Edge container deployment and power distribution service quick response method

Abstract

The invention relates to an edge container deployment and power distribution service quick response method. The method aims at the technical field of power systems and information, fully considers the response time, resource utilization rate, error rate and data consistency of micro-services, constructs micro-service processing quality indexes, expands or shrinks the capacity of a container based on the processing quality of the micro-services in the container, predicts the number of the micro-services by using historical data, and improves the quality of the micro-services. The capacity of all the containers is expanded or shrunk uniformly, so that the system adapts to the long-term change trend, resource overload and performance bottleneck are avoided, and the stability and reliability of the system are improved by reducing over-configured resources and reducing the operation cost. Allocating containers for the app based on the micro-service quantity and the data quantity generated by the app in the current time slot and the container computing resources, and allocating the micro-service quantity processed by each container according to the micro-service processing index of the last time slot of the container, according to the method, the saturation degree of each container under an initial allocation strategy is calculated based on the micro-service data volume, the data complexity, the computing resources, the CPU occupancy rate and the storage resources, and the deployment strategy of the edge containers in the same app is adjusted based on the saturation degree of each container, so that unnecessary container migration is reduced, the continuity and consistency of services are kept, the user experience is improved, and the user experience is improved. Therefore, the quick response of the power distribution service is realized.

Description

Technical Field

[0001] The present invention belongs to the fields of power systems and information technology, and particularly relates to a method for rapid response of edge container deployment and power distribution services. Background Art

[0002] In today's power system, efficient management and real-time response of distribution services are crucial to ensuring grid stability and reliability. With the development of smart grid technology, the application scenarios of distribution services are becoming increasingly complex, involving extensive data collection, processing, and analysis. To meet this demand, microservices architectures are widely used in distribution services to achieve service flexibility, scalability, and rapid iteration.

[0003] The microservices architecture breaks down complex applications into a series of independent, independently deployable and scalable small services. These microservices are typically deployed in containers. Containerization technology allows microservices to quickly start, run, and migrate across different computing environments, thereby improving the agility and resilience of the system.

[0004] 1) Existing technologies often overlook the importance of vertical scaling of containers, that is, dynamically adjusting the computing resources (such as CPU or memory) of a single container according to the processing load of the microservice. This limits the adaptability and elasticity of the system when facing different loads.

[0005] 2) Existing container deployment strategies mainly rely on static resource allocation and simple load balancing, which leads to excessive migration of microservices generated by the same app processed in the container, thus affecting service consistency and user experience. Summary of the Invention

[0006] In view of this, the purpose of the present invention is to propose a method for rapid response of edge container deployment and power distribution services, which enables the system to adapt to long-term changing trends through vertical scaling of containers, avoids resource overload and performance bottlenecks, reduces unnecessary container migration, maintains service continuity and consistency, improves user experience, and thus improves system stability and reliability.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] The method is used for edge container deployment and power distribution service rapid response, and is used in the field of power system and information technology, fully considers response time, resource utilization, error rate and data consistency of microservice, constructs microservice processing degree index, expands or shrinks the container based on the processing degree of microservice in the container, predicts the number of microservices by using historical data, expands or shrinks all containers uniformly, allocates containers for the app based on the number of microservices and data volume generated by the app in the current time slot and container computing resources, allocates the number of microservices processed by each container according to the microservice processing index of the container in the last time slot, calculates the saturation degree of each container under the initial allocation strategy based on the microservice data volume, data complexity, computing resource, CPU occupancy and storage resource, and adjusts the deployment strategy of the edge container in the same app based on the saturation degree of each container. The method specifically comprises the following steps:

[0009] S1: microservice processing degree index construction;

[0010] S2: expanding or shrinking the container based on the processing degree of microservice in the container;

[0011] S3: uniformly expanding or shrinking all containers based on the predicted number of microservices;

[0012] S4: initial allocation based on the microservice processing degree index and microservice performance.

[0013] S5: calculating the saturation degree of each container under the initial allocation strategy;

[0014] S6: adjusting the deployment strategy of the edge container in the same app based on the saturation degree of each container;

[0015] Further, in the step S1, the microservice response time, resource utilization, error rate and data consistency are considered to construct the microservice processing degree index of the container. The shorter the microservice response time, the higher the resource utilization and data consistency, and the lower the error rate, the greater the index value.

[0016] Further, in the step S2, the computing resources of each container in the next stage are vertically adjusted based on the microservice processing degree index of the container in the last stage. For the container with large microservice processing degree index, the container can be appropriately shrunk, and the saved computing resources can be used for other parts. For the container with poor performance, the container should be appropriately expanded to ensure the quality requirements of microservice processing.

