KR20090074447A - Method of changing a switch board - Google Patents

Abstract

A method for changing a switchboard is provided to quickly change a port and continuously offer a service through the network by enabling each service board to change the port to a second switchboard when abnormality occurrence information of a first switchboard is directly received from a chassis manager. When a message is received from a chassis manager, each service board checks whether the message is related to a first switchboard(506). When the message is not related, each service board continuously performs communication through the first switchboard. When the message is related, each service board determines to continuously use the communication through the first switchboard or use the communication through a second switchboard by determining whether the message is related to failure of the first switchboard(516).

Description

How to change the switch board {METHOD OF CHANGING A SWITCH BOARD}

The present invention relates to an ATCA platform device, and more particularly, to a method of changing a switch board of an ATCA platform device.

In general, as network equipment companies release various network equipments, there is a problem of increasing complexity and management costs for managing various products from a carrier's point of view.

To overcome this problem, the telecommunications industry is introducing a method by which the computer industry maximizes efficiency by converting a vertically integrated supply system into a horizontal cooperative model by increasing mutual compatibility through standardization. It is an Advanced Telecom Computing Architecture (ATCA) platform with more than 40 companies participating in standardization. The ATCA platform uses reusable chassis, boards, switches, CPUs, and programmable network processors to enable equipment manufacturers to mass produce standards-based products.

In other words, the ATCA platform enables network systems to be designed in accordance with the PCI Industrial Computer Manufacturers Group (PICMG) standardized specifications, thereby realizing the time savings and economics in developing network systems.

In this ATCA platform, a fabric interface specification for packet forwarding is defined between the switch board and a plurality of service boards, and the fabric interface specification is based on the switch boards such as star topology, dual star topology, and full mesh topology. It is defined.

1 is a block diagram illustrating a structure of a general ATCA platform device.

The ATCA platform device using the ATCA platform as described above, as shown in Figure 1, the chassis management unit (also referred to as the main processor device) (hereinafter, simply referred to as 'chassis management unit') 10, switch board (packet switch 20) and a number of service boards (also referred to as packet processing devices, user matching boards, packet processing units, and I / O devices) (hereinafter simply referred to as 'service boards') (30-1 to 30-m, 40-1 to 40-n). Here, the chassis management unit and the switch board may be of a dual structure (10-1, 10-2, 20-1, 20-2). In addition, the service boards may be divided into the first service board unit 30 and the second service board unit 40 with the switch board as a boundary, and each service board unit is provided with a plurality of service boards.

On the other hand, in the ATCA platform device as described above, communication between service boards is duplicated by a technique called IP bonding (or link aggregation) using ICMP packets. However, the port duplication method implemented by IP bonding has a problem in that it takes a long time to change a port by detecting a failure of a switch board. In other words, due to the nature of ATCA platform devices, all boards mounted in the chassis can be easily ejected / inserted (mounted). This structure has the advantage of being scalable in terms of hardware, but in terms of service availability. In this case, the service down time of the service is long due to the detachment / mounting of the board.

In particular, although the hernia of the switch board which acts as an intermediary of IP communication for all the service boards is a big hit from the service side, the conventional method as described above, the switch in the chassis forming the appearance of the ATCA platform device There is a problem in that it is impossible to accurately determine whether a fault such as a board removal or mounting is performed using an ICMP packet.

Hereinafter, by determining whether the first switch board performing the packet exchange between the service boards (Eject), by changing the first switch board to the second switch board to perform the packet exchange between the service boards The conventional method is described with reference to FIGS. 1 and 2.

Figure 2 is a flow diagram of one embodiment of a conventional switch board change method.

When the ATCA platform device is generally driven without particular obstacles, as shown in FIG. 1, when the mth service board 30-m communicates with the nth service board 40-n, the first switch board IP communication is performed via (72).

Meanwhile, when the first switch board is ejected for replacement or repair, or when the OS of the first switch board is shut down (74), the m-th service board and the n-th service board are ICMP response packets from the first switch board. It is sensed that this is not coming (76). That is, the m-th service board or the n-th service board receives an ICMP response packet from the first switch board to check whether IP communication is performed smoothly. When the first switch board is detached or other failure occurs, Since the same ICMP response packet is not transmitted from the first switch board, the service board determines that the ICMP response packet has not been received after a predetermined time has elapsed. At this time, the wait time for the service board to determine that the above ICMP response packet is not received may be set differently for each system, but generally set to about 10 seconds after packet transmission in consideration of the load of the system itself. It is becoming.

