JP2008225644A - Gateway device, load balancing method for gateway device and load balancing program of gateway device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently distribute load of a relay server without adding a special device. <P>SOLUTION: A load balancing system has a plurality of the relay servers, a client terminal, and the gateway device managing load information of the relay server. The gateway device has; a relay server information management means provided in an intranet, managing the load information of the relay server as periodically updating the load information of the relay server; and a load distribution decision means deciding whether the load of the first relay server is distributed or not based on the load information of the plurality of relay servers and the load information of the first relay server wherein the client terminal is connected to the relay server. The load balancing decision means establishes second connection between the client terminal and a second relay server while establishing first connection between the first relay server and the client terminal when it is decided that the load balancing is effective, cuts the first connection after the second connection establishes. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a gateway apparatus, a load distribution method for a gateway apparatus, a load distribution program for a gateway apparatus, a load distribution system, and a load distribution method for a relay server, and in particular, performs load distribution when a load on the relay server increases. The present invention relates to a gateway device, a load distribution method for a gateway device, a load distribution program for a gateway device, a load distribution system, and a load distribution method for a relay server.

  2. Description of the Related Art Conventionally, with the spread of computers, client server systems in which client terminals and server devices are connected are widely used.

  Here, FIG. 11 shows a general client-server system. In FIG. 11, in the client server system, a client terminal C1, a client terminal C2, and a client terminal C3 are connected to a server apparatus (hereinafter simply referred to as a server) SV1 and a server SV2 via a gateway apparatus GW1. Has been.

  Normally, in such a client system, for example, a processing request transmitted from the client terminal C1 is transmitted to the server SV1 or the server SV2 via the gateway device GW1, and the processing request is processed in the server SV1 or the server SV2. Then, it is returned to the client terminal C1 via the gateway device GW1.

A client server system using such a configuration is described in Patent Document 1.
JP 2002-269061 A

  This Patent Document 1 includes a gateway device GW1 serving as a relay server between a plurality of client terminals and a plurality of servers, and the gateway device GW1 is connected to the server SV1 or the server SV2. A client-server system that manages the server as a load is described.

  Then, the gateway device GW1 assigns to which server the client terminals C1 to C3 are connected depending on the processing load of the server SV1 or the server SV2.

  However, in the client server system described in Patent Document 1, for example, when a processing request from the client terminal C1 is being processed, a heavy processing (for example, arithmetic processing such as image processing) is performed in the server SV1. Or, when the client C1 is disconnected from the server SV1, the gateway device GW1 cannot grasp the load state of the server SV1 in real time.

  That is, the gateway device GW1 only manages the number of connections at the time of processing requests from the client terminals C1 to C3 as a load, and once a client terminal is assigned to a server, the gateway device GW1 relates to the subsequent load status of each server. Therefore, there is a problem that the server to which the client terminal is connected cannot be changed.

  Further, the following load distribution system of relay servers can be considered from the load distribution system using existing relay servers.

  Referring to the configuration diagram of FIG. 12, this conceivable relay server load distribution system includes a client terminal 1101, a network address translation (NAT) 1102, a relay server 1103, a relay server 1104, a relay server 1105, and the like. A mail server 1106, a NAT 1107, a client terminal 1108, and a session management server 1109.

  The relay server load distribution system having such a configuration operates as shown in the flowchart of FIG.

  In the flowchart shown in FIG. 13, when the user performs a connection request operation on the client terminal (step 1211), the client terminal 1101 transmits a session establishment request to the session management server 1109 (step 1212).

  The session management server 1109 notifies the client terminal 1101 of information on the relay server 1103 that can afford to execute packet relay processing (step 1213).

  After establishing a connection with the relay server 1103 (step 1214), the client terminal 1101 transmits a connection start mail to the client terminal 1108 (step 1215).

  The client terminal 1108 periodically checks whether a connection start mail has arrived (step 1216). If the connection start mail has arrived, the client terminal 1108 receives the mail (step 1217) and connects to the relay server 1103. (Step 1218).

  When the connection between the client terminal 1101 and the client terminal 1108 is completed, the relay server 1103 transmits a connection completion message to the client terminal 1101 and the client terminal 1108 (steps 1219 and 1220).

  Upon receiving the connection completion message (connection completion notification) (step 1220), the client terminal 1101 performs packet communication with the client terminal 1108 via the relay server 1103 (step 1222, step 1223).

