JP2018190355A - Resource management method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To inhibit resource shortage for an application of high importance, and improve utilization efficiency of resources, in a server that operates multiple applications.SOLUTION: A server 1 creates APL groups on the basis of a degree of importance of an application (container) activated by the server 1. Further, the server 1 preferentially allocates resources in descending order of the degrees of importance of the APL groups. Each container in an APL group shares the resources allocated to the APL group. Subsequently, the server 1 monitors an amount of resource utilization in the APL groups respectively, and when resource shortage occurs in any of the APL groups, a portion of the allocated resource amount in an APL group with surplus resources is allocated to the APL group in which the resource shortage has occurred.SELECTED DRAWING: Figure 1

Description

  The present invention relates to a resource management method.

  There is a technique for dynamically allocating a resource amount reported from an application when computing resources (resources) are allocated to various applications operating on a server (see Non-Patent Document 1). According to such a technique, server resources can be effectively used.

Benjamin Hindman, et al., "Mesos: A Platform for Fine-Grained Resource Sharing in the Data Center", NSDI. Vol. 11. No. 2011. 2011.

  However, in the above technique, as a result of allocating resources to each application, the resources are depleted, and there is a possibility that resources cannot be allocated to a highly important application. For example, if resources are allocated to a server that needs to run multiple applications with limited resources, such as a server used in edge computing architecture, the less important application uses up resources. This may cause a shortage of resources in highly important applications.

  In the above technique, each application requests the maximum value of the amount of used resources. Therefore, if resources are allocated as requested, the resources may not be used up in each application. For this reason, the utilization efficiency of resources may be lowered.

  Accordingly, an object of the present invention is to solve the above-described problems, to prevent resource shortage of a highly important application in a server that operates a plurality of applications, and to improve resource utilization efficiency.

  In order to solve the above-described problem, the present invention provides a resource management method executed in a resource management apparatus that manages resource allocation amounts for a plurality of applications operating on a server, and accepts the importance of each of the applications. A grouping step for grouping the applications into two or more groups according to the received importance level, a resource allocation step for preferentially allocating resources from the group of high importance applications, A monitoring step for monitoring the resource usage of the application; and, as a result of the monitoring, if a resource usage rate that is the resource usage with respect to the allocated resource amount is detected to be a predetermined value or more, a surplus resource is One group at least a part of the resource allocation amount of the resource utilization, characterized in that the containing, and resource reallocation step of allocating said predetermined value or more groups.

  ADVANTAGE OF THE INVENTION According to this invention, in the server which operate | moves a some application, the resource shortage of a highly important application can be prevented and the utilization efficiency of a resource can be improved.

FIG. 1 is a block diagram illustrating a configuration example of a server according to the present embodiment. FIG. 2 is a diagram illustrating an example of the resource management information of FIG. FIG. 3 is a flowchart showing an outline of the processing procedure of the server of FIG. FIG. 4 is a flowchart showing details of the processing in S1 of FIG. FIG. 5 is a flowchart showing details of the processing of S2 and S3 of FIG. FIG. 6 is a flowchart showing details of the process of S4 of FIG. FIG. 7 is a diagram illustrating a computer that executes a resource management program.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In addition, this invention is not limited to embodiment shown below.

  First, the outline of the server 1 equipped with the resource management apparatus (resource management unit 13) of this embodiment will be described with reference to FIG.

  The server 1 is, for example, a server used as an edge server in the edge computing architecture, and operates a plurality of applications. The server 1 operates, for example, an application having a relatively high importance such as a management application and an application having a relatively low importance such as an application developed by a 3rd party.

  In addition, it is assumed that each application operated by the server 1 has resources (CPU, memory, etc.) that can be used by each application isolated by, for example, container virtualization technology. For example, the server 1 includes a container A for application A, a container B for application B, and a container C for application C.

  The server 1 groups containers A, B, and C based on the importance of each container application. For example, the server 1 groups containers A and B of relatively high importance applications such as management applications into the APL group A, and sets a container C of relatively low importance applications such as applications developed by the 3rd party. Group into APL group B.

