Dynamic configuration controller and dynamic configuration control method

Description: TECHNICAL FIELD [Detailed description of the invention]

【0001】

[Technical field to which the invention belongs]

The present invention relates to configuration control of a Web system whose system scale is difficult to predict, and in particular, dynamic configuration control capable of determining the operating status of a Web server and an application server and dynamically distributing a service configuration to each server. Involved in equipment and dynamic configuration control methods.

【0002】

[Conventional technology]

With the technological innovation of the Internet and the development of network infrastructure, redundancy and load distribution technologies for the safe and smooth operation of Web systems have been established. By the way, in a Web system in which the number of accesses changes drastically due to all circumstances such as current affairs and seasons, it is almost impossible to accurately predict the number of accesses, and it is difficult to determine the timing of system strengthening and cost burden. In such an environment, there is a demand for a mechanism that can operate a Web system of an appropriate system scale that has less investment risk and can prevent loss of market opportunities.

【0003】

A Web system consists of a Web server, an application server, and a database server. The Web server is connected to a network such as the Internet and mainly serves to accept service requests. The application server is connected to the Web server and executes the actual business logic. In addition, the application server is connected to the database server and accesses the database as needed to carry out business.

【0004】

In this way, the access related to the service request of the client is performed in the order of the Web server, the application server, and the database via the network. Depending on the content of the service, the business may be completed only by the Web server, or there may be business that does not require access to the data server. In some cases, one physical server is used for all operations, and in some cases, there is one Web server, one application server, and one database server.

【0005】

Figure 7 shows the initial Web system. Figure 7 shows that the Web system consists of a Web server 11, an application server 12, and a database server 13 and is connected to the network. The Web server 11, the application server 12, and the database server 13 are connected one-to-one-to-one and execute one or more service operations.

【0006】

Early Web systems have a relatively small business scale, which is advantageous when starting a business quickly. As the service business expanded and system scalability and redundancy were required, the Web system became layered in the next stage. Figure 8 shows a layered Web system. In FIG. 8, a load distribution device 14 for allocating individual services to each server is provided, and the Web system is layered between the application server 12 and the database server 13, and the front end Web server 11, the application server 12, and the back are provided. It has a two-layer structure of the database server 13 which is the end, and a plurality of application servers 12 share the database server 13.

【0007】

As the business expanded in this way, the number of servers in the Web system increased, the front end became redundant, and the services performed by each server were distributed. Also, relatively large databases are now shared by multiple applications. Furthermore, by separating the front end and the back end, each business can be specialized, and a large number of accesses can be processed efficiently.

【0008】

The two-tiered Web system in Fig. 9 is another configuration example of the layered Web system. In this way, in a layered Web system, the load distribution device 14, the Web server 11, and the database server 13 may be physically connected to the same network to form a LAN (local area network). When the front end and back end of the Web system are layered in this way and a large number of accesses are efficiently accepted, access to the database from multiple applications becomes a bottleneck. In order to solve this problem, the Web server and the application server were separated in the next stage, and the Web system was made into three layers.

【0009】

Figure 10 shows a three-tiered Web system. In FIG. 10, the Web system is further layered by the application server 12, and has a three-layer structure of the Web server 11, the application server 12, and the database server 13, and a plurality of application servers 12 share the database server 13. The three-tiered Web system in Fig. 11 is another configuration example of the three-tiered Web system. In this way, even in a three-tiered Web system, the load distribution device 14, the Web server 11, the application server 12, and the database server 13 may be physically connected to the same network to form a LAN.

【0010】

By layering between the Web server and the application server and layering the Web system in three layers in this way, each business can be specialized and a large number of accesses can be processed more efficiently. .. The initial Web system, the layered Web system, and the three-layer Web system, and the method of enhancing the entire Web system step by step have been established to a certain level.

【0011】

[Problems to be Solved by the Invention]

However, in either case, it was necessary to determine the system configuration including the server configuration and the service configuration when constructing the system before providing the service. When starting a Web service, determine the system configuration at the design stage, and after starting the Web service, increase the number of servers or services as necessary to maintain the system. The enhancement was done.

【0012】

On the other hand, even if the service is started with the minimum system, some system equipment is required to operate the system that is resistant to both load fluctuations and failures of each server. Therefore, some servers may remain idle until the service operation gets on track. As described above, in the conventional Web system, it is difficult to predict the fluctuation of the system scale, and it is difficult to determine the cost burden and the timing for increasing or degenerating the Web server or application server.