  The configuration for such an operation will be further described with reference to FIG.

  Referring to FIG. 14A, in the relay server load balancing system that operates as described above, when trying to access the internal business server 104 from the external client terminal 101, the internal network is usually protected by the firewall 102. It has been.

  For this reason, the client terminal 101 cannot normally communicate with the business server 104 in the in-house network.

  Referring to FIG. 14B, contrary to FIG. 14A, connection from the client terminal 109 in the company to the server in the outside is often permitted.

  However, communication from the client terminal 109 to the client terminal 105 having a private IP address outside the company is not possible. This is because a private IP cannot uniquely identify a terminal.

  Referring to FIG. 14C, communication from the client terminal 110 to which the private IP address is assigned to the Internet is performed by performing conversion called NAT (network address translation) 111 and converting it to a global IP address. be able to.

  Therefore, a method of relaying and transferring the connection through the NAT 111 of the client terminal 110 using the private IP and the connection from the in-house gateway device 114 by using the relay server 112 having a global IP address is considered. It is done.

  However, when a large number of client terminals use the relay server 112, the load becomes high and the transfer takes time, so that convenience becomes a problem.

  Therefore, in the load distribution system of the relay server that relays communication between the client terminal and the gateway device, which can be considered from the existing relay server, both the session from the server and the session from the client terminal are distributed to other relay servers. It is necessary to install a load balancer and session management server, and there is a problem that costs are increased because special hardware is used to introduce the system.

  The present invention has been made in view of the above problems, and a gateway device that efficiently distributes the load of a relay server without adding a special device, a load distribution method of the gateway device, and a load distribution of the gateway device It is an object to realize a load distribution method for a program, a load distribution system, and a relay server.

  A gateway device according to the present invention is a gateway device that manages load information of a plurality of relay servers, and is provided in an intranet, and manages the load information of the relay server while periodically updating it by communication The relay server information management means includes the first relay server based on the load information of the first relay server connected to the client terminal and the load information of the plurality of relay servers. Load distribution determination means for determining whether or not to perform load distribution of the relay server, and when the load distribution determination means determines that load distribution of the first relay server is effective One of the plurality of relay servers is connected while the first connection between the currently connected client terminal and the first relay server is connected. And connecting the second connection between the second relay server different from the first relay server and the client terminal, and disconnecting the first connection after the second connection is established. Thus, the load distribution of the first relay server is performed.

  In the gateway device according to the present invention, the load distribution determination unit is a relay server having a small number of the client terminals connected to the first relay server among the plurality of relay servers, When the usage rate of the control unit of the relay server is within a predetermined range, it is determined that it is effective to perform load distribution of the first relay server with the relay server as the second relay server. May be.

  Further, the gateway device according to the present invention is a maintenance notice notifying unit that acquires, from the load information of the relay server, the maintenance notice that is transmitted by the relay server information managing means to notify the relay server of the notice of maintenance. When the maintenance notice notice is acquired from the load information of the relay server, the maintenance notice notice means applies the load distribution determining means with the relay server as the first relay server. The client terminal connected to the relay server may be switched from the first relay server to the second relay server so as to be connected to the second relay server.

  In the gateway device according to the present invention, the load distribution determination unit is a relay server having a small number of the client terminals connected to the first relay server among the plurality of relay servers, When the usage rate of the control unit of the relay server is within a predetermined range, the relay server may be switched from the first relay server as the second relay server.

  According to the present invention, the connection between the client terminal and the gateway device is established via the relay server to which the client terminal is connected first, and the gateway device is connected to the plurality of relay servers connected to the own device. Thus, load information can be updated and managed.

  As a result, the gateway device can make a load distribution determination to change the connection destination of the client terminal based on the load information of each relay server, and the gateway device can connect the client terminal and the first relay server. In the established state, the connection can be switched to a second relay server different from the first relay server. Thus, a gateway device that efficiently distributes the load on the relay server without adding a special device, a load distribution method for the gateway device, a load distribution program for the gateway device, a load distribution system, and a load distribution method for the relay server are realized. it can.

Next, the configuration of the embodiment of the present invention will be described in detail with reference to the drawings.
(First embodiment)
(1) Load Balancing Relay System According to First Embodiment (1-1) Configuration of Load Balancing Relay System Referring to FIG. 1, a load balancing relay system according to the first embodiment is an internal business server from outside the company. The client terminal 201 that connects to the 208, the NAT device 218, the gateway device 206 that accepts the connection, the relay server 202 that relays the session between the client terminal 201 and the gateway device 206, and the load on the relay server 202 is distributed A relay server 203, a relay server 205, a firewall 207, a mail server 204, and a business server 208.

