CN103647682B - A Simulation System for Simulating Switched Ethernet Clock Synchronization - Google Patents

Abstract

The invention discloses a kind of analogue system simulating the synchronization of switching Ethernet clock, this analogue system includes Ethernet protocol instruction database, intervalometer, node and link attribute configuration module, synchronizes Topology g eneration module, state management module, State Transferring module, data processing module, fault detect and processing module, clock correcting module and result output module.Ethernet protocol instruction database and intervalometer are to meet in IEEE802.3ISO/IEC8802 3 standard criterion and SAEAS6802 pointed instruction, state relation information.Carry out emulating role's configuration by node and link attribute configuration module and obtain synchronization topology；Then to each nodes for state Role Management in synchronization topology, and data process, clock correction and fault detect are carried out with different conditions role；Thus output state and clock correction result.Analogue system of the present invention, relative to existing analogue system, is set up aspect in network model and is shown higher extensibility, it is simple to Clock Synchronization Procedure is carried out significantly more efficient analysis.

Description

A Simulation System for Simulating Switched Ethernet Clock Synchronization

technical field

The present invention relates to a simulation system based on Ethernet, more particularly, refers to a simulation system for establishing and maintaining clock synchronization by simulating switched Ethernet. Through this system, users can simulate the clock synchronization process of switched Ethernet in different situations according to their needs.

Background technique

Ethernet is a type of computer local area network (technology) that meets a set of standard specifications. The so-called set of standard specifications mainly refers to a series of standards such as IEEE802.3ISO/IEC8802-3, IEEE802.12ISO/IEC8802-12 or IEEE802.11ISO/IEC8802-11, HiperLAN2 and other standards; This set of protocols regulates connected data communication networks. Refer to "Ethernet Technology and Application" edited by Wang Tingyao, first edition in January 2005, pages 4 to 7. In the block diagram of the composition example of the Ethernet, an Ethernet includes at least a server (Server), a switch (Frame Switcher), a router (Router) and an AP node (Access PointEquipment).

Switched Ethernet refers to Ethernet using switching technology, and the connecting nodes are switches. Refer to "10 Gigabit Ethernet and Its Practical Technology", edited by Ao Zhigang, first printing in July 2007, page 82.

As a mission- and safety-critical system, avionics systems require fully deterministic timing guarantees. At present, advanced large aircraft at home and abroad all use Avionics FullDuplex Switched Ethernet (AFDX) as the backbone of the integrated interconnection of airborne electronic systems. The AFDX network proposes a time-deterministic mechanism. And a flow control mechanism is introduced to constrain the rate of communication tasks, which is called rate-limited (Rate Constraint, RC) traffic. However, RC traffic has a certain delay and Latency jitter. Therefore, a time-triggered concept came into being. Time Triggered Ethernet (Time Triggered Ethernet, TTE) introduces the concept of transparent clock and clock synchronization based on the AFDX network, and uses Time Triggered (Time Triggered, TT) traffic to transmit key messages in the avionics system. In a TTE network, each node (the node may be a switch or an end system) maintains its own time schedule for TT traffic, and all behaviors are driven by the fixed time schedule. Since each node in a time-triggered system implements their actions according to their local clocks, clock synchronization among distributed nodes is particularly important.

In order to better study the transmission of TT traffic and the transmission of mixed traffic such as TT traffic and RC traffic in clock-synchronized Ethernet, it is necessary to establish a network that provides time synchronization as a platform for the above traffic transmission. Therefore, it is necessary to carry out the process of time synchronization Effective simulation and emulation. At present, the simulation of the time-triggered Ethernet clock synchronization mechanism is mainly implemented by FPGA or ordinary CPU. These simulation methods usually need to design a fixed model and topology structure. The delay value of the node and the online transmission delay time are determined by the hardware itself. , in the actual transmission process, there will be certain conflicts or packet loss phenomena, only the clock synchronization limited by the hardware level can be established, and the simulation process cannot be modified arbitrarily in a short period of time, and the flexibility is poor.

Contents of the invention

The object of the present invention is to propose a simulation system for simulating the clock synchronization of switched Ethernet. The simulation system realizes the simulation of clock synchronization of switched Ethernet by means of discrete event scheduling. In the simulation system of the present invention, the simulation role model configuration is performed according to the configuration information to obtain a synchronous topology; then state role management is performed on each node in the synchronous topology structure, and data processing, clock correction and fault detection are performed with different state roles; The status and clock correction results are thus output. Compared with the existing simulation system, the simulation system of the present invention shows stronger scalability in the aspect of network model establishment, and is convenient for more effective analysis on the clock synchronization process.

The present invention is a simulation system for simulating switched Ethernet clock synchronization. The simulation system includes an Ethernet protocol command database, a timer, a node and link attribute configuration module (10), a synchronization topology generation module (20), and a state management module. module (30), state conversion module (40), data processing module (60), fault detection and processing module (70), clock correction module (80) and result output module (50).

The node and link attribute configuration module (10) configures the required simulation configuration information D ₁₀ according to the end system, the switch and the physical connection between the end system and the switch in the actual Ethernet.

