KR20040011123A - Internet overload service method and system that take over the overload of an internet application server - Google Patents

Abstract

PURPOSE: An Internet overload servicing system for processing overload of an Internet application server is provided to use remaining resources of an Internet server or to configure an exclusive overload server, and to duplicate a service program of a specific Internet service, then to supply the Internet service from the overload server if the application server is in overload state, thereby reusing resources of the Internet server. CONSTITUTION: An overload server(130) is configured by using remaining resources of an Internet server or an exclusive server system. The overload server(130) duplicates a service program as changing an absolute URL. An application server(120) or the overload server(130) duplicating the service program(122,123) monitors self overload and an Internet service(123,134), monitors a performance state of the overload server(130) or a backup overload server(140), and selects an optimal overload server or an optimal backup overload server(125,136). The application server(120) or the overload server(130) uses an HTTP redirect function, and transfers a service request of a client to the optimal overload or the backup overload server(124,135). The overload server(130) or the backup overload server(140) supplies a corresponding Internet service.

Description

Internet overload service method and system that take over the overload of an internet application server}

Internet servers that provide Internet services, such as web hosting servers, Internet shopping malls, Internet stations, and Internet portal sites, generally include a response server that directly provides Internet services to clients, and a DB (DataBase) server that efficiently processes the data requested by the application server. It consists of. According to the present invention, when the number of service requests of a client exceeds a certain level or the load of an application server exceeds a certain level in the Internet server, that is, when the application server is overloaded, the service quality can be maintained at a low cost. The present invention relates to a method and system for internet overload service that can recycle the remaining resources of a spare internet server to an overload server.

The overload processing methods of the conventional application server are classified into a DNS (Domain Name System) round-robin method and a web switch method.

Figure 1 shows an example of the configuration of the DNS round robin method. The site www.a.com consists of three application servers, which process data efficiently using a separate DB server. Each of the three application servers has the same Internet service program, so the client receives the same service no matter which server they access. These three Internet servers are assigned different IP addresses, 3.3.3.1, 3.3.3.2 and 3.3.3.3. The DNS server registers www.a.com with IP addresses 3.3.3.1, 3.3.3.2 and 3.3.3.3. If the client requires a www.a.com IP address resolution from the DNS server, the DNS server responds round-robin, or alternately, with three IP addresses. As a result, when the first client accesses 3.3.3.1, the second client accesses 3.3.3.2, and the third client accesses 3.3.3.3. In this way, the load on the application server is distributed across three servers. The DNS round robin method is a kind of load balancing method that distributes load among multiple servers, and is somewhat insufficient to effectively handle overload, and requires excessive cost. In addition, in order to maintain a record of products purchased at an internet shopping mall, a session concept that can identify a client on an Internet server is added to track the use of the client's service until the client's access to the service is terminated. The DNS round robin method has difficulty in managing sessions because the Internet server can be changed while the client uses the service.

2 shows an example of a configuration of a web switch method. The site www.a.com consists of three application servers, which use a separate DB server to process data efficiently. Each of the three application servers has the same Internet service program, so the client receives the same service no matter which server they access. These three Internet servers are assigned different internal IP addresses (public or private addresses), that is, 192.168.0.1, 192.168.0.2, and 192,168.0.3, all of which are connected to a web switch assigned an external public IP address 3.3.3.1. do. If the web browser requests to resolve www.a.com as a DNS server, the DNS server responds with an external public IP address of 3.3.3.1. When the web browser approaches 3.3.3.1, the web switch relays the load status and operation status of the three application servers to the appropriate server. In this way, the load on the application server is distributed across three servers. Some Webswitches include a proxy program with session management to ensure that the same server, such as 192.168.0.2, is kept accessible while the client is using the services on the site www.a.com. Such a web switch method is functionally excellent but has a disadvantage in that a large number of servers and web switches are required.

The conventional method as described above requires a large number of servers as well as a web switch or a special DNS server, which requires a lot of cost, and has a disadvantage in that resources of an Internet server that can be spared cannot be recycled.

Therefore, in order to solve the above problem, the present invention configures an overload server that can handle the overload of the application server when the service request is congested, and monitors the function and load to pass the Internet service program to the overload server to the application server If you are overloaded, you can add new service requests to the overload server using HTTP redirect message. If you request too much from the client or the application server is overloaded, It is aimed at internet overload service and system that can maintain quality of service and can recycle remaining resources of internet server with overload server.

Fig. 1 An example of the configuration of an empty scheme by DNS round

2 is a configuration example of a web switch method

3 is a conceptual diagram of the overall configuration of the Internet overload service and system

Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings.