  Further, the server 1 preferentially allocates resources from the APL group of the application (container) having a high priority. For example, the server 1 allocates the allocated resource amount of the application requested by the user as it is to the APL group A of the relatively high importance application, and the user to the APL group B of the relatively low importance application. Allocate a resource that is smaller (for example, about 80%) than the allocated resource amount of the application requested by.

  Thereafter, the server 1 monitors the resource usage of each APL group, and if the resource usage rate is high in any APL group, the server 1 allocates surplus resources of other APL groups. For example, if the resource utilization rate of the APL group A is high and the APL group B has surplus resources, the server 1 allocates a part of the surplus resources of the APL group B. Thereby, the server 1 can prevent the resource shortage of the container of the application with high importance, and can improve the utilization efficiency of the resource.

  The configuration example of the server 1 will be described with reference to FIG. As illustrated in FIG. 1, the server 1 includes, for example, containers A, B, and C, hardware 11, a container control unit 12, and a resource management unit 13.

  The container is a dedicated area of the application where the application runs on the server 1. That is, resources of each application on the server 1 are isolated by the container virtualization technology. For example, the container uses resources within the range of the allocated resource amount set by the container control unit 12. Thereby, the server 1 can restrict | limit the resource usage-amount for every application. Communication between each application is performed via an API (Application Programming Interface). Therefore, the server 1 can easily control the frequency of communication between applications. As a result, the server 1 can prevent the system from being stopped due to a part of applications performing a large amount of communication with important applications.

  In addition, after containers are grouped into APL groups, the APL group determines that the total usage amount of each container resource in the APL group is within the allocated resource amount range of the APL group set by the container control unit 12. Use resources.

  For example, the APL group created by the server 1 is an APL group A having a relatively high importance such as a management application, and an APL group B having a relatively low importance such as an application developed by a 3rd party. Consider the case.

  In this case, even when some applications (containers) in the APL group B use a large amount of resources, the APL group B uses the resources so that the resource usage amount of the APL group B as a whole is within the allocated resource amount. . For example, the server 1 performs overcommit on the resources required by the application of the APL group B so that the resource usage amount as the whole APL group B is within the allocated resource amount. As a result, even when some applications (containers) in the APL group B use a large amount of resources, the server 1 can prevent an application having a relatively high importance from being affected.

  The hardware 11 is a CPU, a memory, a hard disk, a communication interface, an input / output interface, and the like included in the server 1.

  The container control unit 12 performs various controls on the container. For example, the container control unit 12 refers to the resource management information 131 (described later in detail) of the resource management unit 13 and sets the allocated resource amount of the container for each container (application). After each container is grouped into an APL group, the container control unit 12 refers to the resource management information 131 and sets the allocated resource amount of the APL group for each APL group.

  The resource management unit 13 groups containers into APL groups based on the importance of each container, and manages the resource allocation amount of each APL group. The resource management unit 13 includes resource management information 131, a resource allocation unit 132, a resource monitoring unit 133, and a resource reallocation unit 134.

  The resource management information 131 is information indicating, for each container, the allocated resource amount of the container, the APL group to which the container belongs, and the allocated resource amount of the APL group.

  For example, the resource management information 131 shown in FIG. 2 includes the container identification information, the allocated resource amount of the container, the importance of the container (application), the APL group to which the container belongs, and the allocated resource of the APL group. This is information that associates quantities. The resource management information 131 is stored in a storage unit (not shown) included in the server 1.

  The resource allocation unit 132 groups containers into APL groups based on the importance of each container, and allocates resources preferentially from a group of applications with high importance (APL group). Specifically, first, the resource allocation unit 132 receives input of the importance level of each application and the resource amount (requested resource amount) to be allocated to the application of the importance level from the user of the server 1 or the like. Next, the resource allocation unit 132 groups application containers into two or more APL groups based on the importance of each input application.