The synchronous topology generation module (20) is used to generate a synchronous topology structure and output synchronous topology information D ₂₀ .

A state management module (30), configured to manage the execution behavior of node roles and node states.

A state conversion module (40), configured to switch the node states, the node states are integration state, asynchronous state, waiting state, FLOOD state, enabling state and synchronous state.

The data processing module (60) completes the generation, transmission, reception, curing and compression of data frames according to the current node status. And in the process of data processing, it acts on the state management module to notify the state management module whether it currently receives the data frame and the type information of the data frame; after the data is solidified or compressed, it acts on the clock correction module, and the solidified or compressed data The frame is transmitted to the clock correction module; in the process of data processing, it acts on the fault detection and processing module, and transmits the information of the data frame to the fault detection and processing module. It is composed of a data frame generation unit (601), a data frame sending unit (602), a data frame receiving unit (603), a data frame curing unit (604), and a data frame compression unit (605).

The fault detection and processing module (70) is used for checking data frames and performing error processing; detecting and processing clumping; detecting and processing simulation operation status. It is composed of a data frame check error processing unit (701), a synchronous clumping detection and processing unit (702), an asynchronous clumping detection and processing unit (703) and a software exception processing unit (704).

A clock correction module (80), used for correcting the local clock of the node role. It is composed of a data frame selection unit (801), a clock correction calculation unit (802) and a clock correction execution unit (803).

The result output module (50), configured to output the result of the current state information, the size of the clock correction value and the current local clock value of the node in an xml format file or an excel format file.

The advantage of the simulation system of the present invention's simulated switched Ethernet clock synchronization is:

① The present invention uses discrete event scheduling to simulate the process of network clock synchronization, and is not limited by the hardware platform. The topology and scale of the simulation model can be changed as needed, which is flexible, convenient, and easy to implement.

②The present invention realizes the function of clock synchronization in a way that is not limited by specific hardware. When calculating the compression correction value, it can use as many curing time points of the data frames received in the curing queue as possible to achieve a higher clock speed. synchronization accuracy.

③ The present invention sets the time accuracy in the simulation process according to user requirements, and can provide a synchronous network platform with different clock granularities for the upper-layer traffic transmission process, which matches the clock of the application, thereby effectively ensuring the time determinism of TT traffic.

④Through the simulation platform of TTE network clock synchronization, the optimal design of the TTE network clock synchronization process can be realized. Through the optimization of the synchronization topology and the configuration of clock synchronization parameters, the situation of clock synchronization in different situations can be obtained, which can be compared and analyzed , so as to modify the configuration and optimize the performance of TTE network clock synchronization.

⑤Through the TTE network clock synchronization simulation, the critical situation that is difficult to obtain in the actual network can be predicted and analyzed, and the critical situation of TTE network clock synchronization can be obtained.

6. The simulation system designed by the present invention can be used to carry out assumptions based on what-if, simulate different error situations of the network, and analyze the fault-tolerant performance of the network.

⑦The simulation system designed by the present invention can be used to construct various network scenarios (that is, different synchronization topology structures), and different parameter configurations can be used to obtain the convergence and accuracy of network synchronization in each scenario, which is convenient for comparison.

Description of drawings

Fig. 1 is a structural block diagram of the simulation system of the present invention.

Fig. 2 is a diagram of any synchronization topology in the present invention.

Fig. 3 is a process diagram of the data processing module in the present invention.

Fig. 3A is a schematic diagram of the PCF format generated by the data frame generating unit in the present invention.

Fig. 3B is a flowchart of the data frame sending unit in the present invention.

Fig. 3C is a flowchart of the data frame receiving unit in the present invention.

Fig. 3D is a flow chart of the data frame curing unit in the present invention.

FIG. 3E is a specific way of curing the data frame in the present invention.

Fig. 3F is a flowchart of the data frame compression unit in the present invention.

FIG. 3G is a timing diagram of calculating the compression time point in the present invention.

Fig. 4 is a flowchart of the data frame selection unit in the present invention.

Fig. 5 is a flow chart of the state transition module performing state transition according to the corresponding conditions delivered by the state management module in the present invention.

detailed description

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

Referring to shown in Fig. 1, the present invention provides a kind of emulation system of analog switched Ethernet clock synchronization process, this emulation system comprises Ethernet protocol command database, timer, node and link attribute configuration module 10, synchronous topology generation module 20. A state management module 30 , a state conversion module 40 , a data processing module 60 , a fault detection and processing module 70 , a clock correction module 80 and a result output module 50 .

In the present invention, the Ethernet protocol instruction database is constructed according to the IEEE802.3ISO/IEC8802-3 standard specification and the instruction and state association information indicated in SAEAS6802 (time-triggered Ethernet). Different instructions in the Ethernet protocol instruction database, state association information can be used by each module (i.e. node and link attribute configuration module 10, synchronous topology generation module 20, state management module 30, state conversion module 40, data processing module 60 , fault detection and processing module 70, and clock correction module 80).

In the present invention, the timer is constructed according to the IEEE802.3ISO/IEC8802-3 standard specification and the time instruction specified in SAEAS6802 (time-triggered Ethernet). In order to facilitate a more effective analysis of the clock synchronization process, the timer is used to customize the time-related information to ensure the call of the time information during the simulation process.

