CN112202690B - A High-speed Bus Network Based on Switching and Ring Network Redundancy - Google Patents

Abstract

The invention relates to a high-speed bus network based on exchange and ring network redundancy, wherein a network topology structure comprises N exchange nodes and M × N terminal nodes; each exchange node is connected with M terminal nodes end to form a ring network, and N exchange nodes are interconnected according to a star structure to form a composite network structure of exchange + ring network; in the composite network structure, the exchange node is used as a core node of the network to carry out communication scheduling on data interaction and communication interconnection of the whole network, so that the whole network reaches a synchronous stable state in a global uniform time, and the occurrence of transmission conflict is avoided; the terminal node serves as a data terminal of the network for the generation and reception processing of the communication data stream. The invention realizes the high-bandwidth parallel scheduling of the large data block and the high real-time parallel scheduling of the small data block, can meet the requirement of an avionic system on high communication rate, can also ensure the certainty of real-time information, effectively improves the reliability of a bus network and meets the complex requirement of the avionic system.

Description

A High-speed Bus Network Based on Switching and Ring Network Redundancy

technical field

The invention relates to the field of bus technology, in particular to a high-speed bus network based on switching and ring network redundancy.

Background technique

With the in-depth development of the wave of information technology revolution, the data bus technology of avionics system is also being updated rapidly to meet the data transmission and information interconnection requirements of system equipment at all levels.

MIL-STD-1553B, as the first generation of avionics system data bus technology, has good redundancy and fault tolerance and real-time performance (response time is millisecond level), but the transmission rate is only 1Mbps, which cannot meet the requirements of large-capacity data transmission; FC -AE-1553B bus is based on MIL-STD-1553B and optical fiber communication technology. On the basis of being compatible with MIL-STD-1553B bus, it also has the characteristics of high transmission rate and low bit error rate of optical fiber communication. It has been developed rapidly and widely application, but the FC-AE-1553B bus has no time synchronization function and redundancy management function. At the same time, like MIL-STD-1553B, it needs proprietary hardware support, which greatly increases the design cost and design cycle; AFDX (Avionics Full Duplex SwitchedEthernet, The biggest advantage of the avionics full-duplex switched Ethernet) bus is that it can be realized by general-purpose Ethernet hardware. At the same time, virtual link technology and redundancy management technology are used to improve the reliability of data transmission, but the transmission rate is generally 100Mbps. And the real-time performance cannot be guaranteed; TTE (Time Triggered Ethernet) introduces clock synchronization technology and adds time-triggered services on the basis of the traditional Ethernet protocol IEEE802.3, which can not only be fully compatible with traditional Ethernet communication, but also To achieve high real-time time-triggered communication, but the highest rate that TTE can currently achieve is 1Gbps, and proprietary hardware is required; on ordinary Ethernet hardware, TTE can only be realized by software, with a rate of up to 100Mbps. The comparison of the characteristics of the above various buses is shown in the table below.

In the future, with the increase in the complexity of avionics equipment, the data throughput will increase, and the reliability and real-time requirements for data transmission will become higher and higher. Therefore, it is urgent to develop new bus technology to meet the needs of avionics systems. complex needs.

Contents of the invention

In view of the above analysis, the present invention aims to provide a high-speed bus network based on switching and ring network redundancy; to solve the requirements of future avionics systems on data volume, real-time performance and reliability.

The invention discloses a high-speed bus network based on switching and ring network redundancy, which includes N switching nodes and M*N terminal nodes;

Each switching node is connected end-to-end with M terminal nodes to form a ring network, and N switching nodes are interconnected in a star structure to form a composite network structure of switching + ring network;

In the composite network structure, the switching node, as the core node of the network, performs communication scheduling for the data interaction and communication interconnection of the entire network, so that the entire network can reach a synchronous and stable state within a global unified time, avoiding the occurrence of transmission conflicts;

As the data terminal of the network, the terminal node is used for the generation, reception and processing of communication data streams.

Further, the switching nodes and the switching nodes, between the switching nodes and the terminal nodes, and between the terminal nodes are connected by optical fibers and Ethernet twisted-pair wires, and the optical fibers and the Ethernet are mutually redundant.

