A resource elasticity based power edge computing dedicated container engine

Abstract

The application provides a power edge computing dedicated container engine based on resource elasticity, which allocates logical computing resources to power service applications in units of containers after virtualizing hardware computing resources, and the containers created by the dedicated container engine all have initial quotas, and the container resources are allowed to be fine-tuned within a fine-tuning space on the basis of the initial quotas during the running process of the containers. That is, the power edge computing dedicated container engine provided by the application considers a certain resource elasticity, and when the resource elasticity is considered, the resource quota of the container will be offset on the basis of the initial value, so as to adapt to different construction modes and execution logics of power service applications, and can adapt to the development trend of new power system information sensing and efficient processing requirements.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of container resource management, and particularly relates to a special container engine for electric power edge computing based on resource elasticity. BACKGROUND

[0002] Under the drive of the double carbon target, the new power system presents a situation of high proportion of renewable energy access. With the access of distributed power supply, distributed energy storage and intelligent devices for power distribution and utilization, data presents the trend of massive heterogeneity and continuous growth, which puts forward higher requirements for real-time information sensing, rapid processing and decision analysis in real-time and reliability. The running execution efficiency of electric power business application determines the real-time processing capacity of information, and the running of electric power business application needs the driving force of algorithm resource provision, and the execution efficiency is largely dependent on the allocation strategy of algorithm resource. The container engine is a software tool for allocating logical algorithm resource to electric power business application in units of containers after virtualizing hardware algorithm resource, but the algorithm resource allocation strategy of the general container engine is relatively single, fixed and static, on the one hand, it cannot well adapt to different construction methods and execution logic of electric power business application, on the other hand, it also lacks a coordination mechanism and an elastic mechanism for the allocation of algorithm resource among multiple containers, and it is difficult to adapt to the development trend of information sensing and efficient processing demand of the new power system. SUMMARY

[0003] Therefore, the application aims to solve the problem that the algorithm resource allocation strategy of the general container engine is relatively single, fixed and static.

[0004] In order to solve the above technical problems, the application provides the following technical solutions:

[0005] A special container engine for electric power edge computing based on resource elasticity, which allocates logical algorithm resource to electric power business application in units of containers, wherein the containers created by the special container engine all have an initial quota, and the container resources are allowed to be fine-tuned within a fine-tuning space on the basis of the initial quota during the running process of the container.

[0006] Further, the fine-tuning within the fine-tuning space includes fine-tuning in an increasing direction and fine-tuning in a decreasing direction.

[0007] Further, the special container engine sets a plurality of positive elastic gears and a plurality of reverse elastic gears to realize fine-tuning of the container resources within the fine-tuning space, and different elastic gears have different fine-tuning amounts.

[0008] Further, under the condition of the initial container resource quota, the running time of the i-th container is t i,0 , and after considering resource elasticity, the running time of the container t i is calculated as follows:

[0009]

[0010] In the formula, a cpu,i is the initial CPU resource quota size of the i-th container; a ram,i is the initial memory resource quota size of the i-th container; a bw,i is the initial bandwidth resource quota size of the i-th container; a' cpu,i is the offset CPU resource quota size, a' ram,i is the offset memory resource quota size, a' bw,i is the offset bandwidth resource quota size; β cpu,i is the offset percentage of the CPU resource quota, β ram,i is the offset percentage of the memory resource quota, β bw,i is the offset percentage of the bandwidth resource quota; k1, k2, and k3 are weight coefficients of different resources on the running time of the container.

[0011] Further, the dedicated container engine allocates container resources in units of power microservices, and each microservice has a container providing resources for it.

[0012] Further, the objective function of the dedicated container engine when performing coordinated allocation of container resources is as follows:

[0013]

[0014] In the formula, are respectively the total amount of container CPU resources, memory resources, and bandwidth resources at time t; S cpu,p is the total amount of virtualized CPU resources; S ram,p is the total amount of virtualized memory resources; S bw,p is the total amount of virtualized bandwidth resources.