[0017] Furthermore, in step S3, after adjusting the microservice quality processing index, the number of microservices generated is related to the number of user requests and shows a certain volatility over time. The present invention predicts the amount of microservice data in the next stage based on the number of microservices generated in the historical stage, and further uniformly adjusts the computing resources of all containers according to the changes in the predicted number of microservices.

[0018] Furthermore, in step S4, a container is allocated to the app based on the number of microservices and data volume generated by the app in the current time slot and the container computing resources, and the number of microservices processed by each container is allocated according to the microservice processing indicators of the container in the previous time slot.

[0019] Furthermore, in step S5, the saturation of each container under the initial allocation strategy is calculated based on the microservice data volume, data complexity, computing resources, CPU usage, and storage resources. Based on the saturation of each container, the deployment strategy of edge containers in the same app is adjusted to reduce unnecessary container migrations, maintain service continuity and consistency, and improve user experience, thereby achieving rapid response for power distribution services.

[0020] Furthermore, in step S6, the deployment strategy of the edge containers in the same app is adjusted based on the saturation of each container until the saturation of all containers is lower than the threshold, completing the edge container deployment and power distribution business rapid response.

[0021] Compared with the prior art, the present invention has the following advantages:

[0022] 1) The present invention proposes a large-time-scale container vertical scaling adjustment method based on the processing quality of microservices. It fully considers the processing performance of microservices, constructs a microservice processing quality index, expands or shrinks the container based on the processing quality of the microservices in the container, and further uses historical data to predict the number of microservices. All containers are uniformly expanded or shrunk, so that the system can adapt to long-term changing trends, avoid resource overload and performance bottlenecks, reduce operating costs by reducing over-configured resources, and help improve the stability and reliability of the system.

[0023] 2) The present invention proposes a container deployment strategy based on minimized migration based on container saturation. The saturation of each container under the initial allocation strategy is calculated based on the microservice data volume, data complexity, computing resources, CPU occupancy, and storage resources. The deployment strategy of the edge container in the same app is adjusted based on the saturation of each container, thereby reducing unnecessary container migration, maintaining service continuity and consistency, and improving user experience, thereby achieving rapid response of power distribution services. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] Figure 1This is a schematic diagram of the implementation process of the edge container deployment and power distribution service rapid response method according to an embodiment of the present invention; DETAILED DESCRIPTION

[0025] The present invention is described in detail below with reference to the accompanying drawings and specific embodiments. This embodiment is implemented based on the technical solution of the present invention, and provides a detailed implementation method and specific operation process. It should be noted that, unless there is a conflict, the embodiments of the present invention and the features in the embodiments can be combined with each other.

[0026] Figure 1 This is a flow chart of the edge container deployment and power distribution service rapid response method described in this embodiment, which specifically includes the following steps:

[0027] 1) One: A large-scale container vertical scaling method based on the quality of microservice processing

[0028] This paper proposes a large-scale container vertical scaling method based on the performance of microservice processing. This method considers microservice response time, resource utilization, error rate, and data consistency to construct a microservice processing performance indicator. Secondly, the container capacity is expanded or reduced based on the performance of the microservices within it. Finally, the number of microservices is predicted using historical data, and all containers are scaled up or down uniformly. This method enables the system to adapt to long-term trends, avoids resource overload and performance bottlenecks, reduces operating costs by reducing over-provisioned resources, and helps improve system stability and reliability.

[0029] (1) Constructing microservice processing quality indicators based on microservice performance

[0030] The microservices considered in this invention are generated from K power distribution business apps, which are grouped as APP k Represents the kth power distribution business app. Each power distribution business app generates several microservices in each time slot and offloads them to the edge container for processing. Consider T time slots, the length of each time slot is τ0, that is, T0 consecutive time slots constitute a stage, and the set is Stage i Expressed as In each phase, i.e., in a large time scale, the computing resources of the container remain unchanged, and the container is vertically scaled every other phase. k Generate L in time slot t k (t) microservices, the set is in APP for power distribution business k The lth microservice generated at time slot t. Consider M containers, the set is

[0031] Considering microservice response time, resource utilization, error rate and data consistency, we build microservice quality processing indicators for containers. m The microservice quality processing index Zm(t) is expressed as: Where ERm(t) is the container g in time slot t m Handling error rates in microservices. m (t) and TRT m (t) are time slot t container g m Average response time and total response time of processing microservices, RU m (t) and MRU m (t) are time slot t container g m Processing microservice resource utilization and maximum resource utilization, Cm(t) is the time slot t container g m Handles data consistency across microservices. λ1, λ2, and λ3 are adjustment parameters.