Further, in the communication scheduling of the high-speed bus network, a network communication scheduling method combining ordered scheduling and free scheduling is adopted, and each bus network scheduling period is divided into an ordered scheduling period and a free scheduling period;

In the orderly scheduling period, the switching node as the bus controller sends high real-time, small data flow, and low-bandwidth periodic data including control commands to other switching nodes and terminal nodes in an orderly manner;

In the free scheduling period, the nodes in the network are free to transmit low real-time, large data flow, and high bandwidth data including sensor and multimedia data.

Further, the communication scheduling method includes the following steps:

1) Divide the bus time into definite basic cycles;

2) At the beginning of each bus network scheduling cycle, one of the switching nodes is used as the bus controller to send a broadcast message to all network nodes, indicating that it is currently entering the orderly scheduling phase;

3) After the orderly scheduling of each node ends, the bus controller will send another broadcast message, indicating that the current scheduling phase enters the free scheduling phase.

Furthermore, in the ordered scheduling phase, other switching nodes and terminal nodes must respond to the network transmission initiated by the switching node as the bus controller, and other nodes are not allowed to actively initiate network data transmission.

Further, according to the cycle characteristics of the control command sent by the bus controller, the bus controller chooses to only perform free scheduling in the bus network scheduling period when the control command does not change, and does not perform ordered scheduling.

Furthermore, free scheduling adopts point-to-point communication, which allows simultaneous transmission of data in multiple directions in the network.

Furthermore, in the free scheduling phase of each bus network cycle, each network node polls the data flow, and when it finds that there is a data flow, it performs point-to-point data transmission until the end of the bus network cycle; if the data block is large , the transmission time is greater than the free scheduling cycle of a bus network scheduling cycle, the data block is divided into several data sub-blocks according to the maximum data transmission amount of each free scheduling cycle, and the data are transmitted sequentially in the free scheduling phase of the continuous bus network cycle Data word block until complete transfer of the entire data block.

Further, the bus network adopts redundant fault tolerance in the physical layer, link layer, network layer, transport layer and application layer to ensure the reliability of the network.

Further, the redundant fault tolerance includes the following methods:

1) Photoelectric redundancy at the physical layer; through the mutual redundancy of optical fiber and Ethernet, the reliability of transmission is improved by using the anti-vibration and anti-stretch performance of copper cables and the anti-electromagnetic interference performance of optical cables;

2) Frame verification at the link layer to remove redundancy; perform frame verification to remove redundant data messages from message redundancy brought by multiple redundant paths in the network, and upload the only correct data message at the link layer;

3) Data retransmission; configure the transmission timeout period one by one for data messages, provide a retransmission mechanism when a transmission timeout occurs, and support four transmissions including three retransmissions;

4) Fixed response loopback; through online judgment at the transport layer, the fixed response loopback judges the communication status of the transport layer, link layer, and physical layer of the communication, and is used to monitor communication failures;

5) Transmission layer scheduling control switching; set two identical and independent switching units at the switching node, the two are mutually redundant backups, and the backup switching unit redundantly monitors the network scheduling process of the main switching unit continuously through the path. After the main switching unit stops scheduling, the backup switching unit starts immediately to take over the continuous scheduling of the bus network to avoid network paralysis.

6) Application layer operation monitoring; refresh the application layer response loopback, and monitor the application layer operation status.

The present invention at least achieves one of the following beneficial effects:

1. This novel high-speed bus proposed by the present invention has realized the high-bandwidth and high-real-time parallel scheduling of large data blocks (sensor and multimedia data) and small data blocks (control commands), which can satisfy the requirements of avionics systems for high communication rates. Requirements, but also to ensure the certainty of real-time news.

2. The new high-speed bus proposed by the present invention adopts multiple redundant fault-tolerant technologies to strengthen the redundant fault-tolerant design of the network at different protocol levels, thereby effectively improving the reliability of the bus network.

3. The new high-speed bus proposed by the present invention adopts mature physical layer technology and COTS general-purpose devices, which has low design difficulty and cost, high transmission rate, good real-time performance and high reliability, and can meet the combination of high-speed/low-speed multi-protocol The layered network can meet the complex requirements of the avionics system and has broad application prospects.

