KR100826917B1 - A router system with the plural router boxes which are connected with dual tree-type control-paths - Google Patents

Abstract

A single router system through dual tree-type control paths among plural router boxes is provided to bind various router boxes to expand the bound boxes into one single router system and to simplify a control path configuration in accordance with the expansion and a malfunction coping method, thereby increasing availability and reliability of the router system. Power is applied(600), and IDs of boxes are read as an internal processor starts operating(601). It is decided whether to boot to sub boxes or to continue a standby box booting process, depending on whether the IDs of the boxes are IDs of the sub boxes(602,603). If not the IDs of the sub boxes, it is decided whether the IDs are IDs of active boxes(604). If not, an OS(Operating System) suited to standby boxes is booted, and a system state is checked(606). It is decided whether active boxes exist within a system(607). According to the decided results, the active boxes are booted(608) or the standby boxes are initialized while driving applications(609).

Description

A ROUTER SYSTEM WITH THE PLURAL ROUTER BOXES WHICH ARE CONNECTED WITH DUAL TREE-TYPE CONTROL-PATHS}

1 is a diagram illustrating a control path configuration of a single router system made of several router boxes that can be independently operated according to an exemplary embodiment of the present invention.

2 is a diagram illustrating a configuration of dual control paths centering on an active box and a standby box in the router system of FIG. 1;

3 is a diagram illustrating a control path configuration inside an active router box according to an embodiment of the present invention;

4 is an exemplary control path and data transmission and reception of the entire single router system according to an embodiment of the present invention,

5A to 5B are flowcharts illustrating operations for initializing and basic functions of an active box according to an embodiment of the present invention;

6A to 6B are flowcharts illustrating operations for initialization and basic functions of a standby box according to an embodiment of the present invention;

7A to 7B are flowcharts illustrating operations of initialization and basic functions of a sub box according to an embodiment of the present invention.

<Explanation of symbols for main parts of the drawings>

100: single router system

105,104,103,102,101: Router Box

400,105: Active box

410,104: Standby Box

230: control box around the active box

240: control path of the standby box center

319,404,414: core switch

312,316,403,408,418,413: relay switch

The present invention relates to a dual control path configuration and operation, and a method thereof, in which a plurality of router boxes capable of independent router system functions are bundled to form a single router system, and in particular, a router box. One or more Ethernet connection means capable of communicating with an internal general purpose processor and a network processor, a core switching means capable of connecting the Ethernet connection means with another Ethernet box means of another router box, a tree structure, the core switching means and Ethernet By including a means to dually configure radial control paths between router boxes using connecting means, each configured independent router box can function as a single router system by using the dual control paths configured. Easily expand the router's capacity as needed And, to an apparatus and method in a loop configuration of a control path of a dual can not improve the reliability and availability of the system.

In general, a router is a device that connects separate networks using the same transport protocol, and connects network layers to each other, and connects nodes in another network or its own network according to a routing table. In order to determine and send the packet by selecting the most efficient route among the various routes, the router performs flow control, configures several sub-networks within the inter-network, and performs various network management functions. .

Router systems need to increase the number of available interfaces and bandwidth as the number of users increases and the switching capacity to be expanded. The cost of replacing the router system and the downtime due to the replacement of the system can be relatively large. As such, expansion through the addition of system components has many advantages over capacity and interface expansion through system replacement. However, in the case of a box-type router system, unlike the router system of the service card addition method, it is difficult to expand by the component addition method.

As described above, a control path configuration or a processor-to-processor connection configuration method that facilitates an increase in inter-processor connection with the expansion of a router system and copes with an increase in load due to an increase in connection is required. Conventional technologies include a method using a PCI communication method and a method using a shared memory. However, the above methods have limitations in the expansion of a connection configuration and have difficulty in failing due to difficulty in duplication or pluralization. .

On the other hand, in the router-to-processor connection configuration method, the Ethernet-based processor connection enables a radial point-to-point connection through an Ethernet switch, and in the design and expansion of the router system with rich technology, know-how and related parts. Ease can be improved.

Conventionally, in Ethernet-based processor connection configuration, a technique used for duplication or pluralization is to configure a connection between processors in multiple paths so that loop switching on a path may be carried out by a Spanning Tree algorithm (STP) or RST (Rapid Rapid). There is a way to use the Spanning Tree algorithm, but this requires the processor to access all Ethernet switches and perform the relevant protocols, and that one processor must have more than one Ethernet channel and communicate with different Ethernet addresses. Therefore, many functions are required for multi-channel management between processors, and the switching time due to loop occurrence can be lengthened depending on the stability of related protocols, which adds complexity in stability and implementation.

