JP4239910B2 - Object deployment system and method - Google Patents

Description

  The present invention relates to a system and method for deploying objects in a distributed object system.

A distributed object technique is known in which functions of a system are divided and objects are distributed and used on computers on a network.
An example of a system that forms a distributed object and installs the formed distributed object in a distributed object system is disclosed in Patent Document 1.
In the system of Patent Document 1, the object program or user writes only the communication interface portion in the code. The installation means reads such a component and maintains a conversation state with the component that is the object execution location via the object request broker mechanism. Here, in the constructed communication band, an object transfer path is established by the component communication control interface. The server receives the distributed object transmitted from the installation means, creates its object reference, and registers it in the registration service means.
JP 09-54685 A (pages 9-11, FIG. 3)

However, in the system as described above, when installing the distributed object in the system, the first problem is that the user or the system administrator cannot select a free communication technique or a transmission control protocol or the like according to the requirements. There is.
The second problem is that in the conventional installation method, the transmission source and the transmission destination of the distributed object are defined as a one-to-one relationship, and the destination to which the deployment object is transmitted is the execution machine of the object. And not considering installation of deployment objects on multiple computers.
The third problem is that no consideration is given to controlling a communication path between a computer group in which objects are deployed and a client accessing the computer according to a load state of each computer.
In addition, the fourth problem is that a plurality of servers to be installed are planned to be installed (deployment) / uninstalled autonomously starting from the planned installation of deployment objects and the object deployment system. This means that installation (undeployment) processing cannot be performed.

The present invention has been made in view of the above circumstances, and an object of the present invention is to provide a system and method that allow a user such as an operator or an administrator to select various communication methods in deploying objects in a distributed object system. To do.
Another object of the present invention is to provide a system and method suitable for deploying objects to a group of computers.
Another object of the present invention is to provide a system, method, and the like that can control a communication path between a computer group in which an object is deployed and a client in accordance with a load situation.
Another object of the present invention is to enable planned deployment of objects to a group of computers.
Another object of the present invention is to provide a system, method, and the like that allow an object deployment system to autonomously deploy and undeploy objects from a computer group.

Objects deployment system according to a first aspect of this invention,
An object deployment system for deploying objects in a distributed object system,
Resource storage means for storing resource information indicating a computer group to which an object is deployed;
Task plan storage means for storing task reservation information in which a computer group, an object deployed in the computer group, and a reservation operation period of the object are associated;
When requesting object deployment , determine the communication protocol for connection to communicate with the server-side object deployment system . If the determined protocol is specific to the object deployment system, connect to the server side. Means for receiving a system connection operation request from a client requesting deployment of an object after requesting a system connection operation to establish a system connection operation, and calling the system connection operation;
After the system connection operation is invoked and a session with the client side is started, the object deployment request received from the client side is converted into a predetermined format, and the object deployment operation is performed based on the converted object deployment request. Means to call
In accordance with the called object deployment operation, means for specifying a computer group to which the object is deployed, and registering resource information indicating the computer group in the resource storage unit;
Means for selecting a computer group to which a specific object is deployed from a computer group indicated by the resource information stored in the resource storage means;
Means for generating task reservation information in which the selected computer group, the predecessor specific object, and the reservation operation period of the object are associated with each other, and registering the task reservation information in the task plan storage unit;
Means for specifying task reservation information to be changed from the task reservation information stored in the task plan storage means;
For the designated task reservation information, additional registration means for adding additional reservation information related to at least one of resource addition and release and registering in the task plan storage means;
Based on the task reservation information stored in the task plan storage means, means for deploying an object in a computer group for a predetermined period;
It is characterized by providing.

The client may have a stub that sends a request for the system connection operation;
The information processing apparatus may further include registration means for registering the object reference of the stub in a registry service that can be referred to by a client.

The server may have a stub for sending a request for the system connection operation;
Means for uploading the stub to the client may be provided.

The additional reservation information added to the task reservation information by the additional registration means may include operation period information.

Means for receiving load status information indicating the load status of each computer to which the object is deployed;
If the additional reservation information added to the task reservation information by the additional registration means is set with attribute information indicating that the deployment system determines execution of resource addition or release, the load status information and Means for performing at least one of object deployment and object undeployment based on the additional reservation information;
May be further provided.

  The information processing apparatus may further include a unit that distributes a load by distributing a request to each computer based on the load status information about each computer on which the object is deployed.

A means for connecting to a deployment system belonging to another network different from the network to which the object deployment system belongs may be further provided.