In the present invention, the node and link attribute configuration module 10 configures the required simulation configuration information D ₁₀ according to the end system, the switch and the physical connection between the end system and the switch in the actual Ethernet.

In the present invention, the synchronous topology generation module 20 is used to generate a synchronous topology structure and output synchronous topology information D ₂₀ .

In the present invention, the synchronous topology generation module 20 realizes the function of generating synchronous topology according to the physical topology information read in the configuration information reading module, and initializes and configures different node roles, and configures all nodes as synchronous controller roles; Then configure the role of the switching node. The switching node is configured as a compression controller or a synchronization client. Only one compression controller is configured on the same channel, and other switching nodes on the same channel are configured as synchronization clients. The specific execution method is to configure each node from a certain node device on the channel, configure the node as a synchronous controller, and obtain the corresponding configuration information (port rate, online delay, sending and receiving delay) of the node as a synchronous controller , the internal processing time of the port and the clock drift of the simulated device), and then continue to traverse the physical link to configure the device. When it is a switching node and the first switching node on the channel, configure it as a compression controller and as a The compression controller obtains the corresponding configuration information (port speed, line delay, sending and receiving delay, internal processing time of the port and the clock drift of the simulated device) when it is a switching node. For other switching nodes on the channel, it Configured as a synchronous client, also as a synchronous client to obtain the corresponding configuration information of the node as a switching node (port speed, online delay, sending and receiving delay, internal processing time of the port and clock drift of the simulated device).

In the present invention, the state management module 30 is used to manage the execution behavior of node roles and node states.

In the present invention, the state management module 30 generates a timer corresponding to the state configured by the user according to the timer in different states; it acts on the data processing module to process data frames; it acts on the clock correction module to trigger the node local Clock correction; acting on the fault detection and processing module to detect and process possible faults in the simulation system.

In the present invention, the state conversion module 40 is used to switch the node state, and output the node state to the result output module 50 .

In the present invention, the data processing module 60 completes the generation, transmission, reception, curing and compression of data frames according to the current node status. It is composed of a data frame generating unit 601 , a data frame sending unit 602 , a data frame receiving unit 603 , a data frame curing unit 604 and a data frame compressing unit 605 .

In the present invention, the data processing module 60 realizes the generation, transmission, reception, solidification, and compression operations of the data frame according to the effect on it under the corresponding state of the state management module; and acts on the state management module to notify the state during the data processing process Whether the management module currently receives the data frame and the type information of the data frame; after the data is solidified or compressed, it acts on the clock correction module, and passes the solidified or compressed data frame to the clock correction module; during the data processing process, it acts on the clock correction module The fault detection and processing module transmits the information of the data frame to the fault detection and processing module.

In the present invention, the fault detection and processing module 70 is used for checking data frames and performing error processing; detecting and processing agglomeration; detecting and processing simulated operating conditions. It is composed of a data frame check error processing unit 701 , a synchronous clumping detection and processing unit 702 , an asynchronous clumping detection and processing unit 703 and a software exception processing unit 704 .

In the present invention, the clock correction module 80 is used to correct the local clock of the node role. It consists of a data frame selection unit 801 , a clock correction calculation unit 802 and a clock correction execution unit 803 .

In the present invention, the result output module 50 is configured to output the current node status, the clock correction value and the corrected local clock in an xml format file or an excel format file.

The simulation system designed by the present invention realizes interaction with the user through the node and link attribute configuration module 10 and the result output module 50 .

Ethernet protocol command database

In order to make the designed simulation system of the present invention meet IEEE802.3ISO/IEC8802-3 standard specifications, and also meet the needs of avionics system users, the SAEAS6802 (time Instructions, status and other information pointed out in Trigger Ethernet), these information construct the Ethernet protocol instruction database of the simulation system of the present invention.

In the present invention, the node state threshold Including the threshold value of the synchronous controller SM ^B from the integrated state to the synchronous state Threshold value of compression controller CM ^B from integrated state to synchronous state Threshold for synchronizing client SC ^B from integrated state to synchronous state Threshold value of synchronous controller SM ^B from asynchronous state to synchronous state Threshold value of compression controller CM ^B from asynchronous state to synchronous state Threshold value of synchronous controller SM ^B from synchronous state to asynchronous state Threshold value of compression controller CM ^B from synchronous state to asynchronous state Synchronous controller SM ^B maintains the threshold value of the synchronous state Threshold value of compression controller CM ^B to maintain synchronization state Threshold value for synchronization client SC ^B to maintain synchronization state Expressed in aggregate form as $A_0^{\mathrm{User}}=\left\{\begin{array}{c}{\mathrm{ITS}}_{{\mathrm{SM}}^B}^A,{\mathrm{ITS}}_{{\mathrm{CM}}^B}^A,{\mathrm{ITS}}_{{\mathrm{SC}}^B}^A,\\ {\mathrm{ATS}}_{S{m}^B}^A,{\mathrm{ATS}}_{{\mathrm{CM}}^B}^A,\\ {\mathrm{STA}}_{{\mathrm{SM}}^B}^A,{\mathrm{STA}}_{{\mathrm{CM}}^B}^A,\\ {\mathrm{STS}}_{{\mathrm{SM}}^B}^A,{\mathrm{STS}}_{{\mathrm{CM}}^B}^A,{\mathrm{STS}}_{{\mathrm{SC}}^B}^A\end{array}\right\}.$

In the present invention, the node role B ^User includes a synchronization controller SM ^B , a compression controller CMB , and a synchronization client SC ^B ; it is expressed as B ^User ={SM ^B , CM ^B , SC ^B } in ^a set form.