Therefore, there is a need for a method that can increase the ease of design and implementation in the control path connection configuration according to the expansion of the router system, reduce the complexity in considering the fault response ability, and reduce the cost and time in the expansion. .

Therefore, an object of the present invention has been proposed to solve the above problems, one or more Ethernet connection means that the router box capable of independent router system function can communicate with the general purpose processor and the network processor in the box, Ethernet connection means Core switching means which can be connected to the Ethernet connection means of another router box in a tree structure, and means for constructing a plurality of radial control paths between the router boxes by using the core switching means and the Ethernet connection means. Means that each independent router box functions as a single router system, and that the router boxes that constitute a single router system can serve as active router boxes, standby router boxes, or subrouter boxes. By including It allows multiple router boxes to be expanded into a single router system, simplifying the control path configuration and failure response method of expansion, increasing the reliability and availability of the router system, It is an object of the present invention to provide an apparatus and method for improving design and implementation ease.

The present invention for achieving the above object is at least one Ethernet connection means capable of communicating with a general purpose processor and a network processor in the router box capable of independent router system functions; Core switching means capable of connecting the Ethernet connection means to the Ethernet connection means of another router box in a tree structure; Means for constructing a plurality of radial control paths between router boxes using the core switching means and the Ethernet connection means; Means for functioning each independent router box as one single router system using the constructed control path; A router box constituting a single router system is characterized by including means capable of acting as an active router box, a standby router box, or a sub-router box.

The above objects, features and advantages will become more apparent from the following detailed description taken in conjunction with the accompanying drawings, whereby those skilled in the art may easily implement the technical idea of the present invention. There will be.

In addition, in describing the present invention, when it is determined that the detailed description of the known technology related to the present invention may unnecessarily obscure the gist of the present invention, the detailed description thereof will be omitted. Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a diagram illustrating an embodiment of a control path configuration of a single router system made of several router boxes that can be independently operated according to an exemplary embodiment of the present invention.

As shown in FIG. 1, the single router system 100 includes a router box 101, 102, 103, 104, 105 and a power supply 106 that can be operated independently, and control messages, routing packets, and the like between the router boxes. Dual control paths 117 and 118 for data transmission and reception are included.

In this case, the router boxes 101, 102, 103, 104, and 105 are divided into an active box 105, a standby box 104, and a sub box 101, 102, 103 according to a role in the single router system 100. The box 105 and the standby box 104 include a core switching device and connection ports 119 and 118 for establishing a dual control path, and the active box 105, the standby box 104, and the sub boxes 101, 102 and 103 are respectively It includes two Ethernet connection ports 107/112, 108/113, 109/114, 110/115 and 111/116.

Dual control path (117, 118) connection configuration is the Ethernet connection port (107/112, 108/113, 109/114, 110/115, 111/116) and core switch connection port of the router boxes (101, 102, 103, 104, 105). Between (119, 118) is a double radial as shown in the right figure of FIG.

FIG. 2 is a diagram illustrating an exemplary configuration of dual control paths centering on an active box and a standby box according to the embodiment of FIG. 1.

In the single router system, the control path 200 is centered on the control path 1 230 and the standby box 202 having a radial tree structure around the active box 201 as shown in the right figure of FIG. 2. One control path 2 240 has a radial tree structure.

The control path 1 230 in the center of the active box 201 connects the point-to-point connecting means 210, 211, 212, 213, and 214 in a radial tree structure between the core switching device of the active box 201 and the Ethernet connection device of each box 201, 202, 203, 204, and 205. By building. The control path 2 240 at the center of the standby box 202 connects the point-to-point connecting means 220, 221, 222, 223, and 224 in a radial tree structure between the core switching device of the standby box 202 and the Ethernet connection device of each box 201, 202, 203, 204, and 205. By building.

3 is a diagram illustrating an embodiment of a control path configuration inside an active router box according to an embodiment of the present invention.

The router box 300 includes data processors 301 and 302 related to data path and user data packet processing and an application processor 303 related to system maintenance and control packet processing.

The data processor 301, 302 has two Ethernet interface devices 304/307, 305/308, and the application processor 303 also has two interface devices 306/309. The two Ethernet interface devices of the processors 301, 302 and 303 are connected in a point-to-point manner with two different Ethernet switches 312 and 316.