【0013】

Further, in the conventional Web system, since the service configuration is manually changed while monitoring the operating status of the system, a certain delay time occurs. In addition, if the system scale is large, the time lag due to manual intervention will become even larger, and as a result, it will not be possible to cope with the increase in the number of instantaneous wind speed accesses and the concentration of load on some servers. There was a problem.

【0014】

On the contrary, the number of access to the service decreases, and even if the system configuration or service configuration is degraded manually, the system resources can be effectively used until the system configuration or service configuration change is completed. It was hard to say that it was. In view of these points, the present invention is involved in the configuration control of a Web system whose system scale is difficult to predict, determines the operating status of the Web server and the application server, and dynamically distributes the service configuration to each server. It is an object of the present invention to provide a dynamic configuration control device and a dynamic configuration control method that can be used.

【0015】

[Means for solving problems]

The dynamic configuration control device of the present invention is a dynamic configuration control device that determines the operating status of the servers constituting the web system connected to the network and dynamically distributes the service configuration to each server based on the rules. The load distribution means for allocating services to the servers and the operating status of each server are monitored, the service configuration of each server is determined, the service configuration is assigned to the servers according to the operating status, and the system configuration is changed. The traffic management means, the operation monitoring means provided in the server and notifying the traffic management means of the operation status of the service running on the own server, and the operation monitoring means provided in the server and receiving instructions from the traffic management means, the own server It has a configuration management means for changing the service configuration of the server and starting the service, and a resource management means for supplying service resources to the configuration management means.

【0016】

Further, the dynamic configuration control device of the present invention is a dynamic configuration control device that determines the operating status of the servers constituting the web system connected to the network and dynamically distributes the service configuration to each server, and is a rule. Load distribution means for allocating services to servers based on the above, monitoring the operating status of each server, determining the service configuration of each server, reducing the service configuration to the server according to the operating status, and changing the system configuration. A traffic management means for performing the above, an operation monitoring means provided in the server and notifying the traffic management means of the operation status of the service running on the own server, and an operation monitoring means provided in the server and receiving instructions from the traffic management means. It has a configuration management means for changing the service configuration of its own server and releasing the service.

【0017】

Further, the dynamic configuration control method of the present invention is a dynamic configuration control method that determines the operating status of the servers constituting the web system connected to the network and dynamically distributes the service configuration to each server, and is a rule. The step of allocating services to the server based on, the step of monitoring the operating status of each server and determining the service configuration, and when assigning the service configuration to the server, instruct the server to assign the service configuration. It has a step of giving, a step of supplying service resources to the server, a step of the server changing the service configuration and starting the service, a step of changing the rules, and a step of changing the system configuration. When decompressing the configuration, a step to change the rule, a step to give an instruction to release the service to the server, a step to release the service by the server and change the service configuration, and a step to change the system configuration. And have.

【0018】

FIG. 1 is a principle diagram of the present invention, and is a configuration diagram of the dynamic configuration control device of the present invention. 1 in FIG. 1 is a load balancing means, which is called a load balancing device or a load balancer (LB). 2 is a server that acts as a Web server or application server, and includes 21 operation monitoring means and 22 configuration management means. Server 2 is called a Dynamic Link Server (DLS) and runs one or more services. 3 is a traffic management means, which is a server having a storage means for storing the system configuration of the entire Web system and the operating conditions of services. 4 is a resource management means, which is a server having a storage means for storing service resources and service operation information.

【0019】

The load distribution device 1 is connected to the network and each server 2, and services are assigned to each server 2 based on the rules stored in the load distribution device 1. The load distribution device 1 is also connected to the traffic management server 3, and changes the rules according to the instruction from the traffic management server 3. Server 2 executes the directed service. You can run multiple services on one server. In the figure, the part marked running indicates that the service is running, and the part marked idling indicates that the service is not running.

【0020】

The operation monitoring means 21 of the server 2 is connected to the traffic management server 3 and notifies the traffic management server 3 of the operation status of the service running on the local server. The configuration management means 22 of the server 2 is connected to the traffic management server 3 and the resource management server 4, and when the traffic management server 3 receives an instruction to change the service configuration, the resource management server 4 operates the service resources and services. Request information. The resource management server 4 supplies service resources and service operation information to the configuration management means 22.