Description of drawings

The drawings are for the purpose of illustrating specific embodiments only and are not to be considered as limitations of the invention, and like reference numerals refer to like parts throughout the drawings.

Fig. 1 is the topological structure diagram of the high-speed bus in the embodiment of the present invention;

Fig. 2 is a schematic diagram of network scheduling of a high-speed bus in an embodiment of the present invention;

Fig. 3 is a diagram of the protocol layer structure of the high-speed bus in the embodiment of the present invention.

Detailed ways

Preferred embodiments of the present invention will be specifically described below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of the application and are used together with the embodiments of the present invention to explain the principles of the present invention.

This embodiment discloses a high-speed bus network based on switching and ring network redundancy. As shown in FIG. 1 , its network topology includes N switching nodes and M*N terminal nodes;

Each switching node is connected end-to-end with M terminal nodes to form a ring network, and N switching nodes are interconnected in a star structure to form a composite network structure of switching + ring network;

In addition, optical fiber and Ethernet twisted pair are used for connection between switching nodes, between switching nodes and terminal nodes, and between terminal nodes, and the photoelectric mutual redundancy design.

In the composite network structure, the switching node, as the core node of the network, performs communication scheduling for the data interaction and communication interconnection of the entire network, so that the entire network can reach a synchronous and stable state within a global unified time, avoiding the occurrence of transmission conflicts;

Specifically, the switching node includes two completely identical switching units that are independent of each other. The two can serve as redundant backups for each other. The backup switching unit redundantly monitors the network scheduling process of the main switching unit continuously through the path. When it is found that the main switching unit stops scheduling After that, the backup switching unit starts immediately to take over the continuous scheduling of the bus network to avoid network paralysis.

As the data terminal of the network, the terminal node is used for the generation, reception and processing of communication data streams.

Both switching nodes and terminal nodes adopt COTS general-purpose optical transceivers and Ethernet transceiver devices, and provide multi-channel optical fiber interfaces and Ethernet electrical ports, among which optical fibers are used to transmit large-capacity and high-bandwidth data, and can achieve transmission of up to tens of Gbps Speed; the Ethernet port can transmit both large data and time synchronization information.

The switching node includes circuit modules for time synchronization, and realizes time synchronization by implementing the IEEE1588 protocol, so that the entire network maintains a global unified time, and the precision can reach ns level. The terminal node maintains time synchronization with the connected switching node by receiving the time information sent by the switching node. After a period of message exchange and time information processing, the global time of the entire network reaches a stable state of synchronization.

In a specific embodiment, the number of terminal nodes connected to each switching node to form a ring network does not exceed 8, the number of switching nodes does not exceed 16, and the maximum number of communication nodes in the entire network can reach 128.

The communication scheduling of the high-speed bus network in this embodiment adopts a network communication scheduling method combining ordered scheduling and free scheduling, and divides each bus network scheduling cycle into an ordered scheduling cycle and a free scheduling cycle; by periodically scheduling the network, To meet the requirements of combined scheduling of small data blocks and large data blocks.

In the orderly scheduling cycle, a switching node bus controller can be arbitrarily designated; the switching node as the bus controller sends orderly high real-time, small data flow, and low bandwidth cycles including control commands to other switching nodes and terminal nodes data;

In the free scheduling period, the nodes in the network are free to transmit low real-time, large data flow, and high bandwidth data including sensor and multimedia data.

The specific network communication scheduling method is shown in Figure 2.

1) Divide the bus time into definite basic cycles;

The basic cycle can be divided according to specific application scenarios, and the length can be as low as 500us.

2) At the beginning of each bus network scheduling cycle, one of the switching nodes is used as the bus controller to send a broadcast message to all network nodes, indicating that the current order scheduling cycle is entered;

The broadcast message is sent to all network nodes as a high-priority scheduling synchronization instruction packet including control commands to coordinate orderly scheduling of each node.

