CN107315643A - A kind of container resource regulating method - Google Patents

Abstract

The present invention provides a container resource scheduling method, which includes the following steps: S1: scoring nodes; S2: selecting the node with the highest score. Through the above settings, under the premise of resource permitting, the bandwidth of network transmission mirroring and the time required for downloading mirroring are saved, the operating efficiency of the system is improved, and the service quality of the system is guaranteed.

Description

A container resource scheduling method

technical field

The invention relates to the technical field of cloud computing, in particular to a container resource scheduling method.

Background technique

With the rapid development of container technology in cloud computing and the increasing popularity of container technology, more and more enterprises are involved in container technology, and the use of containers in cloud computing infrastructure and services is increasing year by year. A new round of revolution. At present, the following methods are mainly used for the allocation of container resources: one allocation method is to allocate according to the order in which instructions are received, this method does not consider the size of resources required by different containers, which is easy to cause waste of resources; The first allocation method is to allocate according to the different execution functions of the containers. In this way, although the configuration of resources is realized to a certain extent, the containers are not fixed and the functions they perform are also different. Therefore, resource allocation is also different. There will be a waste problem.

Contents of the invention

In order to solve the above problems, a container resource scheduling method is provided, which can allocate corresponding resources for different containers, and greatly improve resource utilization.

A kind of specific embodiment of the present invention adopts following technical scheme:

A container resource scheduling method, said method comprising the following steps:

S1: Score the nodes;

S2: Select the node with the highest score.

Further, the method also includes:

S0: Determine the size of the node's own resources, and remove nodes whose resources are less than the set value.

Further, the size of the set value is 100M.

Further, the specific implementation process of S1 is:

S11: set the mirror image upper limit max;

S12: Obtain the image value X1 required by the application;

S13: Obtain the self-image value X2 of a certain node;

S14: Determine whether X1 is greater than X2, if yes, then the node score = 0; if not, go to S15;

S15: Determine whether X2 is greater than max, if yes, then the node score = 100; if not, go to S16;

S16: Calculate the score of the node according to the formula score=100*(X2-X1)/(max-X1);

S17: Repeat steps S13-S16 to obtain the score of each node.

Further, the method also includes:

S3: Determine whether the number of nodes with the highest score is greater than 1, if yes, go to S4, if not, select the node;

S4: The system randomly selects one of the nodes for resource allocation.

Further, the method also includes:

S3: Determine whether the number of nodes with the highest score is greater than 1, if yes, go to S4, if not, select the node;

S4: Determine whether the highest score is equal to 100 points, if yes, go to S5, if not, the system randomly selects one of the nodes for resource allocation;

S5: Determine whether the mirror value of the node with the highest score is the same, if so, the system randomly selects one of the nodes for resource allocation; if not, go to S6;

S6: Select the node with the highest mirror image value.

The beneficial effect of a kind of specific embodiment of the present invention is:

1. Score the image size of the node. The larger the image, the higher the score of the host, and select the node with the highest score and allocate resources. Through the above settings, under the premise of resource permitting, the bandwidth of network transmission mirroring and the time required for downloading mirroring are saved, the operating efficiency of the system is improved, and the service quality of the system is guaranteed.

2. Before scoring, remove nodes with too small resources in advance, reducing the number of nodes that need to be scored, and improving the efficiency of the entire score.

3. By segmenting the scoring steps, the number of nodes to be calculated is further reduced, and the efficiency of scoring calculations is maximized.

4. Provides two methods for subsequent node selection, one is direct random selection, which simplifies the setting process; the other is segmented judgment, which finally selects the best choice among the highest scores, which improves the quality of selection.

Description of drawings

Fig. 1 is a method flowchart of an embodiment of the present invention;

Fig. 2 is a flowchart of a method in another embodiment of the present invention.

detailed description

In order to enable those skilled in the art to better understand and realize the present invention, the present invention will be described in detail below through specific embodiments.

Explanation of several terms:

Containers are the isolation and division of host computing resources (CPU, memory, disk or network, etc.), such as Docker.

An image is an image similar to an operating system, including a complete operating system. The operating system, basic components, and applications can all be packaged in an image. Containers rely on images to run.

Applications are programs that run in "containers".

A node refers to a container that contains several images.