[0015] Further, the constraint conditions of the dedicated container engine when performing coordinated allocation of container resources are as follows:

[0016]

[0017] In the formula, η cpu , η ram , η bw are respectively the limit value of the proportion of the total amount of container CPU resources, memory resources, and bandwidth resources at time t, and the value range is 0 to 1; β cpu,min , β cpu,max , β ram,min , β ram,max , β bw,min , β bw,maxThese represent the minimum and maximum allowed percentage offsets for the allocation of container CPU resources, memory resources, and bandwidth resources, respectively; T i,串 T i,并 T represents the delay magnitudes of the i-th series and parallel components, respectively; max The delay value for power business applications; t 串，1 t 并，1 These are the container runtimes of the first microservice belonging to the i-th serial and parallel links, respectively.

[0018] Furthermore, the decision variables for dedicated container engines when coordinating and allocating container resources are as follows:

[0019] The offset CPU resource quota size a' cpu,i The size of the memory resource quota after offset, a' ram,i And the size of the bandwidth resource quota after offset, a' bw,i .

[0020] Furthermore, when a dedicated container engine coordinates and allocates container resources, it specifically includes the following steps:

[0021] Step 1: Start the container engine to virtualize the hardware computing resources and obtain the total amount of CPU, memory, and bandwidth resources;

[0022] Step 2: Set the initial container CPU, memory, and bandwidth resource quotas and the forward and reverse elasticity levels;

[0023] Step 3: Set the counter value n = 1, and set the iteration number N;

[0024] Step 4: Update the container's CPU, memory, and bandwidth resource quotas after the nth generation offset;

[0025] Step 5: Calculate the runtime of the nth generation container;

[0026] Step Six: Calculate the objective function value using a container resource coordination and allocation model for power business applications;

[0027] Step 7: Verify the constraints of the container resource coordination and allocation model for power business applications.

[0028] Step 8: Determine if n is greater than N. If the condition is met, proceed to step 9; otherwise, set n = n + 1 and proceed to step 4.

[0029] Step 9: Obtain the Nth generation container resource coordination and allocation scheme;

[0030] Step 10: Output the optimal container resource allocation scheme;

[0031] Step 11: End the process.

[0032] In summary, this invention provides a dedicated container engine for power edge computing based on resource elasticity. This dedicated container engine virtualizes hardware computing resources and allocates logical computing resources to power business applications on a container basis. Each container created by the dedicated container engine has an initial quota, and during operation, the container resources are allowed to be fine-tuned within a certain adjustment space based on the initial quota. In other words, the dedicated container engine for power edge computing provided by this invention considers a certain degree of resource elasticity. When considering resource elasticity, the container's resource quota will shift from its initial value, making it easier to adapt to different construction methods and execution logics of power business applications, and adapting to the evolving needs of information perception and efficient processing in new power systems. Attached Figure Description

[0033] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0034] Figure 1 A schematic diagram illustrating the operation of a container engine in a computer system provided in an embodiment of the present invention;

[0035] Figure 2 This is a schematic diagram illustrating the positive and negative elasticity of container resources provided in an embodiment of the present invention.

[0036] Figure 3 This is a schematic diagram illustrating the resource allocation method of the dedicated container engine for power edge computing provided in an embodiment of the present invention;

[0037] Figure 4 A flowchart illustrating a resource allocation method for dedicated power containers based on resource elasticity, provided in an embodiment of the present invention. Detailed Implementation

[0038] To make the objectives, features, and advantages of this invention more apparent and understandable, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the embodiments described below are only some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0039] Container engines are software tools that virtualize hardware computing resources and allocate logical computing resources to power business applications on a container basis. However, general-purpose container engines have relatively simple, rigid, and static computing resource allocation strategies. On the one hand, they fail to adapt well to the different construction methods and execution logic of power business applications; on the other hand, they lack coordination and elasticity mechanisms for computing resource allocation among multiple containers, making it difficult to adapt to the evolving needs of information perception and efficient processing in new power systems. Therefore, considering the construction requirements of power business applications and the efficient utilization of computing resources, this invention proposes a dedicated container engine for power edge computing based on resource elasticity.