An object deployment method according to the second aspect of the present invention includes:
An object deployment method for deploying an object in a distributed object system, comprising:
A resource storage step for storing resource information indicating a computer group to which the object is deployed;
A task plan storage step for storing task reservation information in which a computer group, an object deployed in the computer group, and a reservation operation period of the object are associated;
On the client side, when the object deployment operation unit requests object deployment , a determination step of determining a connection communication protocol for communicating with the server-side object deployment system ;
If the protocol determined in the determination step is specific to the server-side object deployment system, sending a request for a system connection operation to establish a connection with the server side to the server side;
On the server side, receiving the system connection operation request and calling the system connection operation;
After the system connection operation is invoked and a client and server session is started, sending an object deployment request from the client side;
On the server side, receiving an object deployment request from the client side, converting the deployment request into a predetermined format, and invoking an object deployment operation based on the converted deployment request;
In response to the invoked object deployment operation, designating a computer group as a deployment destination of the object, registering resource information indicating the computer group in the resource storage step,
Selecting a computer group to which a specific object is deployed from a computer group indicated by the resource information stored in the resource storage step;
Generating task reservation information in which the selected computer group, the predecessor specific object, and the reservation operation period of the object are associated with each other, and registering the task reservation information in the task plan storage step;
Designating task reservation information to be changed from the task reservation information stored in the task plan storage step;
For the designated task reservation information, an additional registration step of adding additional reservation information related to at least one of resource addition and release and registering it in the task plan storage step;
Based on the task reservation information stored in the task plan storage step, deploying an object to a computer group for a predetermined period;
It is characterized by providing.

  According to the present invention, a user such as an operator or an administrator can select various communication methods when deploying objects in a distributed object system.

(First embodiment)
FIG. 1 is a physical configuration diagram of a distributed object system according to the first embodiment of the present invention. As illustrated, this system includes a plurality of computers 101, a routing server 103, a router device 107, and the like connected via a network. The computer 101, the routing server 103, the router device 107, and the like include a control unit including a CPU (Central Processing Unit), a ROM (Read Only Memory), a RAM (Random Access Memory), a storage unit including a hard disk device, and a communication control unit. Etc. respectively. A ROM, a storage unit, and the like of the computer 102 store a control program executed by the CPU, objects (encapsulated data and program codes (methods) for operating the data), and the like.

  In the distributed object system, the processes for processing the distributed object in each computer 101 connected to the network are mutually performed in the memory space in the same machine or in the memory space of two or more machines existing in the network. Can communicate. In general, the actual state of an object resides in a server-side process and functions to respond to an operation request from a client.

FIG. 2 is an example of a logical configuration diagram of the distributed object system according to the first embodiment of the present invention. When the CPU of the computer 101 executes a control program stored in a ROM, a storage unit, or the like or a control program acquired via a network, an installation operation unit 200, a first connection unit 201, a server stub 202, a deployment system 210, an object execution unit 240, a first connection service unit 211, a server skeleton 212, a connection skeleton 223, a connection stub 221, a second connection unit 230, a second connection service unit 231, a registry service unit 220, and the like. .
For example, the installation operation unit 200, the first connection unit 201, the server stub 202, the connection stub 221, and the second connection unit 230 are realized in the computer 101 on the client side, and the deployment system 210, the first connection service unit 211, The server skeleton 212, the connection skeleton 223, and the second connection service unit 231 are realized in the server-side computer 101, the object execution unit 240 is realized in the computer 101 on which the object is deployed, and the registry service unit 220 includes the connection stub 221. This is realized in the computer 101 in which object references such as are registered.

The installation operation unit 200 is for requesting deployment (installation) of an object, and functions as a part of a system management tool, for example.
The first connection unit 201 functions so as to link the server stub 202 and the connection stub 221 so that the installation operation unit 200 can connect to the server deployment system 210.
The server stub 202 and the connection stub 221 are linked to the first connection unit 201. The server stub 202 performs processing for transmitting a request to the server side in a format suitable for the distributed object system on the server side. The connection stub 221 performs processing for transmitting an object deployment request to the server side.
The deployment system 210 customizes and manages a distributed object in a configuration suitable for deployment (installation) on a machine to be executed, and deploys the object to the object execution unit 240.
The object execution unit 240 is a unit in which the substance of the distributed object functions as a process.

