CN1889464A - Method and apparatus for realizing communication take-over and control - Google Patents

Abstract

A communication receiving device is featured as receiving route protocol message from neighbor route simultaneously by master processing unit and back-up processing unit i.e. carrying out operation in parallel way by master processing unit and back-up processing unit for raising operation efficient of said device, maintaining local data by both master processing unit and back-up processing unit according to protocol message for ensuring synchronous update of local data on both master processing unit and back-up processing unit.

Description

A method and device for realizing communication takeover

technical field

The present invention relates to the field of communication technology, in particular to a method and device for realizing communication takeover.

Background technique

High Availability (HA: High Availability) refers to the ability of a product or system to continuously serve customers. It is usually carried out through the two indicators of mean time to repair MTTR (Mean Time to Repair) and mean time between failures MTBF (Mean Time Between Failures) logo. MTTR refers to the average time required for a component or equipment to return to normal from failure. In a broad sense, MTTR involves spare parts management and customer service, and is an important indicator in equipment maintenance contracts; MTBF refers to the failure-free performance of a component or equipment. The average running time, which is the time between failures that the user can tolerate, is usually measured in hours.

Existing routers generally adopt a communication takeover solution to achieve high availability of the system. The system generally adopts a dual mainboard design, one of which is the active mainboard (AMB) in working state, and the other is the standby mainboard (SMB) in backup state. The main motherboard runs system software, including routing protocols and routing table maintenance. All routing protocols send and receive packets to and from neighboring routers and process them. The routing protocols of the standby mainboard are also started to receive and process messages. The active main board backs up the data of various routing protocols such as Open Shortest Path First (OSPF: Open Shortest Path First) routing protocol to the standby main board. When the active main board fails, the standby main board immediately takes over the failed main board and becomes the active main board. This process is called "switching", and this switching process can also be driven by the operator through commands.

How to ensure that the switched routers have the same link state database has become a worldwide problem. In order to solve this problem, two solutions in the prior art are introduced below:

The method for realizing the communication takeover of the first scheme includes: providing a router with a main processing unit; connecting a backup processing unit to the router; after connecting the main processing unit to the backup processing unit, transferring the network protocol information from the The active processing unit sends to the standby processing unit to synchronize the link configuration and link protocol status with the active processing unit in the standby processing unit; When there is a fault in the router, switch the router to the standby processing unit.

The main idea of this solution is that the protocol message sent to the router is first sent to the main processing unit, the main processing unit updates the local data according to the message and performs corresponding processing, and then the main processing unit sends the protocol message The message is sent to the standby processing unit, and the standby processing unit updates the local data according to the message. This solution needs to create two hidden interfaces for each area of the OSPF router: one in the active processing unit and one in the standby processing unit. The hidden interfaces of the active processing unit and the standby processing unit are connected through a hidden channel, through which link state data is backed up from the active processing unit to the standby processing unit.

There are certain defects in the above method of realizing communication takeover:

1) When the active processing unit receives a Link State Advertisement (LSA: Link State Advertisement) message from a neighboring router, and responds to the neighboring router and cannot flood the LSA message to the standby processing unit in the future, if the active processing unit at this time If the processing unit and the standby processing unit are switched, the router will lose the LSA message. If a neighbor router can connect to other parts of the network through the router being switched, the neighbor router will also lose the LSA message. Due to the above reasons, the reliability of the above-mentioned existing method is low, and the link state databases of the active processing unit and the standby processing unit may not be kept in sync easily, resulting in data loss and system performance degradation.

2) All received link status notification messages must be connected through the inter-process communication (IPC: Inter-Process Communication) connection between the main processing unit and the standby processing unit, causing a large amount of IPC bandwidth resources to be occupied, and system performance decline.

The method of the second prior art solution is provided below. In this method, the standby processing unit first receives the message and maintains the local data according to the received message, which can solve the problem that the message may be lost during the switching process of the above solution 1 method.

The second scheme is a communication takeover method based on the message mirroring mechanism, which relies on the message mirroring mechanism provided by the Internet Protocol (IP: Internet Protocol) stack. The message received under the message mirroring mechanism first arrives at the standby processing unit, and then the standby processing unit copies the message to the main processing unit; the message sent under the message mirroring mechanism is that the main The active processing unit first sends the message to the standby processing unit, and the standby processing unit actually sends the message; when a failure is detected in the active processing unit, the router is switched to the standby processing unit.

The method for realizing the communication takeover of the above-mentioned existing scheme 2 is described in detail below:

First, when the standby processing unit is started, the standby processing unit backs up link states in batches from the active processing unit. Then, the standby processing unit maintains and updates the link state by combining the explicit information of the active processing unit, the sent protocol message and the received protocol message.

All received OSPF protocol packets first arrive at the standby processing unit, and then are sent by the standby processing unit to the active processing unit; all outgoing OSPF protocol packets are first sent by the active processing unit to the standby processing unit, and then sent by the standby processing unit to an adjacent OSPF router.

During the receiving and sending process, the standby processing unit is a passive information consumer. This information comes from the master processing unit and other routers in the network. The standby processing unit neither generates any protocol packets nor executes the shortest path first algorithm (SPF: Shortest Path First) to calculate routes.

Solution 2 also has certain defects in the method of realizing communication takeover:

1) All protocol messages must be transmitted through the IPC connection between the main processing unit and the backup processing unit, which occupies a large amount of IPC bandwidth resources and degrades system performance.

