JP2015045899A - System and method for allocating virtual machine resource - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a virtual machine resource allocation system and a method capable of using more than a certain number of contracted machines on the same day even if a virtual machine is temporarily required more than the number of contracted machines in an emergency. To do. SOLUTION: When a user urgently needs a virtual machine, a contract (company) that constructs a system that provides virtual machines even if the number of contracted units or more and requires a large amount of virtual machines for emergency response. And, there is a means to coexist companies (contracts) such as no increase in an emergency or no use in an emergency, and at the same time, adjust so that a large amount of virtual machines are not used in all contracts. [Selection diagram] Fig. 2

Description

  The present invention relates to a virtual machine resource allocation system and method.

  As a background art in this technical field, there is JP 2012-226703 A (Patent Document 1). This gazette states that “in a virtualization system, when multiple tasks are executed on one machine, or when one task is executed across multiple machines, resources can be allocated from a business perspective. In response to the problem “To be done”, “work spaces” 124 and 125 are defined as spaces for executing work in the virtual space in the physical machine 100, and the work spaces 124 and 125 are defined. Assign resources to it. The status of the job being executed is acquired, and the resource amount for the business spaces 124 and 125 is calculated in consideration of the business having the job being executed and the priority of the business to which the job belongs. Then, virtual machine replicas 130 and 132 necessary for executing the business are arranged in the business spaces 124 and 125, and these virtual machines 130 and 132 are operated using resources allocated to the business spaces 124 and 125. Let "Solutions are described (see abstract).

JP 2012-226703 A

  Patent Document 1 discloses a resource control method for allocating resources necessary for running a virtual machine on a certain computer to execute a task, the computer including a task, a priority of the task, Job execution information that stores a correspondence relationship with at least one job that constitutes a business, stores a correspondence relationship between the business and a business space that executes the business, and indicates a status of a job that is being executed on the computer From the correspondence between the acquired job execution information, the business space to which the job being executed belongs, and the priority of the job to which the job being executed belongs, the job being executed is A resource amount to be allocated to the business space to which the job belongs is calculated, and a resource to be allocated to a virtual machine on which the job being executed in the business space is operated from the calculated resource amount Calculating a describes a resource control method.

  However, the resource control method of Patent Document 1 does not consider resource allocation in the case where an abnormal situation occurs that exceeds the resource utilization resource limit currently used, for example.

  Normally, for example, when contracting with the virtual machine management side for the use of a virtual machine on the cloud, it is common to contract by designating the number of units used in advance. Therefore, it cannot cope with a rapid increase exceeding the usage range of the contract. For example, there may be a case where a virtual machine is temporarily required to exceed the contracted number in the event of a problem or emergency, but it is usually impossible to use more than the contracted number exceeding the fixed limit on the same day.

  Therefore, in the event of a problem or an emergency such as an emergency, the present invention can use more than the number of contracts exceeding the fixed limit on the same day even if a virtual machine may be temporarily required in excess of the number of contracts. A virtual machine resource allocation system and method capable of

  In order to solve the above problems, the present invention constructs a system that provides a virtual machine even if the number of contracted machines exceeds the contracted number when a user is urgently required, and a large number of virtual machines are provided for emergency response. There is a means for coexisting contracts (corporations) that are required for use with companies (contracts) that do not increase in emergencies or that are not used in an emergency, and at the same time, adjust all contracts so that virtual machines are not used in large quantities .

For example, the virtual machine resource allocation system of the present invention is:
In an information processing system including a data center that provides virtual server resources allocated to a processing server in response to a user request from a client terminal,
The data center is
A first rack that includes one or more processing servers that are assigned and used by the virtual machine resources; and a second rack that includes one or more processing servers that are assigned virtual machine resources and are unused. A management server for managing virtual machine resources allocated to the processing servers of the first and second racks;
The management server
Means for managing a plurality of users so that a plurality of users can use virtual machine resources in the first and second racks according to desired specification cost and performance / size;
In an emergency, the virtual machine resource of the processing server in use is kept to the minimum resource, a free virtual machine resource is secured, and the user can respond to the virtual machine resource request, according to the user virtual machine resource request And a means for moving the allocated virtual machine resource to an unused processing server that can secure the required virtual machine resource and reallocating the virtual machine resource. And