In recent years, information and communication services have been reorganized around the Internet. Servers used in the construction of Internet shopping malls, Internet broadcasting stations, Internet portal sites, and web hosting servers that provide services on the Internet are called Internet servers. The service is requested by the HyperText transport Protocol procedure. The Internet server can be configured only with an application server that includes both the service control logic and data processing portion of an application that provides the Internet service. However, in recent years, the response server that directly provides the Internet service to the client by mounting only the service control logic portion of the program is provided. And the DB server that processes the data requested by the application server effectively.

The procedure for providing an Internet service using an Internet server by an Internet service provider or an individual is generally as follows; First, a domain name or IP address to provide a service is assigned, and a development program and a test environment of a service program for providing an Internet service are composed, and a necessary program is developed and self-tested. Then, if the service program developed in the leased or prepared Internet server is distributed, the client uses the service as a URL (Uniform Resource Location) called http: // domain name or IP address. Here URL is a complex of protocols and paths to be accessed. On the other hand, Internet service program is accessed by URL, so if only the route information in URL is changed to new domain name or IP address, service is provided with changed domain name or IP address, and service program is a portable application language such as HTML language and Java. Even if the server system is changed, for example, the service program is mounted on the overload server.

An Internet server that provides an Internet service with an Internet service program can be provided by a company or individual with its own server computer and can be used by accessing the Internet. It also uses a dedicated Internet server facility called Center. Internet data center is dedicated to reliable internet server equipped with highly reliable server equipment, high-speed internet dedicated line and complete auxiliary facilities (UPS, constant temperature and humidity, emergency power generation), and security facilities (firewall, motion monitoring camera, access control system, etc.). Means equipment.

There are two aspects to the service quality felt by Internet service users, that is, the client, which is the reliability of providing the service reliably and reliably, and the performance of providing the service with fast response time. Stability can be fully achieved by using an Internet data center, which is a dedicated Internet server facility. However, in general, the Internet service has a much stronger congestion characteristic in which the client uses the Internet service at a time than the conventional communication such as a telephone, so it is very difficult to stably maintain the quality of service. Allocating enormous server resources to cope with the occasional congestion of service usage is costly. When the service usage is congested, the Internet server transitions to an overload condition. An overload is a traffic overload that causes bottlenecks due to excessive traffic on the communication channel with the Internet, and a server overload that causes the server's ability to actually process the service request of the client to be insufficient. Can be broken down into The program that provides the Internet service is composed of the service control logic to provide the service and the part that processes the data required by the service control logic. To effectively accommodate such a program, the server generally performs the service control logic. The recent trend is to configure a DB (Data Base) server that effectively processes the data requested by the application server and the application server. The DB server receives a data processing request in a standardized language called SQL and provides a result. Since the overload problem of the DB server is not the object of the present invention, a detailed description thereof will be omitted.

The Internet server is generally operated by a separate service provider from an operator who manages and manages individual Internet services mounted on the server. Therefore, the Internet server is equipped with the Internet service program of many individual providers, and the Internet service is provided to the client by a separate domain name or IP address for each Internet service. For example, in the case of an internet shopping mall, an e-commerce operator operating and managing a shopping mall server and a virtual shop owner operating a virtual shop in an internet shopping mall may be separate operators. Typically, a virtual shop owner leases tens to hundreds of Mbytes or more of disk space to distribute his service software, ie, enters the shop and provides services by his domain name or IP address.

Figure 1 shows the overall configuration of the Internet overload service and system according to a preferred embodiment of the present invention. The overall structure is as follows; A client 110 requesting a service to an application server or an overload server using a web browser as a terminal such as a computer, a PDA, a wireless LAN terminal, and a mobile phone; An application server 120 which provides an Internet service and hands over a service program to be performed by an overload server to the overload server, and monitors the overload and hands over a newly input service request to the overload server when the overload server is overloaded; An overload server 130 that receives a service request input from a client and provides an Internet service instead by loading a service program passed from the application server 120; A backup overload server 140 which provides an Internet service to be provided by the overload server when the overload server 130 transitions to an overload state; DB server 150 for processing the data requested by the application server or the overload server and backup overload server; And a development test environment 160 for developing and testing an Internet service program and spreading it to an application server or an overloaded server.

In the preferred embodiment of the present invention, the HTTP re-request 111 processing method in the client 110 that is operated when the application server 120 under the overload state returns an HTTP redirect response to the client is as follows. same; Attempt the service request again with the domain name or IP address of the new destination in the HTTP redirect response message.

In a preferred embodiment of the present invention, the application server 120 is configured with the functions of the server 121, the replication skipping 122, load monitoring 123, HTTP redirection 124 and overload server management 125. The server 121 receives and processes the HTTP request of the client 110 and responds with the result. The server executes the service control logic of the Internet service program, and requests the DB server 150 to process necessary data.