Example 1

As shown in Figure 1, a specific embodiment is provided, which provides a container resource scheduling method, which includes the following steps:

S1: Score the nodes, the specific implementation process is:

S11: Set the mirror image upper limit max, in practical applications, you can set max to 2G, which is 2048M;

S12: Obtain the image value X1 required by the application;

S13: Obtain the self-image value X2 of a certain node;

S14: Determine whether X1 is greater than X2, if yes, then the node score = 0; if not, go to S15;

S15: Determine whether X2 is greater than max, if yes, then the node score = 100; if not, go to S16;

S16: Calculate the score of the node according to the formula score=100*(X2-X1)/(max-X1);

S17: Repeat steps S13-S16 to obtain the score of each node.

S2: Select the node with the highest score.

S3: Determine whether the number of nodes with the highest score is greater than 1, if yes, go to S4, if not, select the node.

S4: The system randomly selects one of the nodes for resource allocation.

Example 2

As shown in Figure 2, a specific embodiment is provided, which provides a container resource scheduling method, which includes the following steps:

S1: Score the nodes, the specific implementation process is:

S11: Set the mirror image upper limit value max In practical applications, you can set max to 2G, which is 2048M;

S12: Obtain the image value X1 required by the application;

S13: Obtain the self-image value X2 of a certain node;

S14: Determine whether X1 is greater than X2, if yes, then the node score = 0; if not, go to S15;

S15: Determine whether X2 is greater than max, if yes, then the node score = 100; if not, go to S16;

S16: Calculate the score of the node according to the formula score=100*(X2-X1)/(max-X1);

S17: Repeat steps S13-S16 to obtain the score of each node.

S2: Select the node with the highest score.

S3: Determine whether the number of nodes with the highest score is greater than 1, if yes, go to S4, if not, select the node.

S4: Determine whether the highest score is equal to 100 points, if yes, go to S5, if not, the system randomly selects one of the nodes for resource allocation.

S5: Determine whether the mirror value of the node with the highest score is the same, if yes, the system randomly selects one of the nodes for resource allocation; if not, go to S6.

S6: Select the node with the highest mirror image value.

Example 3

On the basis of Embodiment 1, the following steps are added before step S1:

S0: Determine the size of the node's own resources, and remove nodes whose resources are less than the set value.

By setting step S0, nodes that do not meet the requirements can be removed in advance before scoring, reducing the number of nodes that need to be scored, and improving efficiency. In the actual application, an example is given, and the size of the set value is set to 100M.

Example 4

On the basis of embodiment 2, the following steps are added before step S1:

S0: Determine the size of the node's own resources, and remove nodes whose resources are less than the set value.

By setting step S0, nodes that do not meet the requirements can be removed in advance before scoring, reducing the number of nodes that need to be scored, and improving efficiency. In the actual application, an example is given, and the size of the set value is set to 100M.

Although the specification, drawings and embodiments have described the invention in detail, those skilled in the art should understand that the invention can still be modified or replaced in an equivalent manner; and everything that does not depart from the spirit and scope of the invention The technical solutions and their improvements are all included in the scope of protection of the invention patent.

Claims (6)

1. A container resource scheduling method, characterized in that said method comprises the following steps: S1: Score the nodes; S2: Select the node with the highest score. 2. A container resource scheduling method according to claim 1, wherein said method further comprises: S0: Determine the size of the node's own resources, and remove nodes whose resources are less than the set value. 3. A container resource scheduling method according to claim 2, wherein the size of the set value is 100M. 4. A container resource scheduling method according to claim 1 or 2, wherein the specific implementation process of S1 is: S11: set the mirror image upper limit max; S12: Obtain the image value X1 required by the application; S13: Obtain the self-image value X2 of a certain node; S14: Determine whether X1 is greater than X2, if yes, then the node score = 0; if not, go to S15; S15: Determine whether X2 is greater than max, if yes, then the node score = 100; if not, go to S16; S16: Calculate the score of the node according to the formula score=100*(X2-X1)/(max-X1); S17: Repeat steps S13-S16 to obtain the score of each node. 5. A container resource scheduling method according to claim 4, characterized in that said method further comprises: S3: Determine whether the number of nodes with the highest score is greater than 1, if yes, go to S4, if not, select the node; S4: The system randomly selects one of the nodes for resource allocation. 6. A container resource scheduling method according to claim 4, characterized in that said method further comprises: S3: Determine whether the number of nodes with the highest score is greater than 1, if yes, go to S4, if not, select the node; S4: Determine whether the highest score is equal to 100 points, if so, go to S5, if not, the system randomly selects one of the nodes for resource allocation; S5: Determine whether the mirror value of the node with the highest score is the same, if so, the system randomly selects one of the nodes for resource allocation; if not, go to S6; S6: Select the node with the highest mirror image value.