Further, the virtual machine resource allocation method of the present invention includes:
In an information processing method including a data center that provides virtual server resources allocated to a processing server in response to a user request from a client terminal,
The data center is
A first rack that includes one or more processing servers that are assigned and used by the virtual machine resources; and a second rack that includes one or more processing servers that are assigned virtual machine resources and are unused. A management server including a control unit that manages virtual machine resources allocated to the processing servers of the first and second racks;
The control unit of the management server
Usually, a step of managing so that a plurality of users can use virtual machine resources in the first and second racks according to desired specification cost and performance / size,
In an emergency, the virtual machine resource of the processing server in use is suppressed to the minimum resource, a free virtual machine resource is secured, and the virtual machine resource request of the user can be handled,
When there is a shortage in response to the user's virtual machine resource request, the allocated virtual machine resource is moved to an unused processing server that can secure the required virtual machine resource, and the virtual machine resource is re-allocated. And a step of performing.

ADVANTAGE OF THE INVENTION According to this invention, when a virtual machine is urgently needed, the virtual machine resource allocation system and method which can utilize a virtual machine more than the contracted number can be provided.
Problems, configurations, and effects other than those described above will become apparent from the following description of embodiments.

It is a figure which shows the structure of the information processing system which provided the virtual machine resource allocation system of this invention. It is a figure which shows the mode of reallocation of a virtual machine resource at the time of emergency. It is a figure which shows the resource usage condition in a rack unit at the time of reallocation of a virtual machine resource at the time of emergency. It is a functional block diagram of a management server. It is a flowchart which shows the procedure of whether to perform virtual machine resource allocation processing. It is a flowchart which shows a virtual machine resource allocation processing procedure.

  Hereinafter, the present embodiment will be described with reference to the drawings. In this example, the description will be given on the assumption of “Hadoop”, which is an open source platform that can easily process a large amount of data distributed to a plurality of machines. In this example, a user and a customer are used synonymously, and means a contractor user who can use a virtual machine in an emergency. The general user is a user who has not yet contracted the number of options.

  Data centers strive to minimize the amount of unused resources in order to effectively use resources and eliminate waste. This means the contradiction of not creating or having spare resources.

  Today, however, these old-style fixed DC operation policies cannot respond to resource requirements for a cloud environment, a stock distribution center (hereinafter referred to as DC) compatible site, and a BigData compatible environment. Therefore, simultaneously with normal resource management, a parallel computing environment such as Hadoop is operated in the DC, and the Hadoop environment is released as a general VM resource in response to an emergency resource request from a customer.

  At this time, the lack of Hadoop resources use the resources of the external DC for emergency evacuation. Note that parallel distributed processing resources (virtual machine VM resources) such as Hadoop aim to improve calculation processing performance by minimizing network delays and the like. For this purpose, these VMs are generally allocated to one place.

  More specifically, for example, a virtual machine resource provided on a contract basis is defined by three examples of Small / Midium / Large (hereinafter referred to as S, M, and L). The lower limit / upper limit of the CPU capacity of each machine resource, the lower limit / upper limit of the memory size, and the lower limit / upper limit of the HDD capacity are defined corresponding to the price, the quantity that can be secured, and the performance. S, M, and L are considered as units, and their simulated sizes are expressed by numerical values such as 1, 5, and 10 (CPU speed, memory size, etc.).

  As an example, it is assumed that an environment for L for five machines can be secured on one physical server. Let this size be 1000. The virtual machine VM resource for L is a service definition, and the basic is 1000/10 = 100, 100/2 = 50, and the maximum is 50.

  Next, the virtual machine VM resource for M is defined as a service definition, with the maximum performance being 100/5 = 20 and the basic being 20/2 = 10. Similarly, the virtual machine VM resource for S is configured to provide a resource with a basic 1 and a maximum of 2.

  With such a management method, when there is a request for a large amount of virtual machine resources in an emergency, the use of MAX from the virtual machine VM resource of the general user is prohibited in response to the request (restricted to the minimum resource of the virtual machine resource). ) Based on the premise, it is possible to consolidate and secure large-scale S virtual machine resources in one place by packing S, M, and L into the assignable areas of S, M, and L virtual machine resources. It becomes possible.

  FIG. 1 is a diagram showing a configuration example of an information processing system to which a virtual machine resource allocation system of the present invention is applied.

  In the figure, the information processing system 7 includes a plurality of data centers (first, second, and third data centers) 6a, 6b, and 6c, and a client terminal 8.