The overload server 130 may be configured for each application server, for each Internet service mounted on the application server, or may be configured to accommodate a plurality of Internet services mounted on several application servers. Therefore, the overload server 130 may use as many domain names or IP addresses as the number of Internet services provided by the overload server. Multiple overload servers can also be configured to increase reliability, increase availability, and distribute load effectively. For example, you can configure several overload servers for a specific application server. The overload server management function 125 monitors and manages overload servers related to the mounted application server to determine whether there is an optimal overload server for each Internet service installed in the application server, and if so, the domain name or IP address to be used by the overload server. It's responsible for telling you what it is. The overload server management 125 performs an operation of determining whether each overload server is ready to operate by duplicating an Internet program to determine an optimal overload server, and whether the overload server is loaded and operating. .

The replication turnover 122 function is set up to handle the overload of a specific Internet service installed in the application server, and is automatically operated when the service program of the corresponding Internet service is changed or under the intention of the application server operator. It is operated in cooperation with the replica receiving 131 of the server. Replicating turns over the Internet service program to the domain name or IP address to be used by the overload server indicated by the overload server management function 125. The scope of the Internet program to be replicated is determined by the scope of the work to be performed by the overload server, which can be divided into three types: replication, service control logic replication, and on-demand replication; Replicating all duplicates all programs for the Internet service, that is, service control logic and data processing, so that the overload server can completely handle the Internet service. There is no need for a separate DB server, and there is no data shared between the second application server and the corresponding Internet service program distributed to the overload server. Third, there is no problem in service even if the client accesses either the application server or the overload server. It can be used only for such Internet services.

Service control logic replication only replicates the service control logic part of the service program executed in the application server.In this case, the data processing required by the service control logic mounted on the overload server requires the use of the DB server in the Internet server. The DB server must be open to the external network, the Internet. On-demand replication is a method of replicating only the service control logic and data processing related to the service to be delivered to the overloaded server. This method is mainly applicable to the static server, not the session-based server, due to the difficulty of session management.

The operation of the load monitoring 123 function is as follows; The load scale is monitored during the operation of the application server, and if it is over the specified threshold value, it is determined as an overload. The load scale monitored here is classified into the load and quality of the corresponding Internet service, and the load and quality of the application server. The former includes the number of concurrent users and the processing time of the service request. It includes the total number of concurrent users, CPU time of each web server and DB server, and processing time of service requests of the entire Internet server.

The HTTP Redirection 124 function, which is operated when an application server is under heavy load, creates an HTTP redirect message containing the domain name or IP address of the overload server for a newly entered HTTP request. 110). The domain name or IP address of the appropriate overload server is determined and informed by the overload server management (125) function.

In the preferred embodiment of the present invention, the overload server 130 is replicated 131, virtual server 132, replication skip 133, load monitoring 134, HTTP redirection 135 and backup overload server management ( 136). The overload server 130 may be configured as a dedicated system, but any Internet server may use the remaining resources thereof to mount the function of the overload server.

The virtual server 132 provides a corresponding Internet service by executing a program replicated from the application server 120. The virtual server 132 also has a virtual web hosting function that provides a service by mounting a plurality of virtual web servers. The virtual web hosting method allocates a separate IP address for each virtual web server and uses IP address-based virtual web hosting to identify the virtual web server by IP address, and domain name-based virtual to identify the virtual web server by domain name. Accepts all web hosting methods.

The overload server 130 replicates the Internet service program mounted on the backup overload server 140, and if the overload server itself is in an overload state, requests the service request from the client as an HTTP redirect message to the backup overload server. If you pass it to the backup overload server, the backup overload server can be operated to provide the service. The backup overload server has the same functional structure as the overload server 130, and the relationship between the overload server 130 and the backup overload server 140 is the same as the relationship between the application server 120 and the overload server 130.

The backup overload server 140 may be configured for each overload server, for each Internet service mounted on the overload server, or may be configured to accommodate a plurality of Internet services mounted on various overload servers. Backup overload servers can be configured in multiple numbers to increase reliability, increase availability, and distribute load effectively. For example, you can configure several backup overload servers for specific overload servers. The backup overload server management function (136) monitors and manages the backup overload servers related to the overload server to determine whether there is an optimal backup overload server for each Internet service in the overload server, and what domain name or IP address to use for the backup overload server. It is responsible for the function that informs the recognition. The backup overload server management function (136) determines whether each backup overload server is ready to operate by duplicating an Internet program to determine the optimal backup overload server, and how the load state and operation state of each backup overload server are. Perform the action.