In the present invention, the data frame transmission role MB ^User includes the synchronization controller SSM ^B of the source node, the compression controller SCM ^B of the source node, the synchronization controller DSM ^B of the destination node, the compression controller DCM ^B of the destination node, the destination node The synchronous client DSC ^B of the node and the synchronous client MSC ^B of the intermediate node are expressed as MB ^User = {SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } in a set form.

In the present invention, the node physical delay time Including distribution delay time TD ^C , sending delay time TS ^C , online transmission delay time TW ^C , receiving delay time TR ^C ; expressed in aggregate form as $C_0^{\mathrm{User}}=\{{\mathrm{TD}}^C,{\mathrm{TS}}^C,{\mathrm{TW}}^C,{\mathrm{TR}}^C\}.$

In the present invention, node status V ^User includes integrated status asynchronous state Data information receiving process status waiting state enable state sync status Expressed in aggregate form as $V^{\mathrm{User}}=\{Z_{\mathrm{IN}}^{\mathrm{User}},Z_{\mathrm{AS}}^{\mathrm{User}},Z_{\mathrm{FL}}^{\mathrm{User}},Z_{\mathrm{WT}}^{\mathrm{User}},Z_{\mathrm{EA}}^{\mathrm{User}},Z_{\mathrm{Sy}}^{\mathrm{User}}\}.$

timer

In the present invention, in order to ensure the time synchronization of the simulation system, the timer is used to customize the time to manage the operation of the state.

Node and link attribute configuration module 10

Referring to shown in Fig. 1, in the present invention, node and link attribute configuration module 10 first aspect transfers in the Ethernet protocol instruction database to configure connection information

In the present invention, according to the end system in the actual Ethernet, the switch and the physical connection between the end system and the switch, configure the configuration connection information required by the simulation system of the present invention That is, refer to the schematic diagram of the topology Ethernet structure shown in FIG. 2 .

The second aspect of node and link attribute configuration module 10 calls in the Ethernet protocol instruction database TD ^C ,TS ^C ,TW ^C ,TR ^C , Perform initial assignment to obtain the simulation configuration information after assignment $P_{10}^{\mathrm{User}},{\mathrm{D.}}_{10}^{\mathrm{User}},{\mathrm{D.}}_{\max -10}^{\mathrm{User}},f_{10}^{\mathrm{User}},{\mathrm{TD}}_{10}^C,{\mathrm{TS}}_{10}^C,{\mathrm{TW}}_{10}^C,{\mathrm{TR}}_{10}^C,{\mathrm{E.}}_{10}^{\mathrm{User}},G_{10}^{\mathrm{User}};$ Expressed in aggregate form as ${\mathrm{D.}}_{\mathrm{ass}}^{10}=\{P_{10}^{\mathrm{User}},{\mathrm{D.}}_{10}^{\mathrm{User}},{\mathrm{D.}}_{\max -10}^{\mathrm{User}},f_{10}^{\mathrm{User}},{\mathrm{TD}}_{10}^C,{\mathrm{TS}}_{10}^C,{\mathrm{TW}}_{10}^C,{\mathrm{TR}}_{10}^C,{\mathrm{E.}}_{10}^{\mathrm{User}},G_{10}^{\mathrm{User}}\};$

In the present invention, the information output by the node and link attribute configuration module 10 to the synchronous topology generation module 20 is denoted as configuration information D ₁₀ for simulation, expressed as

Synchronous topology generation module 20

Referring to Fig. 1, in the present invention, the first aspect of synchronous topology generation module 20 is used to receive ${\mathrm{D.}}_{10}=\{{\mathrm{D.}}_{\mathrm{link}}^{10},{\mathrm{D.}}_{\mathrm{ass}}^{10}\};$

In the second aspect, the synchronous topology generation module 20 calls the node role B ^User and Configure node roles in a synchronized manner A synchronous topology is obtained, as shown in Figure 2;

The synchronization method refers to the synchronization role configuration of the end system and the switch required for the simulation, that is, the end system is configured as the role of the synchronization controller, and the switch is configured as the compression controller or the synchronization client; and then configured according to the node role Combined with the Ethernet structure, a virtual link is generated

In the present invention, if there are multiple switches on the same channel, one of the switches on the same channel is configured as a compression controller, and the rest of the switches on the same channel are configured as synchronization clients.

Referring to shown in Figure 2, in the topological structure diagram of the simulation system in the present invention, the first synchronous controller is marked as 211, the second synchronous controller is marked as 212, the third synchronous controller is marked as 213, and the fourth synchronous controller 214, the fifth synchronization controller 215, the sixth synchronization controller 216, the first synchronization client 221, the second synchronization client 222, and the compression controller 231.