The first connection service unit 211 performs connection reception processing for the deployment system 210 to accept a connection from the installation operation unit 200 by linking with the server skeleton 212 and the connection skeleton 223.
The server skeleton 212 receives a request from the server stub 202 in a format suitable for the distributed object system, and requests the first connection service unit 211 responsible for connection reception processing to process it.
The server skeleton 212 and the connection skeleton 223 are linked to the first connection service unit 211. The server skeleton 212 corresponds to the server stub 202 and performs processing for receiving a request from a client in a format suitable for the distributed object system on the deployment system 210 side and passing it to the first connection service unit 211. The connection skeleton 223 corresponds to the connection stub 221 and performs processing for passing an object deployment request from a client to the first connection service unit 211.
The registry service unit 220 registers object references.

The connection stub 221 may be located in the registry service unit 220, for example. In this case, for example, the first connection service unit 211 registers the information of the connection stub 221 in the registry service unit 220 in the specific computer 101 on the network 100 so as to be disclosed as a reference point from the client. Also good.
For example, before the installation operation unit 200 accesses the deployment system 210, the connection stub 221 is located in the first connection service unit 211, and when communication from the installation operation unit 200 is started, One connection service unit 211 may be uploaded to the server stub 202 via the server skeleton 212 and linked to the first connection unit 201.

The second connection unit 230 adopts a communication method that conforms to internationally standardized communication protocols or industry standard standards, and includes the first connection unit 201, the server stub 202, and the connection stub 221. It has a function for connecting to the server side in place of the communication execution means.
The second connection service unit 231 employs a communication method that conforms to internationally standardized communication protocols or industry standard standards, and includes a first connection service unit 211, a server skeleton 212, and a connection skeleton. Instead of the communication execution means configured by H.223, it has a function for connecting to the client side.
The second connection unit 230 designates a connection route with the second connection service unit 231 by a protocol-specific method adopted by a user such as an operator or an administrator in the installation operation unit 200, and performs an operation command such as deployment. Or establish a transmission path.

Next, object deployment processing in the distributed object system according to the first embodiment will be described with reference to the flowchart of FIG.
First, the installation operation unit 200 in one computer 101 on the network receives a connection address including information such as the host name, IP address, and communication port number of the computer 102 (server) on which the object is deployed in response to an input operation by the user. Information is input (step A1). Note that the connection address information may include, for example, information such as a server group name and a domain name including a plurality of server groups on which objects are deployed.
Next, the installation operation unit 200 analyzes the input connection address information and selects a connection communication protocol (step A2). Then, the installation operation unit 200 determines whether the selected protocol is a protocol unique to the deployment system 210 (step A3).

If it is determined that the selected protocol is a communication unit unique to the deployment system 210 (step A3: YES), the installation operation unit 200 opens a path to the first connection unit 211 as an access destination to the server. In response to this, the first connection unit 211 connects to the server skeleton 212 in the other computer 101 (server) via the network via the server stub 202 based on the connection address information, and the first connection A session with the service unit 211 is started (step S4). The server stub 202 and the server skeleton 212 serve to start and end the connection between the first connection unit 201 and the first connection service unit 211.
The first connection service unit 211 attaches to the deployment system 210 and monitors a connection request from the first connection unit 201. The first connection unit 201 is usually located in a different memory space and a different machine from the deployment system 210, but is not limited thereto, and may be in the same machine, for example.

When a session between the first connection unit 201 and the first connection service unit 211 is started, an operation call preparation process is performed (step A5). For the operation call, a distributed object that defines an operation interface between a client and a server is used. The connection stub 221 of the distributed object is linked in advance to the first connection unit 201, for example.
Note that the connection stub 201 may be prepared by other methods. For example, when the connection stub 221 is not linked to the first connection unit 201 in advance and is stored in the first connection service unit 211, for example, in the initial process that the deployment system 210 starts, the first The connection service unit 211 may register with the registry service 220. At this point, the connection stub 221 can be referenced from any client. For example, the first connection unit 201 realized by the client program looks up the object reference of the connection stub 221 registered in the registry service unit 220 according to a method unique to the distributed object system currently in use. When the lookup is completed, it is loaded into the process space of the first connection unit 201, and the deployment operation via the operation interface becomes a callable state.
When the connection stub 221 is stored in the first connection service unit 211, for example, when the session between the server stub 202 and the server skeleton 212 starts, the first connection service unit 211 is connected by the session. The stub 221 may be transferred to the first connection unit 201. The first connection unit 201 dynamically loads the first connection unit 201 into its own process space when reception of the connection stub 221 is completed.

