KR20090063122A - System and method for managing work load in the multi system - Google Patents

Abstract

A multiple system workload management system and a method thereof for dynamically assigning virtual resources to each service according to requirement are provided to perform an integration management for utilizing the resource of distributed heterogeneous systems with various forms. A plurality of physical nodes allocates virtual resource corresponding to the request of user terminals connected on internet. A plurality of physical nods provides a service to the user terminal. A service controller controls the physical node in order to provide the service requested by the user terminal by performing reconfiguration in consideration of state and resource of the physical nodes. A virtual node pool(21) includes a plurality of virtual nodes in which service is mounted.

Description

SYSTEM AND METHOD FOR MANAGING WORK LOAD IN THE MULTI SYSTEM}

TECHNICAL FIELD The present invention relates to a multi-system workload management system. In particular, in a distributed heterogeneous environment, a new virtual layer is established between physical computing resources and services to virtualize all resources, and to request virtualized resources. The present invention relates to a system and method for managing a workload based on a dual system that can be dynamically allocated or reclaimed according to each service.

The present invention is derived from the research conducted as part of the IT new growth engine core technology development project of the Ministry of Information and Communication and the Ministry of Information and Telecommunications Research and Development. [Task Management Number: 2006-S-035-02, Task name: for distributed heterogeneous server environment Development of open SW-based virtual infrastructure implementation technology].

In general, the method of utilizing resources in a distributed system generally includes a method of managing a plurality of server systems as a single virtual system and dividing a large server system into several smaller virtual systems. have.

However, when such a conventional virtual system uses a method of integrating and managing a plurality of server systems to operate as a single virtual system, there is a limited problem in the binding situation of a server system existing according to an execution environment of a service.

In addition, when a conventional virtual system uses a method of dividing one server system into multiple virtual systems, there is a problem in that it cannot be supported considering the scalability of a service with limited resources.

The present invention has been created in view of the above problems, and in a distributed heterogeneous environment, a new virtual layer is established between physical computing resources and services to virtualize all resources, and each virtualized resource according to requirements. Its purpose is to provide a multi-system workload management system and method for dynamically allocating or retrieving services.

In addition, an object of the present invention is to use a multi-distributed heterogeneous system by complex operation even when one system is divided into a plurality of small scale virtual system to configure a single system consisting of a plurality of server systems to integrate management To provide a multi-system workload management system and method that can be integrated and managed to utilize their resources.

In addition, another object of the present invention is to dynamically allocate services and computing resources as needed in a virtual infrastructure environment, to implement a new service-oriented infrastructure resource management that allocates and uses as many resources as needed when services are needed. It provides a system and workload management system and method.

Another object of the present invention is to provide a multi-system workload management system and method for improving the performance of the system in the infrastructure.

Another object of the present invention is to provide a multi-system workload management system and method for minimizing system management costs by maximizing the utilization of system resources in the infrastructure.

Multi-system workload management system according to an aspect of the present invention for achieving the above object is divided into a virtual system to configure a system into a plurality of server systems to manage a multi-system workload in the Internet, A plurality of physical nodes allocating virtual resources in response to requests of connected user terminals to provide services to the user terminals; A service control unit configured to control the physical node to provide a service requested by a user terminal by reconfiguring in consideration of the state and resources of the physical node; And a virtual node pool including a plurality of virtual nodes on which the service is mounted.

In addition, a multi-system workload management method according to another aspect of the present invention for achieving the above object is a method for managing a multi-system workload by configuring a system into a plurality of server systems by dividing into a small virtual system Reconfiguring resources of the physical node in consideration of the status and resources of a plurality of physical nodes in response to requests of user terminals connected to the Internet; Providing a service desired by the user terminals through the physical node reconstructed the resource.

By the above-mentioned problem solving means, the present invention is that most of the major services are configured and executed in a complex structure in the virtual infrastructure environment. In such a virtual infrastructure environment, by dynamically allocating services and computing resources as needed, the technology for resource virtualization, on-demand provisioning, and automatic management of policy-based workloads can be improved.

In other words, in terms of resource virtualization, heterogeneous resources in the infrastructure are defined as virtualized resources, and a resource pool is formed so that all services can share resources.

In addition, in terms of on-demand provisioning, there is an effect of providing optimal virtual computing by dynamically allocating or retrieving virtualized resources to each service according to service requirements by minimizing administrator intervention.

In addition, in terms of automatic policy-based workload management, it is possible to provide optimal virtual computing by automatically allocating or reclaiming resources as needed for each service according to a predetermined policy and current situation.