If the synchronization controller (211-216) is used as the source node, the synchronization client (221, 222) is used as the intermediate node, and the compression controller (231) is used as the destination node, the information (protocol control frame) is transmitted. Said sync-compressed protocol control frame has at least transparent clock

If the compression controller (231) is used as the source node, the synchronization clients (221, 222) are used as the intermediate nodes, and the synchronization controllers (211-216) are used as the destination nodes for information transmission. The compressed-synchronized protocol control frame has at least a transparent clock

The third aspect of synchronous topology generation module 20 calls the node state threshold value in the Ethernet protocol instruction database combine Get the state threshold value of each node after assignment

Synchronization topology generation module 20 is based on the fourth aspect ${\mathrm{D.}}_{\mathrm{ass}}^{10}=\{P_{10}^{\mathrm{User}},{\mathrm{D.}}_{10}^{\mathrm{User}},{\mathrm{D.}}_{\max -10}^{\mathrm{User}},f_{10}^{\mathrm{User}},{\mathrm{TD}}_{10}^C,{\mathrm{TS}}_{10}^C,{\mathrm{TW}}_{10}^C,{\mathrm{TR}}_{10}^C,{\mathrm{E.}}_{10}^{\mathrm{User}},G_{10}^{\mathrm{User}}\}$ middle combined with the computational delay of the compression controller Generate timing duration TO and window size W _b ;

In the present invention, the timing duration TO includes all the timing names recorded in the timer, and is expressed as TO={T _A , T _B , T _C , T _D , T _E , T _F , T _G , T _H , T _I , T _J }.

In the present invention, the synchronous topology information output by the synchronous topology generation module 20 is denoted as D ₂₀ , expressed as ${\mathrm{D.}}_{20}=\{{\mathrm{D.}}_{\mathrm{VL}}^{20},{\mathrm{D.}}_{\mathrm{queue}}^{20},{\mathrm{D.}}_{\mathrm{state}}^{20},\mathrm{TO},W_b\}.$

In the present invention, according to the configuration information of the node and link attribute configuration module 10, the synchronous topology generation module 20 generates a synchronous topology structure, and the startup time and physical parameters of each node in the synchronous topology structure are assigned according to the read configuration information, The physical parameters include the port rate, the online delay of the connected channel, the sending and receiving delay, the internal processing time of the port, and the clock drift of the simulated node. At the same time, a certain number of data queues are initialized for each node to facilitate Subsequent operations, the data queues maintained in the nodes include data sending queues, data receiving queues, VL receiving queues, data curing queues, data compression queues (for nodes configured as compression controllers), and data collection queues. During the initialization process, the queue is empty, and it will be left in the subsequent process to perform specific operations on each queue. The queue adopts a first-in-first-out model. Each element in the queue includes the encapsulated data frame, the current time value of entering the queue, and the data queue belongs to. The type and name of the node. For each channel containing multiple synchronous controllers, it is necessary to produce twice as many VLs (Virtual Link, virtual link).

State Management Module 30

Referring to Fig. 1, Fig. 3, Fig. 3A ~ Fig. 3G, the state management module 30 acts on the timer in different states; the state management module 30 is based on the node role (generated by the synchronous topology generation module 20) and the node state (state transition module 40) acts on the data processing module 60 to process data frames; acts on the clock correction module 80 to trigger the correction of the local clock of the node; acts on the fault detection and processing module 70 to detect and process possible faults in the simulation system.

In the present invention, the management of state by state management module 30 is:

In the first aspect, according to the information in the D ₂₀ of the synchronous topology generation module 20, the state transition module 40, the data processing module 60, the fault detection and processing module 70 and the clock correction module 80, comprehensively analyze and generate the nodes of the state transition module 40 The state matches the execution behavior.

In the second aspect, transferring the execution behavior obtained by the first aspect to the matching execution behavior refers to that when the node state transferred by the state transition module 40 is the integrated state: the execution behavior of the synchronization controller SM ^B in the integrated state Execution Behavior of Compression Controller CM ^B in Integrated State Execution behavior of synchronization client SC ^B in integrated state

When the node state delivered by the state transition module 40 is a synchronous state: the execution behavior of the synchronous controller SM ^B in the synchronous state Execution Behavior of Compression Controller CM ^B in Synchronous State Execution behavior of sync client SC ^B in sync state

When the node state delivered by the state conversion module 40 is an asynchronous state: the execution behavior of the synchronous controller SM ^B in the asynchronous state Execution Behavior of Compression Controller CM ^B in Asynchronous State

When the node state transferred by the state transition module 40 is the waiting state: the execution behavior of the synchronization controller SM ^B in the waiting state Execution Behavior of Compression Controller CM ^B in Waiting State