  For example, the resource allocation unit 132 groups containers of applications having the same or similar importance of applications and different types of resources used by the application into the same APL group.

  For example, the resource allocation unit 132 groups application containers having a large memory usage amount and applications having a high CPU usage rate into the same APL group among application containers having the same or similar importance. Thereby, since the applications of the containers belonging to the same APL group do not compete for the resources allocated to the APL group, efficient resource allocation can be performed.

  Further, the resource allocation unit 132 allocates a requested resource amount with priority from an APL group having high importance.

  For example, the resource allocation unit 132 allocates a resource amount according to the requested resource amount from the user to the highly important APL group, but about 80% of the requested resource amount from the user to the less important APL group. Allocate the amount of resources.

  That is, the resource allocation unit 132 increases the ratio of the actual allocated resource amount to the APL group with respect to the user's requested resource amount for the higher importance APL group, and the user request for the APL group with the lower importance level. The ratio of the resource amount actually allocated to the APL group with respect to the resource amount is lowered.

  When the resource allocation unit 132 determines the allocated resource amount of the container of each APL group as described above, it records the determined allocated resource amount in the resource management information 131 (see FIG. 2).

  The resource monitoring unit 133 monitors the resource usage of the containers of each APL group, and determines whether or not resource reassignment (redistribution) is necessary for each APL group.

  For example, the resource monitoring unit 133 monitors the resource usage of each APL group, and as a result, the resource usage rate (the ratio of the resource usage to the allocated resource) in the APL group A having a high priority is a predetermined value (for example, 80% ) Consider the case above. In this case, the resource monitoring unit 133 determines that it is necessary to reallocate (redistribute) resources if there is a surplus resource in the APL group B with low priority. Then, the resource monitoring unit 133 instructs the resource reallocation unit 134 to allocate a part of the allocated resource amount in the APL group B (for example, 20% of the allocated resource amount of the APL group B) to the APL group A. .

  The resource reallocation unit 134 reallocates resources between APL groups in accordance with instructions from the resource monitoring unit 133. For example, in accordance with an instruction from the resource monitoring unit 133, the resource reallocation unit 134 determines a part of the allocated resource amount of the APL group B having a low priority and a surplus resource that has a high priority and a resource utilization rate equal to or higher than a predetermined value. Assign to APL group A. Specifically, the resource reallocation unit 134 reduces a part of the allocated resource amount of the APL group B in the resource management information 131 and allocates the reduced resource amount to the APL group A.

  According to the server 1 described above, it is possible to prevent resource shortage in an application with high importance (priority), and to improve resource utilization efficiency.

  In the above example, the case where the server 1 allocates the surplus resource of the APL group B with low priority to the APL group A with high priority (performs resource movement) is described, but the present invention is not limited to this. For example, the server 1 may assign the surplus resource of the APL group A having the higher priority to the APL group B when the resource usage rate of the APL group B having the lower priority becomes a predetermined value or more. In this case, a condition is set such that the surplus resource of the APL group A having a high priority is 30% or more of the allocated resource amount of the APL group A. By doing so, it is possible to prevent a shortage of resources of the APL group having a low priority while securing surplus resources of the APL group having a high priority.

  Next, the processing procedure of the server 1 will be described. First, the outline of the processing procedure of the server 1 will be described with reference to FIG.

  First, the resource allocation unit 132 of the server 1 groups applications (containers) (S1). As a result, a plurality of APL groups are created. After that, the resource monitoring unit 133 monitors the resource usage of each APL group (S2), and when it is determined that resource reassignment is necessary (Yes in S3), the resource reassignment unit 134 Are reassigned (S4). On the other hand, when the resource monitoring unit 133 determines that resource reassignment is unnecessary (No in S3), the process returns to S2.

  Next, the process of S1 in FIG. 3 will be described in detail with reference to FIG. 4 with a specific example.