It not only improves the performance of systems in the infrastructure, but also provides scalability and reliability.

In addition, the system management cost can be minimized by maximizing the utilization of system resources in the infrastructure.

The multi-system workload management system and method of the present invention virtualizes all resources by establishing a new virtual layer between physical computing resources and services in a distributed heterogeneous environment, and virtualizes each service according to requirements. The technical gist is to allow dynamic allocation or retrieval of a.

Hereinafter, with reference to the accompanying drawings, preferred embodiments of the present invention will be described in detail, focusing on the parts necessary to understand the operation and action according to the present invention.

In the following description, specific details of the multi-system workload management system and method of the present invention are presented to provide a more comprehensive understanding of the present invention, which, without these specific details, may also be readily implemented by these modifications. It will be apparent to one skilled in the art.

Hereinafter, a multi-system workload management system and method according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a block diagram showing a schematic configuration of a multi-system workload management system of the present invention. In FIG. 1, a configuration for providing a service utilizing each user terminal 10 connected through the Internet 12 is illustrated.

Referring to FIG. 1, the client requests a service control unit (a) connected through the Internet 12 for a desired service through the user terminal 10. Then, the service control unit (a) controls the physical nodes (15a-15n) to provide a service desired by the user in the optimal system by reconfiguring in consideration of the state and resources of the physical nodes (15a-15n).

The service control unit a includes a service management unit (Admin node) 13 and a service request unit (dispatcher system) 14.

In this case, the service manager 13 controls the virtual node pool 21 to reconstruct resources of the physical nodes 15a to 15n so as to service in an optimal state according to the states and resources of the physical nodes 15a to 15n. The service request unit 14 is controlled according to the reconfiguration. In other words, the service management unit 13 integrates and manages distributed heterogeneous systems.

The service request unit 14 analyzes the user's request for the service under the control of the service manager 13 and utilizes the virtual node when the service is mounted on the virtual node in advance or does not satisfy the user's requirements. If not, the virtual node pool 21 arranges the virtual node loaded with the service requested by the user to the optimal physical node 15a-15n and provides the service to the terminal 10 of the user requesting the service. The request signal is distributed to the virtual node in which the corresponding service disposed in the physical nodes 15a to 15n is mounted. Then, the physical nodes 15a to 15n to which the distributed virtual node belongs support the state and resources of the virtual node to be reconfigured. The virtual node transmits a service response to the terminal 10 and provides the service desired by the user. To get started.

An operation of providing a service desired by a user by an optimal virtual node 17, 18, 19... Arranged in the physical nodes 15a to 15n is as follows.

Virtual node pools (VN Pools) 21 on which the desired services are mounted are prepared by the install node 22, and the services (e.g., media server 17, Web) are available based on available resources. The virtual node equipped with the server 18 and the terminal server 19 is disposed in each physical node 15a to 15n and provided to the terminal 10 requesting the corresponding service.

At this time, in order to ensure service continuity according to the state of the virtual nodes in service, a virtual node migration (VN Migration) 16 of the physical nodes 15a to 15n is provided.

The concept between the physical nodes 15a to 15n and the virtual node pool 21 will be described below.

2 is an exemplary diagram illustrating the concept of a physical node and a virtual node pool of the present invention.

Here, the virtual infrastructure environment (VINE) 30 of FIG. 2 represents an environment of a virtual infrastructure system that allows a user to have a consistent and convenient logical structure, and a new virtual layer is established between physical computing resources and services. Resources are virtualized, and services are environments where virtualized resources can be shared through virtual layers.

Referring to FIG. 2, the physical layer 33 including the physical nodes includes a bare metal node 34 and a physical OS 35 to be mounted on the bare metal node 34. It consists of. Here, the bare metal node unit 34 is composed of computer system nodes in which no SW including an OS is installed, and the physical OS unit 35 is constituted of OSs basically installed in the computer system node.

The virtual layer 32 constituting the virtual server directly includes a virtual OS unit 36 and a virtual server unit 37. The virtual OS unit 36 is composed of virtual OSs of the basic elements constituting the virtual server to provide a virtual operating environment, and configure the virtual OS unit 36 through a hypervisor such as Xen. . In addition, the virtual server unit 37, which is constructed for each service to provide resources for each service, is composed of a single virtual OS or a plurality of virtual OSs.

The service layer 31, which has a service to an actual user, is a type of service provided to end users, and each service is dynamically provided with as many resources as necessary through a corresponding virtual server.

Next, the operation according to the internal configuration of the service manager 13 and the physical nodes 15a to 15n will be described.

