KR0153931B1 - Dual system and duplication method of parallel common bus type high speed packet exchanging apparatus - Google Patents

Abstract

The present invention relates to a redundancy system and a redundancy method of a parallel common bus type high speed packet exchange device, comprising a high speed packet exchange device based on a parallel common bus and providing one-to-one communication and one-to-many communication between nodes connected through the device. Communication between the two devices is supported, and the I / O subscriber unit, arbitration exchange unit, and parallel common bus unit are provided so that packets from each node can be delivered quickly and accurately through the device. The invention is directed to a redundant configuration and manner.

Description

Redundancy System and Redundancy Method of Parallel Common Bus High-speed Packet Switching System

1 is a redundancy overall configuration diagram

2 is a configuration diagram of a link state monitoring unit.

3 is the overall structure of the arbitration exchange and the DCC-bus structure.

4 is a state signaling monitoring procedure in the arbitration exchange and control information exchange procedure through the DCC-bus.

The present invention relates to a duplication system and a duplication method of a parallel common bus type high speed packet switch.

In the prior art, a packet switching device has a single structure, and thus is weak in reliability due to a blockage in which the operation of the entire device is interrupted or a communication is prevented even by a failure of some of the device components.

The present invention devised in order to solve the above problems in the related art has a single structure having two single structures and configured in redundancy so that a service can be continued through the other even when one side fails. The purpose is to provide a redundancy method.

In order to achieve the above object, the present invention provides a plurality of subscriber input and output means connected to a parallel common bus; A single structure having an arbitration exchange means connected to the plurality of subscriber input / output means and having an arbitration exchange means is doubled by forming two single structures using the same single structure, wherein the arbitration exchange of the redundant switching apparatus It is characterized by including a DCC bus for mutual redundancy control between the means.

In addition, in a method for achieving the object of the present invention, each arbitration exchange means comprises: a first step of initializing all link variable values of a state table managed by itself at the same time as starting; A second step of inspecting the link state of each subscriber input / output unit through the M-bus after performing the first step, and changing and updating the variable value of the link in which the failure occurred in the state table; A third step of determining whether the arbitration switching means or the parallel common bus is normally operated after performing the second step; A fourth step of outputting a system failure signal as normal after performing the third step and outputting a system failure signal as abnormal if abnormal and going to a recovery standby state; A fifth step of checking a system failure signal of the other party after performing the fourth step; After performing the fifth step, if the counterpart is abnormal, the counterpart value of all link variables in the state table is processed abnormally. If the counterpart is normal, all counterpart link variable values are initialized to normal and the self-side identified in the first step. A sixth step of transmitting the number of the failed link among the link variable values of the other through the DCC-bus to update the faulty link numbers notified from the other party to its own state table; After performing the sixth step, it is determined whether there has been a change or update of the state table, and if the state table has not been updated, the second step or less is performed, and if there is an update, to notify all subscriber nodes having the change. And a seventh step of performing broadcasting through a parallel common bus and repeatedly performing the second step or less.

Based on the parallel common bus, the switching unit is constructed based on a single structure in which n I / O units are connected and one arbitration exchange unit in charge of the arbitration exchange is connected. Single structures are redundant. For convenience of explanation, one side is called the exchanger A and the other side is called B. FIG. At this time, the DCC-bus for mutual redundancy control exists between the arbitration exchange in exchange A and the arbitration exchange in exchange B.

Each subscriber is connected to the switching device in a serial communication manner and a subscriber I / O matching unit supporting the switching device exists in the subscriber node configuration. In order to receive reliable service through the redundant switching device, each subscriber node also configures the subscriber I / O matching part with redundancy, one subscriber I / O matching part is connected to the subscriber I / O part in the switching device A and the other is connected to the subscriber I / O part in the switching device B. Characterized in that the structure is connected to the subscriber input and output unit.

In addition, the subscriber I / O unit periodically transmits an alive packet to determine whether the link connected to the subscriber side is normal and whether the subscriber I / O matching unit is normally operated, and periodically receives the alive packet returned through the loopback within a predetermined time period. It is provided with a link state monitoring unit for monitoring whether or not it is received, and has a function of transmitting an error signal to the arbitration exchange unit when an abnormality is detected.

Process of initial state and normal operation

In principle, the two exchangers A and B operate in a redundant sharing mode of load sharing. That is, in the initial state, both A and B are capable of normal operation. Accordingly, the subscriber node can transmit and receive packets to either A or B.

As a method for designation of this initial state and recognition at the subscriber node side, Arbitration Switch A and Arbitration Switch B initialize the state table and broadcast the state table to all the I / O matching units. Initialize the state table.

The subscriber node must first decide whether to use A or B in order to output a packet to the subscriber I / O matching unit and make a transmission request. Referring to the state table loaded from the arbitration exchange, either A or B The normal link should be used. If both are normal, load balancing should be used alternately.