  For example, the resource allocation unit 132 of the server 1 creates a plurality of APL groups (S11). For example, the resource allocating unit 132 creates three APL groups of application importance: high, application importance: medium, and application importance: low. After that, the resource allocation unit 132 receives an input of the application (container) to be activated, the importance of the application, and the requested resource amount from the user of the server 1 (S12).

  Next, the resource allocation unit 132 allocates a resource amount obtained by multiplying the requested resource amount received in S12 by an arbitrary magnification to each APL group and activates it (S13). For example, the resource allocation unit 132 identifies the APL group corresponding to the importance received in S12. Then, the resource allocating unit 132 determines the requested resource amount received in S12 for the allocated resource amount of the APL group * an arbitrary multiplication rate (for example, “1.0” if the importance of the application is high, : If it is medium, increase it by “0.8”. After that, the resource allocation unit 132 activates the application input in S12 as the APL group of the importance level input in S12. The arbitrary multiplication rate is set to a higher value as the importance of the application input in S12 is higher, and to a lower value as the importance of the application is reduced. The value of this arbitrary multiplication rate can be set by the user.

  By doing in this way, the server 1 can group each application into an APL group according to the importance of each application (container). Further, the server 1 can perform resource allocation preferentially from the APL group of the application having high importance. As a result, the server 1 can prevent a resource shortage in a highly important application.

  Next, the processing of S2 and S3 of FIG. 3 will be described in detail with reference to FIG.

  First, the resource monitoring unit 133 of the server 1 monitors the resource usage of each APL group (S21). For example, the resource monitoring unit 133 periodically monitors the resource usage of each APL group. Then, the resource monitoring unit 133 has an APL group with a resource usage rate of a predetermined value (for example, 80%) or more and an APL group with surplus resources (for example, a resource usage rate of 50% or less). If so (Yes in S22), the resource reallocation unit 134 is requested to reallocate resources (S23).

  For example, a part of the allocated resource amount of an APL group with surplus resources (for example, 20% of the allocated resource amount of the APL group) is given to the resource reallocation unit 134 with a resource utilization rate of a predetermined value (for example, 80 %) Request to be assigned to the above APL groups. Then, the process returns to S21.

  On the other hand, when the resource monitoring unit 133 does not have an APL group with a resource utilization rate equal to or greater than a predetermined value, or there is no APL group with surplus resources even if there is an APL group with a resource utilization rate greater than or equal to a predetermined value (No in S22), no resource reallocation is requested to the resource reallocation unit 134, and the process returns to S21.

  Next, with reference to FIG. 6, the process of S4 in FIG. 3 will be described in detail with specific examples. When the resource reallocation unit 134 of the server 1 receives a resource reallocation request from the resource monitoring unit 133 (Yes in S41), the resource reallocation unit 134 refers to the resource management information 131 and determines whether the resource can be reallocated. (S42). For example, the resource reallocation unit 134 refers to the resource management information 131 and uses a part of the allocated resource amount of an APL group with surplus resources (for example, 20% of the allocated resource amount of the APL group) as a resource usage rate. Is determined to be able to be assigned to an APL group having a predetermined value (for example, 80%) or more.

  In S42, when the resource reassignment unit 134 determines that the resource can be reassigned (Yes in S42), the resource reassignment unit 134 executes resource reassignment (S43). That is, the resource reallocation unit 134 changes the allocated resource amount between APL groups in the resource management information 131. For example, the resource reallocation unit 134 reduces a part of the allocated resource amount of the APL group with surplus resources in the resource management information 131, and uses the reduced allocated resource amount for the APL group whose resource usage rate is equal to or greater than a predetermined value. Add to the allocated resource amount. Although not shown here, after S43, the container control unit 12 sets the allocated resource amount of each APL group based on the updated resource management information 131.

  Further, the resource reallocation unit 134 does not accept the resource reallocation request from the resource monitoring unit 133 (No in S41), or determines that the resource reallocation is impossible in S42 (No in S42). ), The process is terminated.

  By doing in this way, the server 1 can allocate the resource of the APL group with a surplus resource, when the resource shortage occurs in any APL group. As a result, the server 1 can prevent resource shortage of each application and improve resource utilization efficiency.