  The first connection service unit 211 can establish a connection simultaneously with a plurality of different first connection units 201. However, the first connection unit 201 can be connected to only one first connection service unit 211. The installation operation unit 200 can have a plurality of connections with connection service units (a first connection service unit 211 and a second connection service unit 231) existing in different computers 102.

  When the preparation for the operation call is completed, a deployment operation session is started between the installation operation unit 200 and the deployment system 210 (step A6). Thereafter, using this session, various operations are performed on the deployment system for deploying the distributed object, and the object is deployed to the designated machine (step A7).

  If it is determined in step A3 that the selected protocol is not a communication means unique to the deployment system 210 (step A3: NO), whether or not another standard communication protocol linked by the installation operation unit 200 can be used. Determine (step A8). At this time, if other standard communication protocols cannot be used and the category of the protocol that can be recognized by the installation operation unit 200 is exceeded (step A8: NO), the process ends as an error (step A10). If it is determined that another standard communication protocol can be used (step A8: YES), the second connection unit 230 constructs a message message suitable for the protocol and transmits it to the second connection service unit 231. This establishes a session (step A9). Thereafter, using this session, various operations are performed on the deployment system for deploying the distributed object (step A7).

  As described above, in the first embodiment, not only a distributed component object deployment method (protocol) provided by the deployment system 210 but also a user such as an operator or an administrator can select various communication protocols. It is configured. Since the user can select a communication protocol, a developer can create a pluggable connector or adapter whose function is expanded according to the application, and realize a highly scalable management architecture.

(Second Embodiment)
FIG. 4 is a physical configuration diagram of a distributed object deployment system according to the second embodiment of the present invention. As shown in the figure, in this system, a computer 420 that stores a program of the deployment system, a computer 410 that stores an operation program connected to the computer 420 via a network, and a computer 420 that connects to the computer 420 via a network. Cluster A430 and cluster B440. Here, a cluster is a group of computers that are grouped as one logical unit. The cluster A 430 includes, for example, a plurality of computers 431 (431a, 431b, 431c, 431d) and the like. The cluster B 440 includes, for example, a plurality of computers 431 (431d, 431e, 431f, 431g) and the like. Each of the computers 410, 420, and 431 includes a control unit including a CPU, ROM, RAM, a storage unit including a hard disk device, a communication control unit, and the like. Control programs and objects to be executed are stored.
The operation program in the computer 410 and the deployment system program in the computer 420 may be included in the same computer (computer 410 or 420). Further, the network to which the computer 410 and the computer 420 are connected may be the same as the network to which the computer 420 is connected to the cluster A 430 and the cluster B 431. In addition, the computers and the number of computers that the cluster A and the cluster B respectively have are only examples, and the present invention is not limited thereto.

  FIG. 5 is an example of a logical configuration diagram of a distributed object system according to the second embodiment of the present invention. In this system, an installation operation unit 500 is realized by the computer 410 executing a control program stored in a ROM, a storage unit, etc. in the computer 410, and a CPU is stored in the ROM, storage unit, etc. in the computer 420. A deployment system 510 is provided that is realized by executing a control program or the like.

  The installation operation unit 500 interactively requests the deployment system 510 to execute a command, and the connection service unit 511 to be described later is operated by an operation flow that is theoretically equivalent to that of the installation operation unit 200 in the first embodiment. connect. In addition to this connection function, the number of installation operation items increases according to the function added to the deployment system 510.

  The deployment system 510 includes a connection service unit 511, a deployment unit 512, a cluster management unit 520, and a storage unit 530. Furthermore, the cluster management unit 520 includes a resource registration unit 521, a task creation unit 522, a task reservation unit 523, and a schedule registration unit 524. The storage unit 530 includes a resource storage unit 531 and a task plan storage unit 532.

The connection service unit 511 has the same functions as the first connection service unit 211 and the second connection service unit 231 in the first embodiment, and executes request reception monitoring processing.
The deployment unit 512 has substantially the same function as the deployment system 210 in the first embodiment. The component object is deployed to the object execution unit in the deployment destination server computer to activate the object, which is the same as in the first embodiment, but the object is deployed according to the information about the server computer stored in the storage unit 530. This is different from the first embodiment.
The resource registration unit 521 inputs information about a server computer (deployment destination computer) that operates the component object, outputs information about the grouped computer group (cluster), and stores the information in the resource storage unit 531.
The task creation unit 522 associates and outputs the component object to be a task and the processing target cluster. The list of clusters is read from the resource storage unit 531.
The task reservation unit 523 generates task reservation information by adding the time information indicating the start time and stop time of the task to the output data (the object and the cluster are associated with each other) output by the task creation unit 522 And stored in the task plan storage unit 532.
The schedule registration unit 524 adds and releases resources to the already reserved tasks registered by the task reservation unit 523.