  Here, the internal configuration of the client terminal 201 is shown in FIG.

  As shown in FIG. 2, the client terminal 201 includes an operating system 309, an SMTP (simple mail transfer protocol) unit 304, a gateway device / relay server communication unit 305, a control unit 306, an address table 307, and an application unit. 308.

  The operating system 309 includes a MAC (media access control) unit 301, an IP (internet protocol) unit 302, and a TCP / UDP (transmission control protocol / user datagram protocol) unit 303.

  The MAC unit 301 is a standard protocol processing unit that transmits and receives packets on Ethernet (registered trademark).

  The IP unit 302 is a standard protocol processing unit that performs packet routing.

  The TCP / UDP unit 303 is a standard protocol processing unit that performs reliable packet communication.

  The SMTP unit 304 is a standard protocol processing unit that transmits electronic mail.

  The gateway device / relay server communication unit 305 is a part that communicates with the gateway device 206 to acquire control information, and communicates with the relay server 202, the relay server 203, and the relay server 205.

  The control unit 306 calls a function of the SMTP unit 304 or calls a function for receiving the request of the relay server 202 received by the gateway device communication unit 305 in order to transmit mail.

  The address table 307 stores mail account information for the gateway device 206 and the IP address of the default relay server to be connected first.

  The application unit 308 communicates with the control unit 306 in order to exchange packets with the application unit of the business server 208.

  The relay server 202, the relay server 203, and the relay server 205 are assigned global IP addresses. By relaying the TCP connection connection request from the client terminal 201 and the TCP connection connection from the gateway device, Means for transferring packets to and from the gateway device 206 is included.

  Next, the relay server 202, the relay server 203, and the relay server 205 will be described. Since the internal configurations of the relay server 202, the relay server 203, and the relay server 205 are all the same, only the relay server 202 will be described.

  As shown in FIG. 3, the relay server 202 includes an operating system 410, a packet identification unit 404, a relay transfer unit 405, a gateway device communication unit 406, a server management unit 407, a relay table 408, and a load information table. 409.

  Further, since the operating system 410 is the same as the operating system 309, description thereof is omitted.

  The packet identification unit 404 distributes the packet to the relay transfer unit 405 and the gateway device communication unit 406 according to the type of the received packet.

  The relay transfer unit 405 has an operation of waiting for a connection from the client terminal 201 and the gateway device 206 and a function of transferring those packets.

  The gateway device communication unit 406 has a function of receiving a request from the gateway device 206, acquiring information from the load information table 409, and returning a response.

  The server management unit 407 periodically collects load information of the relay server 202 and writes it in the load information table 409.

  The relay table 408 is a table having information for connecting the communication between the client terminal 201 and the gateway device 206 and the current number of relay transfers.

  The server management unit 407 obtains the number of relay transfers from the relay server table 408 at regular intervals, and stores the CPU (central processing unit) usage rate at that time in the load information table 409.

  Next, the mail server 204 will be described. The internal configuration of the mail server 204 is shown in FIG.

  As shown in FIG. 4, the mail server 204 includes a MAC unit 401, an IP unit 402, a TCP / UDP unit 403, an SMTP unit 411, a POP (post office protocol) unit 412, and a mail spool 413. ing.

  Here, the MAC unit 401, the IP unit 402, the TCP / UDP unit 403, and the SMTP unit 411 are the same as the parts provided in the client terminal 201, and thus the description thereof is omitted.

  The POP unit 412 is a protocol for receiving mail from a server storing electronic mail.

  The mail spool unit 413 is an area for storing delivered mail.

  Next, the gateway device 206 will be described. The internal configuration of the gateway device 206 is shown in FIG.

  As shown in FIG. 5, the gateway device 206 includes an operating system 509, a control information communication unit 504, a transfer unit 505, a POP unit 506, a control unit 507, and a relay server information table 508.

  Since the operating system 509 is the same as the operating system 309 and the operating system 410, description thereof is omitted.

  The control information communication unit 504 communicates with the relay server to transmit and receive load information requests.