As described above, when the ICMP response packet is not received even after the preset waiting time elapses (76), the m-th service board and the n-th service board determine to communicate through the second switch board, and internal communication Change the port to the second switch board (78).

However, the conventional method as described above not only takes a long time to change the communication port to the second switch when a failure of the first switch board occurs, but also simply changes the switch board using no ICMP packet response. There is a problem in that the communication port cannot be changed by accurately determining whether the service board is removed or mounted.

For example, in an actual ATCA platform device operating environment, when a large number of network packets are passed through a switch board, ICMP packet loss between service boards becomes severe. In the conventional method as described above, A simple ICMP packet loss phenomenon is regarded as a failure of the switch board, and the switch board is unnecessarily changed. That is, in the conventional method as described above, even if the switch board is not hermetic and only the network traffic increases and the waiting time elapses, the switch board is in the hermetic or fault state from the standpoint of the service board. In this case, changing the switch board does not solve the fundamental problem on the network. Therefore, unnecessary switch board change is made.

An object of the present invention for solving the above problems is, when the chassis management unit receives the failure signal of the switch board, and transmits the failure signal to the service board, the service board changes the switch board to perform communication One way to do this is to provide a switchboard change method.

In order to achieve the above object, the present invention provides a switch board change method applied to an ATCA platform device including a chassis management unit, a plurality of service boards, an activated first switch board, and an inactive second switch board. When each of the service boards receives a message transmitted from the chassis manager, determining whether the message is a message related to the first switch board; If the message is not related to the first switch board, as a result of determining whether the message is related, continuing each communication board via the first switch board; And if the message is a message related to the first switch board, determining whether or not the message is related to a failure of the first switch board. Determining whether to continue using communication via the board or change to communication via the second switch board.

According to the present invention, the service boards receive the error occurrence information of the first switch board directly from the chassis management unit and then directly change the communication port to the second switch board. There is an excellent effect that can be made without interruption.

In addition, the present invention has an excellent effect that each service board receives the failure information of the first switch board directly from the chassis management unit to perform the port change accordingly, thereby preventing unnecessary port changes.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

3 is a block diagram illustrating a structure of an ATCA platform apparatus to which the present invention is applied. In addition, Figure 4 is an exemplary view showing a state of use of the ATCA platform device to which the switch board change method according to the present invention is applied, it is an exemplary view showing a state that each board is mounted in the chassis in a slot manner.

ATCA platform device to which the present invention is applied, as shown in Figure 3, the chassis management unit 110 for performing the overall control function of the routing information, various state information, ATCA platform device for packet delivery in the ATCA platform device, A switch board 120 performing a high-speed packet exchange between a plurality of service boards, determines a forwarding path for IP packets received from a network, changes packet headers, and forwards them to the switch board, and vice versa from the switch board. A plurality of service boards 130-1 to 130-m and 140-1 to 140-n for reassembling the received packet and performing control on packet traffic for transmission of the packet to the network; It includes a detection unit 150 for detecting whether the switch board is mounted (Eject) / mounting (Insert) and transmitted to the chassis management unit. Here, the chassis management unit 110 and the switch board 120 may be formed in a dual structure. That is, in a dual-chassis chassis management unit or switch board, if one chassis management unit or switch board is active and being used, the other chassis management unit or switch board may not operate normally. In the case of being replaced, it is connected to the slot 161 of the chassis 160 in a standby state. In addition, the service boards may be divided into a first service board unit 130 and a second service board unit 140 with the switch board as a boundary, and each service board unit is provided with a plurality of service boards.

The switch board 120 performs a high speed packet switching function among a plurality of service boards and is configured as an active / stand-by redundancy structure. Since the switch board is generally difficult to upload a program, when a program upload is required while in use while being mounted in the slot 161 of the chassis 160 as shown in FIG. 4, the switch board is ejected from the chassis. Replaced with a new switch board. Meanwhile, the switch board communicates with the chassis manager through an intelligent platform management bus (IPMB) to transmit failure information of the switch board to the chassis manager. Hereinafter, the first switch board 120-1 is activated and the second switch board 120-2 is deactivated.