3 is a diagram illustrating a multi-system workload management process of the present invention.

Referring to FIG. 3, in order to manage a workload, the manager 50 establishes a management policy through the GUI 51 and an interface, and manages the workload accordingly.

Managing the actual workload is targeted to the service 59, and the execution of the service is performed in the virtual server 57. The virtual server 57 is a virtual server system that is constructed for each service to provide resources for each service. Each virtual server system includes a single virtual OS or a plurality of virtual OSs 54a to 54c. These virtual OSs 54a to 54c are managed by a virtual resource pool 55.

The virtual resource pool 55 manages virtual resources, and the virtual resources are resources that are shown to users by converting physical resources into logical resources. The virtual CPU, virtual memory, This is the case for virtual networks and virtual disks. The virtual resource pool 55 physically utilizes a physical resource pool 53 including a bare metal node 51 and local OSs 52a to 52c operated by the bare metal nodes 51a to 51c. .

Workload management in Figure 3 has a resource management function that can manage the system remotely.

The resource pool management function 56 defines a computing resource in an infrastructure as a virtualized resource and provides a manager with a virtual image of all resources by integrating and managing all resources so that all the services can be shared. The function of resource provisioning 60 not only installs an operating system on the bare metal node, but also dynamically performs all processes until the actual service can be driven up to the installation of various software required for the service. The infrastructure monitoring function 61 is based on the international industry standard DMTF-CIM / WBEM to monitor and analyze the resource operation status and performance in the infrastructure.

Infrastructure coordination management function 58 dynamically allocates or reclaims virtualized resources according to service needs.

4 is a block diagram showing the internal configuration of the service management unit and the physical node of the present invention. In FIG. 3, reference numeral '100' denotes a service manager 13 as a management node, and reference numeral '200' denotes physical nodes 15a to 15n as execution nodes.

Referring to FIG. 4, workload management is performed in the management node 100 and the execution node 200. For workload management, the following means (101. 102., 103, 104, 105, 106, 107, 108, 109) are required.

First, the coordinator 101 is a module for interworking a user with a graphic user interface (GUI) 51 and an interface. The coordinator 101 manages coordination of overall operations of a multi-system workload management system.

The VNI-Virtual Node Interface 102 provides a physical interface that the coordinator 101 needs to manage a virtual node. The VNI-Virtual Node Interface 102 receives a request for a virtual node from the coordinator 101 and is located at a remote physical node. The virtual infrastructure manager (VIM) 106 causes the operation related to the virtual node to be performed.

The resource pool manager (RPM) 103 dynamically supports virtual resource management and information collection for processors, memory, etc. according to workloads generated on virtual nodes for statically limited static resources. .

The multi-system manager 104 sets overall operations according to multi-system workload management, monitors the state of the system, and dynamically controls system resources according to events that occur. In addition, the multi-system management unit 104 has an autonomous system management function to minimize the intervention of the system administrator.

The cluster manager 105 has a configuration and configuration management function of a cluster using a plurality of virtual nodes.

The virtual infrastructure management unit 106 manages interworking between heterogeneous systems for the purpose of standards-based resource management. That is, a system-based virtual infrastructure is implemented for the purpose of managing nodes, virtual nodes, and resources of the execution node 200.

The virtual node manager (VNM) 107 controls a hypervisor (not shown), such as Xen, that allows the resources of a physical node to be used by multiple virtual OSs, and the virtual node manager 107 also controls the virtual node. It controls provisioning and virtual node resources and directly manages the state of the virtual nodes running in the node.

The resource monitor 108 supports standards-based system monitoring monitoring to enable heterogeneous interworking of resources throughout the system.

In FIG. 4, the connection between the management node 100 and the execution node 200 is made of a client / server model, and each device is managed and communicated in the form of a client 114 and a server 115.