  The transfer unit 505 transfers communication between the inside and outside the company. In this example, communication with the business server 208 (FIG. 1) and the client terminal 201 (FIG. 1) is transferred.

  Similar to the POP unit 411 of the mail server 204, the POP unit 506 receives a connection start mail from the mail server.

  The control unit 507 calls a function of the control information communication unit 504, the transfer unit 505, or the POP unit 506, or stores received data in the relay server information table 508. Further, it has means for determining a connection destination relay server based on information in the relay server information table 508.

  In FIG. 1, the management content of the load information table 409 (FIG. 3) provided in the relay server 202 is illustrated as the load information table 209.

  In the load information table 209, the time 211, the number of connections 212, and the CPU usage rate 213 are described.

  Also, in FIG. 1, the management content of the relay server information table 508 (FIG. 5) provided in the gateway device 206 is illustrated as the relay server information table 210.

In the relay server information table 210, a host name 214, an IP address 215, the number of connections 216, and a CPU usage rate 217 are described.
(1-2) Connection Destination Establishment Processing Procedure of Load Balancing Relay System Next, referring to the configuration diagram of FIG. 1 and the flowchart of FIG. 6, the connection destination establishment processing of the load distribution relay system according to the first embodiment The procedure will be described in detail.

  First, the user 601 of the client terminal 201 instructs the client terminal 201 to start connection (step 611).

  Then, the client terminal 201 establishes a TCP connection with the relay server 202 (step 612).

  The client terminal 201 transmits a connection start mail to the mail account of the gateway device 206 (step 613).

  The gateway device 206 periodically checks whether a connection start mail has arrived at the mail server 204 (step 614).

  If a connection start mail exists in the mail server 204, the gateway device 206 receives the connection start mail (step 615).

  The gateway device 206 opens a TCP connection with the relay server 202 (step 616).

  Next, when the gateway device 206 issues a connection completion notification to the user 601 (step 617), it reaches the client terminal 201 via the relay server 202 (step 618), and the connection to the user 601 is completed. (Step 619).

  Thereby, the first connection is established.

  Next, the control unit 507 (FIG. 5) of the gateway device 206 transmits to the relay server 202, the relay server 203, and the relay server 205 registered in the relay server information table 508 (FIG. 5) via the control information communication unit 504. On the other hand, a load information provision request (referred to as status information request in FIG. 6) such as the currently established number of connections 212 and CPU usage rate 213 is issued (step 620).

  When the packet identification unit 404 (FIG. 3) determines that the load information provision request from the gateway device 206 is a control packet, the relay server 202, the relay server 203, and the relay server 205 determine that the gateway device communication unit 406 The control packet is passed to (FIG. 3). In addition, each relay server stores the load information acquired by the server management unit 407 (FIG. 3) and the number of connections described in the relay table 408 (FIG. 3) in the load information table 409 (FIG. 3). Store while updating regularly. Then, the gateway device communication unit 406 (FIG. 3) of each relay server transmits the latest connection number 212 (FIG. 1) and the CPU usage rate 213 (FIG. 1) to the gateway device 206 (step 621).

  Then, the gateway device 206 receives the load information (referred to as a status response in FIG. 5) from the relay server 202, the relay server 203, and the relay server 205 that can be received via the control information communication unit 504 and the control unit 507. Stored in the relay server information table 508.

  Thereby, the gateway device 206 can update the relay server information table 210 shown in FIG. 1 and can hold the latest connection number 212 and the CPU usage rate 213 as load information.

  Next, the control unit 507 (FIG. 5) refers to the relay server information table 508 and examines the connection destination of the load distribution operation (this is called a changed connection destination determination processing procedure, which will be described later).

  Here, when the control unit 507 (FIG. 5) of the gateway device 206 finds a new connection destination, the control information communication unit 504 (FIG. 5) of the gateway device 206 issues a connection destination change request to the client terminal 201 ( Step 622).

  Receiving the connection destination change request, the client terminal 201 opens a TCP connection to the change destination relay server 203 while maintaining the connection of the first connection (step 623).

  The relay server 203 establishes a TCP connection as the relay server 203 to be changed from the gateway device 206 simultaneously with the establishment of the TCP connection by the client terminal 201 (step 624).

  When the connection between both the client terminal 201 and the gateway device 206 is established, a connection completion notification is issued from the gateway device 206 to the client terminal 201 via the newly established connection (that is, via the relay server 203) (steps 625 and 626). ).