【0021】

Figure 2 shows the operation flow diagram of the dynamic configuration control device. First, the traffic management server 3 that monitors the operation status of each server 2 receives a notification about the operation status from the operation monitoring means 21 of each server 2. (Step S1) The traffic management server 3 determines the operating status of whether the server is in the steady state, the high load state, or the idle state according to the service configuration of the notified server 2, and if it is not in the steady state, the server 2 is notified. On the other hand, assign or degenerate the service configuration. If it is in a steady state, the process returns to step S1. (Step S2) When allocating the service configuration, the traffic management server 3 instructs the server 2 that can accept the additional allocation to that effect. (Step S3) Upon receiving an instruction from the traffic management server 3 to allocate the service configuration, the acceptable configuration management means 22 of the server 2 requests the resource management server 4 to acquire the service resource and the operation information of the service (step S4). The resource management server 4 that receives the acquisition request supplies the service resource and the operation information of the service to the configuration management means 22 that is the request source. (Step S5) The configuration management means 22 starts the service based on the service resource and the operation information of the service. After starting the service, the configuration management means 22 notifies the traffic management server 3 to that effect. (Step S6) When the traffic management server 3 that receives the notification instructs the load distribution device 1 to change the rules for service distribution, the load distribution device changes the rules. (Step S7) Next, the traffic management server 3 changes the system configuration of the entire Web system. (Step S8) The load distribution device 1 distributes the service to each server 2 according to the changed rule. Then, the process returns to step S1.

【0022】

On the other hand, when degenerating the service configuration, the traffic management server 3 instructs the load distribution device 1 to change the rules for service distribution. (Step S9) The load distribution device 1 distributes the service to each server 2 according to the changed rule. Subsequently, the traffic management server 3 instructs the configuration management means 22 of the server 2 to release the service. (Step S10) Upon receiving an instruction to release the service configuration from the traffic management server 3, the configuration management means 22 releases the service. (Step S11) After the service is released, the configuration management means 22 notifies the traffic management server 3 to that effect. Upon receiving the notification, the traffic management server 3 changes the system configuration of the entire Web system. Then, the process returns to step S1. (Step S12) In this way, when it becomes possible to automate the start and release of services, the appropriate number of servers and the service configuration are dynamically determined according to the usage status after the start of service operation. It is possible to eliminate the conventional manual delay time for system expansion or degeneration, and the efficiency of the server, which is overloaded due to a rapid increase in the number of accesses or load concentration, or conversely, is idle with few accesses. This has the effect of enabling effective use of system resources.

【0023】

As described above, in the present invention, in a Web system in which it is difficult to predict fluctuations in the system scale, the service configuration of the server is operated according to the load status of the server when the Web server and the application server constituting the Web system are used. Can be assigned to the target. This makes it possible to operate a system that is resistant to both load fluctuations and failures.

【0024】

BEST MODE FOR CARRYING OUT THE INVENTION

Examples of the present invention will be described in detail with reference to the drawings. Figure 3 is a configuration diagram of a dynamically linked Web system. 1 in Fig. 3 is a load balancer and is called a load balancer (LB). Reference numeral 2 denotes a server such as a Web server (W1, W2, Wn) or an application server (A1, A2 ,,), which includes the operation monitoring means 21 and the configuration management means 22 of the present invention.

【0025】

Server 2 constitutes a three-tier Web system, and each server operates because it configures a dynamic link that does not depend on business applications between the Web server (W1, W2, Wn) and the application server (A1, A2 ,,). It is called a target link server, and one unit executes multiple services. 3 is a traffic management server, which has a storage means for storing the system configuration of the entire Web system and the operating conditions of services. Reference numeral 4 is a resource management server, which has a storage means for storing service resources and service operation information.

【0026】

To determine whether the server is in a steady state, a high load state, or an idle state based on the performance and capacity of the server, the type of service, and the configuration of the service from the system configuration and service operating conditions. , The policy for each server, that is, the distribution policy is determined. From the distribution policy, the rule for allocating the service when the Web system receives a processing request, that is, the distribution rule is determined.

【0027】

The load distribution device 1 is connected to the network and the server 2, relays the destination according to the distribution rule stored in the load distribution device 1, and distributes the service to the server 2. The load distribution device 1 is also connected to the traffic management server 3, and the distribution rule is dynamically changed according to the instruction from the traffic management server 3. The Web system in Figure 4 shows that the dynamically linked Web system is connected by LAN. Server 2 logically constitutes a three-tiered Web system with a Web server (W1, W2, Wn), an application server (A1, A2 ,,) and a database server (D1, D2 ,,), but load distribution. It is physically connected to the same network together with device 1, traffic management server 3, and resource management server 4, and constitutes a LAN.