The uplinks 313 and 317 of the two Ethernet switches 312 and 316 above are connected to the control path connection ports 320 and 323 in the back plate 321 of the router box 300, which are connected to the ports 320 and 323. The one 320 is connected 322 to the connection port 325 of the core switch 319 of the active box, and the other 323 is connected 324 to the core switch of the standby box.

The Ethernet switches 312 and 316 and the core switch 319 in the router box 300 are connected to the switch monitoring interfaces 314, 318 and 310 of the application processor 303, and the application processor 303 controls each switch through the interface. And maintain it.

4 is a diagram illustrating an embodiment of transmitting and receiving control paths and data of the entire single router system according to the present invention.

The active box 400 includes an active application processor that performs functions such as management and routing packet processing of a single router system as a whole, and control path access means for transmitting and receiving management messages and routing packets of the entire system by the application processor. Two Ethernet interfaces 401 and 402 are used. The data processor also accesses the entire system's control path through two Ethernet interfaces, like the application processor. The active box 400 includes an active core switch 404 for system-wide control path configuration, and point-to-point each box and control path through a core switch connection port on the back plate of the box. Connection configurations 409 and 407 in a radial manner.

The standby box 410 includes an application processor serving as a backup function of an application box of an active box, and uses two Ethernet interfaces 411 and 412 for accessing a control path of the application processor. The standby box 410 includes an active core switch 414 for configuration of a control path separate from the system control path centered on the core switch 404 in the active box 400, and on the back plate of the box. Each box and control path are connected (419, 417) in a point-to-point radial fashion via a core switch connection port.

The active box 400, the standby box 410, and the sub box 420, 430, 440 boxes all include two relay switches 403, 408, 418, 413, 421, 425, 431, 435, 441, 445 in addition to the core switch. In addition, all processors in the box go through the above two relay switches before connecting the entire control path of the system. One relay switch (403, 418, 421, 431, 441) is a control path (409, 407) centered on the active box. ) Relays the Ethernet interfaces 405, 415, 423, 433, 443 of the processor, and another relay switch 408, 413, 425, 435, 445 is a control path 419, 417 configured around the standby box. Relays the processor's Ethernet interfaces (406, 416, 424, 434, 444).

The core switches 422, 432, and 442 may also be mounted in the sub boxes 420, 430, and 440. However, when the functions are limited to the sub boxes, the core switches are removed or the functions are down.

5A to 5B are flowcharts illustrating an example of operation of initializing and basic functions of an active box according to an exemplary embodiment of the present invention.

As shown in FIGS. 5A to 5B, the active box initialization method according to the present invention is first powered on (500), and the ID of the box is read (501) while the processor starts to operate. It is determined whether the booting into the sub-box (503) or the active box booting process (402) according to the ID of the.

In this case, if it is determined that it is not a sub box ID, it is determined whether it is an active box ID (504). If it is not an active box ID, the standby box is booted (505). And check the system status (506). Subsequently, it is determined whether the active box already exists in the system according to whether or not a message indicating the presence of the active box exists in the system (507), and booting to the standby box (508) or initializing the active box and starting the application (509) according to the determination result )do.

After the OS booting and basic initialization of the active box are completed as described above, a function 510 is performed to receive and process an open interface (OPEN) request from another box, which waits for an open request from another box (511). In operation 512, the box information, the processor information, and the interface information are stored. Then, it sends a response message for the open request and starts transmitting and receiving a heartbeat message (513), and when all processors of the open request box have completed opening the control path (514), the open request box is approved and notified of the system ( 515).

In addition, after the OS boot and basic initialization are completed, the active box starts a function 520 of checking the operation state of the box and the control path in which the system subscription is completed. If it is determined that the operation is not performed properly (521), and if an abnormality occurs as a result of determination, the box information, processor information, and interface information are stored (522). Then, based on this information, if all of the interfaces of the error-prone processor fail (523), the box is defined as a fault condition, all the boxes in the system are notified and the system stops functioning (524). ).

In addition, the active box starts a function 530 to check whether the box is abnormal after OS boot and basic initialization. This function continuously checks the self-state of power, clock, interface, function, etc. in the box and checks for abnormality. In operation 531, if a failure or abnormality occurs, the existence of a standby is confirmed (532). Thereafter, the standby box is notified of the failure of the active box (533), the function of the active box is interrupted (534) and the power down (535).

6A to 6B are flowcharts illustrating an operation of an embodiment of initialization and basic functions of a standby box according to the present invention.