2) Some additional steps are required to process each OSPF protocol message. Each received OSPF protocol message has to go through an additional internal step from the standby processing unit to the active processing unit. Every sent OSPF protocol message has to go through an additional internal step from the active processing unit to the standby processing unit. This reduces the efficiency of transmission and reception.

3) Depending on the IP stack or the operating system to provide a packet mirroring mechanism. However, the general IP stack does not provide a packet mirroring mechanism. This makes it more difficult to realize, has high requirements on the system, and has poor versatility.

Contents of the invention

The purpose of the present invention is to provide a method and device for realizing communication takeover, which are applied in the field of communication technology. The method and device can save a large amount of IPC bandwidth resources, and make the system run more efficiently, more stably and reliably.

In order to achieve the purpose of the present invention, the technical solution adopted is: a method for realizing communication takeover, comprising: the main processing unit and the standby processing unit synchronously receive routing protocol messages from neighbor routers; the main processing unit and the standby processing The units maintain their respective local data according to the message; when the main processing unit fails, the backup processing unit takes over the communication of the main processing unit.

Preferably, the process of maintaining local data of the main processing unit and the backup processing unit according to the message includes: the backup processing unit maintains the local data of the backup processing unit according to the message; the backup processing unit sends and receives the The response message of the message is sent to the main processing unit; the main processing unit receives the response message, and maintains the local data of the main processing unit according to the message.

Preferably, if the main processing unit does not receive a response message from the standby processing unit for receiving the message within a preset time after receiving the message, then send the message to the standby processing unit and A warning message is generated.

Preferably, if the main processing unit does not receive the message within a preset time after receiving the response message from the backup processing unit, the method further includes: the main processing unit Request the message from the standby processing unit; the standby processing unit sends the message to the main processing unit, and generates an alarm message.

Preferably, the routing protocol is an open shortest path first routing protocol; the message is a link state notification message; the main processing unit and the backup processing unit maintain their respective local data according to the link state notification message The process includes: the standby processing unit updates the local link state database of the standby processing unit according to the link state notification message; the standby processing unit sends a response message to the main processing unit after receiving the link state notification message; the main processing unit The unit receives the response message, updates the local link state database of the main processing unit according to the link state notification message; runs the shortest path first algorithm to calculate the best path to the destination and updates the routing table and forwarding table according to these paths .

Preferably, the active processing unit sends a response message to the neighbor router after receiving the response message of the link state notification message received by the standby processing unit.

Preferably, the process of maintaining local data by the standby processing unit according to the link state notification message further includes: for each received link state notification message, if the standby processing unit determines that the local router sends the message to a neighbor When the router floods, the backup processing unit puts the link state notification message into the retransmission table of the neighbor router locally recorded by the backup processing unit; When the reply message of the notification message is sent, the link state notification message is deleted from the retransmission table of the neighbor router recorded locally.

Preferably, the routing protocol is an open shortest path first routing protocol; the message is a connection message; the process of maintaining local data by the backup processing unit according to the connection message includes: the backup processing unit maintains the local data according to the connection message The state information of the neighboring router locally recorded by the standby processing unit.

Preferably, the routing protocol is an open shortest path first routing protocol; the message is a database description message; the main processing unit and the backup processing unit maintain their own local data according to the database description message. The process includes: standby The processing unit maintains the request table of the neighbor router locally recorded by the backup processing unit according to the database description message; the backup processing unit sends a response message receiving the database description message to the main processing unit; the main processing unit receives The response message, according to the database description message, maintains the locally recorded request table of the neighbor router and the status information of the neighbor router, and sends the database description message to the neighbor router after receiving the database description message. response message to the text.

Preferably, the process of maintaining the request table of the neighbor router locally recorded by the standby processing unit includes: checking whether each link state advertisement description information in the database description message is higher than that in the link state database of the standby processing unit The corresponding link state notification is new, if the standby processing unit puts the description information of the link state notification into the request table of the neighbor router; when the standby processing unit receives the link state notification from the neighbor router When describing the link state advertisement message corresponding to the information, the link state advertisement description information is deleted from the request table of the neighbor router.

Preferably, when the standby processing unit is started, the data of the active processing unit is backed up to the standby processing unit to complete the initial data synchronization.

Preferably, after the initial data synchronization is completed, if a local state change occurs in the active processing unit, the local state change information is transmitted to the standby processing unit; the local data of the standby processing unit is also maintained according to the local change information.

Preferably, the local state change information includes: configuration change information, interface state change information, and neighbor router state change information.

Preferably, when a hardware or software failure triggers the failure of the main processing unit, the process for the standby processing unit to take over the communication of the main processing unit includes the following steps: the standby processing unit reads in all network interface information from the interface layer; performs active-standby switching, and the standby processing unit Replace the main processing unit; for any neighbor router whose state is not full, the standby processing unit re-establishes the adjacent relationship with it; starts the relevant timer; introduces the path of other protocols according to the configuration and regenerates the link state from the local router Notify the message; the standby processing unit runs the shortest path first algorithm to calculate the best path to the destination and updates the routing table and forwarding table according to these paths.