When the node state delivered by the state transition module 40 is the FLOOD state: the execution behavior of the synchronization controller SM ^B in the FLOOD state When the node state transferred by the state transition module 40 is the enabled state: the execution behavior of the compression controller CM ^B in the enabled state Expressed in aggregate form as ${\mathrm{D.}}_{\mathrm{ACT}}^{30}=\left\{\begin{array}{c}{\mathrm{ACT}}_{{\mathrm{SM}}^B}^{\mathrm{IN}},{\mathrm{ACT}}_{{\mathrm{CM}}^B}^{\mathrm{IN}},{\mathrm{ACT}}_{{\mathrm{SC}}^B}^{\mathrm{IN}},\\ {\mathrm{ACT}}_{{\mathrm{SM}}^B}^{\mathrm{Sy}},{\mathrm{ACT}}_{{\mathrm{CM}}^B}^{\mathrm{Sy}},{\mathrm{ACT}}_{{\mathrm{SC}}^B}^{\mathrm{Sy}},\\ {\mathrm{ACT}}_{{\mathrm{SM}}^B}^{\mathrm{AS}},{\mathrm{ACT}}_{{\mathrm{CM}}^B}^{\mathrm{AS}},\\ {\mathrm{ACT}}_{{\mathrm{SM}}^B}^{\mathrm{WT}},{\mathrm{ACT}}_{{\mathrm{CM}}^B}^{\mathrm{WT}},\\ {\mathrm{ACT}}_{{\mathrm{SM}}^B}^{\mathrm{FL}},{\mathrm{ACT}}_{{\mathrm{CM}}^B}^{\mathrm{EA}}\end{array}\right\}.$

Specifically, according to the different node role configurations, the sets are introduced in turn The specific content of each element in .

First, for a node whose role is configured as a sync controller:

(A) The execution behavior of the synchronization controller SM ^B in the integrated state Means: in the timing TA of _A , the state management module 30 judges the elements in the data curing queue Q _perm delivered by the data processing module 60, judges whether there are data elements therein, and if so, further judges the received data frame FR type: if it is a CA frame, trigger the state transition module 40 to transition the node state to a waiting state, and trigger the synchronous cluster detection and processing unit 702 to set M _sync to 1. If it is an IN frame, continue to judge whether the synchronization member variable reaches the state threshold If reached, the trigger state conversion module 50 converts the node state into a synchronous state, and starts to trigger the data processing module 60 to record the above-mentioned IN frame and utilize the data processing module 60 to perform processing on the IN frame; if not reached, the trigger state The transition module 50 transitions the node state into a waiting state. If after _A timing _TA ends, the data curing queue Q _perm is still empty, then the trigger state transition module 40 converts the node state to an asynchronous state, sets the timing duration to TB, and triggers the data frame generation in the data processing module 60 Unit 601 starts to generate a CS frame, and then operates in sequence according to each unit in the data processing module 60, and initiates a handshake process.

(B) The execution behavior of the synchronous controller SM ^B in the asynchronous state Refer to: in _B timing TB, judge the element in the data curing queue Q _perm that data processing module 60 transmits, judge whether there is data element wherein, if have then further judge the data frame type that receives: if be CS (ColdStart Frame) frame, then the trigger state conversion module 40 is converted to the FLOOD state by the node state; M _sync is set to 1; if it is an IN frame, continue to judge whether the synchronization member variable has reached If the node synchronization member variable reaches Then the trigger state transition module 40 converts the node state into a synchronous state, and starts to trigger the data processing module 60 to record it and execute the processing to the IN frame in the data processing module 60; if not reached, then the trigger state transition module 40 converts the node state The state transitions to a wait state. If after _B timing TB ends, the data curing queue Q _perm of the node is still empty, then the trigger state conversion module 40 converts the node state to an asynchronous state, at this time, triggers the data frame generation unit 601 in the data processing module 60, Start to generate the CS frame, and then operate according to the order of each unit in the data processing module 60, and initiate a handshake process.

(C) The execution behavior of the synchronization controller SM ^B in the FLOOD state It means that the state management module 30 sets the timing duration according to the current variable configuration information of the node.

The variable configuration information includes flood_recv and closed_window (these two values are bool variables, which are set to 0 when initialized).

In the timing T _C of C, judge the elements added in the data curing queue Q _perm delivered by the data processing module 60, judge whether there are data elements in it, and if there is, then further judge the received PCF frame type: if it is a CS frame , then the trigger state transition module 40 sets the node state to the FLOOD state again;

If it is a CA frame, and in the case of a configuration in which the value of flood_recv is 1 and the value of closed_window is 0, the state transition module 40 is triggered to transform the node state into a waiting state.

After C timing T _C ends, the data curing queue Q _perm of the node is still empty, then judge the flood_recv value and closed_window value, if both are 0, then trigger the state transition module 40 and set the node state to the FLOOD state again, and trigger The data processing module 60 sends the CA frame, and at the same time, assigns flood_recv to 1;

If the flood_recv value is 1, and the closed_window value is 0, then within _I timing T1, continue to judge the elements in the data curing queue Q _perm delivered by the data processing module 60, judge whether there is a data element in it, if not, then The trigger state transition module 40 is set to the FLOOD state again with the node state, and does not carry out state transition;

If both the flood_recv value and the closed_window value are 1, then within J timing T _J , continue to judge the elements in the data curing queue Q _perm delivered by the data processing module 60, and judge whether there is a data element in it, if not, then trigger The state conversion module 40 converts the node state into an asynchronous state, and assigns the flood_recv value and closed_window value to 0.