In the orderly scheduling stage, the periodic data such as control commands are mainly transmitted, which is fixedly transmitted once in each bus network cycle. A certain switching node acts as the bus controller to initiate network transmission, and other switching nodes and terminal nodes must carry out the transmission initiated by the bus controller. Responses, rather than actively initiating network data transfers. In this stage, all data streams of the bus network will not have any competition, and the data is executed under strict time definition, with good time determinism, which is suitable for transmitting highly reliable and high real-time control data.

Specifically, the control instructions include time synchronization information, sensor parameter configuration commands and other custom commands;

Specifically, the content of the response of each node is time synchronization information, the current parameter configuration of the sensor, whether the working status is normal or whether the command configuration is successful, etc.; and the response message of each node contains destination address information;

When a node receives a response message from other nodes, it judges the destination address information. When it finds that the destination address is not its own, it forwards the response message. In this way, all the response messages will eventually converge after being forwarded by each node. to the bus controller; the bus controller can judge the response status of each node according to the source address (node address) and destination address information of each node response message, after the bus controller receives the response messages of all nodes, it can be considered Or, if no response message is received from the node within the set time range, the in-order scheduling cycle is considered to be over. The set time range may be determined according to specific scheduling and response requirements.

Although the above-mentioned control commands and response message data are transmitted every cycle, the amount of data is not large, and the occupied bandwidth is relatively small. It is also possible to schedule once every several bus network cycles according to the cycle characteristics of the control command, so as to further reduce the occupied bandwidth.

3) When the ordered scheduling period is over, the bus controller will send another broadcast message, indicating that the current scheduling phase enters the free scheduling phase.

The broadcast message includes the low-priority scheduling synchronization instruction packet of the free scheduling instruction, so that each node can clearly enter the free scheduling phase.

In the free scheduling stage, it mainly transmits periodic data with large data flow and high bandwidth, such as sensor and multimedia data. In the second half of each bus scheduling cycle, that is, after the ordered scheduling commands are processed, the network enters the free scheduling phase. Free scheduling generally adopts point-to-point communication, and data in multiple directions can be transmitted in the network at the same time, thereby realizing the doubling of the bus bandwidth. In the free scheduling stage, error recovery measures such as network redundancy, communication confirmation, and retransmission are still effective, which is suitable for transmitting high-bandwidth audio and video data.

In the free scheduling phase, each node to send data polls multiple possible large data streams in each bus network cycle, and if there is data, a certain amount of data is transmitted until the end of the bus network cycle. If the data block is very large, the bus does not require or allow the transmission of the entire data block within one bus network cycle, but divides the data block into several data sub-blocks according to the maximum data transmission amount of each free scheduling cycle, and continuously In the free scheduling phase of the bus network cycle, the data blocks are transmitted sequentially until the transmission of the entire data block is completed.

In addition, it can be seen from Figure 2 that the priority of ordered scheduling is high, and the priority of free scheduling is low.

Further, the bus controller of the bus network in this embodiment also provides two other event triggers, which are high-priority aperiodic scheduling and low-priority aperiodic scheduling, which are used for emergency or emergency event transmission.

In the periodic scheduling process of bus network scheduling, it is allowed to insert high-priority aperiodic messages and network interruptions throughout the process; in the free scheduling time period, it is allowed to insert low-priority aperiodic messages; to improve the real-time response of the bus in an emergency state ability.

Preferably, the priority of the terminal nodes in the bus network is classified. Under normal circumstances, the data of each terminal node is transmitted through periodic scheduling. When an emergency occurs, it is necessary to prioritize the transmission of high-priority terminal node data, then In the whole process of periodic scheduling, by inserting high-priority non-periodic messages and network interruptions, the high-priority terminal node data should be transmitted preferentially; if the low-priority terminal node data needs to be transmitted preferentially, it can be in In the free scheduling time period, low priority aperiodic messages are inserted to improve the real-time response capability of the bus in an emergency state.

For example, sensors in different positions in the avionics system play different roles in the operation of the entire system. Some sensors such as position, attitude, key position temperature, air pressure and other data are very important, so they need to be set as high priority; other non-critical sensors such as brightness inside the system, temperature outside the system, air pressure, etc. are low priority.