Preferably, when the manual switching triggers the failure of the main processing unit, the process of the backup processing unit taking over the communication of the main processing unit includes: the backup processing unit reads in all network interface information from the interface layer; waits for the status of all neighbor routers to become full; Active-standby switching, the standby processing unit replaces the main processing unit; the standby processing unit uses the remaining time of the timer obtained from the main processing unit to set and start the relevant timer; according to the configuration, the path of other protocols is introduced and regenerated from the local Router's link state notification message; the standby processing unit runs the shortest path first algorithm to calculate the best path to the destination and updates the routing table and forwarding table according to these paths.

Preferably, after waiting for the preset time, if there is still a neighbor router state that cannot become full; then carry out the steps: carry out active-standby switching, and the standby processing unit replaces the main processing unit; for any neighbor router that is not full in state, the standby The processing unit re-establishes the adjacent relationship with it; starts the relevant timer; introduces the path of other protocols according to the configuration and regenerates the link state notification message from the local router; the standby processing unit runs the shortest path first algorithm to calculate the route to the destination. optimal paths and update routing and forwarding tables based on these paths.

A device for realizing communication takeover provided by the present invention includes a message receiving unit, a main processing unit, a standby processing unit and a main/standby control unit; the message receiving unit is used to receive routing protocol messages from neighboring routers, and send The main processing unit and the standby processing unit send the message at the same time; the main processing unit and the standby processing unit maintain their respective local data according to the message; the main and standby control unit is used to control the standby processing unit to take over the main processing unit communication.

Preferably, the active/standby control unit includes: a takeover control unit, configured to control the standby processing unit to take over the communication of the active processing unit when the active processing unit fails; the active/standby control unit also includes a data backup control unit, configured to Controlling the backup of the data of the active processing unit to the standby processing unit.

Preferably, the backup processing unit includes: a backup receiving unit, configured to receive the protocol message sent by the message receiving unit; a backup data maintenance unit, used to maintain local data according to the message received by the backup receiving unit; a response message sending unit After the backup data maintenance unit maintains the local data according to the received message, it sends a response message of receiving the message to the response message receiving unit; the main processing unit includes: a main receiving unit for receiving The protocol message sent by the message receiving unit; the response message receiving unit, receiving the response message from the received message sent by the response message sending unit; the main data maintenance unit, used to receive the response message at the response message receiving unit Afterwards, the local data is maintained according to the message received by the main receiving unit.

Preferably, the main processing unit further includes: a main message forwarding unit, configured to send the message to the sending unit when the response message receiving unit does not receive the response message within a preset time after the main receiving unit receives the message When receiving the response message of the message, send the message to the standby receiving unit, and generate an alarm message to the user; the standby receiving unit is also used to receive the message forwarded by the main message forwarding unit.

Preferably, the main processing unit further includes: a request unit, configured to, after the response message receiving unit receives a response message from the received message sent by the response message sending unit, when the main processing unit receives within a preset time When the unit does not receive the message sent by the message receiving unit, it requests the message from the standby message forwarding unit; the standby processing unit also includes: a standby message forwarding unit, which is used to After processing the request of the unit, send the message requested by the requesting unit to the main receiving unit, and generate an alarm message; the main receiving unit is also used to receive the message forwarded by the standby message forwarding unit.

Adopt above-mentioned technical scheme, the beneficial technical effect of the present invention is:

1) The present invention synchronously receives the routing protocol message from the neighbor router through the main processing unit and the standby processing unit, and the parallel working mode of the main processing unit and the standby processing unit avoids all protocol messages in the prior art Through the IPC connection between the main processing unit and the standby processing unit, the problems of a large amount of IPC bandwidth being occupied and resource waste in the prior art are effectively solved, and the efficiency of system operation is improved.

2) The present invention synchronously receives the routing protocol message from the neighbor router through the main processing unit and the standby processing unit, on the basis that the main processing unit and the standby processing unit work in parallel, after the standby processing unit maintains the local data according to the protocol message , by sending a short response message to the active processing unit, and the active processing unit maintains the local data according to the protocol message, it can ensure that the local data of the active processing unit and the standby processing unit are updated synchronously and consistent, so that the switching process Higher reliability.

3) The present invention does not depend on any special improvement of IP stack or operating system, and has universal applicability.

Description of drawings

The present invention will be further described in detail below through specific embodiments and in conjunction with the accompanying drawings.

FIG. 1 is a flowchart of a method for realizing communication takeover according to Embodiment 1 of the present invention;

FIG. 2 is a flowchart of a method for realizing communication takeover according to Embodiment 2 of the present invention;

FIG. 3 is a flowchart of a method for implementing communication takeover according to Embodiment 3 of the present invention;

FIG. 4 is a flowchart of a method for realizing communication takeover according to Embodiment 4 of the present invention;

Fig. 5 is the flow chart of unplanned switching when the inventive method applies OSPF agreement;

Fig. 6 is a flow chart of planned switching when the method of the present invention is applied to the OSPF protocol;

7 is a schematic structural diagram of a device for implementing communication takeover according to Embodiment 5 of the present invention;

FIG. 8 is a schematic structural diagram of an apparatus for realizing communication takeover according to Embodiment 6 of the present invention.