The copy skip (133) function is configured to handle the overload of a specific Internet service loaded on the overload server by the backup overload server, and is automatically operated when the service program of the corresponding Internet service is changed or by the intention of the overload server operator. It is operated in cooperation with the replication receiving function of the backup overload server 140. The replication transfers the Internet service program to the domain name or IP address to be used in the backup overload server indicated by the backup overload server management function 136. The range of Internet programs to be replicated is the same as the range of Internet programs to be replicated in the replication page 132 in the application server.

The operation of the load monitoring 134 function is as follows; The load scale is monitored during the operation of the overload server and it is determined to be an overload if it exceeds the predetermined threshold. The load scale monitored here is classified into the load and quality of the Internet service and the load and quality of the overload server. The former includes the number of concurrent users and the processing time of service requests. It includes the total number of concurrent users, CPU time of each web server and DB server, and processing time of service requests of the entire Internet server.

The HTTP Redirection (135) feature, which works when the overload server is in an overloaded state, creates an HTTP redirect message containing the domain name or IP address of the backup overload server for incoming HTTP requests. Answer 110. The domain name or IP address of the appropriate backup overload server is determined and informed by the backup overload server monitoring (125) function.

Replication overload (133), load monitoring (134), HTTP path redirection (135) and backup overload server management (136) functions that are mounted on the overload server 130, replication overload (123), load in the application server 120 The monitoring 125, the HTTP path redirection 125 and the overload server management 126 can be configured with the same program.

The copy receiving 131 function operates through the communication with the copy turning 122 function in the application server 120, and the main role is to copy a program necessary for the Internet service to be transferred from the application server and change the URL. To duplicate a program, the overload server creates the same directory structure as the application server and copies the program passed by the application server to the directory. Since the Internet service program has a high portability that can be performed on other domains or IP addresses by changing the domain name or IP address, that is, changing the URL, the overload server can easily provide the Internet service provided by the application server only by changing the URL. have. There are two ways to use URLs: absolute URLs and relative URLs. Absolute URLs are all types of absolute values of a path, for example, "http://www.a.com/internet/HTML/index.html". Relative URLs write only relative paths from the current path. For example, when the current directory is internet in www.a.com, it is in the form of "HTML / index.html". Relative URLs depend on the absolute URLs defined above, so there is no need to change them when cloning Internet programs.

In the overload server replica 131 function, the URL change is performed as follows; Change the domain or IP address that corresponds to the application server among the absolute URLs in the passed program to the domain name or IP address to be used by the overload server. Let's take an example of a URL change. If you duplicate the service program of the Internet service provided by the application server as www.a.com to the overload server that provides the overload service under the domain name of www.abcom, all the domain name www.a.com in the Internet service program Just change to www.abcom.

In an exemplary embodiment of the present invention, another overload server may be operated in preparation for the overload of the backup overload server 140.

The present invention duplicates the service program of a specific Internet service by using the remaining resources of the Internet server or by configuring a dedicated overload server, and then the application server that provides the Internet service causes the corresponding Internet service due to excessive service request or server overload. The present invention relates to an internet overload service method and system that provides an internet service provided by an application server when the system transitions to an overload condition that cannot guarantee the quality of the service. By using such an Internet overload service method and system, it is effective to maintain the Internet service quality at low cost in case of overload, and to manage the overload of the remaining Internet server resources by overloading other Internet servers. It works.

Claims (2)

In an Internet service environment where a client requests a service through an HTTP procedure and an Internet server such as a web hosting server, an internet shopping mall, an internet station, or an internet portal site provides a service, After using the remaining resources of the Internet server that can afford performance or using a dedicated server system to configure the overload server 130, and receives the service program of the Internet service to provide a service instead of changing the absolute URL (131) step; The application server 120 or the overload server 130 which duplicates the Internet service program (122, 133) monitors the Internet service and its overload (123, 134), the overload server 130 or the backup overload server 140 Monitoring (125, 136) an optimal overload server or backup overload server by monitoring the operation or performance status If the application server 120 or the overload server 130 determines that itself or the corresponding Internet service is overloaded, the client requests the service request of the client to the optimal overload server or the backup overload server using the HTTP redirect function (124). , 135) step; And a step of providing the Internet service by the overload server (130) or the backup overload server (140) receiving the service request from the client. In claim 1, Replicating Internet service programs by full replication, service control replication, and on-demand replication on the overload server and backup overload server; A backup overload server 140 for overload processing of the overload server 130 and a method for configuring another overload server for overload processing of the backup overload server; A method of configuring a plurality of overload servers and a plurality of backup overload servers to improve the reliability, availability, and load sharing of the overload server and the backup overload server; To provide multiple Internet services replicated by virtual web hosting in virtual server 132