Under normal circumstances, the data of each sensor is transmitted through periodic scheduling. When an emergency occurs, it is necessary to focus on monitoring a certain position sensor, or when it is necessary to obtain more information other than periodic transmission data, it can be used in periodic scheduling. Throughout the process, the communication of this critical sensor data is guaranteed by inserting high-priority acyclic messages and network interruptions. Or, if the data of a non-critical sensor needs to be obtained preferentially outside the periodic scheduling, it is also possible to insert a low-priority non-periodic message during the free scheduling time period to ensure the communication of the non-critical sensor data, so as to improve the emergency response time of the bus. real-time responsiveness in the state.

It can be seen from the above that the bus realizes the high bandwidth of large data blocks (sensor and multimedia data) and high real-time parallel scheduling of small data blocks (control commands), which can not only meet the requirements of avionics systems for high communication rates, but also can Guarantee the certainty of real-time news.

Due to the unreliability of the network, a series of redundant fault-tolerant measures must be adopted in order to carry out reliable information transmission. The method often used in practice is based on physical redundancy backup. However, in the complex application environment of avionics systems, Only relying on the redundancy of physical links is not enough to meet the reliability index. More means and multi-level methods are needed to ensure the correctness of data and improve the redundancy and fault tolerance performance of the network.

Fig. 3 is the protocol structure of the high-speed bus proposed by the present invention. The protocol structure includes five layers, which are physical layer, link layer, network layer, transport layer and application layer.

The bus network of this embodiment adopts redundant fault-tolerant design at the above different protocol levels to ensure the reliability of the network. details as follows:

1) The photoelectric redundancy technology is adopted at the physical layer, through the mutual redundancy of optical fiber and Ethernet, the anti-vibration and anti-stretch performance of copper cables and the anti-electromagnetic interference performance of optical cables are used to make up for the shortcomings of a single medium and improve transmission reliability;

2) In the link layer, the frame verification de-redundancy technology is used to perform frame verification on the message redundancy brought by multiple redundant paths in the network to remove redundant data messages, and upload the only correct data message in the link layer arts;

3) Using retransmission technology; configure the transmission timeout period for data packets one by one, provide a retransmission mechanism when a transmission timeout occurs, and support four transmissions including three retransmissions;

4) Fixed response loopback; through online judgment at the transport layer, the fixed response loopback judges the communication status of the transport layer, link layer, and physical layer of the communication, and is used to monitor communication failures;

5) The transmission layer scheduling controller is switched, and two identical and independent switching units are set on the switching node. After finding that the main switching unit stops scheduling, the backup switching unit starts immediately to take over the continuous scheduling of the bus network to avoid network paralysis.

6) Application layer operation monitoring; refresh the application layer response loopback, and monitor the operation status of the application layer (general software).

The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present invention can easily think of changes or Replacement should be covered within the protection scope of the present invention.

Claims (6)

1. A high-speed bus network based on exchange and ring network redundancy is characterized by comprising N exchange nodes and M × N terminal nodes;

each exchange node and M terminal nodes are connected end to form a ring network, and N exchange nodes are interconnected according to a star structure to form a composite network structure of exchange + ring network;

in the composite network structure, one switching node is used as a core node of the network to carry out communication scheduling on data interaction and communication interconnection of the whole network, so that the whole network reaches a synchronous stable state in a global unified time, and transmission conflict is avoided;

the switching node comprises two completely identical and mutually independent switching units which are mutually redundant and backup, the redundancy of the backup switching unit continuously monitors the network scheduling process of the main switching unit through a path, and when the main switching unit stops scheduling, the backup switching unit immediately starts to take over the continuous scheduling of the bus network to avoid network paralysis;

the terminal node is used as a data terminal of the network and used for generating and receiving processing of communication data flow;

in the communication scheduling of the high-speed bus network, a network communication scheduling method combining ordered scheduling and free scheduling is adopted, and each bus network scheduling period is divided into an ordered scheduling period and a free scheduling period;

the communication scheduling method comprises the following steps:

1) Dividing the bus time into certain basic cycles;

