ∑m + ∑n;

[0067] The service analysis module sorts the edge model and the central model according to the priority from small to large, and the first Rn computing service models are deployed on the edge server after optimization, Rn is the maximum number of computing service models that the loading execution unit can run, the central server sends the computing service model that needs to be replaced to the edge server, and the loading execution unit of the edge server deletes the replaced computing service model and loads a new computing service model;

[0068] The service detection module is provided with a timer, when the service detection module performs statistics on the detected data, the timer starts to work after setting the time length, when the count of the timer reaches 0, the service detection module sends the statistical data to the service analysis module, the service analysis module optimizes the deployment according to the statistical data, when the edge model in the edge server is replaced, the service detection module clears the statistical detection data, and starts to detect and count again after setting the timer, when the edge model in the edge server is not replaced, the service detection module retains the statistical data, and detects and counts on the basis of the original statistical data after setting the timer;

[0069] The service detection module obtains statistical data by identifying data entering and exiting the data flow unit, the data sent by the data flow unit includes the number of the computing service model, the execution result data received by the data flow unit includes the number of the computing service model, the service detection module identifies the number of the computing service model and obtains the processing frequency data of each computing service model, and the service detection module obtains time T1 or T2 by calculating the difference between the time stamp of the sent data and the time stamp of the received data with the same computing service model number.

[0070] The above disclosed content is only the preferred feasible embodiment of the present application, and does not limit the protection scope of the present application, so any equivalent technical change made by applying the content of the present application specification and drawings is included in the protection scope of the present application, and in addition, the elements can be updated as technology develops.

Claims

1. An edge-computing-based service deployment optimization system, characterized by, The system comprises an edge server, a central server, a data transmission module, a service detection module and a service analysis module, the edge server is configured to run a computing service model locally, the data transmission module is configured to transmit data between the edge server and the central server, the service detection module is configured to detect and count running data of the computing service model, the service analysis module is configured to obtain an optimized deployment result according to the running data of the computing service model, and the central server is configured to change the computing service model deployed in the edge server according to the optimized deployment result, the computing service model is configured to execute a computing task; The computing service model running in the edge server is referred to as an edge model, and the computing service model running in the central server is configured to execute a computing task not executed in the edge server, and the computing service model running in the central server is referred to as a central model; The service analysis module calculates a priority P1 of the edge model according to the following formula: Wherein, n is the processing number of the edge model, N is the processing number of all edge models and central models, T0 is the standard demand time of the computing task corresponding to the computing service model, and T1 is the time difference between the data flow unit sending data to the edge model and receiving the execution result from the edge model; The service analysis module calculates a priority P2 of the central model according to the following formula: Wherein, m is the processing number of the central model, T2 is the time difference between the data flow unit sending data to the central model and receiving the execution result from the central model, and t is the one-way transmission time of information between the edge server and the central server; The service analysis module sorts the edge models and the central models according to the priorities from small to large, and the first Rn computing service models are deployed in the edge server as the computing service models after optimization, and Rn is the maximum number of computing service models that can be run in the edge server.

2. The edge computing based service deployment optimization system of claim 1, wherein, The central server comprises a prototype library, a copying unit, an execution unit and an interaction unit, the prototype library stores original data of all service computing models, the copying unit copies the service computing models in the prototype library, the execution unit is configured to run the copied service computing models, and the interaction unit is configured to process data received from the data transmission module.

3. The edge computing based service deployment optimization system of claim 2, wherein, The data transmission module comprises a data channel and a model channel, the model channel is configured to transmit the service computing model to the edge server, the model channel is a one-way channel, the data channel is configured to transmit ordinary data that is not a service computing model, and the data channel is a two-way channel.

4. The edge computing based service deployment optimization system of claim 3, wherein, The edge server comprises a loading execution unit and a data flow unit, the loading execution unit is configured to receive, load and execute the service computing module, and the data flow unit is configured to interact with the loading execution unit and the data channel.

5. The edge computing based service deployment optimization system of claim 4, wherein, The system comprises two stages when processing a computing task, a first stage is a simple deployment stage, and a second stage is an optimized deployment stage, the edge server submits a model application to the central server in the simple deployment stage, the central server sends a corresponding computing service model to the edge server, and after the computing service model running on the edge server reaches an upper limit, the central server optimizes and deploys the computing service model deployed in the edge server at a fixed time interval.

Images