  In addition, although the resource management part 13 demonstrated as an example the case where it is equipped with the server 1, it is not limited to this. For example, the resource management unit 13 may be realized by a device separate from the server 1. In this case, the resource management unit 13 may manage the resources of the server 1 via the network.

(program)
Further, it can be implemented by installing a program for realizing the function of the server 1 described in each embodiment in a desired information processing apparatus (computer). For example, the information processing apparatus can function as the server 1 by causing the information processing apparatus to execute the program provided as package software or online software. The information processing apparatus referred to here includes a desktop or notebook personal computer. In addition, the information processing apparatus includes mobile communication terminals such as smartphones, mobile phones and PHS (Personal Handyphone System), PDA (Personal Digital Assistants), and the like. Further, the server 1 may be mounted on a cloud server.

  An example of a computer that executes the program (resource management program) will be described with reference to FIG. As illustrated in FIG. 7, the computer 1000 includes, for example, a memory 1010, a CPU 1020, a hard disk drive interface 1030, a disk drive interface 1040, a serial port interface 1050, a video adapter 1060, and a network interface 1070. These units are connected by a bus 1080.

  The memory 1010 includes a ROM (Read Only Memory) 1011 and a RAM (Random Access Memory) 1012. The ROM 1011 stores a boot program such as BIOS (Basic Input Output System). The hard disk drive interface 1030 is connected to the hard disk drive 1090. The disk drive interface 1040 is connected to the disk drive 1100. A removable storage medium such as a magnetic disk or an optical disk is inserted into the disk drive 1100, for example. For example, a mouse 1110 and a keyboard 1120 are connected to the serial port interface 1050. For example, a display 1130 is connected to the video adapter 1060.

  Here, as shown in FIG. 7, the hard disk drive 1090 stores, for example, an OS 1091, an application program 1092, a program module 1093, and program data 1094. Various data and information described in the above embodiment are stored in, for example, the hard disk drive 1090 or the memory 1010.

  Then, the CPU 1020 reads out the program module 1093 and the program data 1094 stored in the hard disk drive 1090 to the RAM 1012 as necessary, and executes the above-described procedures.

  Note that the program module 1093 and the program data 1094 related to the resource management program are not limited to being stored in the hard disk drive 1090. For example, the program module 1093 and the program data 1094 are stored in a removable storage medium and the CPU 1020 via the disk drive 1100 or the like. May be read. Alternatively, the program module 1093 and the program data 1094 related to the above program are stored in another computer connected via a network such as a LAN or a WAN (Wide Area Network) and read by the CPU 1020 via the network interface 1070. May be.

1 server 11 hardware 12 container control unit 13 resource management unit 131 resource management information 132 resource allocation unit 133 resource monitoring unit 134 resource reallocation unit

Claims (4)

A resource management method executed in a resource management device that manages resource allocation amounts for a plurality of applications that operate on a server,
A receiving step for receiving the importance of each of the applications;
A grouping step of grouping the applications into two or more groups according to the received importance;
A resource allocation step of preferentially allocating resources from the group of applications with high importance;
A monitoring step of monitoring resource usage of each group application;
As a result of the monitoring, when the resource usage rate, which is the resource usage amount relative to the allocated resource amount in any group, is equal to or greater than a predetermined value, at least a part of the resource allocation amount of the group having surplus resources is used as the resource usage amount. A resource reallocation step of allocating to a group having a rate equal to or greater than the predetermined value;
A resource management method comprising: In the resource allocation step,
The group of applications having higher importance increases the amount of resources allocated to the group of applications, and the amount of resources allocated to groups of applications decreases as the group of applications having lower importance. The resource management method according to 1. In the resource allocation step,
2. The resource management method according to claim 1, wherein applications having similar importance levels and different types of resources mainly used by the applications are grouped into the same group. . The resource management method according to claim 1, wherein a resource that can be used by the application is isolated by container virtualization technology.

Images