2) At the initial time of each bus network scheduling period, one of the exchange nodes is used as a bus controller, and a broadcast message is sent to all the network nodes to indicate that the current bus network enters an ordered scheduling stage;

in the ordered scheduling period, the switching node serving as a bus controller orderly sends periodic data including control commands, such as high real-time data streams, small data streams and low bandwidth, to other switching nodes and terminal nodes;

the control commands comprise time synchronization information, sensor parameter configuration commands and other self-defined commands;

the content of each node response is time synchronization information, the current parameter configuration and the working state of the sensor are normal or the command is successfully configured, and the response message of each node contains destination address information;

when one node receives the response messages of other nodes, the destination address information is judged, and when the destination address is not found, the response message is forwarded, so that all the response messages are finally converged to the bus controller through the forwarding of each node; the bus controller judges the response condition of each node according to the node address and the destination address information of each node response message, and after receiving the response messages of all the nodes, the bus controller considers that the ordered scheduling period is finished; or, the response message of the node is not received within the set time range, namely the ordered scheduling cycle is considered to be finished;

3) After the ordered scheduling of each node is finished, the bus controller can send a broadcast message again to indicate that the current scheduling stage enters a free scheduling stage;

in a free scheduling period, the transmission of low real-time, large data flow and high-bandwidth data including sensor and multimedia data is freely carried out among all nodes in the network;

the broadcast message indicating the current scheduling stage to enter the free scheduling stage comprises a low-priority scheduling synchronous indication packet of a free scheduling instruction, so that each node definitely enters the free scheduling stage;

a free scheduling stage, which is mainly used for transmitting the periodic data with large data flow and high bandwidth; the free scheduling adopts point-to-point communication, and data in multiple directions can be transmitted in the network at the same time; if the data block is very large, the bus does not require or allow the transmission of the whole data block to be completed in one bus network period, but divides the data block into a plurality of data sub-blocks according to the maximum data transmission quantity of each free scheduling period, and sequentially transmits the data sub-blocks in the free scheduling stage of the continuous bus network period until the transmission of the whole data block is completed.

2. The high speed bus network of claim 1, wherein the switching nodes and the switching nodes, the switching nodes and the termination nodes, and the termination nodes are connected by optical fiber and ethernet twisted pair, the optical fiber and the ethernet being redundant to each other.

3. The high speed bus network of claim 1,

in the ordered scheduling phase, the network transmission initiated by the switching node as the bus controller must be responded by other switching nodes and terminal nodes, and other nodes are not allowed to actively initiate network data transmission.

4. The high speed bus network of claim 3,

according to the cycle characteristic of the control command sent by the bus controller, the bus controller selects the bus network scheduling cycle with unchanged control command to only perform free scheduling and not perform ordered scheduling.

5. The high speed bus network of claim 1, wherein the bus network employs redundant fault tolerance at the physical layer, link layer, network layer, transport layer, and application layer to ensure reliability of the network.

6. The high speed bus network of claim 5, wherein the redundant fault tolerance comprises the method of:

1) Opto-electrical redundancy of the physical layer; the transmission reliability is improved by utilizing the vibration resistance and the tensile resistance of the copper cable and the anti-electromagnetic interference performance of the optical cable through mutual redundancy of the optical fiber and the Ethernet;

2) Frame check of a link layer removes redundancy; carrying out frame check on message redundancy brought by a plurality of redundant paths in the network to remove redundant data messages, and uploading the only correct data message of a link layer;

3) Retransmitting data; carrying out transmission overtime time configuration on the data message one by one, providing a retransmission mechanism when transmission overtime occurs, and supporting four transmissions including three times of retransmission;

4) A fixed response loop; the communication conditions of a transmission layer, a link layer and a physical layer of communication are judged by a fixed response loop through online judgment on the transmission layer, so that the monitoring of communication faults is realized;

5) Scheduling control switching of a transmission layer; two completely identical and mutually independent exchange units are arranged at an exchange node, wherein the two exchange units are mutually redundant and backup, the redundancy of the backup exchange unit continuously monitors the network scheduling process of the main exchange unit through a path, and when the fact that the main exchange unit stops scheduling is found, the backup exchange unit is immediately started to take over the continuous scheduling of a bus network, so that network paralysis is avoided;

6) Monitoring the operation of an application layer; and refreshing the response loop of the application layer and monitoring the running condition of the application layer.

Images