  The plurality of data centers (first, second, and third data centers) 6a, 6b, and 6c each have a router, and are connected via the router.

  The client terminal 8 is connected to the third router 4c of the first data center 6a. The data center, for example, the first data center 6a includes first and second racks (physical servers) 5a and 5b.

The first rack 5a (/ doc1 / rack1) includes a first router 4a, a management server 1, a first processing server 2a (node1.xxx.co.jp), and a second processing server 2b (node2.xxx. co.jp).
The router 4a is connected to the third router 4c, the management server 1, and the first and second processing servers 2a and 2b.

  One virtual machine VM resource 3a is allocated to the first processing server 2a (node1.xxx.co.jp), and two virtual machines are allocated to the second processing server 2b (node2.xxx.co.jp). Machine VM resources 3b and 3c are allocated.

The second rack 5b (/ doc1 / rack2) includes a second router 4b, a third processing server 2c (node3.xxx.co.jp), and a fourth processing server 2d (node4.xxx.co.jp). ), And a fifth processing server 2e (node5.xxx.co.jp).
The router 4b is connected to the third router 4c and the third, fourth, and fifth processing servers 2c, 2d, and 2e.

  One virtual machine VM resource 3d is allocated to the third processing server 2c (node3.xxx.co.jp), and two virtual machines are allocated to the fourth processing server 2d (node4.xxx.co.jp). Machine VM resources 3e and 3f are allocated, and two virtual machine VM resources 3g and 3h are allocated to the fifth processing server 2e (node5.xxx.co.jp).

  The management server 1 manages the allocation of the virtual machine VM resource to each processing server and the allocation. In an emergency, the management server 1 releases the resource of the second rack 5b, that is, the resource for Hadoop. And a function of reallocation as a general virtual machine resource.

  In such an information processing system, first, normally, a plurality of customers (users) use virtual machine VM resources in any of the racks 5a and 5b depending on the cost and performance / size of using S, M, and L. Make it possible.

  The management server 1 performs virtual machine VM resource management based on user contract conditions (basic size / maximum size / maximum CPU performance / minimum CPU performance, etc.).

  In such a situation, a virtual machine resource that exceeds the contracted range that the user (client terminal 8) has contracted due to a large number of telecommuters occurring due to an emergency such as a problem or emergency, such as traffic interruption or new influenza When the number is required, the following problems arise.

That is, in an emergency, a huge resource request is generated compared to a normal time. In response to this enormous resource request, it takes enormous time to perform resource allocation in an arbitrary free resource portion in the management server 1 with enormous data search processing.
The present invention corrects such a problem. In such an emergency, the management server 1 assigns virtual machine VM resources described below.

  FIG. 2 is a diagram illustrating how virtual machine resources are reallocated in an emergency. In the figure, at the time of the resource request in the above-mentioned emergency, first, each currently allocated resource is temporarily suppressed to the minimum resource, and free resources are secured in a short time.

  Thereafter, in response to a resource request of a general user, if there is a shortage, these virtual machine VM resources are placed in racks (processing servers 2c to 2e in the second rack in this example) that can secure necessary resources. Move (3d-3g) and re-allocate resources. In other words, the Hadoop resource of the second rack is released, and processing corresponding to the general user virtual machine resource request is performed.

  FIG. 3 is a diagram illustrating a resource usage status in a rack unit when a virtual machine resource is reassigned in an emergency.

  In the figure, in the virtual machines VM3a to VM3h, the shaded portion indicates being used, and the dotted line portion represents an unused portion. For example, the maximum size of the resource size L is in use for the virtual machine VM3a, the basic size of the resource size L is in use for the virtual machines VM3b and 3c, and the basic size of the resource size M is in the virtual machine VM3d. The virtual machine 3e is in use, indicating that the basic size and the maximum size of the resource size L are unused.

  FIG. 4 is a functional block diagram of the management server. In FIG. 1, the management server 1 includes a management device (management control means) 11, a storage device 12, and an input / output device 13.

  The management apparatus 11 includes a virtual resource management unit and a virtual resource allocation unit.

  The virtual resource allocation unit allocates virtual machine VM resources to the processing servers 2a to 2e of the first and second racks 5a and 5b.

The virtual resource management unit manages those virtual machine VM resources.
The storage device includes a virtual resource allocation program unit and a virtual resource contract information storage unit.

  The virtual resource allocation program unit stores a program operated by the virtual resource management unit, and the virtual resource contract information storage unit stores information indicating a contract condition concluded with a customer (user).