6A to 6B, the standby box initialization method according to the present invention, the power is first applied (600), while the internal processor starts to read the ID of the box (601), the box It is determined whether to boot into the sub box (603) or to continue the standby box boot process (602) according to the ID of the. In addition, as a result of the determination, if it is not the sub box ID, it is determined whether it is an active box ID (604). If the active box ID is correct, boot into the active box (605), and if it is not the active box ID, the OS corresponding to the standby box is determined. Boot and check 606 the system status. Then, it is determined whether the active box already exists in the system (607) according to whether or not a message indicating the presence of the active box in the system, booting to the active box (608) or initializing the standby box and starting the application according to the determination result. (609).

As described above, after the OS booting and basic initialization of the standby box are completed, the active box performs an open request for all Ethernet interfaces of the box and joins the system (610). This function selects a target processor and an interface to make a control path open request and sends a control path open request for the interface to the active box (611), waiting for the active box open request (612), and the open request Receives a response to manage the information of the interface and starts the step of transmitting and receiving the heartbeat message (613). Then, after confirming that the control paths of all the processors in the box are open (614) and waiting for the system subscription approval notification (615), when the subscription is completed, data transmission and reception with other boxes are started through the open control path ( 616).

In addition, the standby box starts a function 620 of checking the operation state of the control path in the state where the system subscription is completed after the OS boot and basic initialization are completed. This function determines whether or not the heartbeat message transmission and reception is not operating properly in the system subscription complete state (621), and if the abnormality occurs as a result of the determination, the processor information and interface information is stored (622). In addition, based on this information, if all of the interfaces of the error-prone processor fail (623), the box is notified of all failures, the fault information is stored, and the system is removed (624). It is set as the state (625).

In addition, the standby box starts a function 630 to check whether the box is abnormal after the OS boot and basic initialization are completed. This function checks the self-state of power, clock, interface, and functions in the box. It continuously checks and checks for abnormality (631). At this time, if failure or abnormality is confirmed, the active box is notified of the failure of the active box (632), the function of the box is interrupted and the failure information is stored, and the system is removed from the system (633). Down (634).

In addition, the standby box starts a function 640 that replaces the function of the active box when the active box fails and requires switching after the OS boot and basic initialization are completed. When the function is notified of the failure from the active box (641), the active box responds to the failure notification (642), then notifies the other box of the failure of the existing active box, stops the function of the existing active box, and enters the standby state. Switch the function of the box to the active state (643).

7A to 7B are flowcharts illustrating operations of an embodiment of initialization and basic functions of a sub box according to an embodiment of the present invention.

As shown in FIG. 7A to FIG. 7B, in the initialization method of the sub-box according to the present invention, power is first applied 700, and the ID of the box is read (701) while an internal processor starts operation. Subsequently, it is determined whether to boot into the active / standby box or continue the sub box booting process according to the ID of the box (702), and if the sub box ID is correct, boot the OS corresponding to the sub box and check the system status (704). )do. Then, it is determined whether the active box already exists in the system according to whether or not a message indicating the presence of the active box in the system (705), and if the determination results in the initialization of the sub-box and the application is started (706). If there is no active box, it will not boot and wait for the active box.

As described above, after OS booting and basic initialization of the sub box are completed, an open request for all Ethernet interfaces of the box to the active box is performed and a function 710 for joining the system is performed. This function selects a target processor and an interface to make a control path open request and sends a control path open request for the interface to the active box (711), waits for an active box open request (712), and the open request. Receives a response to the management of the information of the interface and begins the step of transmitting and receiving a heartbeat message (713). Subsequently, after confirming that the control paths of all the processors in the boxes have been opened (714), the system waits for a system subscription approval notification (715). When the subscription is completed, data transmission and reception with other boxes are started through the open control path ( 716).

In addition, after the OS booting and basic initialization are completed, the sub box starts a function 720 of checking an operation state of a control path in a state where a system subscription is completed. This function determines whether or not the heartbeat message transmission and reception in the system subscription complete state is not properly operated (721), if the abnormality occurs as a result of the determination, and stores the processor information and interface information (722). Based on this information, if all of the interfaces of the error-prone processor fail (723), the failure status of the box is notified to all boxes, the fault information is saved, and the system is removed (724). It is set as a state (725).

In addition, the sub box starts a function 730 to check whether the box is abnormal after the OS boot and basic initialization are completed. This function continuously checks the self-state of the power supply, clock, interface, function, etc. in the box and checks whether there is an error (731), and notifies the active box of the failure of the active box (732) when a failure or abnormality occurs. The box then ceases to function, stores the fault information, exits 733 from the system, and then powers down 734.