Detailed ways

The purpose of the present invention is to provide a method and device for realizing communication takeover, the core idea of which is that the present invention synchronously receives routing protocol messages from neighboring routers through the main processing unit and the standby processing unit, and the main processing unit and the standby processing unit The way of working in parallel avoids that in the prior art, all protocol messages have to go through the IPC connection between the main processing unit and the backup processing unit, effectively solving the problem that a large amount of IPC bandwidth is occupied and resources are wasted in the prior art problem, improving the efficiency of system operation. Simultaneously, the present invention synchronously receives the routing protocol message from the neighbor router through the main processing unit and the standby processing unit, on the basis that the main processing unit and the standby processing unit work in parallel, after the standby processing unit maintains the local data according to the protocol message, By sending a short response message to the main processing unit, after receiving the response message, the main processing unit maintains the local data according to the protocol message, which can ensure the synchronous update of the local data of the main processing unit and the standby processing unit And consistent. A method and device for implementing communication takeover provided by the present invention will be described in detail below.

Embodiment 1. A method for realizing communication takeover, the flowchart of which is shown in FIG. 1 , including:

A1, the active processing unit and the standby processing unit receive routing protocol packets from neighboring routers synchronously;

In the present invention, the synchronous reception of the routing protocol message by the main processing unit and the standby processing unit may be that the message receiving unit first receives the routing protocol message from the neighbor router, and sends the message to the main processing unit and the standby processing unit in a multicast manner. The processing units send at the same time, and the active processing unit and the standby processing unit are regarded as members of the multicast group.

In this embodiment, the routing protocol is the OSPF protocol. Similarly, the method of the present invention can also be applied to other routing protocols. What this method emphasizes is to realize the redundant backup of the communication function, and the specific routing protocol does not constitute a limitation to the present invention. limit.

In this embodiment, the message is a Link State Advertisement (LSA: Link State Advertisement) message. It can be understood that, under the OSPF routing protocol, the message can also be a connection (Hello) message or a database Description (DD: Database Description) message or link state request (Link StateRequest) message or link state response (Link State Ack) message.

It can be understood that the present invention needs to synchronize the initial data of the active processing unit and the standby processing unit in advance. It is emphasized here that the active processing unit and the standby processing unit have the same local data before synchronously receiving protocol packets from neighboring routers. In the present invention, the synchronization of the local data of the main processing unit and the backup processing unit can be performed by the operator in advance. It can be understood that the synchronization method can also be that the data of the main processing unit is backed up in batches when the backup processing unit is started. Spare handling unit.

Wherein, the active processing unit and the standby processing unit are software or hardware modules that have the same processing capability for routing protocol packets. The synchronized local data may include data for routing control such as configuration information, interface information, neighbor router state information, and link state information.

A2, the active processing unit and the standby processing unit maintain their respective local data according to the message;

The maintenance of local data by the main processing unit and the standby processing unit is to analyze the information carried by the message, and maintain the local data according to the analysis results, for example: maintain the local link state database, neighbor router status or request table wait.

A3. When the active processing unit fails, the standby processing unit takes over the communication of the active processing unit.

In the present invention, the failure of the primary processing unit is triggered by software and hardware faults, manual switchover, or a preset time interval.

In this embodiment, the switching process in which the standby processing unit takes over the communication of the active processing unit may adopt an existing conventional implementation manner, which will not be repeated here.

Embodiment 1 Through the message receiving unit to send protocol messages to the main processing unit and the standby processing unit at the same time, the main processing unit and the standby processing unit work in parallel, avoiding the need for all protocol messages in the prior art Through the IPC connection between the main processing unit and the standby processing unit, the problems of a large amount of IPC bandwidth occupied and resource waste in the prior art are effectively solved, and the efficiency of system operation is improved.

Embodiment 2. A method for realizing communication takeover, the flowchart of which is shown in FIG. 2 , including:

B1, the active processing unit and the standby processing unit receive LSA messages from neighboring routers synchronously;

In this embodiment, the received message is an LSA message under the OSPF routing protocol.

B2, the standby processing unit updates the local link state database of the standby processing unit according to the LSA message;

B3, the standby processing unit sends a response message of receiving the LSA message to the active processing unit;

B4, the active processing unit receives a response message from the standby processing unit;

B5, the active processing unit updates the local link state database of the active processing unit according to the LSA message;

B6. The main processing unit sends a response message for receiving the LSA message to the neighbor router;

B7, the main processing unit runs the SPF algorithm to calculate the best path to the destination and updates the routing table and forwarding table according to these paths;

B8. When the active processing unit fails, the standby processing unit takes over the communication of the active processing unit.

In this embodiment, step B5 and step B4 and step B7 and step B4 have no absolute time sequence. The specific execution sequence should not be construed as a limitation of the present invention. It is emphasized here that step B6 is after step B4 and step B5, that is, after the active processing unit receives the LSA message and receives the response message of receiving the LSA message sent by the standby processing unit, it sends it to the neighbor router The response message of the LSA message is received. Its purpose is to ensure that both the active processing unit and the standby processing unit send a reply after receiving the message sent by the neighbor router. In a network with a retransmission mechanism, if the neighbor router does not receive the response message, it will send the LSA message again.

The difference between the second embodiment and the first embodiment is that the message receiving unit sends the protocol message to the main processing unit and the backup processing unit at the same time, and on the basis that the main processing unit and the backup processing unit work in parallel, the backup processing unit receives After the protocol message, by sending a short response message to the active processing unit, it can ensure that the local data of the active processing unit and the standby processing unit are updated synchronously, so that the reliability of the switching process is higher.

For each received LSA message, if the backup processing unit determines that the local router floods the message to a certain neighbor router, the backup processing unit puts the LSA message into the local record of the neighbor router. In the retransmission table: when the backup processing unit receives a response message from the neighbor router to the LSA message, delete the LSA message from the retransmission table of the neighbor router recorded locally.