Next, operations in the distributed object deployment system according to the second embodiment will be described with reference to FIG.
First, the installation operation unit 500 and the deployment system 510 perform connection processing in the same manner as in the first embodiment. In a connected state, the installation operation unit 500 requests a deployment operation from the deployment system 510, and in response to this request, In response, the deployment system 510 registers the computer resources that can be deployed first in cluster units (step B1). For example, values of “cluster A = server C1, server C2” and “cluster B = server C3” are input and stored in the resource storage unit 531.

  Next, the deployment system 510 creates a component object processing task (step B2). Further, the cluster name for executing the created task is acquired from the resource storage unit 531. Then, by inputting the information of the created task (task ID, etc.), the acquired cluster name, the name of the component that requests processing to the cluster, and the information of the start / end time, and generating and registering the task reservation information, Reservation is executed (step B3). The input value is, for example, “task 1 = shopping component, 10: 0 to 16:30, cluster A”. The task reservation information is stored in the task plan storage unit 532.

  Further, when changing the execution machine configuration of the component object, the deployment system 510 reads the registered task from the task plan storage unit 532 and adds or releases computer resources. That is, the deployment system 510 selects a target task from the registration data in the task plan storage unit 532 (step B4), performs necessary input, and updates the registration content (step B5). For example, when adding a resource for a certain task, target task reservation information is specified, and an additional cluster name and start / end time are input for the task (for example, “task 1 = shopping component, 13 : 00-14: 30, cluster B "). The input value is additionally stored in the task plan storage unit 532 as additional reservation information. As a result, information related to the task plan is registered in the task plan storage unit 532 and is changed as necessary.

  Then, when the deployment unit 512 in the deployment system 510 reaches the corresponding time (start time set in the task reservation information) according to the task reservation information registered in the task plan storage unit 532, the component is transferred to the server computer under the cluster. The deployment process is started (step B6). In addition, when the corresponding time (end time set in the task reservation information) is reached, the deployment unit 512 performs a process of canceling the deployment of the object.

  As described above, according to the second embodiment, component objects can be deployed to a plurality of server computers (clusters). In addition, since the function to deploy distributed component objects to server components and the function to manage deployable server computers are configured separately in the deployment system, operational costs are reduced by integrating computer systems. be able to. In addition, since the operation period of the component object can be set, computer resources can be systematically allocated, and in addition, business continuity with respect to a load that can be predicted by the resource schedule reservation function can be maintained.

(Third embodiment)
FIG. 7 is a physical configuration diagram of an object deployment system according to the third embodiment of the present invention. As illustrated, the system includes a computer 710 that stores an operation program, a computer 720 that stores a program of a deployment system, a cluster A 730, and a cluster B 740, which are connected via a network. The cluster A 730 includes, for example, a computer 731 (731a, 731b, 731c, etc.) and a load balancer 732. The cluster B 740 includes, for example, a computer 731 (731b, 731c, etc.). Cluster B 740 is integrated with cluster A 730, for example. The configuration of this system is substantially the same as that of the second embodiment except that a load distribution device 732 is provided.

  FIG. 8 is an example of a logical configuration diagram of an object deployment system according to the third embodiment of the present invention. Of the components of this system, the same components as those of the second embodiment (for example, the installation operation unit 500, the connection service unit 511, etc.) will not be described.

The deployment system 810 includes a connection service unit 511, a cluster management unit 820, a storage unit 830, a deployment unit 813, a load information reception unit 814, and a load distribution request unit 815.
The connection service unit 511 is the same as that in the second embodiment.
The cluster management unit 820 includes a resource registration unit 521, a task creation unit 522, a task reservation unit 523, and a schedule registration unit 824 similar to those in the second embodiment. In addition to the function of the schedule registration unit 524 in the second embodiment, the schedule registration unit 824 has a function of inputting and setting parameters (additional parameters) for starting and ending the deployment process based on the determination of the deployment system 810. Prepare.
The storage unit 830 includes the second embodiment, a resource storage unit 531, and a task plan storage unit 832. The task reservation information registered in the task plan storage unit 832 has a data structure in which an additional parameter (attribute information) field is further added to the task reservation information registered in the task plan storage unit 532 in the second embodiment. .