The service boards 130-1 to 130-m and 140-1 to 140-n determine the forwarding path for IP packets received from the outside, perform packet header change functions, and transfer them to the switch board, and receive them from the switch board. It has a packet reassembly function and a control function for packet traffic. On the other hand, the service board communicates with the chassis management unit through an Intelligent Platform Management Interface (IPMI), thereby providing failure occurrence information such as switch board shutdown, OS shutdown, and load generation. In addition, since the service board is capable of uploading a program while being mounted in the slot 161 of the chassis 160 as shown in FIG. 4, once mounted in the slot 161, except for a special case, from the chassis It does not eject.

The chassis manager 110 performs routing information, various state information, and overall control functions inside the device, and is configured as an active / stand-by duplex structure. In addition, the chassis manager may check whether the service board or the switch boards are mounted and the operation state, and control the power on / off state. That is, the chassis management unit manages the overall functions of the device, and is a management system that is monitored by a user although not shown in the drawing, and may perform a function of providing overall operation state information of the device. Meanwhile, as described above, the chassis manager may communicate with the switch board and the service board using IPMB and IPMI. That is, the IPMB interface is connected in series to all slots in the chassis 160 to manage status information, sensor values, and power on / off of each board. Hereinafter, the first chassis manager 110-1 is activated and the second chassis manager 110-2 is deactivated (standby state).

The sensing unit 150 detects whether the switch board is mounted or inserted into the chassis management unit, and may be configured by using a latch of a slot as illustrated in FIG. 4. have. That is, when the user releases the latch 150 of the slot for exchanging the first switch board, the release information of the latch is transmitted to the chassis management unit through the IPMB, and the chassis management unit transmits the release information to the service board through the IPMI. The service board receiving the release information changes the communication port from the first switch board to the second switch board. That is, the release of the latch 150 means that the first switch board can be mounted, so that the service board changes the switch board for smooth communication. On the other hand, the sensing unit may be configured with a sensor that can sense the movement of the switch board in the slot, as well as the latch as described above. That is, the sensing unit may be configured in various forms that can detect hernia / mounting of the switch board.

On the other hand, Figure 4 is an exemplary view showing a state of use of the ATCA platform device to which the switch board change method according to the present invention is applied, each component of the board form constituting the ATCA platform device slot 160 in the form of a chassis 160 This is an exemplary view showing a state mounted on the board.

The chassis 160 includes a switch slot positioned at the center and node slots formed at both sides of the switch slot.

Here, a redundant switch board 120 is mounted in the switch slot, and a redundant chassis management unit 110 and a plurality of switch boards 130 and 140 are mounted in the node slot.

Meanwhile, in FIG. 4, the chassis manager is inserted into the slot in the form of a switch board, but the chassis manager may be placed in the horizontal direction at the top of the slot.

In addition, in the case of a larger and more complex ATCA platform device, the chassis as shown in FIG. 4 may be overlapped and used. In this case, the upper chassis is self 1 (Shelf1) and the lower chassis is self2 (Shelf2). It can be said. At this time, the Chassis Management Module (CMM) may be disposed on the shelf 1, and may be referred to as a self manager because it manages two shelves.

5 is a flowchart illustrating one embodiment of a method for changing a switch board according to the present invention.

As described above, the ATCA platform apparatus to which the present invention is applied may be composed of a plurality of boards, which are mounted in the slot 161 of the chassis 160 as shown in FIG. 4, and the chassis management unit. Controlled by

On the other hand, fault information such as HW, OS, etc. occurring in all boards and detach / mounting information are notified to the first chassis management unit (hereinafter, simply referred to as 'chassis management unit') through IPMB, and the chassis management unit of the collected chassis is In addition to having management information, it also shares information to boards that require management information. The logical communication channel (interface) used is IPMI. In other words, the chassis manager can transmit fault information of each board collected in real time to all boards through the IPMI interface.

A process of implementing the switchboard changing method according to the present invention in an ATCA platform device including the configuration and function as described above is as follows.

When the ATCA platform device is generally driven without any special obstacles, as shown in FIG. 3, the m-th service board 130-m of the first service board unit 130 is connected to the second service board unit 140. When communicating with the n-th service board 140-n, IP communication is performed through the first switch board (502). That is, each service board is exchanging packets through the first switch board in an active state, and the second switch board is waiting in an inactive state to assist the first switch board.

In a state where communication is performed between each service board via the first switch board, when each service board receives an IPMI event message (or an SNMP trap message) from the chassis management unit (504), the message is a message related to the first switch board. It is determined whether or not (506). That is, as described above, the chassis manager performs a function of collecting the management information of the entire chassis 160 through the IPMB. The dismounting or moving information of the first switch board is also transmitted from the first switch board through the IPMB. After processing the hernia or movement information of the transmitted first switch board and transmitting it to all the service boards through IPMI, each service board first determines whether the message transmitted through the IPMI is related to the first switch board. do.