Meanwhile, the embodiments of the present invention described above have been described in detail, but various modifications can be made without departing from the scope of the present invention. Therefore, the scope of the invention should be determined by the claims rather than by the described embodiments.

As described above, the present invention provides a router box capable of an independent router system function, wherein at least one Ethernet connection means capable of communicating with an internal general-purpose processor and a network processor, and the Ethernet connection means are connected to the Ethernet connection means of another router box. By using a dual control path configured to include a plurality of core switching means that can be connected in a tree structure, and means for configuring a plurality of radial control paths between router boxes using the core switching means and Ethernet connection means. Each independent router box can function as a single router system, facilitating the expansion of the router's capacity as needed, and increasing the reliability and availability of the system by configuring a loopless control path. There is.

Claims (15)

Single router system with dual radial control path configuration between multiple router boxes, Providing one or more Ethernet connection devices capable of communicating with a general purpose processor and a network processor in a box, the switches connecting the Ethernet connection devices to the Ethernet connection devices of other router boxes in a tree structure, wherein the switch and Ethernet A plurality of router boxes capable of forming a plurality of radial control paths between router boxes by connecting devices; Box recognition means for checking the role of each router box and recognizing it as an active box that manages and operates the entire system, a standby box as its backup box, and a sub box controlled by the active box; Storage means for storing an OS and an application image corresponding to the recognition result Router system comprising a. The method of claim 1, The control path is a router system, characterized in that used as a data transmission and reception path for tying the independent router box to operate as a single router system. The method of claim 1, The Ethernet connection between the plurality of router boxes is connected in a dual radial point-to-point connection configuration around the core switch in the active box and the standby box. The method of claim 1, The active box receives an open request to open a control path from the standby box and the sub-box, stores the requested box information, processor information, and interface information, and then gives a control path open request response, and provides a corresponding box / processor / interface. And a heartbeat message to start receiving and sending, if all the interfaces of the requested box is opened, the box is subscribed to one of the router system, characterized in that notifying the box. The method of claim 1, If the heartbeat message does not come from the box of the router system subscription state, the active box stores the box information, the processor information, and the interface information, and then continues to transmit and receive data through the alternate interface. A router system, characterized in that the box is defined as a fault state and the other box in a normal state is notified of the fault state of the box to stop the function of the fault box in the system. The method of claim 1, The active box checks whether the router box is abnormal, and if an error occurs, notifies the standby box of a failure, passes the active box function to the standby box, and stops the function of the box. The method of claim 1, And the plurality of router boxes are initialized as active boxes, standby boxes, and sub boxes that perform an active, standby, or sub function through an ID indicating the identity of each router box. The method of claim 7, wherein The ID indicating the identity of the box is stored in a physical jumper setting or a nonvolatile memory. The method of claim 7, wherein The active box is initialized to a standby state when the active box detects the presence of a box that is already functioning as an active box through an active box notification message coming into the Ethernet interface of the box. . The method of claim 1, The standby box and the sub-box open to open the control path to the active box, and upon receiving the control path open request response message, start the open interface information management and give a heartbeat message through the active box and the corresponding interface. After repeating the request for request to open the control path to the active box until all interfaces of the corresponding standby box or subbox are open, all the interfaces are opened and a system subscription approval message is received from the active box. A router system, characterized by initiating an exchange. The method of claim 1, The standby box and the sub-box, if the heartbeat message does not come out in the state that the router system subscription is finished, or if a specific interface is a failure state, and stores the failure state of the interface, and continues to communicate with the active box through an alternate interface, And if the replacement interface fails, the router box is in an error state, and the router box function is suspended. The method of claim 1, The standby box and the sub-box checks whether the router box is abnormal, and if an error occurs, notifies the active box of a failure and stops the function of the router box. The method of claim 1, When the standby box is notified of the failure state of the active box from the active box, the function of the switch to the active box, router system, characterized in that to stop the function of the existing active box. The method of claim 1, The standby box detects the existence of a router box that already functions as an active box during booting into the standby box through an active box notification message coming into the Ethernet interface of the box, and initializes the active box if there is no active box. Router system, characterized in that. The method of claim 1, The sub box detects the existence of a router box that already functions as an active box during booting into the sub box through an active box notification message coming into the Ethernet interface of the box, and if the active box does not exist, the active box completes booting. Router system, characterized by waiting until.

Images