  The input / output device includes a reception unit that receives a virtual resource allocation request from a user in an emergency.

  The billing server 9 charges the virtual machine that is temporarily used on a daily or hourly basis for use outside the contract.

  FIG. 5 is a flowchart showing a procedure for determining whether to perform virtual machine resource allocation processing.

In the figure, when the virtual machine resource allocation process is started in step S110, it is determined in step S112 whether or not the option increase contract is present between the customer and the virtual machine manager.
If there is a contract at this step, the process proceeds to the virtual machine resource allocation processing step S113 of FIG.

  FIG. 6 is a flowchart showing a virtual machine resource allocation processing procedure.

  In the figure, when the virtual machine allocation process is started in step S120, it is confirmed in step S121 whether or not there is a free area in the used device (server / storage).

If there is an empty area in this step, a virtual machine is created in the using device in step S124.
If not, in the next step S122, it is determined whether or not it is possible to move an unused virtual machine of another device 1 (use system) different from the use device to the use device and to add a free space. To do.

If it can be added in this step, a virtual machine is created in step S124.
If it is not possible to add, in the next step S123, it is determined whether or not a virtual machine of another device 2 (new server / storage) can be assigned to the use device, and then in step S124, the virtual machine Create.

According to the present embodiment, the number of virtual machines can be increased using the administrator's self-portal site from the Web.
As a result, it is possible to provide a contract that can increase the number of virtual machines suddenly, assuming an emergency.

  Also, normally, a newly applied virtual machine is used after 2 to 3 business days, but according to this system, it can be used immediately.

  In addition, when a virtual machine is suddenly needed for work-at-home due to traffic interruption, the system administrator can “create and start a virtual machine” without adding a contract.

If the number of units is equal to or greater than the initial contract, billing will be made at a later date by the number of units used x days x unit price.
Contracts that can have a sudden increase in number can be selected as an option.

  Temporary use of virtual machines may be billed on a daily or hourly basis for use outside the contract.

In addition, this invention is not limited to the Example mentioned above, Various modifications are included. For example, the above-described embodiments have been described in detail for easy understanding of the present invention, and are not necessarily limited to those having all the configurations described.
Further, the control lines and information lines indicate what is considered necessary for the explanation, and not all the control lines and information lines on the product are necessarily shown. Actually, it may be considered that almost all the components are connected to each other.

1 Management Server 2 Processing Server 3 Virtual Machine VM (Resource)
4 router 5 rack 6 data center 7 information processing system 8 client terminal 9 billing server

Claims (2)

In an information processing system including a data center that provides virtual server resources allocated to a processing server in response to a user request from a client terminal,
The data center is
A first rack that includes one or more processing servers that are assigned and used by the virtual machine resources; and a second rack that includes one or more processing servers that are assigned virtual machine resources and are unused. A management server for managing virtual machine resources allocated to the processing servers of the first and second racks;
The management server
Means for managing a plurality of users so that a plurality of users can use virtual machine resources in the first and second racks according to desired specification cost and performance / size;
In an emergency, a means for keeping the virtual machine resource of the processing server in use to a minimum resource, a means for securing a free virtual machine resource and making it possible to respond to the virtual machine resource request of the user, a request for the virtual machine resource of the user And a means for moving the allocated virtual machine resource to an unused processing server that can secure a required virtual machine resource and re-allocating the virtual machine resource. Feature virtual machine resource allocation system. A virtual machine resource allocation method in an information processing system including a data center that provides a virtual server resource allocated to a processing server in response to a user request from a client terminal,
The data center is
A first rack that includes one or more processing servers that are assigned and used by the virtual machine resources; and a second rack that includes one or more processing servers that are assigned virtual machine resources and are unused. A management server including a control unit that manages virtual machine resources allocated to the processing servers of the first and second racks;
The control unit of the management server
Usually, a step of managing so that a plurality of users can use virtual machine resources in the first and second racks according to desired specification cost and performance / size,
In an emergency, the virtual machine resource of the processing server in use is suppressed to the minimum resource, a free virtual machine resource is secured, and the virtual machine resource request of the user can be handled,
When there is a shortage in response to the user's virtual machine resource request, the allocated virtual machine resource is moved to an unused processing server that can secure the required virtual machine resource, and the virtual machine resource is re-allocated. A virtual machine resource allocation method comprising the steps of: