JP2009194787A - Gateway apparatus - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To continue important communication by a standby system even when a fault is generated in an active system in a gateway apparatus having a configuration composed of standby systems of which the number is smaller than the number of a plurality of active systems. <P>SOLUTION: In user authentication, priority is set in each user or service. Out of session information managed by the active systems, only session information with high priority is copied to the standby systems. The frequency of copying session information is changed in accordance with the priority of each session. By reducing an information transfer amount from the active systems to the standby systems, a redundant configuration reducing processing loads can be achieved. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a redundant configuration in a gateway device.

  In mission-critical communication, even when a failure occurs in a communication path or communication device, it is required to continue communication and to quickly recover from the failure.

  Generally, in a network device, communication is continued even when a failure occurs in a certain component by providing a plurality of components that perform the same processing (redundancy). Network device redundancy is realized by combining the system used for actual operation (hereinafter referred to as the active system) with the system that takes over processing when a failure occurs in the active system (hereinafter referred to as the standby system). To do.

As an example of redundancy, Japanese Patent Application Laid-Open No. 2007-251809 shows a configuration in which an active system and a standby system using the same hardware as the active system correspond one-to-one.
Japanese Patent Application Laid-Open No. 2004-120062 discloses a switching method of an N + 1 redundant configuration relating to a line termination unit.
JP 2007-251809 A JP 2004-120062 JP

  The IEEE 802.16e standard is a wireless connection method that supports a function of continuing communication while switching a base station to which a terminal is connected while moving. The wireless terminal receives user authentication from the authentication server by a user ID, a password, or the like before starting communication. In addition, by supporting mobile IP (Internet Protocol) defined by IETF (Internet Engineering Task Force), communication can be performed without changing the address of the terminal even in a situation where the connection destination base station is switched as the terminal moves. Can be continued.

  A gateway device installed at the boundary between a radio access network and a backbone network is responsible for processing for establishing a session, such as user authentication and IP address assignment to the terminal, when the mobile terminal accesses the network. The gateway device manages information related to a session, that is, an IP address assigned to a mobile terminal, an address of a partner for setting a tunnel in the mobile IP, and the like.

  In order to continue communication when a failure occurs in the processing unit that manages the session in the gateway device, it is necessary for the standby system to take over information related to the session held by the active system of the redundant processing unit. Session information takeover processing is performed by copying session information held by the active system to the standby system as needed.

  When a single standby system is provided for one active system, as in JP 2007-251809, the active system and the standby system are configured by configuring the active system and the standby system with the same hardware. Information synchronization can be achieved relatively easily. With such a configuration, it is possible to improve reliability, but the processing capacity available for communication in the entire apparatus is 50%.

  On the other hand, as shown in Japanese Patent Application Laid-Open No. 2004-120062, by providing a plurality of active systems with a smaller number of standby systems than the number of active systems, it is possible to increase the available processing capacity of the entire apparatus. become.

  However, in this configuration, if the active system and the standby system are realized by the same hardware, there is a possibility that all session information in the active system cannot be taken over. That is, there is a possibility that all session information held by the active system cannot be copied to the standby system due to restrictions on the amount of memory or CPU capacity. In addition, since the redundancy processing is a burden, the processing capacity of the active system may be reduced. For example, in the case of a configuration with four active systems capable of accommodating a maximum of 1000 sessions per hardware and a single standby system, copy the session information held by all active systems to the standby system. The standby system needs to accommodate up to 4000 sessions.

  If the hardware configurations of the active system and the standby system are the same, it is impossible to store all the session information held by all the active system hardware with the standby system hardware. When a failure occurs, a session that cannot continue communication occurs.

  An object of the present invention is to realize a system switching process according to the importance of a session in a gateway apparatus configured by a smaller number of standby systems than the number of active systems for a plurality of active systems.

  In order to solve the above problems, a gateway device according to the present invention comprises a plurality of active processing units and a plurality of active processing units and one or more standby processing units fewer than the active processing units. When each active processing unit establishes a user session, a table for managing session information is created, and an identifier corresponding to the priority of the session is included in each user session entry in the table for managing session information. Among the entries for each user session, the identifier is used to determine whether or not to copy the entry to the one or more standby processing units, and the identifier indicating copying is set in the determination. Only the entry of the registered user session is copied to the processing unit of the standby system. .

The gateway device according to the present invention is a table for managing session information provided in the standby processing unit when the failure occurs in the active processing unit in the gateway device. Session information relating to the processing unit of the active system in which the failure has occurred is taken out, and the processing of the processing unit of the active system in which the failure has occurred is taken over.

In order to solve the above problems, a gateway device according to the present invention is a gateway device configured redundantly by a plurality of active processing units and one or more standby processing units smaller than the active processing units. When a plurality of active processing units establish a user session, a table for managing the session information of the user session is created, and the priority of the session is added to the entry for each user session in the table for managing the session information. An identifier is set according to the degree, and when the entry for each user session is copied to the one or more standby processing units, the entry is copied and is determined in advance corresponding to the identifier. Is determined by the identifier set in the entry, and the session of the user session is determined. If the elapsed time of the emissions information from the time that the last copy is greater than the determined copy cycle for that entry, characterized by copying the entry to the processing unit of the standby system.

The gateway device according to the present invention is a table for managing session information provided in the standby processing unit when the failure occurs in the active processing unit in the gateway device. Session information relating to the processing unit of the active system in which the failure has occurred is taken out, and the processing of the processing unit of the active system in which the failure has occurred is taken over.

  According to the present invention, it is possible to reduce the amount of information transferred from the active system to the standby system and realize a simple redundant configuration with a reduced processing load. In addition, reliability according to the user and service grade can be secured.

  Embodiments of the present invention will be described below with reference to the drawings.

  FIG. 1 illustrates an example of a network configuration to which the present invention is applied.

  The base station 3 is connected to a terminal (MS) 2 via radio. The gateway device 1 is connected to the base station 3 and functions as a gateway between the access network 7 and the backbone network 8. In the backbone network 8, a DHCP (Dynamic Host Configuration Protocol) server 6, an authentication server (AAA) 5, and a mobile IP home agent (HA) 4 defined by IETF RFC3344 are installed. The gateway device 1 performs user authentication processing with the authentication server 5. The gateway device 1 performs mobile IP processing with the home agent 4. The DHCP server assigns an IP address to the terminal 2 via the gateway device 1.

  In FIG. 1, the gateway device 1 has a plurality of CPU boards that perform processing for establishing a user session. Further, the gateway device 1 performs processing for determining the priority of a session at the time of user authentication, for example, with communication by the terminals 2-1 and 2-3 as a priority session and communication by the terminal 2-2 as a non-priority session.

  Although FIG. 1 shows an example of network access using radio, the network in which the gateway device to which the present invention is applied is not necessarily limited to radio access.

  FIG. 2 shows a configuration example of the gateway apparatus 1 to which the present invention is applied.

  The gateway device 1 performs a packet transfer process, a process for establishing a session, a process for managing session information, a CPU board 11, an input / output process with a physical line, and a distribution process to the CPU board. The switch board 12, the CPU board 11, and the control unit 13 that performs processing for controlling the switch board 12 are configured. By mounting a plurality of CPU boards 11, load distribution of processing on the CPU board and redundancy for switching processing between the CPU boards when a failure occurs are realized.

  FIG. 3 shows a configuration example of the CPU board 11 to which the present invention is applied.

  The CPU board 11 includes a switch I / F (interface) 111 that is an input / output interface with the switch board 12, a packet buffer 112 that stores transmission / reception packets processed by the CPU board 11, transfer processing for transmission / reception packets, user session establishment, user The CPU 113 that performs session management, a work area of the CPU 113, a memory 114 that holds a session management table that manages information related to a session, and a control unit I / F (interface) 115 that is an interface with control.

  FIG. 11 is a configuration example of a session management table 1000 provided on the active CPU board 11.

  The session management table 1000 includes a session identifier 1001 for identifying a session, an IP address assigned to the terminal 2, key information of a GRE (Generic Routing Encapsulation) tunnel set between the base station 3 and the gateway device 1, mobile IP Time at which the session information related to the session is updated in the user session information 1002 and the session management table 1000, which are information necessary for managing the session, such as the IP address of the home agent that is the connection destination, information on the base station, etc. , An entry including a field for setting an identifier 1004 indicating the priority of the session is provided for each session.

  FIG. 12 shows a configuration example of the session management table 1100 provided on the standby CPU board.

  The session management table 1100 provided in the standby system is configured by copying information of the session management table 1000 provided in the active CPU board shown in FIG.

  The session management table 1100 provided in the standby system includes a CPU number 1101 indicating which CPU board the information is copied from, a session identifier 1102 for identifying the session, an IP address assigned to the terminal, a base Information necessary for managing the session such as GRE (Generic Routing Encapsulation) tunnel key information set between the station 2 and the gateway device 1, the IP address of the home agent to which the mobile IP is connected, information on the base station, etc. The user session information 1103 is a field for setting the time 1104 when the session information is updated by copying the session information related to the session.

  FIG. 4 is a flowchart for explaining a procedure in which the active CPU board 11 creates the session management table 1000 when a user session is established.

  The gateway device 1 performs user authentication by communicating with the authentication server (AAA) 5 according to the sequence shown in FIG. 6 (S110).

  In the user authentication process, the terminal 2 transmits a user ID and password to the gateway device 1, and the gateway device 1 sends the user ID and password to the authentication server (AAA) 5 as shown in the sequence of FIG. And the authentication server (AAA) 5 returns the user attribute information. As a user authentication protocol used between the gateway device 1 and the authentication server (AAA) 5, for example, RADIUS (Remote Authentication Dial In Service) is used.

  After the user authentication process is performed in step S110, a session entry is created in the session management table 1000 (S120). The gateway device 1 acquires session priority class information from the user attribute information of the corresponding session acquired by the user authentication processing of FIG. 6 performed with the authentication server (AAA) 5 (S130). In the subsequent session priority determination step (S140), the value of the identifier (1004) indicating the priority of the session is set in the entry of the session management table 1000 based on the priority class information of the session. Is finished (S170).

  Here, in general, when three or more classes are provided as session priority classes, in the session priority determination step (S140), an identifier (1004) indicating the session priority based on the session priority class. To the entry. Here, the number of identifiers (1004) indicating the session priority is less than or equal to the number of classes of the priority class. For example, if there are 5 classes as session priority classes, the identifier “1” for class 1 and the identifier “2” for classes 2 and 3 are used as the identifier (1004) indicating the session priority. Can be set to the entry of the session for class 4 and 5.

In the session priority determination step (S140), the identifier (1004) indicating the session priority can be determined and set as only two patterns of priority / non-priority. FIG. 4 is a flowchart showing the procedure in this case. In FIG. 4, when the session is a priority session, the value of a flag (1004) serving as an identifier indicating the priority of the session is set to 1 (S160). If the session is a non-priority session, the value of a flag (1004) serving as an identifier indicating the priority of the session is set to 0 (S160). Then, the value of the flag is set in the entry of the session management table 1000 created in step S120, and the process ends (S170).

  FIG. 5 shows a procedure for copying session information from the active CPU board to the standby CPU board in order to realize redundancy among a plurality of CPU boards 11, and a failure in one active CPU board. FIG. 11 is a sequence diagram for explaining a procedure for switching a system to a standby CPU board when an error occurs. For the sake of simplicity, FIG. 5 shows an example of a redundant configuration including two active CPU boards and one standby CPU board.

  In the initial state, the two CPU boards of CPU # 1 (11-1) and CPU # 2 (11-2) are used as active CPU boards (2000-1, 2000-2). Further, the CPU board of CPU # 3 (11-3) is set as a standby CPU board (2000-3).

  CPU # 1 (11-1) and CPU # 2 (11-2), which are active CPU boards, periodically store session information held by the CPU board as CPU #, which is a standby CPU board. 3 (11-3) is copied (2100-1, 2100-2).

  If a failure occurs in the active CPU board CPU # 1 (11-1) (2200), the standby CPU board CPU # 3 (11-3) fails in the CPU # 1 (11-1). Is detected (2250). CPU # 3 (11-3) detects that a failure has occurred in CPU # 1 (11-1) by, for example, sending a signal notifying the occurrence of a failure transmitted from CPU # 1 (11-1) to CPU # 1 (11-1). 3 (11-3) or the standby CPU board (11-3) receives a signal periodically transmitted by the active CPU board CPU # 1 (11-1). It is done by disappearing.