[0032] (2) Expand or shrink the container based on the processing quality of the microservices in the container

[0033] In order to maintain the consistency of microservice processing in the same app and minimize microservice migration, we should strive to ensure that the microservices of the app are handled by as few fixed containers as possible. In this scenario, performing vertical scaling operations on the container, that is, adjusting the computing resources of a single container, can bring significantly better results than re-matching all containers and microservices in each time slot. This method not only optimizes resource utilization, but also improves the overall performance and stability of the system, while ensuring the continuity of services and consistency of user experience. Therefore, the present invention vertically adjusts the computing resources of each container in the next stage based on the microservice quality processing index of the container in the previous stage. The container g in stage i+1 m The computing resources χ′ m The (i+1) adjustment process is expressed as: Where, is the container g in stage i m Average microservice quality processing indicators, They are respectively the container g in stage i m The maximum and minimum values ​​of the microservice quality processing indicators. thAdjust the threshold for performance. Containers with high microservice performance metrics can be scaled down appropriately to free up computing resources for other tasks. Containers with poor performance should be scaled up appropriately to ensure they meet the quality requirements for microservice processing. Vertically scaling containers allows the system to adapt to long-term trends rather than simply responding to short-term fluctuations. This prevents resource overload and performance bottlenecks, reduces over-provisioned resources, lowers operating costs, and helps improve system stability and reliability.

[0034] (3) Predict the number of microservices based on historical data and scale all containers up or down uniformly

[0035] After adjusting the microservice quality processing index, the number of generated microservices is related to the number of user requests and shows a certain volatility over time. Therefore, the present invention predicts the amount of microservice data in the next stage based on the number of microservices generated in the historical stage, and obtains the predicted value of the total number of microservices in stage i+1. Based on step 2, the computing resources of all containers are further adjusted uniformly according to the predicted changes in the number of microservices. m The ultimate computing resources The adjustment process is expressed as in, The total number of microservices generated in phase i.

[0036] 2. Minimizing Migration Container Deployment Strategy Based on Container Saturation

[0037] The present invention proposes a container deployment strategy that minimizes migration based on container saturation. First, containers are allocated to the app based on the number of microservices and data volume generated in the current time slot of the app and the container computing resources, and the number of microservices processed by each container is allocated according to the microservice processing indicators of the previous time slot of the container; then, the saturation of each container under the initial allocation strategy is calculated based on the microservice data volume, data complexity, computing resources, CPU occupancy, and storage resources; finally, the deployment strategy of the edge container in the same app is adjusted based on the saturation of each container. This reduces unnecessary container migration, maintains service continuity and consistency, improves user experience, and thus achieves rapid response of power distribution services. The specific process is as follows:

[0038] (4) Initial allocation is performed based on microservice processing quality indicators and microservice performance.

[0039] Based on the number of microservices and data volume generated by the app in the current time slot and the container computing resources, the app is allocated a container, and the number of microservices processed by each container is allocated according to the microservice processing indicators of the container in the previous time slot. k Generate L in time slot t k (t) microservices, with a container quota of Q allocated to them k (t) is expressed as: Where, is the average computing resource of each container in time slot t, expressed as is the ceiling function, and Microservices The data volume and computational complexity of the power distribution business APP k The more microservices are generated, the larger the amount of microservice data and computational complexity, and the more containers are allocated. π is a tuning parameter used to balance the order of magnitude relationship between the amount of microservice data and the number of containers, preventing the result from being too large or too small. The number of microservices processed by each container is allocated based on the microservice processing indicators of each container in the previous time slot. The kth power distribution business APP k The number of microservices assigned to the qth container Expressed as

[0040] (5) Calculate the saturation of each container under the initial allocation strategy.

[0041] Based on the microservice data volume, data complexity, computing resources, CPU occupancy, and storage resources, the saturation of each container under the initial allocation strategy is calculated. k The saturation of the qth container Expressed as: Where, and They are respectively power distribution business APP k The CPU and memory usage of the qth container in time slot t are calculated. α1, α1, and α1 are the weight coefficients of each indicator and are adjusted according to actual conditions and business needs. Based on the saturation of each container, the deployment strategy of edge containers in the same app is adjusted to reduce unnecessary container migration, maintain service continuity and consistency, and improve user experience, thereby achieving rapid response of power distribution services.