In addition to the function of the deployment unit 512 in the second embodiment, the deployment unit 813 performs the deployment and dedeployment of the distributed component object under the instruction of the load information reception unit 814, so that the operating component resource Add, degenerate.
The load information receiving unit 814 receives information indicating the load increase / decrease of the server computer from the load monitoring unit 821, autonomously determines the number of server computers to be operated according to the load status, and sends the information to the deployment unit 813. Request resource scale-up / scale-down.
When the load information receiving unit 814 detects a load imbalance state of the server computer in a situation where it is difficult to scale up the server computer, the load distribution request unit 815 sends the load distribution request unit 815 to the server computer in an overloaded state. The balance adjustment configuration information indicating the weighting value of each computer that restricts the inflow of client access and increases the inflow of client access to the server computer in a low load state is generated, and the generated information is transmitted to the load balancer 732. Implement dynamic load balancing by sending and setting.

In the present embodiment, each computer 731 in the cluster A and the cluster B includes a load monitoring unit 821 that transmits computer resource load information such as a CPU usage rate and a free memory amount to the load information receiving unit 814.
The load distribution device 732 performs load distribution by dynamically switching the path from the client to the server according to the load status of the server computer. For example, a device such as a load balancer or an L4 / L7 switch may be used, and a program for performing load balancing may be stored in one computer 731 and executed.

Next, object deployment processing in the distributed object deployment system according to the third embodiment will be described with reference to FIG.
First, in the deployment system 810, it is assumed that task reservation information that can be assigned to computers in cluster A from 10:00 to 16:30 is created and registered as an initial reservation. The deployment system 810 makes an additional reservation for computer resources in preparation for scale-up for this task (step C1). At this time, the input value is appended with an OnDemand parameter indicating that the object is to be scaled up, such as “task 1 = shopping component, OnDemand, cluster B”.
Then, it is assumed that the task 1 registered in the deployment system 810 starts at 10:00, and the server computer S1 in the cluster A detects an overload at a certain time. Then, the load monitoring unit 821 in the computer S1 notifies the load information receiving unit 814 of the deployment system 810 of the load increase (Step C2). Recognizing the increase in load, the load information receiving unit 814 instructs the deployment unit 813 to deploy the component object to the cluster B in order to scale up the server computer.

  If, for example, only the server computer S2 of the cluster B is notified of an overload at a certain time (step C3), the load information receiving unit 814 processes the load balance in the load distribution requesting unit 815. Thus, the reconfigured distribution configuration is set in the load distribution device 732 (for example, “server S1, S3 → inflow increase, S2 → inflow decrease” is set in the load distribution device 732). Later, when a low load is detected in all the server computers, the server computer is instructed to cancel the component deployment (for example, to cancel the deployment of one server in the cluster B) (step C4). Furthermore, if there is a low load detection, the component deployment cancellation of other server computers in cluster B is instructed in the same manner as in step C4 (step C5).

As described above, according to the third embodiment, since the deployment system and the server computer are configured to share information, the processing target computer of the distributed component object is autonomously scaled up and scaled down. it can. As a result, the operation cost due to the improved operating rate is reduced. In addition, business continuity against unexpected loads is improved.
Further, according to the present embodiment, since the load of each computer operated by the component object is configured to be monitored, the load is automatically sent to the computer to another server as the response time of the computer resource decreases. When the server computer group is regarded as one system, it is possible to provide a stable and unified access point for a server operating in a cooperative manner.

(Fourth embodiment)
FIG. 10 is a schematic diagram of a distributed object system according to the fourth embodiment of the present invention. As illustrated, the distributed object system according to the present embodiment has a configuration in which two or more distributed object systems are connected to the third embodiment via a network. When the system according to the present embodiment is referred to as one hosting environment, the system according to the third embodiment includes, for example, three hosting environments: a hosting environment A, a hosting environment B, and a hosting environment C.

  The deployment system 1020 in each hosting environment has substantially the same function and configuration as the deployment system 810 in the third embodiment, but is connected from a connection unit 1022 for connecting to another hosting environment from another hosting environment. And a connection service unit 1023 for the purpose. The connection unit 1022 and the connection service unit 1023 are the connection unit (first connection unit and second connection unit) and the connection service unit (first connection service unit and second connection service unit) in the first embodiment. It has the same function and configuration. The mutual deployment system 1020 performs self-diagnosis, other self-diagnosis, etc. on the life cycle state (load status of each server machine, deployment / cancellation status of each object, etc.) At the same time, a deployment system 1020 is linked to another deployment system 1020, thereby forming a grid system that shares information on server machines and component objects that are geographically distributed. The configuration of each hosting environment and the processing function of each component are substantially the same as in the third embodiment.