【0028】

The start of the service in FIG. 5 is a sequence diagram in the case of allocating additional services. The dynamically linked server in Figure 5 is either a web server or an application server, or both. The dynamic link server has an operation monitoring means and a configuration management means. The operation monitoring means of the dynamic link server is connected to the traffic management server, and notifies the traffic management server of the operation status of the service running on the own server.

【0029】

Based on the distribution policy, the traffic management server determines whether the server is in a high load state according to the service configuration of the notified server, and continues monitoring if there is no problem in the operating state. Whether the server is in a steady state, a high load state, or an idle state depends on the server performance such as the BUSY rate of the CPU and the operating capacity such as the number of accesses to the services performed by the server. If the server that received the notification is in a high load state, the service configuration is assigned to other servers that have free operating capacity.

【0030】

When the traffic management server assigns the service configuration, it instructs the configuration management means of the dynamic link server. Upon receiving an instruction from the traffic management server to add or change the service configuration, the configuration management means requests the resource management server for service resources and service operation information. The resource management server supplies the requested service resources and service operation information to the requesting configuration management means.

【0031】

The configuration management means starts the service based on the service resource and the operation information of the service. After starting the service, the configuration management means notifies the traffic management server that the service has started. Upon receiving the notification, the traffic management server instructs the load distribution device to change the distribution rule. The load distribution device changes the distribution rule and adds the service destination.

【0032】

The traffic management server then changes the system configuration of the entire Web system. The load distribution device distributes the service to each destination server according to the changed distribution rule. The release of the service in FIG. 6 is a sequence diagram in the case of degenerating the service. First, the operation monitoring means of the dynamic link server notifies the traffic management server of the operation status of the service running on the own server to the traffic management server.

【0033】

Based on the distribution policy, the traffic management server determines the operating status of the server according to the service configuration of the notified server. If it is determined that the load of a specific service is low for the server that received the notification, or if it is determined that the number of running servers for a specific service is excessive, the server that degenerates the service operation is determined.

【0034】

When the traffic management server degenerates the service configuration, it instructs the load distribution device to change the distribution rule. The load distribution device changes the distribution rule and reduces the destination of the service. After that, the load distribution device distributes the service to each destination server according to the changed distribution rule. Subsequently, the traffic management server instructs the configuration management means of the dynamically linked server to release the service. The configuration management means releases the service when it receives an instruction to release the service configuration from the traffic management server.

【0035】

After the service is released, the configuration management means notifies the traffic management server that the service release is completed. The traffic management server that receives the notification changes the system configuration of the entire Web system. In this way, in a Web system including a Web server and an application server, it is possible to allocate a specific service to another server before the specific server is overloaded. On the other hand, if the load on the server is low or the number of running servers is excessive, the service operation is degraded.

【0036】

[Effect of the invention]

As is clear from the above description, in the dynamic configuration control device and the dynamic configuration control method of the present invention, even in a Web system in which load fluctuation is large and it is difficult to predict the system scale, the dynamic configuration control device and the dynamic configuration control method can be used immediately according to the load status. It has the effect of making it possible to change the service configuration dynamically.

【0037】

That is, according to the present invention, it is possible to ensure smooth operation on an appropriate scale without obstacles by a mechanism for dynamically allocating resources when the system is used. For Web system operators, it is possible to avoid wasteful investment in specific services, avoid overinvestment, and instantly change the service configuration without relying on human resources, resulting in loss of market opportunities. It can be avoided. On the other hand, for Web system users, it has the effect of contributing to the elimination of the acceptance waiting time.

[Simple explanation of drawings]

[Figure 1]

Principle of the present invention [Figure 2]

Operation flow diagram of dynamic configuration controller [Fig. 3]

Dynamic linking web system [Fig. 4]

Web system [Fig. 5]

Service startup [Fig. 6]

Release of service [Fig. 7]

Early web system [Fig. 8]

Hierarchical web system [Fig. 9]

2-tier web system [Fig. 10]

3-tier web system [Fig. 11]

3-tier web system [Explanation of symbols]

1 Load distribution means, load distribution device, LB 2 server 3 Traffic management means, traffic management server 4 Resource management means, resource management server 11 Web server 12 Application server 13 Database server 14 Load distributor, LB 21 Operation monitoring means 22 Configuration management means