[0042] (6) Adjust the deployment strategy of edge containers in the same app based on the saturation of each container.

[0043] Adjust the deployment strategy of edge containers in the same app based on the saturation of each container. k , calculate whether there is a container whose saturation exceeds the threshold B among its Qk(t) containers th If it exists, migrate the last microservice in the container that exceeds the saturation to the APP k Process the data in the container with the lowest saturation. Continue to adjust until the saturation of all app containers does not exceed threshold B. th , thus completing the edge container deployment and rapid response of power distribution services.

[0044] Although the specific embodiments and drawings of the present invention are disclosed for illustrative purposes, and are intended to facilitate understanding and implementation of the present invention, those skilled in the art will appreciate that various substitutions, changes, and modifications are possible without departing from the spirit and scope of the present invention and the appended claims. Therefore, the present invention should not be limited to the preferred embodiments and the drawings, and the scope of protection claimed by the present invention shall be determined by the scope defined in the claims.

Claims

1. A method for rapid response of edge container deployment and power distribution services, characterized in that: This method fully considers microservice response time, resource utilization, error rate, and data consistency, constructs a microservice processing performance indicator, and scales containers up or down based on the performance of the microservices within them. Historical data is used to predict the number of microservices, allowing all containers to be scaled up or down uniformly. This allows the system to adapt to long-term trends, avoid resource overload and performance bottlenecks, and reduce operational costs by reducing over-provisioned resources, thereby improving system stability and reliability. Containers are allocated to apps based on the number of microservices and data volume generated by the app in the current time slot, as well as the container's computing resources. The number of microservices processed by each container is allocated based on the container's microservice processing indicators from the previous time slot. The saturation of each container under the initial allocation strategy is calculated based on the microservice data volume, data complexity, computing resources, CPU utilization, and storage resources. Based on each container's saturation, the edge container deployment strategy within the same app is adjusted, reducing unnecessary container migrations, maintaining service continuity and consistency, and improving user experience, thereby achieving rapid response for power distribution services.

2. A method for rapid response of edge container deployment and power distribution services according to claim 1, the method specifically comprising the following steps: S1: Construction of microservice processing quality indicators; S2: Expand or shrink the container based on the processing quality of the microservices in the container; S3: All containers are scaled up or down based on the predicted number of microservices. S4: Initial allocation is performed based on microservice processing quality indicators and microservice performance. S5: Calculate the saturation of each container under the initial allocation strategy; S6: Adjust the deployment strategy of edge containers in the same app based on the saturation of each container.

3. The microservice processing quality indicator according to claim 2 is characterized in that: Consider microservice response time, resource utilization, error rate, and data consistency to build a container's microservice quality processing index. The shorter the microservice response time, the higher the resource utilization and data consistency, and the lower the error rate, the larger the index value.

4. The method of expanding or shrinking a container based on the processing quality of microservices in the container according to claim 2, wherein: Based on the container's microservice performance metrics from the previous phase, the computing resources allocated to each container in the next phase are adjusted vertically. Containers with high microservice performance metrics can be scaled down appropriately to free up computing resources for other uses. Containers with poor performance should be scaled up appropriately to ensure they meet the microservice quality requirements.

5. The method of claim 2, wherein all containers are uniformly expanded or reduced in capacity based on the predicted number of microservices. After adjusting the microservice quality processing indicators, the number of generated microservices is related to the number of user requests and shows a certain volatility over time. The present invention predicts the amount of microservice data in the next stage based on the number of microservices generated in the historical stage, and further uniformly adjusts the computing resources of all containers according to the changes in the predicted number of microservices.

6. The method of performing initial allocation based on microservice processing quality index and microservice performance according to claim 2, characterized in that: Containers are allocated to apps based on the number of microservices and data volume generated by the app in the current time slot and the container computing resources. The number of microservices processed by each container is allocated based on the microservice processing indicators of the container in the previous time slot.

7. The method for calculating the saturation of each container under the initial allocation strategy according to claim 2, characterized in that: Calculate the saturation of each container under the initial allocation strategy based on microservice data volume, data complexity, computing resources, CPU utilization, and storage resources. Adjust the deployment strategy of edge containers within the same app based on the saturation of each container to reduce unnecessary container migrations.

8. The method of claim 2, wherein the method comprises: Based on the saturation of each container, the deployment strategy of the edge container in the same app is adjusted until the saturation of all containers is lower than the threshold, completing the rapid response of edge container deployment and power distribution services.

Images