  A user such as an administrator of this object deployment system or an administrator of a distributed component object (server-side application program) deploys an object using the operation unit 1010 realized by starting an operation program stored in a computer. Control. Specifically, for example, the user sets the accommodation plan and the advance plan for the network, server computer, application, etc. in the hosting environment A. At this time, machines in different hosting environments B and C are set and registered in the distribution system 1020 in addition to the cooperation resources. When such other hosting environments are also added to the resource, an authentication mechanism may be used to authenticate the user in order to maintain security. When the setting for such cooperation is completed, the deployment system 1020 of the hosting environment A sends the server computers included in both the hosting environment B and the hosting environment C to any timing (reservation time) as described above. Or the timing determined by the judgment of the distribution system).

  For example, when it is detected that an increase in the processing load of a server computer that is currently operating in one hosting environment (for example, hosting environment A) at a certain point in time, or it is difficult to continue processing by detecting some kind of failure. When the determination is made, the hosting environment deployment system 1020 connects to the connection service unit 1023 of another hosting environment (for example, hosting environment B) via the connection unit 1022. Once the connection is established and ready, the hosting environment A deployment system 1020 requests the hosting environment B deployment system to deploy additional component objects. In response to this, the deployment system of the hosting environment B immediately allocates a machine from the server computers under the hosting environment and deploys the component. Since access from the client terminal is distributed to the two hosting environment machines, the business can be continued.

  As described above, the feature of the deployment system having a plurality of connection portions can be used for other hosting environments to be connected, for example, information resources can be interchanged between projects and departments. This is because the task information and load information registered in the deployment system, and the computer configuration and the resource information of the computer itself can mutually refer to each other between the deployment systems. In such a grid-structured system, distributed component objects can be operated at a global position, and a large-scale calculation program that requires enormous processing costs and time can be completed in a short time. . Furthermore, from the viewpoint of resource management, it is easy to integrate and grasp resources managed separately by each management cluster or domain, department, and organization.

  As described above, according to the present invention, a device that deploys distributed objects between clusters of application servers that are connected to each other, an operation management middleware that monitors job adjustment between server clusters and the load on each server, etc. Applicable to usage. It can also be applied to Internet data center operators that provide hosting services and housing services.

The system of the present invention can be realized using a normal computer system, not a dedicated system. For example, a program for executing the above operation is stored in a computer-readable recording medium (FD, CD-ROM, DVD, etc.) and distributed, and the program is installed in the computer to execute the above processing. Computers 101, 410, 420, 710, 720, 731, 732 and the like may be configured. Alternatively, it may be stored in a disk device of a server device on a network such as the Internet and downloaded to a computer, for example.
In addition, when the OS realizes the above functions by sharing the OS or jointly with the OS and the application, etc., only the part other than the OS may be stored and distributed in the medium, or may be downloaded to the computer. Good.

It is a figure which shows the physical structural example of the distributed object system which concerns on the 1st Embodiment of this invention. It is a figure which shows the logical structural example of the distributed object system which concerns on the 1st Embodiment of this invention. It is a flowchart for demonstrating the processing operation of the distributed object system which concerns on 1st Embodiment. It is a figure which shows the physical structural example of the distributed object system which concerns on the 2nd Embodiment of this invention. It is a figure which shows the logical structural example of the distributed object system which concerns on the 2nd Embodiment of this invention. It is a figure for demonstrating the processing operation of the distributed object system which concerns on the 2nd Embodiment of this invention. It is a figure which shows the example of a physical structure of the distributed object system which concerns on the 3rd Embodiment of this invention. It is a figure which shows the logical structural example of the distributed object system which concerns on the 3rd Embodiment of this invention. It is a figure for demonstrating the processing operation of the distributed object system which concerns on the 3rd Embodiment of this invention. It is a schematic diagram of the distributed object system which concerns on the 4th Embodiment of this invention.