[0040] The following is a detailed description of an embodiment of the present invention: a dedicated container engine for power edge computing based on resource elasticity.

[0041] This embodiment provides a dedicated container engine for power edge computing based on resource elasticity. The dedicated container engine allocates logical computing resources to power business applications on a container basis. Each container created by the dedicated container engine has an initial quota, and the container resources are allowed to be fine-tuned within the fine-tuning space based on the initial quota during the operation of the container.

[0042] A container engine is a software tool that virtualizes hardware computing resources and allocates logical computing resources to power business applications in units of containers. For a computer system, the container engine resides between the operating system layer and the application layer, such as... Figure 1 As shown, the container engine provides the necessary environment and resources for the operation of power business applications by driving and creating containers. It compiles the power business applications so that the operating system and underlying hardware can recognize them, ultimately enabling program execution. Taking a Java-formatted power business application as an example, the Java file is compiled into a class file by the container engine through the JVM, then passed to the operating system for DDL linking to obtain an object file. Finally, the system compiles it into an executable file (exe), which is then read and executed by the hardware CPU.

[0043] For power business applications located at the application layer, they can only see the size of the container resources allocated to them, while hardware computing resources are invisible. This is the result of the container engine virtualizing hardware computing resources, which can be quantitatively characterized mathematically by the following formula:

[0044]

[0045]

[0046]

[0047] In the formula: s cpu,is represents the size of the i-th hardware CPU resource; ram,i Let s be the size of the i-th hardware memory resource; bw,i S represents the size of the i-th hardware bandwidth resource. cpu,p S represents the total amount of CPU resources after virtualization. ram,p S represents the total amount of virtualized memory resources. bw,p This represents the total bandwidth resources after virtualization.

[0048] A container engine is a software tool that allocates logical computing resources to power business applications on a container basis. Therefore, given a fixed total amount of virtualized CPU, memory, and bandwidth resources, a portion of these resources can be allocated as container resources. Assuming that m containers are created using the container engine, each with a certain initial allocation of CPU, memory, and bandwidth resources, then these m containers must satisfy the following formula:

[0049]

[0050]

[0051]

[0052] In the formula: a cpu,i Let a be the initial CPU resource quota for the i-th container; ram,i Let a be the initial memory resource allocation size for the i-th container; bw,i Let be the initial bandwidth resource quota for the i-th container.

[0053] Once each container has its initial CPU, memory, and bandwidth resource quotas set, the container can begin running. During operation, container resources are allowed to be fine-tuned around the initial quota size; this fine-tuning space is called resource elasticity. When container resources are fine-tuned in an increasing direction from the initial quota size, it is called positive resource elasticity; when fine-tuned in a decreasing direction, it is called negative resource elasticity. Because the magnitude of these increases and decreases must be limited within a certain range, positive and negative elasticity levels are set, such as... Figure 2 As shown.

[0054] When considering resource elasticity, the resource quota of a container will be offset from its initial value. This offset will affect the runtime of the container. Let the runtime of the i-th container be t under the initial container resource quota. i,0 Considering resource elasticity, the container runtime t i The formula is as follows:

[0055] a' cpu,i =(1+β) cpu,i )a cpu,i(7)

[0056] a' ram,i =(1+β) ram,i )a ram,i (8)

[0057] a' bw,i =(1+β) bw,i )a bw,i (9)

[0058]

[0059] k1+k2+k3=1 (11)

[0060] In the formula: a' cpu,i The size of the CPU resource quota after offset; a' ram,i This represents the size of the memory resource quota after the offset; a' bw,i β represents the amount of bandwidth resource quota after offset. cpu,i This represents the percentage offset from the CPU resource quota, and can take values ​​such as +5%, 10%, +15%, -5%, -10%, -15%, etc.; β ram,i This represents the percentage offset of the memory resource quota, and the value is determined similarly; β bw,i The value is the percentage offset of the bandwidth resource quota, and the values ​​are taken in the same way; k1, k2, and k3 are the weighting coefficients of the impact of different resources on container runtime.