(D) The execution behavior of the synchronization controller SM ^B in the waiting state Refer to: in _E timing TE, judge the element in the data solidification queue Q _perm that data processing module 60 transmits, judge whether there is data element wherein, if have, then further judge the data frame type that receives: if be If it is a CS frame, the trigger state transition module 40 will convert the node state to the FLOOD state; if it is a CA frame, then the trigger state transition module 40 will continue to stay in the waiting state with the node state, and trigger the synchronous clumping detection and processing unit 702 to M _sync Set to 1; if the data receiving queue Q _recv is empty, and the local member value of the node reaches the state threshold Then the trigger state conversion module 40 converts the node state into a synchronous state, and simultaneously triggers the data frame generation unit 601 in the data processing module 60 to start generating IN frames (new members in the frame=local member value of the node) to initiate the synchronization process; If the node's local membership value does not reach the state threshold Then the state conversion module 40 is triggered to convert the node state into an asynchronous state.

(E) The execution behavior of the synchronization controller SM ^B in the synchronization state It means: in this state, the cold start handshake has been successful, and the clock difference of each node is kept within the range of the set clock accuracy. The state management module 30 triggers the data processing module 60 at the starting point of each integration cycle to generate an IN frame, and performs operations such as sending, receiving, and curing it; the trigger clock correction module 80 performs clock correction after receiving the solidified IN frame; Trigger the fault detection and processing module 70 to monitor the running process, utilize the synchronous clumping detection and processing unit 702 and the asynchronous clumping detection and processing unit 703 in the fault detection and processing module 70 to check whether there is clumping in the synchronous process Detect and deal with it. In the synchronous state, the data processing module 60 feeds back the elements added in the data solidification queue Q _perm delivered to the state management module 30 for judgment, if it is an IN frame, then continue to remain in the synchronous state, and continue subsequent operations; if it is a CA frame, It indicates that there is currently an out-of-sync situation, and the trigger state transition module 40 switches the node state into a waiting state, and starts to wait for re-initiation of the synchronization process. Fault detection and processing module 70 reports the detected problem to state management module 40, when synchronous clumping detection and processing unit 702 or asynchronous clumping detection and processing unit 703 detects the situation of clumping, the situation that occurs will be reported to The state management module 30, the state management module 30 triggers the state transition module 40 to transform the node state into an asynchronous state. The clustering situation mentioned is specified in SAEAS6802 (Time-Triggered Ethernet).

For a node configured as a compression controller: discuss according to different state information transmitted by the state transition module 40 .

(A) Execution behavior of the compression controller CM ^B in the integrated state Refer to: in D timing T _D , judge the elements in the data curing queue Q _perm delivered by data processing module 60, judge whether there are data elements in it, if there is, then further judge the received data frame type: if it is CS Or CA frame, it is regarded as the situation that should not occur, and the data frame is directly discarded; if it is an IN frame, continue to judge whether the synchronization member variable reaches the state threshold If the synchronization member variable reaches the state threshold Then the trigger state transition module 40 converts the node state into a synchronous state, and starts to trigger the data processing module 60 to record it and utilize the data processing module 60 and the clock correction module 80 to perform corresponding operations; if not reached, the data frame is discarded . If the data solidification queue Q _perm of the node is still empty after the D timing T _D ends, the trigger state conversion module 40 is entered to convert the node state to an asynchronous state.

(B) The execution behavior of the compression controller CM ^B in the asynchronous state Refers to: judge the elements in the data solidification queue Q _perm of the node delivered by the data processing module 60, judge whether there are data elements in it, and if there is, then further judge the received data frame type: if it is a CS frame, or it is a CA frame and the member value information of the received CA frame is greater than the status threshold Then the trigger state transition module 40 converts the node state to the enable state, and the trigger data processing module 60 uses the data frame generating unit 601 to generate a CS or CA frame, and generates and sends the data frame according to the steps in the data processing module 60 . If it is an IN frame, and the value of the new member of the IN frame is greater than the status threshold Then the state conversion module 40 is triggered to convert the node state into a synchronous state, and the data processing module 60 is triggered for subsequent processing. If the data curing queue Q _perm is always empty, the node stays in the asynchronous state.

The steps in the described data processing module 60 include: because the generated data frame is basically consistent with the received data frame format, therefore, when generating a new data frame, take the method of changing the received PCF internal structure , modify the transparent clock field of the frame to be 0, and add new VL (swapping its source address and destination address) information during data processing, and send it back to the corresponding destination node according to the subsequent steps in the data processing module 60 .

(C) Execution behavior of the compression controller CM ^B in the enabled state Refer to: in the timing TF of _F , judge the elements in the data curing queue Q _perm delivered by the data processing module 60, judge whether there are data elements in it, if there is, then further judge the received data frame type: if it is CS Or IN frame, then regard as the situation of error, directly discard data frame, do not operate; If receive CA frame, then will carry out forwarding operation to this CA frame, trigger data processing module 60 to generate CA frame, and according to data The steps in processing module 60 transmit the data frame. In this state, each time a CA frame is received, the forwarding behavior is performed. Since the format of the generated data frame is basically the same as that of the received data frame, when generating a new data frame, the received PCF content is taken The method for making changes is to modify the transparent clock domain of the frame, and add a new VL (exchange its source end address and destination end address) during data processing, and send it back according to the subsequent steps in the data processing module 60 For the corresponding destination node, in this case (that is, the case of receiving a data frame), the node continues to stay in the current state and does not perform state transition. If the node data receiving queue Q _recv is still empty after the _F timing TF ends, that is, no new data is received, the trigger state transition module 40 will switch the node state to a waiting state.