  The spare CPU board CPU # 3 (11-3) performs a system switching process for functioning as the active system in place of the CPU board CPU # 1 (11-1) in which the failure has occurred (2300).

  The CPU # 3 (11-3) notifies the CPU # 2 (11-2) which is the active CPU board that the system has been switched and started as the active system (2400). Thereafter, CPU # 3 functions as an active CPU board (2500).

  The CPU # 1 (11-1), which is the CPU board in which the failure has occurred, replaces the hardware or restarts the software installed in the CPU board, and then performs the recovery process (2600), and then the standby CPU It functions as a board (2700).

  The CPU # 1 (11-1) notifies the active CPU board CPU # 2 (11-2) and CPU # 3 (11-3) that it has been activated as a standby system (2800- 1, 2800-2). Thereafter, the active CPU board CPU2 (11-2) and CPU # 3 (11-3) copy the session information to the CPU # 1 (11-1), which is the standby CPU board ( 2900-1, 2900-2).

  FIG. 7 is a flowchart for explaining a system switching process performed by the standby CPU board when a failure occurs in the active CPU board.

  First, the standby CPU board detects that a failure has occurred in the active CPU board (S210).

  With respect to the entries in the session management table provided in the standby system, only entries relating to the active CPU board in which the failure has occurred are left, and entries relating to other active CPU boards in which no failure has occurred are deleted (S220).

  The switch of the system from the standby system to the active system is notified to the active CPU board where no failure has occurred (S230), and the CPU board is switched to the active system (S240). Thereafter, it functions as the active system by taking over the session information of the CPU board in which the failure has occurred (S250).

  Next, a processing procedure for copying session information from the active system to the standby system will be described with reference to FIG. FIG. 8 shows the case where the identifier (1004) indicating the priority of the session is determined as two priority / non-priority patterns and the identifier flag is set in the entry in the session priority determination step (S140) of FIG. It is a flowchart explaining the process for copying session information from a system to a standby system. If the session is a priority session, the value of the flag (1004) indicating the priority of the session is 1, and if the session is a non-priority session, the flag (1004) is an identifier indicating the priority of the session ( The value of 1004) is 0.

  First, the CPU board compares the time when the session information was previously copied to the standby CPU board with the update time of the session (S310). In step 310, if the session information of the session has not been updated since the previous copy time, the process ends (S340). This process reduces the processing load by copying only information that has been updated since the previous copy.

  If the session has been updated since the previous copy, it is next determined whether or not the session is a priority session (the identifier flag value indicating the priority / non-priority of the session is 1 or 0). (S320). If the value of the identifier flag indicating the priority / non-priority of the session is “1”, the session information is copied assuming that the session is a priority session (S330), and the process is terminated (S340). On the other hand, if the value of the identifier flag indicating priority / non-priority of the session is “0” in step S320, the session information is not copied and the process ends without assuming that the session is not a non-priority session ( S340).

  In the present invention, in the process of copying session information from the active system to the standby system, the frequency of copying can be changed according to the priority of the session.

Here, a process of performing copying by changing the frequency of taking over session information from the active system to the standby system according to the priority of the session will be described. FIG. 9 is a flowchart for explaining the above process.
When the CPU board that inherits the session information performs a copy process, first, the session information is copied to the standby CPU board last time and the update time of the session is compared. If the session information is not updated after the previous copy time, the process of ending the process is the same as in the above embodiment (corresponding to (S310) in FIG. 8).

  First, an identifier (1004) indicating the priority of the session is determined (S400). Here, for each identifier, a corresponding session information copy cycle is determined in advance, and the process branches so that the session information copy cycle differs in accordance with the identifier. For example, when the identifier is “0, 1, 2,..., N” and the session information copy period corresponding to the identifier is determined to be short in order from the smallest value, Then, a determination process is performed to determine whether or not a predetermined copy cycle has elapsed (S420 to S423).