[0061] like Figure 3 As shown, the container resource allocation method of the general container engine is based on the power monolithic application, while the power edge computing dedicated container engine proposed in this invention is based on power microservices. Both power monolithic applications and power microservices are ways to build power business applications. The difference is that a power monolithic application packages all functions into one application, while a power microservice is equivalent to breaking down the various functions of this monolithic application into multiple microservices. Multiple microservices can be executed serially or in parallel, effectively improving computing speed.

[0062] The power edge computing dedicated container engine proposed in this invention allocates container resources on a unit basis for power microservices. Each microservice has a container to provide resources for it, and multiple containers can coordinate resources to better utilize computing resources efficiently. The container resource coordination and allocation model of the power edge computing dedicated container engine constructed in this invention is shown in Equation (12). The objective function of this model is to minimize the maximum value of the total proportion of CPU, memory, and bandwidth resources allocated at any time. The constraints of this model include the total allocation of container resources, the allocation offset of container resources, and the latency constraint of power business applications. The decision variables of this model are the allocation of container CPU, memory, and bandwidth resources after considering the elastic offset of resources.

[0063]

[0064]

[0065] Decision variable: a' cpu,i ,a' ram,i ,a' bw,i (12)

[0066] In the formula: Let be the total allocated CPU resources, memory resources, and bandwidth resources of the container at time t, respectively, and let be the objective function. To minimize the maximum percentage of the total allocation of these three resources; η cpu η ram η bw β represents the total allocation limit of container CPU resources, memory resources, and bandwidth resources at time t, respectively, with values ​​ranging from 0 to 1; cpu,min β cpu,max β ram,min β ram,max β bw,min β bw,max These represent the minimum and maximum allowed percentage offsets for the allocation of container CPU resources, memory resources, and bandwidth resources, respectively; T i,串 T i,并 T represents the delay magnitudes of the i-th series and parallel components, respectively; max The delay value for power business applications; t 串，1 t 并，1 These are the container runtimes of the first microservice belonging to the i-th serial and parallel links, respectively.

[0067] Based on the above settings, the process of the resource allocation method for dedicated power containers based on resource elasticity in this embodiment is as follows (e.g.) Figure 4 As shown):

[0068] Step 1: Start the container engine to virtualize the hardware computing resources and obtain the total amount of CPU, memory, and bandwidth resources;

[0069] Step 2: Set the initial container CPU, memory, and bandwidth resource quotas and the forward and reverse elasticity levels;

[0070] Step 3: Set the counter value n = 1, and set the iteration number N;

[0071] Step 4: Update the container's CPU, memory, and bandwidth resource quotas after the nth generation offset;

[0072] Step 5: Calculate the runtime of the nth generation container;

[0073] Step Six: Calculate the objective function value using a container resource coordination and allocation model for power business applications;

[0074] Step 7: Verify the constraints of the container resource coordination and allocation model for power business applications.

[0075] Step 8: Determine if n is greater than N. If the condition is met, proceed to step 9; otherwise, set n = n + 1 and proceed to step 4.

[0076] Step 9: Obtain the Nth generation container resource coordination and allocation scheme;

[0077] Step 10: Output the optimal container resource allocation scheme;

[0078] Step 11: End the process.

[0079] This embodiment provides a dedicated container engine for power edge computing based on resource elasticity. This dedicated container engine virtualizes hardware computing resources and allocates logical computing resources to power business applications on a container basis. Each container created by the dedicated container engine has an initial quota, and the container resources are allowed to be fine-tuned within a certain range based on the initial quota during operation. In other words, the dedicated container engine for power edge computing provided by this invention considers a certain degree of resource elasticity. When considering resource elasticity, the container's resource quota will shift from its initial value, making it easier to adapt to different construction methods and execution logic of power business applications, and adapting to the evolving needs of information perception and efficient processing in new power systems.