о link #i (A) or subscriber I / O #i (A) or subscriber I / O #i (A)

Failures and failures of any link #i, I / O #i or I / O #i can be viewed as the same fault group. That is, if only one of the three failures occurs, follow the same troubleshooting procedure. This failure detection method is understood later in the description of Figure 2, and when such a failure is detected, the arbitration exchanges A and B update the state table to broadcast the state table. In order to maintain the same state table, the A and B arbitration exchanges send and receive mutual state information in the form of messages through the DCC-bus to ensure consistency.

Then, if you want to send a packet destined for a certain node #i, refer to the status table at each node. If A side is broken and B side is normal, the packet is sent through the B side path.

оTroubleshooting in the event of an intermediary exchange (A) or parallel common bus (A) failure

In case of failure of Arbitration Exchange or Parallel Common Bus, all subscriber nodes and I / O units in the state are inoperable, so all nodes in the status table are processed and blocked. This state table is broadcasted through the steady-state switching device side. The detection method of the failure of the arbitration exchange or the parallel common bus is described in FIG.

Then, from this point on, all the nodes refer to the state table for packet communication. The A side is all blocked and the B side is normal, so the B side exchange device communicates.

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

1 is an overall configuration diagram of the present invention. In the drawings, 1 and 2 represent an arbitration exchange unit, and if there is a packet to be transmitted from each subscriber I / O unit connected to a parallel common bus, mediation of transmission request signals enables packet transmission one by one. It plays a role. 3 and 4 represent parallel common buses that consist of a group of data transmission and control signals. 5 and 6 represent M-buses for reading the status signals of each subscriber station. 7,8 represents the subscriber I / O unit, and receives the serial transmission packet from the subscriber side, waits until the parallel common bus license is obtained from the arbitration exchange unit, and transmits it to the desired destination subscriber I / O unit and receives it through the parallel common bus. If the packet is destined for itself, it receives the packet and delivers it to the subscriber. A plurality of subscriber input / output units having the same function are present in the switching device (7-1 to 7-n, 8-1 to 8-n). 9-1, 2 and 10-1, 2 denote subscriber I / O matching units which are implemented when the subscriber nodes 9 and 10 are configured to match subscriber I / O units of the switching device. 3) Receives the packet generated from the system node (9-4, 10-4) in the subscriber node and transmits it to the switching device, and sends the packet received from the switching device to the subscriber main control through the system bus. Do it. 11 shows a serial communication link between the subscriber I / O unit of the switching device and the subscriber I / O matching unit of the subscriber node. 12 is a DCC bus as a communication link for transmitting and receiving redundancy control information between each arbitration switching part of two switching devices.

2 is a block diagram of the subscriber I / O matching unit and the link state monitoring unit existing in the subscriber I / O unit according to the present invention, wherein the alive packets are generated at regular intervals from the periodic alive packet generator of the subscriber I / O unit to the serial transmitter means. The serial transmitter means serializes it and transmits it to the serial receiver means of the subscriber I / O matching section. The serial receiver of the subscriber input / output matching unit packetizes serially received data and outputs the packetized data to a sorter. The classifier classifies whether the packet is an alive packet or a pure data packet and outputs the pure data packet to the host and outputs the alive packet to the loopback circuit. The loopback circuit periodically outputs a cycle time interval to the monitoring timer unit to check whether an alive packet is received, and transmits it through a serial transmitter to form a closed circuit. If the signal is not received periodically, the monitoring timer of the subscriber input / output matching unit outputs an abnormal state signal to the host by determining that the link is disconnected or a failure of the other subscriber input / output unit. The alive packet transmitted serially through the serial transmitter is received through the serial receiver of the subscriber I / O unit. The received packet is passed through a sorter and the pure data is output to the transmission waiting queue to be transmitted through the parallel common bus. The alive packet is output to a watchdog timer for monitoring whether the packet comes back periodically. If the live packet does not return within the specified cycle time, the watchdog timer outputs an abnormal status signal as a result of link disconnection or failure of the partner subscriber I / O matching unit.

The arbitration exchange polls the subscriber I / O in turn via the M-bus and reads the status signals of each subscriber I / O to determine the presence of abnormal status. The M-bus consists of a data bus, an address bus, and a read signal, and each subscriber's input / output unit decodes the address bus when the read signal is activated to determine whether it is selected, and outputs the status signal value to the data bus in the read operation. do.

3 is a block diagram of the arbitration exchange unit, which is composed of a parallel common bus matching unit, an M-bus master unit, a control unit, and a state table unit, and is composed of a DCC-bus unit to exchange control messages between the redundant arbitration exchange units.

The controller is responsible for the software that monitors the status signal of each subscriber's input and output through the M-bus master, and sends and receives the redundancy control message through the DCC-bus to determine the redundancy status and update the status table, and then through the parallel common bus matching unit. Output loads the table to all subscriber nodes. DCC bus consists of serial communication link, system fault A signal and B signal.