If it is determined that the relevance decision result 506 is not a message related to the first switch board, each service board continuously performs communication via the first switch board (502).

If it is determined that the relevance determination result 506 is not a message related to the first switch board, each service board determines whether the message is a message related to an ejection of the first switch board (508). That is, as described above, a latch or a sensor is provided in the slot of the chassis in which the first switch board is mounted so that the latch is released or the first switch board is externally demounted for dismounting the first switch board. When the card is moved inside the slot, the hernia or movement information is transmitted to the chassis management unit via IPMB. At this time, the chassis management unit processes the information through the IPMI and transmits the information to each service board. Therefore, each service board determines whether the message received through the IPMI is hernia or movement information of the first switch board.

If the determination result of the hernia or movement information 508, the message is not related to the hernia or movement information of the first switch board, each service board is the shutdown of the operating system (OS) of the first switch board (Shutdown) It is determined whether the message is related to () (510). That is, when the present invention changes the first switch board, which is used for packet exchange between service boards, to the second switch board, not only the case where the first switch board is removed from the chassis as described above, but also the first switch board is used. It also includes a case where the OS of the switch board is shut down so that the first switch board cannot perform normal operation. Accordingly, each service board determines whether the message is a shutdown message indicating that the first switch board does not perform a normal operation unless the message is information related to a hernia of the first switch board. Meanwhile, the ATCA platform apparatus to which the present invention is applied is configured such that when the shutdown information of the first switch board is transmitted to the chassis manager through IPMB, the chassis manager can transmit the shutdown information to each service board.

As a result of determining whether the shutdown information is 510, if the message is not related to the shutdown of the first switch board, each service board determines whether the message is a message related to the network load of the first switch board (512). That is, the present invention allows the first switch board to be changed to the second switch board even when it is determined that the first switch board cannot perform normal operation due to the load. If it is not related to the board's dismounting or shutdown, then it is finally determined whether it is related to network load. Information related to the network load is also transmitted from the chassis manager to each service board through the above process.

As a result of determining whether the load is related (512), if it is not a message related to the load of the first switch board, each service board continuously performs communication via the first switch board (502).

Meanwhile, as a result of determining whether the hernia or movement information is determined as a message related to a hernia or movement information of the first switch (508), or as a result of determining whether the shutdown information is determined as a message related to shutdown of the first switch (510) or determining whether the load is related As a result of the message related to the load of the first switch (512), each service board determines whether the second switch board is in an available state through information transmitted from the chassis manager (514). That is, the chassis manager not only transmits the various information of the first switch board in the activated state to all the service boards, but also collects and transmits the information about the second switch board in the standby mode in the inactive state to all the service boards. do. Since the slot of the chassis in which the second switch board is mounted is provided with the sensing unit configured as the latch or the sensor as described above, the chassis manager can detect whether the second switch is mounted and also detect whether the OS is operating normally. Can be. Meanwhile, in the above, the service boards determine whether the second switch board is in a usable state through information transmitted from the chassis manager, but another method may be applied. That is, each service board may inquire whether the second switch board is available or not by using the IPMI or SNMP Get message to the chassis management unit, and then receive and confirm a response message from the chassis management unit. have.

As a result of determining the use state of the second switch board 514, when the second switch board is in a usable state, each service board changes the communication port from the first switch board to the second switch board, so as to communicate with other switch boards. Communicate. That is, the present invention allows each service board to directly receive the failure information of the first switch board from the chassis management unit, and then directly change the communication port to the second switch board, thereby providing service through the network without interruption. Have

As a result of determining the use state of the second switch board 514, when the second switch board is not in a usable state, each service board transmits failure occurrence information of the second switch to the chassis manager and receives failure occurrence information. The chassis manager may transmit the failure information to the monitoring system so that the administrator can directly check the state of the second switch board. However, the state of the second switch board is frequently monitored by the chassis management unit as described above, and when the monitoring results in a failure (hernia, OS shutdown, power off, etc.) of the second switch board, the chassis management unit itself is monitored. Since fault information can be transmitted to the monitoring system, the above-described processes 514 and 518 can be omitted. That is, in the present invention, when each determination result 508, 510, 512 related to the first switch board, or a message related to a hermetic switch, shutdown, or load of the first switch, each service board is a communication port without a separate determination process. May be changed from the first switch board to the second switch board to communicate with other switch boards.