  This establishes a second connection.

  Thereby, the gateway device 607 issues a disconnection request for the first connection (step 627).

  Also, a TCP connection disconnection request is issued from the client terminal 602 simultaneously with the TCP connection disconnection request of the gateway device 206 (step 628).

In this way, communication can be started from the client terminal 201 to the in-house business server 208 (FIG. 1) using the second connection.
(1-3) Change Connection Destination Determination Processing Procedure of Load Balancing Relay System Next, the change connection destination determination processing procedure will be described with reference to the flowchart of FIG.

  First, the gateway device 206 that has collected and stored the status information response in step 621 (FIG. 6) uses the relay server 202 (FIG. 1) used in the first connection from the relay server information table 508 (FIG. 5). It is searched whether there is a relay server having a smaller number of connections than the number of connections 212 (FIG. 1) (step 701).

  If there is no relay server having a smaller number of connections than the number of connections of the relay server 202 in the relay server information table 508, the connection destination changing operation is stopped (step 702).

  On the other hand, when the relay server information table 508 includes a relay server having a smaller number of connections than the number of connections of the relay server 202, the CPU usage rate 213 (FIG. 1) of the relay server having the smallest number of connections is 80%. It is checked whether or not the following is true (step 703).

  This is because if the CPU usage rate is too high, even if the number of connections is small, the transfer rate of the relay server is slowed down, leading to a decrease in transfer rate. In order to prevent the transfer rate from decreasing, if the CPU usage rate is 80% or less, the relay server is set as the connection destination of the load distribution operation (step 707).

  In step 703, if the CPU usage rate is not 80% or less, the CPU usage rate of the relay server with the next smallest number of connections is checked (step 704).

  If the CPU usage rate of the relay server is 80% or less, the relay server is set as the connection destination of the load distribution operation (step 707).

  If the CPU usage rate of the relay server is not 80% or less, the operation after step 703 is recursively repeated to search for a connection change destination relay server. If still not found, the relay server with the smallest number of connections is set as the connection change destination (step 706).

  In this way, the connection change destination can be determined from the established relay server to another relay server connected to the gateway device 206 by the change connection destination determination processing procedure.

As a result, the connection destination can be changed to the connection destination relay server determined by the above connection destination establishment processing procedure.
(1-4) Operation and Effect of Load Balancing Relay System In the present embodiment, the client terminal 201 and the gateway device are once passed through a default (standard setting) relay server (the relay server 202 in the above example). This can be realized by message communication in which the gateway device 206 inquires each relay server about the load status or issues a connection destination change instruction to the client terminal 201 after establishing a session with the client 206.

  According to the present embodiment, load distribution can be performed without using a special device such as a load balancer, and therefore a reduction in system introduction cost can be expected.

  Further, the conventional method using the session management server is disadvantageous in terms of security because the session management server is installed outside the firewall, but in this embodiment, each relay is connected to the gateway device 206 in the firewall. Server information can be stored.

Thereby, according to this Embodiment, a load distribution function can be operated comparatively safely rather than the conventional system.
(Second Embodiment)
(2) Load Balancing Relay System According to Second Embodiment (2-1) Configuration of Load Balancing Relay System Next, a load balancing relay system according to the second embodiment will be described.

  Referring to FIG. 8, in the load distribution relay system according to the second embodiment, an administrator terminal 818 managed by an administrator is added to the configuration of the first embodiment.

  The internal configuration of each part is as shown in FIGS. 2 to 5 as in the first embodiment.

  A global IP address is assigned to the relay server 802, the relay server 803, and the relay server 805, and the client terminal is relayed by relaying the TCP connection connection request from the client terminal 801 and the TCP connection connection request from the gateway device 806. The relay transfer of the packet between 801 and the gateway device 806 is performed.

  The relay server 802, the relay server 803, and the relay server 805 have a load information table 809, the number of connections 812 between the client terminal 801 currently established via the self and the gateway device 806, and the CPU usage rate 813. Are collected and stored regularly.

The gateway device 806 has a relay server information table 810 that stores the host names and IP addresses of the relay server 802, the relay server 803, and the relay server 805.
(2-2) Connection Destination Establishment Processing Procedure of Load Balancing Relay System Next, referring to the configuration diagram of FIG. 8 and the flowchart of FIG. 9, the connection destination establishment processing of the load distribution relay system according to the second embodiment. The procedure will be described in detail.