[0080] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A dedicated container engine for power edge computing based on resource elasticity, characterized in that, The dedicated container engine allocates logical computing resources to power business applications on a container basis. Each container created by the dedicated container engine has an initial quota, and the container resources are allowed to be fine-tuned within the fine-tuning space based on the initial quota during the operation of the container. Assuming the initial container resource quota is given, the first... The runtime of each container is Then, considering resource elasticity, the container runtime... The specific calculation formula is as follows: ； In the formula, For the first The initial CPU resource quota for each container; For the first The initial memory resource allocation size for each container; For the first The initial bandwidth resource quota for each container; The size of the CPU resource quota after offset. This represents the size of the memory resource quota after the offset. This represents the size of the bandwidth resource quota after the offset. This represents the percentage offset from the CPU resource quota. The percentage offset of the memory resource quota. This represents the percentage offset of the bandwidth resource quota. , , Weighting coefficients for the impact of different resources on container runtime; The dedicated container engine allocates container resources on a per-power microservice basis, with each microservice having a container to provide resources for it. The objective function of the dedicated container engine when coordinating and allocating container resources is as follows: ； In the formula, , , They are respectively The total amount of CPU resources, memory resources, and bandwidth resources allocated to the container at any given time; This represents the total amount of CPU resources after virtualization. This represents the total amount of memory resources after virtualization. This represents the total bandwidth resources after virtualization.

2. The dedicated container engine for power edge computing based on resource elasticity according to claim 1, characterized in that, Fine-tuning within the fine-tuning space includes fine-tuning in an increasing direction and fine-tuning in a decreasing direction.

3. The dedicated container engine for power edge computing based on resource elasticity according to claim 2, characterized in that, The dedicated container engine is equipped with several positive elasticity levels and several negative elasticity levels to enable container resources to be fine-tuned within the fine-tuning space. Different elasticity levels have different fine-tuning limits.

4. The dedicated container engine for power edge computing based on resource elasticity according to claim 1, characterized in that, The specific constraints imposed by the dedicated container engine when coordinating and allocating container resources are as follows: ； In the formula, , , They are respectively The limit value for the total allocation of container CPU resources, memory resources, and bandwidth resources at any given time, with a value range from 0 to 1; , , , , , These are the minimum and maximum allowed percentage values ​​for the allocation offset of container CPU resources, memory resources, and bandwidth resources, respectively. , The first The delay of each series and parallel component; The delay magnitude for power business applications; , They belong to the first The container runtime of the first microservice in a series and parallel process.

5. The dedicated container engine for power edge computing based on resource elasticity according to claim 4, characterized in that, The specific decision variables for the dedicated container engine when coordinating and allocating container resources are as follows: CPU resource quota size after offset The size of the memory resource quota after offset and the size of the bandwidth resource quota after offset. .

6. The dedicated container engine for power edge computing based on resource elasticity according to claim 1, characterized in that, The dedicated container engine performs the following steps when coordinating and allocating container resources: Step 1: Start the container engine to virtualize the hardware computing resources and obtain the total amount of CPU, memory, and bandwidth resources; Step 2: Set the initial container CPU, memory, and bandwidth resource quotas and the forward and reverse elasticity levels; Step 3: Set the counter value n=1, and set the iteration number N; Step 4: Update the container's CPU, memory, and bandwidth resource quotas after the nth generation offset; Step 5: Calculate the runtime of the nth generation container; Step Six: Calculate the objective function value using a container resource coordination and allocation model for power business applications; Step 7: Verify the constraints of the container resource coordination and allocation model for power business applications. Step 8: Determine if n is greater than N. If the condition is met, proceed to step 9; otherwise, set n = n + 1 and proceed to step 4. Step 9: Obtain the Nth generation container resource coordination and allocation scheme; Step 10: Output the optimal container resource allocation scheme; Step 11: End the process.

Images