Meanwhile, although the present invention has been described as sequentially determining whether the first switch board is hermetic, shutdown, or loaded, the present invention is not limited thereto. Accordingly, the present invention may be configured to perform at least one of the three determination processes.

Those skilled in the art will appreciate that various changes and modifications can be made without departing from the technical spirit of the present invention. Therefore, the technical scope of the present invention should not be limited to the contents described in the detailed description of the specification but should be defined by the claims.

1 is a block diagram showing the structure of a general ATCA platform device.

Figure 2 is a flow diagram of one embodiment of a conventional switch board change method.

Figure 3 is an exemplary diagram showing a structure of an ATCA platform device to which the present invention is applied.

4 is an exemplary view showing a state of use of the ATCA platform device to which the switch board change method according to the present invention is applied.

5 is a flowchart illustrating one embodiment of a method for changing a switch board according to the present invention;

<Description of Major Symbols in Drawing>

110: chassis management unit 120: switch board

130: first service board unit 140: second service board unit

130-1 to 130-m, 140-1 to 140-n: service board

150: detection unit 160: chassis

161: slot

Claims (8)

In the switch board change method applied to the ATCA platform device including a chassis management unit, a plurality of service boards, the first switch board in the active state, the second switch board in the inactive state, When each of the service boards receives a message transmitted from the chassis manager, determining whether the message is a message related to the first switch board; If the message is not related to the first switch board, as a result of determining whether the message is related, continuing each communication board via the first switch board; And As a result of the relevance determination, when the message is a message related to the first switch board, each of the service boards determines whether the message is related to a failure of the first switch board, and thus, the first switch board. Determining whether to continue to use communication via or change to communication via a second switch board Switch board change method comprising a. The method of claim 1, The disorder is, At least one of a dismounting of the first switch board from the chassis on which the first switch board is mounted, shutdown of the OS of the first switch board, and load generation of the first switch board. How to change the switchboard. The method of claim 1, The switch board change step, As a result of the determination, if the message is a message related to the first switch board, each of the service boards may include a message indicating a disconnection of the first switch board from a chassis on which the first switch board is mounted. Determining whether the message is related; If the message is not determined that the message is related to a hernia of the first switch, the service boards continuously performing communication via the first switch board; And As a result of the determination whether the hernia is determined, when the message is a message related to the hernia of the first switch, the second communication board connects the communication port where the respective service boards are connected to the first switch board to the second switch board. Steps to perform communication via the switch board Switch board change method comprising a. The method of claim 3, wherein If the message is not a message related to hernia of the first switch, the step of continuously performing communication via the first switch board, As a result of the determination whether the hernia is determined, if the message is not a message related to the hernia of the first switch, each of the service boards determines whether the message is a message related to the shutdown of the OS of the first switch board. Making; If the message determines that the shutdown is not related to the shutdown of the OS of the first switch, each of the service boards continuously performing communication via the first switch board; And As a result of determining whether to shut down, when the message is a message related to the shutdown of the OS of the first switch, the communication ports connecting the respective service boards to the first switch board are connected to the second switch board, Performing communication via the second switch board Switch board change method comprising a. The method of claim 4, wherein If the message is not a message related to the shutdown of the OS of the first switch, continuously performing communication via the first switch board, Determining whether the message is a message related to a load of the first switch board, when the message is not related to the shutdown of the OS of the first switch as a result of the shutdown determination; As a result of the load determination, if the message is not a message related to the load of the first switch, continuously performing communication via the first switch board by the respective service boards; And As a result of the determination whether the load is the message related to the load of the first switch, the second service board connects the communication port where the respective service boards are connected to the first switch board to the second switch board. Steps to perform communication via the switch board Switch board change method comprising a. The method according to any one of claims 1 to 5, The message related to the failure of the first switch board, And the information transmitted from the first switch board to the chassis manager through the intelligent platform management bus (IPMB) is processed by the chassis manager and then transmitted to each service board. The method of claim 6, The message is The switch board changing method, characterized in that the transmission from the chassis management unit to each service board through an intelligent platform management interface (IPMI). The method of claim 6, If the failure of the first switch board is related to hernia of the first switch board from a chassis in which the first switch board is mounted, the message is: And a latch or a sensor provided in the slot in which the first switch board is mounted among the slots of the chassis is transmitted to the chassis manager through the intelligent platform management bus.

Images