  However, portions overlapping with the operations of the first embodiment will be omitted.

  First, when the user 902 of the client terminal 801 performs a connection request operation on the client terminal 801 (step 911), the client terminal 801 and the gateway device 806 are connected as in the case of the first embodiment. A first connection via the relay server 802 is established (steps 911 to 919).

  Further, the gateway device 806 periodically issues a request for status information of each relay server to each relay server 802, relay server 803, and relay server 805 (step 920).

  Here, in the second embodiment, when the administrator of the relay server issues a maintenance notice to the relay server 802 from the administrator terminal 818 (step 922), the load information table 809 provided in the relay server 802 is updated. Only the maintenance flag 819 becomes 1.

  In this embodiment, the relay server 802, the relay server 803, or the relay server 805 that has received the periodic status information request in step 920 receives the number of connections 812 in the load information management table 809 provided in each relay server, A response in which maintenance flag information 819 is further added in addition to the CPU usage rate 813 is returned (step 921).

  The gateway device 806 confirms the response of the relay server 802, the relay server 803, or the relay server 805, and stores the number of connections 816 and the CPU usage rate 817 in the relay server information table 810.

  At this time, if there is a relay server in which the maintenance flag 819 is 1, the connection established via the relay server is switched to the connection via another relay server.

  Specifically, since the maintenance notice is issued from the administrator terminal 818 to the relay server 802, only the maintenance flag 819 of the relay server 802 becomes 1, and the gateway device 806 establishes a connection via the relay server 802. The existing connection to the connection via another relay server.

  When a new connection destination is determined, a request for connection to the relay server 803 serving as a new connection destination is issued from the gateway device 806 to the client terminal 801 (step 923).

  Receiving it, the client terminal 801 establishes a TCP connection to the relay server 803 of the change destination while maintaining the connection of the first connection (step 924).

  At the same time, the gateway device 806 opens a TCP connection with the relay server 803 as the change destination (step 925).

  When both the relay server 802 and the gateway device 806 are connected to the new connection destination, the gateway device 806 issues a connection completion notification to the client terminal 801 via the newly established connection (steps 926 and 927). .

  Thereby, the second connection is established.

Accordingly, the gateway device 806 issues a first connection disconnection request to the relay server 802 (step 928). At the same time, the client terminal 801 issues a disconnection request to the relay server 802 (step 929).
(2-3) Change Connection Destination Determination Process of Load Balancing Relay System Next, the change connection destination determination process will be described with reference to the flowchart of FIG.

  The flowchart of FIG. 10 is a procedure for determining the connection destination relay server to be changed.

  First, the gateway device 806 refers to the relay server information table 810 and checks whether or not the CPU usage rate 817 of the minimum relay server with the number of connections 816 is 80% or less (step 1001).

  Here, if the CPU usage rate of the minimum relay server is 80% or less, the relay server is determined as a new connection destination target (step 1004).

  On the other hand, if the CPU usage rate is not 80% or less, a relay server with the next smallest number of connections is searched to check whether the CPU usage rate is 80% or less (step 1002).

  If the CPU usage rate is 80% or less, the relay server is set as a new connection destination target (step 1004).

  On the other hand, when the CPU usage rate is not 80% or less, a search for a relay server having a smaller number of connections and a CPU usage rate of 80% or less is recursively performed.

If there is no relay server having a CPU usage rate of 80% or less, a relay server other than the maintenance target relay server 802 and having the smallest number of connections 816 is set as a new connection destination target (step 1005).
(2-4) Effects of Load Balancing Relay System In the load balancing relay system according to the second embodiment, the administrator issues a maintenance notice to the relay server 802 connected using the administrator terminal 901. “1” is written in the maintenance flag 819 of the load information table 809 provided in the relay server 802.

  For this reason, the gateway device 806 can recognize the relay server that switches the connection based on the maintenance flag 819 in which “1” is written, which is sent from the relay server. Therefore, the gateway device 806 determines the next relay server whose connection destination is to be changed from the relay server 802 and reconnects it.

  As described above, the client terminal 801 starts communication using the second connection to the in-house business server. As a result, the administrator can change the relay server through which the user is not conscious.

  ADVANTAGE OF THE INVENTION According to this invention, the load distribution of a relay server can be efficiently performed in the system using the relay server which opens a connection to a relay server from both a server and a client. In addition, the administrator can reduce the burden of managing the relay server.