  In the determination process, the time (1104) when the session information related to the session was copied last time is compared with the time when the copy process is performed, and the elapsed time since the last copy is obtained. If the elapsed time from the previous copy is longer than the copy period of the session information determined corresponding to the identifier, it is determined that the copy period has passed, while the copy period is longer than the copy period. If the elapsed time is small, it is determined that the predetermined copy cycle has not elapsed.

  Then, according to the determination process, if the copy cycle has elapsed, the session information is copied (S410), and the process ends (S430). If the copy cycle has not elapsed, copying is not performed, and the process ends (S430).

  In the session priority determination step (S140), when the identifier (1004) indicating the priority of the session has only two patterns of priority / non-priority, the session information is copied from the active system to the standby system. Processing to be performed will be described. FIG. 10 is a flowchart showing the procedure in this case.

  First, in step S500, it is determined whether or not the session is a priority session (the identifier flag value indicating priority / non-priority of the session is 1 or 0). If the session is a priority session (the value of the identifier flag indicating the priority / non-priority of the session is 1), the session information is copied (S510), and the process ends (S530). On the other hand, if it is determined in step S500 that the session is not a priority session (the identifier flag value indicating the priority / non-priority of the session is 0), it is determined whether the copy period corresponds to the non-priority session. (S520).

  Here, the session information copy cycle for the non-priority session is determined in advance. In the determination process, the time (1104) at which the session information related to the session was previously copied is compared with the time at which the copy process is performed. , Find the elapsed time since the last copy. If the elapsed time from the previous copy is larger than the copy cycle determined for the non-priority session, it is determined that the copy cycle has elapsed, while the elapsed time is smaller than the copy cycle. In this case, it is determined that the predetermined copy cycle has not elapsed.

  If the copy period corresponding to the non-priority session has elapsed according to the determination process, the session information is copied (S510), and the process is terminated (S530). On the other hand, if the copy period corresponding to the non-priority session has not elapsed in step 520, the process ends (S530).

Example of network configuration to which the present invention is applied Configuration example of gateway device to which the present invention is applied Configuration example of control unit to which the present invention is applied Session management table creation flowchart when establishing a session Sequence between CPU boards for redundancy processing Authentication sequence between gateway and AAA Flow chart for operation of standby CPU when failure occurs Flow chart showing an example of a procedure for copying to a redundant system (example of determining whether copying is possible with priority / non-priority) Flow chart showing an example of the copy procedure to the redundant system (example of changing the copy frequency) Flow chart showing an example of a copy procedure to a redundant system (example of changing copy frequency between priority and non-priority) Configuration example of the active session management table Standby session management table

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 Gateway apparatus 2 Terminal 3 Base station 4 Home agent 5 Authentication server 6 DHCP server 7 Access network 8 Backbone network 11 CPU board 12 Switch board 13 Control part 111 Switch interface 112 Packet buffer 113 CPU
114 Memory 115 Control unit interface 1000 Session information management table 1100 possessed by active CPU board Session information management table possessed by standby CPU board

Claims (3)

In a gateway apparatus configured redundantly by a plurality of active processing units and one or more standby processing units smaller than the active processing units,
When each of the multiple active processing units establishes a user session, create a table to manage the session information of the user session.
An identifier corresponding to the priority of the session is set in the entry for each user session in the session information management table,
Of the entries for each user session, it is determined whether to copy the entry to the one or more standby processing units based on the identifier, and the user session entry in which the identifier indicating copying is set Is copied to the standby processing unit. In a gateway apparatus configured redundantly by a plurality of active processing units and one or more standby processing units smaller than the active processing units,
When each of the multiple active processing units establishes a user session, create a table to manage the session information of the user session.
An identifier corresponding to the priority of the session is set in the entry for each user session in the session information management table,
When copying the entry for each user session to the one or more standby processing units, a cycle for copying the entry, which is set in advance corresponding to the identifier, is set in the entry. Discriminated by the identifier,
When the elapsed time from the time when the session information of the user session is copied last time is longer than the determined copy period for the entry, the entry is copied to the standby processing unit. Gateway device. In the gateway device according to claim 1 or 2,
When a failure occurs in the active processing unit,
The standby processing unit extracts session information relating to the active processing unit in which the failure has occurred from a table for managing session information provided in the standby processing unit, and the active processing unit in which the failure has occurred A gateway device that takes over the processing of the above.