Explanation of symbols

101, 410, 420, 431, 710, 720, 731 Computer 103 Routing server 107 Router device 200, 500 Installation operation unit 201 First connection unit 202 Server stub 210, 510, 810, 1020 Deployment system 240 Object execution unit 211 No. One connection service unit 212 Server skeleton 223 Connection skeleton 221 Connection stub 230 Second connection unit 231 Second connection service unit 220 Registry service unit 430, 730 Cluster A
440, 740 Cluster B
511 Connection service unit 512, 813 Deployment unit 520, 820 Cluster management unit 521 Resource registration unit 522 Task creation unit 523 Task reservation unit 524, 824 Schedule registration unit 530, 830 Storage unit 531 Resource storage unit 532 Task plan storage unit 814 Load information Reception unit 815 Load distribution request unit 820 Cluster management unit 832 Task plan storage unit 1022 Connection unit 1023 Connection service unit

Claims (8)

An object deployment system for deploying objects in a distributed object system,
Resource storage means for storing resource information indicating a computer group to which an object is deployed;
Task plan storage means for storing task reservation information in which a computer group, an object deployed in the computer group, and a reservation operation period of the object are associated;
When requesting object deployment , determine the communication protocol for connection to communicate with the server-side object deployment system . If the determined protocol is specific to the object deployment system, connect to the server side. Means for receiving a system connection operation request from a client requesting deployment of an object after requesting a system connection operation to establish a system connection operation, and calling the system connection operation;
After the system connection operation is invoked and a session with the client side is started, the object deployment request received from the client side is converted into a predetermined format, and the object deployment operation is performed based on the converted object deployment request. Means to call
In accordance with the called object deployment operation, means for specifying a computer group to which the object is deployed, and registering resource information indicating the computer group in the resource storage unit;
Means for selecting a computer group to which a specific object is deployed from a computer group indicated by the resource information stored in the resource storage means;
Means for generating task reservation information in which the selected computer group, the predecessor specific object, and the reservation operation period of the object are associated with each other, and registering the task reservation information in the task plan storage unit;
Means for specifying task reservation information to be changed from the task reservation information stored in the task plan storage means;
For the designated task reservation information, additional registration means for adding additional reservation information related to at least one of resource addition and release and registering in the task plan storage means;
Based on the task reservation information stored in the task plan storage means, means for deploying an object in a computer group for a predetermined period;
An object deployment system comprising: The client has a stub for sending a request for the system connection operation;
Registration means for registering the object reference of the stub in a registry service that can be referred to by the client;
The object deployment system according to claim 1 . The server has a stub for sending a request for the system connection operation;
Means for uploading the stub to the client;
The object deployment system according to claim 1 . The additional reservation information added to the task reservation information by the additional registration means includes information on an operation period.
The object deployment system according to any one of claims 1 to 3, wherein Means for receiving load status information indicating the load status of each computer to which the object is deployed;
If the additional reservation information added to the task reservation information by the additional registration means is set with attribute information indicating that the deployment system determines execution of resource addition or release, the load status information and Means for performing at least one of object deployment and object undeployment based on the additional reservation information;
The object deployment system according to any one of claims 1 to 4 , further comprising: Further comprising means for distributing a load by distributing a request to each computer based on the load status information for each computer on which an object is deployed,
The object deployment system according to claim 5 . Means for connecting to a deployment system belonging to another network different from the network to which the object deployment system belongs,
Objects deployment system according to any one of claims 1 to 6, characterized in that. An object deployment method for deploying an object in a distributed object system, comprising:
A resource storage step for storing resource information indicating a computer group to which the object is deployed;
A task plan storage step for storing task reservation information in which a computer group, an object deployed in the computer group, and a reservation operation period of the object are associated;
On the client side, when the object deployment operation unit requests object deployment , a determination step of determining a connection communication protocol for communicating with the server-side object deployment system ;
If the protocol determined in the determination step is specific to the server-side object deployment system, sending a request for a system connection operation to establish a connection with the server side to the server side;
On the server side, receiving the system connection operation request and calling the system connection operation;
After the system connection operation is invoked and a client and server session is started, sending an object deployment request from the client side;
On the server side, receiving an object deployment request from the client side, converting the deployment request into a predetermined format, and invoking an object deployment operation based on the converted deployment request;
In response to the invoked object deployment operation, designating a computer group as a deployment destination of the object, registering resource information indicating the computer group in the resource storage step,
Selecting a computer group to which a specific object is deployed from a computer group indicated by the resource information stored in the resource storage step;
Generating task reservation information in which the selected computer group, the predecessor specific object, and the reservation operation period of the object are associated with each other, and registering the task reservation information in the task plan storage step;
Designating task reservation information to be changed from the task reservation information stored in the task plan storage step;
For the designated task reservation information, an additional registration step of adding additional reservation information related to at least one of resource addition and release and registering it in the task plan storage step;
Based on the task reservation information stored in the task plan storage step, deploying an object to a computer group for a predetermined period;
An object deployment method comprising:

Images