It is a block diagram which shows the structure of the load distribution relay system by 1st Embodiment in this invention. It is a block diagram of the client terminal by the 1st Embodiment in this invention. It is a block diagram of the relay server by the 1st Embodiment in this invention. It is a block diagram of the mail server by the 1st Embodiment in this invention. It is a block diagram of the gateway apparatus by the 1st Embodiment in this invention. It is a flowchart showing the connection destination establishment processing procedure by the 1st Embodiment in this invention. It is a flowchart showing the change connection destination determination processing procedure by 1st Embodiment in this invention. It is a block diagram which shows the structure of the load distribution relay system by 2nd Embodiment in this invention. It is a flowchart showing the connection destination establishment process sequence by the 2nd Embodiment in this invention. It is a flowchart showing the change connection destination determination processing procedure by the 2nd Embodiment in this invention. It is a block diagram of the prior art known conventionally. This is a load distribution relay system conceivable from the prior art known from the past. It is a flowchart showing operation | movement of the load distribution relay system considered from the prior art conventionally known. It is the block diagram which showed the structure considered from the prior art conventionally known.

Explanation of symbols

201 Client terminal 202, 203, 205 Relay server 204 Mail server 206 Gateway device 207 Firewall 208 Business server 301 MAC unit 302 IP unit 303 TCP / UDP unit 304 SMTP unit 305 Gateway device / relay server communication unit 306 Control unit 307 Address table 308 Application unit 309 Operating system 401 MAC unit 402 IP unit 403 TCP / UDP unit 404 Packet identification unit 405 Relay transfer unit 406 Gateway device communication unit 407 Server management unit 408 Relay table 409 Load information table 410 Operating system 411 SMTP unit 412 POP unit 413 Mail spool 501 MAC unit 502 IP unit 503 TCP / UDP unit 504 Control information communication unit 50 Transfer unit 506 POP unit 507 Control unit 508 Relay server information table 801 Client terminal 802, 803, 805 Relay server 804 Mail server 806 Gateway device 807 Firewall 808 Business server 818 Administrator terminal 901 User C1, C2, C3 Client terminal SV1, SV2 server 1101, 1108 Client terminal 1102, 1107 NAT
1103, 1104, 1105 Relay server 1106 Mail server 1109 Session management server 101, 105, 109, 110 Client terminal 102, 107, 113 Firewall 103, 108, 114 Gateway device 104, 115 Business server 106, 111 NAT
112 Relay server

Claims (13)