The following table 1 shows an example of the state table, which is composed of the link number and failure state variable connected to each subscriber node, and the state table of the counterpart switching device is also configured. Although the state variable value of each link in the state table is normally normal at first, the process of checking the status signal of each subscriber I / O through the M-bus of the arbitration switch enables the failed link to be changed and updated abnormally.

4 is a flow diagram of a state signaling monitoring procedure in the arbitration exchange and a control information exchange procedure through the DCC bus, and shows a flowchart of control information for informing each other of redundant control information and contents of its link state table.

Each arbitration exchange A and B transfers the mutual control information through the DCC bus while going through the same process of their own. At the beginning, each link variable of the state table managed by itself is managed. Initializing all the values to normal (20 and 40), and checking the link status of each subscriber's input / output unit through the M-bus to change and update the variable values of the failed link in the status table (21 and 41). In the first step consisting of, the process of judging the normal operation of the arbitration switching unit or the parallel common bus unit (22 and 42), if it is normal, the process of outputting the system fault A signal to normal 23 and the process of outputting the system fault B signal to normal The second stage of the process of process 43 to 25 and 44 to 45 to process the process of 43 and, if abnormal, outputs system faults A and B signals abnormally and goes to the recovery standby state. If the counterpart is abnormal through the process of checking fault signals A and B (26 and 46), the counterpart process of all the link variables in the state table is abnormally processed (30 and 50). Is initialized to normal and goes through the processes (27 and 47), and the DCC-bus is sent to the other party's arbitration exchange with the number of the failed link among the link variable values of its own side identified in the first step (20 and 40). The process (28 and 48) is carried out and the fault link numbers notified from the state are updated into their own state tables (29 and 49). It is determined whether there has been an update, and if there is no update of the state table (31 and 51), the process returns to steps (21 and 41) and repeats the previous step if there is an update. Redundancy status monitoring and control processed in the fourth step of returning to the process (21 and 41) through the process of broadcasting through the parallel common bus (32 and 52) to notify all the slow subscriber nodes. Procedure.

According to the present invention configured and operated as described above, two packet switching devices having a single structure are installed, a DCC bus is provided between them, and duplication control information is exchanged between the two. It is possible to detect and process status, and at the subscriber node side, the subscriber I / O matching unit is duplexed so that one link is connected to one switching device and the other is connected to the other switching device. It can be used as a packet switching device in a communication system that requires a reliability to receive a communication service.

Claims (4)

Subscriber I / O means connected to a plurality of parallel common buses; A single structure having an arbitration exchange means connected to the plurality of subscriber input / output means and having an arbitration exchange means is doubled by forming two single structures using the same single structure, wherein the arbitration exchange of the redundant switching apparatus A redundancy system of a parallel common bus type high speed packet switching device comprising a DCC bus for mutual redundancy control between means. 2. The subscriber input / output matching unit according to claim 1, wherein each subscriber is connected to the switching device in a serial communication manner and a subscriber input / output matching unit supporting the same exists in the subscriber node configuration. 1. A redundancy system of a parallel common bus type high speed packet switching device, characterized in that one is connected to a subscriber input / output unit in one switching device and the other is connected to a subscriber input / output unit in the other switching device. The receiver of claim 2, wherein the subscriber I / O unit transmits a cyclic alive packet and receives an alive packet returned through loopback to determine whether a link connected to the subscriber side is normal and whether the subscriber I / O matching unit is normally operated. And a link state monitoring unit for monitoring whether or not it is periodically received within a predetermined time period, and transmitting an error signal to the arbitration switching unit when an abnormality is detected. Each arbitration exchange means includes: a first step (20, 40) of initializing all link variable values of a state table managed by itself at the same time as starting; A second step (21, 41) of checking and updating the link state of each subscriber input / output unit through the M-bus after performing the first step (20, 40) and changing the variable value in which the failure occurred in the state table; A third step (22,42) of determining whether the arbitration exchange means or the parallel common bus are normally operated after performing the second step (21, 41); After performing the third step (22, 42), if the normal output the system failure signal, if abnormal, output the system failure signal abnormally and go to the recovery standby state (4, 24, 25, 44, 45) ; A fifth step (26, 46) of checking a system failure signal of the other party after performing the fourth step (24, 25, 44, 45); After performing the fifth steps 26 and 46, if the counterpart is abnormal, the counterpart value of all the link variables in the state table is abnormal.If the counterpart is normal, all counterpart link variable values are initialized to normal and the first step ( 20, 40) transmits the number of the faulty link among the link variable values of one's own side through the DCC bus to the other party's arbitration exchange means to update the faulty link numbers notified from the other side to its own state table. Sixth step (27-30,47-50); After performing the sixth steps 27 to 30 and 47 to 50, it is determined whether there is a change or update of the state table, and if there is no update of the state table, the second steps 21 and 41 or less are performed, and the update is performed. And the seventh step (32, 52), which broadcasts through the parallel common bus and repeats the second step (21, 41) or less in order to notify all the subscriber nodes with the change, if any. A redundancy method of a parallel common bus type high speed packet exchange apparatus characterized by the above-mentioned.