Embodiment 3. A method for realizing communication takeover, the flowchart of which is shown in FIG. 3 , including:

C1, the active processing unit and the standby processing unit receive Hello packets from neighboring routers synchronously;

In this embodiment, the received message is a Hello message under the OSPF routing protocol;

C2, the standby processing unit maintains the state information of the neighbor router locally recorded by the standby processing unit according to the Hello message;

The present invention applies the OSPF protocol, and the states of neighbor routers include:

Down state, Attempt state, Initial state, Two-way state, ExStart state, Exchange state, Loading state, Full state .

C3, the active processing unit maintains the state information of the neighbor router locally recorded by the active processing unit according to the Hello message;

C4. When the main processing unit fails, the standby processing unit takes over the communication of the main processing unit.

In the present invention, the failure of the primary processing unit may be triggered by software and hardware faults, manual switchover, or a preset time interval.

Embodiment three is the processing mode of the present invention to the Hello message, if the standby processing unit will process the Hello message, it needs to run the Hello timer and the Dead timer of the neighbor router, by receiving the Hello message of the neighbor router The contents and time intervals of the interface status changes determine the status of neighboring routers and record them. This embodiment emphasizes the processing capability of the standby processing unit for the Hello message. It can be understood that there is no absolute sequence between step C2 and step C3. It is also possible to carry out step C3 first, and then carry out step C2, and the order of the specific steps does not constitute a limitation to the present invention.

Embodiment 4. A method for realizing communication takeover, the flowchart of which is shown in FIG. 4 , including:

D1, the active processing unit and the standby processing unit receive DD messages from neighboring routers synchronously;

In this embodiment, the received message is a DD message under the OSPF routing protocol.

D2. The backup processing unit maintains the request table of the neighbor router and the state information of the neighbor router locally recorded by the backup processing unit according to the DD message;

The process for the standby processing unit to maintain the request table of the neighbor router is: compare each LSA description information in the DD message with the corresponding LSA in the link state database of the standby processing unit; The corresponding LSA in the state database is new, then the backup processing unit puts the LSA description information into the request table of the neighbor router; when the backup processing unit receives the LSA message corresponding to the LSA description information from the neighbor router , delete the LSA description information from the request table of the neighbor router.

D3, the standby processing unit sends a response message of receiving the DD message to the active processing unit;

D4, the active processing unit receives a response message from the standby processing unit;

D5. The active processing unit maintains the request table of the neighboring router and the state information of the neighboring router locally recorded by the active processing unit according to the DD message;

D6. The main processing unit sends a response message for receiving the database description message to the neighbor router for the database description message;

D7. When the active processing unit fails, the standby processing unit takes over the communication of the active processing unit.

Embodiment 4 embodies the processing method of the present invention for DD messages. The processing of DD messages enables the neighboring routers to retransmit unreceived DD messages after switching is realized, so that the seamless switching process more perfect.

Embodiment 3 and Embodiment 4 maintain the state information of the neighbor router recorded locally according to the Hello message and the DD message. It can be understood that the standby processing unit can also obtain the state information of the neighbor router by using an IPC connection between the main processing unit and the standby processing unit to transfer the state information of the neighbor router when the main processing unit judges that the state of the neighbor router changes. The status change is notified to the standby processing unit. In this way, the standby processing unit may not run the relevant timer, but it will occupy a little more IPC bandwidth resources. Specifically how to implement the technical personnel can choose the processing method according to the actual situation of the system.

The above-mentioned embodiment 2 to embodiment 4 describe the processing methods of the LSA message, Hello message and DD message under the OSPF protocol by applying the OSPF protocol in the present invention. Compared with the conventional processing flow, the processing method of LSA packets and DD packets adds a response message sent from the standby processing unit to receive LSA packets and DD packets, so that the active processing unit and the standby processing unit can Received packets update the local data and make them consistent to prepare for the switchover.

In order to further improve the solution of the present invention, if the main processing unit receives the LSA message and the DD message multicast by the message receiving unit, it does not receive the message from the standby processing unit within the preset time. When responding to the unexpected situation of the message, the main processing unit sends the message to the backup processing unit through the IPC between the main processing unit and the backup processing unit and generates an alarm message, prompting the user that the backup processing unit may not be operating normally.

If the main processing unit receives the response message from the standby processing unit to receive the LSA message and the DD message, but does not receive the message within the preset time, the main processing unit sends a message to the standby processing unit. The unit requests the message; after receiving the request from the main processing unit, the backup processing unit sends the message to the main processing unit through the IPC connection between the main processing unit and the backup processing unit, and generates an alarm message , prompting the user that the active processing unit may not be working properly.

In the above-mentioned Embodiment 2 to Embodiment 4, after the primary processing unit and the standby processing unit complete the initial data synchronization through data batch backup or other means, if a local state change occurs in the primary processing unit, the local state change information is sent to A standby processing unit, the local data of the standby processing unit is also maintained according to the local state change information.

The local state change information includes: motherboard configuration change information, interface state change information, and neighbor router state change information. When the active processing unit has a configuration change, an interface state change, or a neighbor router state change, it sends local state change information to the standby processing unit through the IPC connection between the active processing unit and the standby processing unit, and the standby processing unit according to the partial state Change information maintains local data to ensure consistency with the data of the primary processing unit. Because the occurrence probability of the local state change is relatively low and the amount of information is small, the IPC bandwidth resources are occupied very little, and the user does not need to worry about the problem that the IPC bandwidth is occupied too much.