In FIG. 4, the workload management control flow proceeds around the coordinator device 101. Physical node management, virtual node management, resource pool management, cluster management, and policy management are via the coordinator device 101. The management of the physical nodes 15a to 15n by the manager transmits the command dispatched from the coordinator 101 to the resource pool management unit 103 at the management node 100 and manages the physical nodes collected through the resource monitoring unit 108. Understand the status and proceed. The virtual node management forwards the command dispatched from the coordinator 101 to the virtual node interface 102 and to the virtual infrastructure management device 106 that manages the virtualized resources through the server of the executing node 200, which is actually Execution takes place via the virtual node management apparatus 107 that is the implementer. In order for execution to occur through the hypervisor, the contents of the virtual node are set in the region setting table 110 stored in the storage 110. Resource pool management delivers the command dispatched from the coordinator 101 to the resource pool management unit 103, to identify and manage the status of all the virtual resources collected through the resource monitoring unit 108 to proceed. Managed data is collected in the storage unit 113. Cluster management transmits the command dispatched from the coordinator 101 in the managed node 100 to the cluster manager 105, checks the status of the virtual nodes through the storage 113, and the corresponding node in the execution node 200. The storage unit 111 stores a configuration table for executing a cluster service through the cluster management agent 109 of the virtual node and executes the storage 111. Policy management transmits the command dispatched from the coordinator 101 to the multi-system management unit 104 in the management node 100, analyzes the transmitted command, establishes a management policy regarding policy conditions and policy control, and intervenes system administrators. Implement autonomic system management that minimizes

Meanwhile, in the detailed description of the present invention, specific embodiments have been described, but various modifications are possible without departing from the scope of the present invention. Therefore, the scope of the present invention should not be limited to the described embodiments, but should be determined not only by the scope of the following claims, but also by the equivalents of the claims.

1 is a block diagram showing a schematic configuration of a multi-system workload management system of the present invention.

2 is an exemplary diagram illustrating the concept of a physical node and a virtual node pool of the present invention.

3 illustrates a multi-system workload management process of the present invention.

Figure 4 is a block diagram showing the internal configuration of the service management unit and the physical node of the present invention.

Claims (10)

In a system for managing a multi-system workload by dividing into a virtual system to configure a system into a plurality of server systems, A plurality of physical nodes allocating virtual resources in response to requests of user terminals connected to the Internet to provide services to the user terminals; A service control unit configured to control the physical node to provide a service requested by a user terminal by reconfiguring in consideration of the state and resources of the physical node; Virtual node pool containing multiple virtual nodes with services Multi-system workload management system comprising a. The method of claim 1, wherein the service control unit, A service request unit for distributing a service request signal to a virtual node disposed in the physical node to provide the service to the user terminal requesting the service; A service manager configured to reconfigure resources of the physical node by controlling the virtual node pool according to the state and resources of the physical node and to control the service request unit according to resource reconfiguration of the virtual node Multi-system workload management system comprising a. The method of claim 2, wherein the service management unit, A coordinator for coordinating the operation of the multi-system workload management system; A virtual node interface providing an interface between the coordinator and the physical node; A resource pool manager that dynamically manages virtual resources and collects information according to a workload occurring on a virtual node with respect to physically limited static resources; A multi-system manager configured to set an operation of the multi-system workload management system according to the multi-system workload management, and monitor a state of the multi-system workload management system and control system resources according to an event generated; A storage unit for storing state information of the collected virtual nodes Multi-system workload management system comprising a. The method of claim 3, wherein the service management unit, A cluster management unit that receives a command dispatched from the coordinator and performs a configuration and configuration management function of a cluster using a plurality of virtual nodes. Multi-system workload management system further comprising. The system of claim 3, wherein the multi-system manager comprises: A multi-system workload management system characterized by performing autonomous system management functions without the involvement of a system administrator. The method of claim 1, wherein the physical node, A virtual infrastructure management unit for managing heterogeneous systems to interwork for the purpose of standards-based resource management; A virtual node manager that controls the resources of the physical node to be used by various virtual systems and manages the states of the virtual nodes installed and operated in the physical node; A resource monitoring unit configured to perform standard-based system monitoring monitoring to enable heterogeneous interworking of resources of the entire system; A storage unit configured to store information for managing the states of the virtual nodes Multi-system workload management system comprising a. The system of claim 1, wherein the multi-system workload management system is And a physical node providing an environment of a virtual node performing migration in order to ensure service continuity according to the state of the virtual nodes in service. A method of managing a multi-system workload by dividing the virtual system into smaller virtual systems and configuring one system into multiple server systems, Reconfiguring resources of the physical node in consideration of a state and resources of a plurality of physical nodes in response to a request of user terminals connected to the Internet; Providing a service desired by user terminals through a virtual node mounted on the physical node in which the resource is reconfigured Multi-system workload management method comprising a. The method of claim 8, wherein reconfiguring resources of the physical node comprises: Collecting resource status of the physical node; Allocating and mounting a server of a virtual resource to the corresponding physical node in consideration of the resource state of the collected physical node; Multi-system workload management method comprising a. The method of claim 8, wherein the providing of the desired service by the user terminals comprises: Performing cluster management by checking a state of a virtual node disposed in the physical node; Multi-system workload management method further comprising.

Images