A gateway device that manages load information of a plurality of relay servers,
Provided in the intranet, comprising relay server information management means for managing the load information of the relay server while periodically updating by communication,
The relay server information management means includes
Based on the load information of the first relay server to which the client terminal is connected to the relay server and the load information of the plurality of relay servers, it is determined whether or not load distribution of the first relay server is to be performed. Equipped with load balancing means
The load distribution determining means includes
If it is determined that load distribution of the first relay server is effective,
A second relay that is one of the plurality of relay servers and is different from the first relay server, with the first connection between the currently connected client terminal and the first relay server connected. A second connection between the server and the client terminal is connected, and after the second connection is established, the load is distributed to the first relay server by disconnecting the first connection. Gateway device. The load distribution determining means includes
Among the plurality of relay servers, when the number of the client terminals connected to the first relay server is small and the usage rate of the control unit of the relay server is within a predetermined range The gateway apparatus according to claim 1, wherein it is determined that it is effective to perform load distribution of the first relay server with the relay server as the second relay server. The relay server information management means includes
A maintenance advance notice means for acquiring a maintenance advance notice transmitted to notify the relay server of the advance notice from the load information of the relay server,
The maintenance notice means is
When the maintenance notice is acquired from the load information of the relay server, the load distribution determination unit is applied using the relay server as the first relay server, and the client terminal connected to the first relay server is The gateway device according to claim 1, wherein the gateway device is switched from the first relay server to the second relay server so as to be connected to the second relay server. The load distribution determining means includes
Among the plurality of relay servers, when the number of the client terminals connected to the first relay server is small and the usage rate of the control unit of the relay server is within a predetermined range The gateway device according to claim 3, wherein the relay server is switched from the first relay server as the second relay server. A load balancing method in a gateway device that manages load information of a plurality of relay servers,
A relay server information management step that is provided in an intranet and manages load information of the relay server while being regularly updated by communication;
The relay server information management step includes:
Based on the load information of the first relay server to which the client terminal is connected to the relay server and the load information of the plurality of relay servers, it is determined whether or not load distribution of the first relay server is to be performed. A load balancing step,
The load distribution determining step includes:
If it is determined that load distribution of the first relay server is effective,
A second relay that is one of the plurality of relay servers and is different from the first relay server, with the first connection between the currently connected client terminal and the first relay server connected. A second connection between the server and the client terminal is connected, and after the second connection is established, the load is distributed to the first relay server by disconnecting the first connection. Load balancing method. The relay server information management step includes:
A maintenance advance notice step for obtaining a maintenance advance notice transmitted to notify the relay server of the advance notice from the load information of the relay server;
The maintenance notice notification step includes:
When the maintenance notice is acquired from the load information of the relay server, the load distribution determination step is applied with the relay server as the first relay server, and the client terminal connected to the first relay server is The load distribution method according to claim 5, wherein switching from the first relay server to the second relay server is performed so as to be connected to the second relay server. A load distribution program in a gateway device that manages load information of a plurality of relay servers,
Provided in the intranet, comprising a relay server information management procedure for managing the load information of the relay server while periodically updating by communication,
The relay server information management procedure includes:
Based on the load information of the first relay server to which the client terminal is connected to the relay server and the load information of the plurality of relay servers, it is determined whether or not load distribution of the first relay server is to be performed. It has a load balancing judgment procedure,
The load distribution determination procedure includes:
If it is determined that load distribution of the first relay server is effective,
A second relay that is one of the plurality of relay servers and is different from the first relay server, with the first connection between the currently connected client terminal and the first relay server connected. A computer is connected to distribute the load of the first relay server by disconnecting the first connection after the second connection between the server and the client terminal is established and the second connection is established. A load balancing program characterized by being executed. The relay server information management procedure includes:
A maintenance advance notice procedure for obtaining a maintenance advance notice transmitted to notify the relay server of the advance notice from the load information of the relay server;
The maintenance notice notification procedure is as follows:
When the maintenance notice is acquired from the load information of the relay server, the load distribution determination procedure is applied using the relay server as the first relay server, and the client terminal connected to the first relay server is The load distribution program according to claim 7, wherein the computer executes switching from the first relay server to the second relay server so as to be connected to the second relay server.   A load distribution system comprising: the gateway device according to claim 1; a plurality of relay servers; and a client terminal connected to the relay server. A load distribution method in a load distribution system comprising a plurality of relay servers, a client terminal connected to the relay server, and a gateway device that manages load information of the relay server,
The gateway device is
A relay server information management step that is provided in an intranet and manages load information of the relay server while being regularly updated by communication;
The relay server information management step includes:
Based on load information of the first relay server to which the client terminal is connected to the relay server and load information of the plurality of relay servers, it is determined whether or not load distribution of the first relay server is to be performed. Load balancing judgment step to
The load distribution determining step includes:
If it is determined that load distribution of the first relay server is effective,
A second relay that is one of the plurality of relay servers and is different from the first relay server, with the first connection between the currently connected client terminal and the first relay server connected. A second connection between the server and the client terminal is connected, and after the second connection is established, the load is distributed to the first relay server by disconnecting the first connection. Load balancing method. The load distribution determining step includes:
Among the plurality of relay servers, when the number of the client terminals connected to the first relay server is small and the usage rate of the control unit of the relay server is within a predetermined range The load distribution method according to claim 10, wherein it is determined that load distribution of the first relay server is effective with the relay server as the second relay server. It further comprises an administrator terminal that performs maintenance of the relay server,
The relay server information management step includes:
A maintenance advance notice step for obtaining a maintenance advance notice transmitted from the manager terminal to the relay server for advance notice from the load information of the relay server,
The maintenance notice notification step includes:
When the maintenance notice is acquired from the load information of the relay server, the load distribution determination step is applied with the relay server as the first relay server, and the client terminal connected to the first relay server is The load distribution method according to claim 10, wherein switching from the first relay server to the second relay server is performed so as to be connected to the second relay server. The load distribution determining step includes:
Among the plurality of relay servers, when the number of the client terminals connected to the first relay server is small and the usage rate of the control unit of the relay server is within a predetermined range The load distribution method according to claim 12, wherein the relay server is switched from the first relay server as the second relay server.

Images