The process in which the standby processing unit takes over the communication of the main processing unit is called switching. The switching methods of the present invention can be divided into unplanned switching and planned switching. The unplanned switching and planned switching can be applied to the above-mentioned embodiments During the switching process from the second to the fourth embodiment.

Please refer to Fig. 5, it is the flowchart of unplanned switching when the method of the present invention applies OSPF agreement, comprises:

r1, the standby processing unit reads all network interface information from the interface layer;

r2, perform active-standby switchover, and the standby processing unit replaces the active processing unit;

The active-standby switchover here refers to the interface where the standby processing unit takes over the active processing unit.

r3, for any neighboring router whose status is not Full, the standby processing unit re-establishes the adjacent relationship with it;

In this embodiment, the neighbor relationship is re-established starting from the Down state. It can be understood that, if the standby processing unit performs the processing in Embodiment 4 on the DD message before the switchover, the method of re-establishing the adjacent relationship here can also be: for any state that is not full and not Exchange The standby processing unit establishes an adjacency relationship following its saved state; for any neighbor router whose state is switching, establishes an adjacency relationship from the ExStart state. This makes the seamless switchover process faster. Specifically how to implement the technical personnel can choose the processing method according to the actual situation of the system, and the specific processing method is not enough to limit the present invention.

r4, start the relevant timer;

The started timers include a Hello timer, a Dead timer and an LSA retransmission timer.

r5, import the path of other protocols according to the configuration and regenerate the LSA message from the local router;

r6, the standby processing unit runs the SPF algorithm to calculate the best path to the destination and updates the routing table and forwarding table according to these paths.

Unplanned switchover is generally aimed at the situation that software and hardware faults trigger the failure of the main processing unit, and the main processing unit cannot work normally.

Please refer to Fig. 6, it is a flow chart of planned switching when the method of the present invention applies the OSPF protocol, including:

s1, the standby processing unit reads in all network interface information from the interface layer;

s2, the active processing unit and the standby processing unit wait for the state of all neighboring routers to become Full;

It is understandable that if the active processing unit or the standby processing unit waits for a preset time, and the status of all neighbor routers still cannot become Full, then a similar unplanned switching step can be used for switching. For the specific process, refer to step r2 in Figure 6 Go to step r6.

s3, perform active-standby switchover, and the standby processing unit replaces the active processing unit.

s4, the standby processing unit uses the remaining time of the timer obtained from the main processing unit to set and start the relevant timer;

s5, import the path of other protocols according to the configuration and regenerate the LSA message from the local router;

s6, the standby processing unit runs the SPF algorithm to calculate the best path to the destination and updates the routing table and the forwarding table according to these paths.

Planned switchover is generally aimed at the situation that the main processing unit can work normally, and the manual execution of the switch triggers the failure of the main processing unit or the situation that the user presets a certain time interval to trigger the failure of the main processing unit, so that the standby processing unit can take over the main processing unit. The process is smoother with processing units.

It is understandable that there is no absolute chronological order between the steps in the above-mentioned unplanned switching and planned switching processes, and even some steps can be executed at the same time, as long as the switching requirements are met, the specific execution order does not be construed as a limitation of the invention.

Embodiment 5. A device for realizing communication takeover. The structure diagram of the device is shown in FIG. Unit 400 : includes a data backup control unit 410 and a takeover control unit 420 .

The data backup control unit 410 is configured to control the backup of the data of the active processing unit 200 to the standby processing unit 300 when the standby processing unit 300 is started.

The message receiving unit 100 is configured to receive a routing protocol message from a neighbor router, and send the message to the active processing unit 200 and the standby processing unit 300 at the same time.

In this embodiment, the message receiving unit 100 sends the message in a multicast form, and the active processing unit 200 and the standby processing unit 300 are regarded as members of the multicast group.

The standby processing unit 300 is used to provide routing protection for the active processing unit. After receiving the routing protocol message multicast by the message receiving unit 100, the standby processing unit 300 maintains the local data of the standby processing unit 300 according to the message.

The main processing unit 200 is used to provide the main routing function of the router; after receiving the routing protocol message multicast by the message receiving unit 100, the main processing unit maintains the local data of the main processing unit 200 according to the message.

The takeover control unit 420 is configured to control the standby processing unit 300 to take over the communication of the main processing unit 200 when the main processing unit 200 fails.

Embodiment 6. A device for realizing communication takeover. The structure diagram of the device is shown in FIG. The unit 200 includes: a main receiving unit 210, a main data maintenance unit 220, a response message receiving unit 230, a main message forwarding unit 240 and a request unit 250; the backup processing unit 300 includes: a backup receiving unit 310, a backup data The maintenance unit 320 , the response message sending unit 330 and the backup message forwarding unit 340 .

Each unit is described in detail below.

The packet receiving unit 100 is configured to receive routing protocol packets from neighboring routers, and send the packets to the active receiving unit 210 on the active processing unit 200 and the standby receiving unit 310 on the standby processing unit 300 simultaneously.

The standby receiving unit 310 is configured to receive the protocol packet sent by the packet receiving unit 100 ; and is also configured to receive the packet forwarded by the active packet forwarding unit 240 .

A standby data maintenance unit 320, configured to maintain local data according to the message received by the standby receiving unit 310;

A response message sending unit 330, configured to send a response message receiving the message to the response message receiving unit 230 after the standby data maintenance unit 320 maintains the local data according to the received message;

The active receiving unit 210 is used to receive the protocol message sent by the message receiving unit 100 ; and is also used to receive the message forwarded by the standby message forwarding unit 340 .

The main data maintenance unit 220 is configured to maintain the local data according to the message received by the main receiving unit 210 after the response message receiving unit 230 receives the response message;

The response message receiving unit 230 receives a response message from the received message sent by the response message sending unit 320;

The main message forwarding unit 240 is configured to receive the message sent by the response message sending unit 330 within the preset time after the main receiving unit 210 receives the message. When responding to a message, send the message to the standby receiving unit 310, and generate an alarm message to the user.

The request unit 250 is used to receive the response message from the received message sent by the response message sending unit 330 after the response message receiving unit 230, when the main receiving unit 210 does not receive the message receiving unit within the preset time When the message is sent from 100, request the message from the standby forwarding unit 340;

The backup message forwarding unit 340 is configured to send the message requested by the requesting unit 250 to the main receiving unit 210 after receiving the request from the main processing unit, and generate an alarm message.

The method and device for realizing communication takeover provided by the present invention have been introduced in detail above. In this paper, specific examples are used to illustrate the principle and implementation of the present invention. The description of the above embodiments is only used to help understand the present invention. method and its core idea; at the same time, for those of ordinary skill in the art, according to the idea of the present invention, there will be changes in the specific implementation and application scope. Invention Limitations.

Claims (21)

1. a method that realizes communicating by letter and taking over is characterized in that, comprising:

Main with processing unit and the synchronous routing protocol packet that receives from neighbor router of standby processing unit;

Main processing unit and the standby processing unit used is according to described message maintenance local data separately;

When the master used the processing element failure, standby processing unit was taken over the main processing unit communication of using.

2. the method that realization communication as claimed in claim 1 is taken over is characterized in that,

Described master safeguards that according to described message the process of local data separately comprises with processing unit and standby processing unit:

Standby processing unit is safeguarded standby processing unit local data according to described message;

Standby processing unit sends the response message of receiving described message and uses processing unit to the master;

The master receives described response message with processing unit, safeguards the main processing unit local data of using according to described message.

3. the method that realization communication as claimed in claim 2 is taken over, it is characterized in that, further comprise: if after the master receives described message with processing unit, within preset time, do not receive the response message of receiving described message from standby processing unit, then send described message to standby processing unit and produce a warning message.

4. the method that realization communication as claimed in claim 2 is taken over is characterized in that, if main with processing unit receive receive the response message of described message from standby processing unit after, within preset time, do not receive described message, then described method further comprises:

The main processing unit of using is to the described message of standby processing unit request;

Standby processing unit sends described message and uses processing unit to main, and produces a warning message.

5. the method that realization communication as claimed in claim 1 is taken over is characterized in that described Routing Protocol is an OSPF Open Shortest Path First; Described message is the LSA message;

Described master safeguards that according to the LSA message process of local data separately comprises with processing unit and standby processing unit:

Standby processing unit upgrades standby processing unit link-local slip condition database according to described LSA message;

Standby processing unit sends the response message of receiving this LSA message and uses processing unit to the master;

The master receives response message with processing unit, upgrades main with processing unit link-local slip condition database according to described LSA message; The operation shortest-path first algorithm calculates the optimal path that arrives the destination and according to these routing update routing tables with transmit.

6. the method that realization communication as claimed in claim 5 is taken over, it is characterized in that, after described master receives that with processing unit standby processing unit is received the response message of this LSA message, send the response message of receiving described LSA message for described neighbor router.

7. the method that realization communication as claimed in claim 5 is taken over, it is characterized in that, standby processing unit safeguards that according to the LSA message process of local data further comprises: for each LSA message that receives, if standby processing unit determines local router described message is carried out inundation to certain neighbor router, then standby processing unit is put into described LSA message in the re-transmission table of this neighbor router of standby processing unit local record; When standby processing unit receives from this neighbor router the response message of this LSA message, described LSA message is deleted from the re-transmission table of this neighbor router of local record.

8. the method that realization communication as claimed in claim 1 is taken over is characterized in that described Routing Protocol is an OSPF Open Shortest Path First; Described message is for connecting message;

Described standby processing unit safeguards that according to connecting message the process of local data comprises:

Standby processing unit is safeguarded the state information of the described neighbor router of standby processing unit local record according to described connection message.

9. the method that realization communication as claimed in claim 1 is taken over is characterized in that described Routing Protocol is an OSPF Open Shortest Path First; Described message is a database description packet;

Described master safeguards that according to database description packet the process of local data separately comprises with processing unit and standby processing unit:

Standby processing unit is safeguarded the required list of the described neighbor router of standby processing unit local record according to described database description packet;

Standby processing unit sends the response message of receiving this database description packet and uses processing unit to the master;

Main receive described response message, safeguard the state information of the required list of described neighbor router of local record and described neighbor router according to described database description packet and give described neighbor router transmission for this database description packet and receive the response message of this database description packet with processing unit.

10. the method that realization communication as claimed in claim 9 is taken over is characterized in that the process of the required list of the described neighbor router of the standby processing unit local record of described maintenance comprises:

Check whether each the LSA descriptor in this database description packet is newer than the link corresponding state advertisement in the standby processing unit LSD, if then standby processing unit is put into described LSA descriptor in the required list of this neighbor router; When standby processing unit is received described LSA descriptor link corresponding state advertisement message from described neighbor router, described LSA descriptor is deleted from the required list of described neighbor router.

11. the method that realization communication as claimed in claim 1 is taken over is characterized in that, when standby processing unit started, main data with processing unit are backuped to standby processing unit, and to finish primary data synchronous.

12. the method that realization communication as claimed in claim 11 is taken over is characterized in that, finish primary data synchronously after, if mainly change, then transmit described local state change information and give standby processing unit with processing unit generation local state; The local data of described standby processing unit is also safeguarded according to described localized variation information.

13. the method that realization communication as claimed in claim 12 is taken over is characterized in that described local state change information comprises: configuration variation information, Interface status change information and neighbor router state change information.

14., it is characterized in that when hardware and software failure triggered the master with the processing element failure, standby processing unit was taken over main process with the processing unit communication and comprised step as any method that described realization communication is taken over of claim 5 to 10:

Standby processing unit reads in the all-network interface message from interface layer;

Carry out active and standby switching, standby processing unit replaces the main processing unit of using;

For any one state is not full neighbor router, and standby processing unit rebulids neighbouring relations with it;

Start relevant timer;

Introduce the path of other agreements and produce the LSA message that comes from local router again according to configuration;

Standby processing unit operation shortest-path first algorithm calculates the optimal path that arrives at the destination and according to these routing update routing tables with transmit.

15. the method that any described realization communication is taken over as claim 5 to 10 is characterized in that, triggers mainly when handling element failure when manually starting to switch, standby processing unit is taken over main process with the processing unit communication and is comprised:

Standby processing unit reads in the all-network interface message from interface layer;

It is full to wait for that all neighbor router states become;

Carry out active and standby switching, standby processing unit replaces the main processing unit of using;

Standby processing unit is used from the timer that be provided with remaining time and startup is relevant of leading the timer that obtains with processing unit;

Introduce the path of other agreements and produce the LSA message that comes from local router again according to configuration;

Standby processing unit operation shortest-path first algorithm calculates the optimal path that arrives at the destination and according to these routing update routing tables with transmit.

16. the method that realization communication as claimed in claim 15 is taken over is characterized in that,

After the time that wait is preset, if it is full to still have the neighbor router state not become; Then carry out step:

Carry out active and standby switching, standby processing unit replaces the main processing unit of using;

For any one state is not full neighbor router, and standby processing unit rebulids neighbouring relations with it;

Start relevant timer;

Introduce the path of other agreements and produce the LSA message that comes from local router again according to configuration;

Standby processing unit operation shortest-path first algorithm calculates the optimal path that arrives at the destination and according to these routing update routing tables with transmit.

17. a device of realizing communicating by letter and taking over comprises message receiving element, main with processing unit, standby processing unit and main-slave control cell, it is characterized in that:

The message receiving element is used to receive the routing protocol packet from neighbor router, and sends described message to main simultaneously with processing unit and standby processing unit;

Main processing unit and the standby processing unit used is according to described message maintenance local data separately;

Main-slave control cell is used to control standby processing unit and takes over main with handling unit communication.

18. the device that realization communication as claimed in claim 17 is taken over is characterized in that described main-slave control cell comprises: take over control unit, be used for when the master uses the processing element failure, control standby processing unit and take over the main processing unit communication of using;

Described main-slave control cell also comprises the data backup control unit, is used to control the data that will lead with processing unit and backups to standby processing unit.

19. the device that realization communication as claimed in claim 17 is taken over is characterized in that,

Described standby processing unit comprises:

Standby receiving element is used to receive the protocol massages that the message receiving element sends;

The preliminary data maintenance unit is used for safeguarding local data according to the message that standby receiving element receives;

The response message transmitting element is used for sending and receiving that the response message of described message is to the response message receiving element after the preliminary data maintenance unit is safeguarded local data according to the message that receives;

Described master draws together with the processing unit pack:

The master uses receiving element, is used to receive the protocol massages that the message receiving element sends;

The response message receiving element receives the response message of receiving message that sends from the response message transmitting element;

The master uses the data maintenance unit, is used for safeguarding local data according to the main message that receives with receiving element after the response message receiving element is received described response message.

20. the device that realization communication as claimed in claim 19 is taken over is characterized in that described master also comprises with processing unit:

The main message retransmission unit of using, be used for after the master receives message with receiving element, when response message receiving element in preset time do not receive that the response message transmitting element sends receive the response message of described message the time, send described message to standby receiving element, and produce a warning message to the user;

Described standby receiving element also is used to receive the main message of transmitting with the message retransmission unit.

21. the device as claim 19 or 20 described realization communications adapters is characterized in that,

Described master also comprises with processing unit:

Request unit, be used for after the response message receiving element is received the response message of receiving message that sends from the response message transmitting element, when main when not receiving the described message that the message receiving element sends within preset time, to the described message of standby message retransmission unit request with receiving element;

Described standby processing unit also comprises:

Standby message retransmission unit is used for receiving the master with after the request of processing unit that the message that sends the request unit request is used receiving element to main, and produces a warning message;

Described master also is used to receive the message that standby message retransmission unit is transmitted with receiving element.

Images