(D) The execution behavior of the compression controller CM ^B in the wait state Refer to: in G timing T _G , judge the elements in the data curing queue Q _perm delivered by the data processing module 60, judge whether there are data elements in it, if there is, then further judge the received data frame type: if it is CS Or a CA frame, it is regarded as an error, and the data frame is discarded directly, and no operation is performed; if it is an IN frame, and the value status threshold of the new member of the IN frame Then the trigger state conversion module 40 converts the node state into a synchronous state, and triggers the data processing module 60 and the clock correction module 80 to operate on the received data frame. If the data solidification queue Q _perm of the node is always empty after the G timer T _G ends, that is, no new data is received, the trigger state conversion module 40 will convert the node state to an asynchronous state.

(E) The execution behavior of the compression controller CM ^B in the synchronous state It means: when in this state, it means that the cold start handshake is successful, and the clock difference of each node is kept within the range of the set clock accuracy. State management module 30 carries out corresponding operation according to the data frame fed back by data processing module 60: if receive the CS or CA frame after solidification, then regard as the situation of error, directly discard data frame, do not operate; The cured IN frame then triggers the data frame compression unit 605 in the data processing module 60 to collect and compress the IN frame, and utilizes the clock correction module 80 to perform clock correction, utilizes the synchronization in the fault detection and processing module 70 The clumping detection and processing unit 702 and the asynchronous clumping detection and processing unit 703 detect whether clumping occurs during the synchronization process, and make processing. Fault detection and processing module 70 reports the detected problem to state management module 30, when synchronous clumping detection and processing unit 702 or asynchronous clumping detection and processing unit 703 detects the situation of clumping, the situation that occurs will be reported to State management module 30, state management module 30 triggers state transition module 40 to convert current state into integrated state, thus waits again to start the process of handshake synchronization; The local configuration information of the node is updated correspondingly when data is processed in the data processing module 60 .

For a node configured as a synchronous client: discuss according to different state information delivered by the state transition module 40 .

(A) The execution behavior of the synchronization client SC ^B in the integrated state Refer to: first judge the elements in the data curing queue Q _perm delivered by the data processing module 60, judge whether there are data elements in it, if there is, then further judge the received data frame type: if it is an IN frame, then further judge its Whether the synchronization member variable reaches the state threshold If yes, then the trigger state conversion module 40 will convert the node state into a synchronous state, and trigger the data processing module 60, process it according to the method when receiving the IN frame in the data processing module 60, and perform certain local information on the node update; if not, the state threshold has not been reached It means that the state of synchronization has not yet been reached, and the data processing module 60 is continuously triggered to perform corresponding data forwarding operations. If the received PCF type is a CS or CA frame, the data processing module 60 is also continuously triggered to perform corresponding data forwarding operations.

(B) The execution behavior of the synchronization client SC ^B in the synchronization state Refers to: the state management module 30 triggers the data processing module 60, the clock correction module 80, etc. to process the received IN frame, and corrects the current local clock of the node. If it occurs, the state transition module 40 is triggered to convert the node state to an integrated state; otherwise, it remains in the synchronous state. At the same time, the current configuration information of the node is modified accordingly.

State Transition Module 40

As shown in FIG. 5, the state transition module 40 experiences a total of 6 node states during the synchronization establishment process, and the node states are respectively integration states asynchronous state waiting state FLOOD state enable state (not shown in Figure 5) and the synchronous state The first five states are all asynchronous states. When in the above state, the state transition module 40 starts to function, completes the state transition, and waits for the next command to trigger the state transition. The information about the judgment conditions shown in the figure is transmitted from the state management module 30 to the state transition module 40 .

Referring to shown in Fig. 5, in the present invention, synchronous controller SM ^B enters integrated state at first after node startup, judges whether _A receives data frame in timing TA, if receive, then continue to judge its type; If not When received, it enters the asynchronous state. If the received data frame type is an IN frame, continue to judge whether the local member value of the node reaches the state threshold If it is not an IN frame, then judge whether it is a CA frame. For the above case where the received PCF is an IN frame, if the local member value of the node reaches the status threshold Then enter the synchronous state; if not, enter the waiting state. For the above-mentioned situation that the received data frame is not an IN frame, if the received PCF is a CA frame, it enters a waiting state; if not, it enters an asynchronous state.

For the state transition in the asynchronous state: it is also judged whether the data frame is received at the timing TB of _B , and if it is received, the type of the data frame received is further judged; if it is not received, it is still in the asynchronous state. If it is a CA frame, it enters a waiting state; if not, it further judges whether it is a CS frame. If it is a CS frame, enter the FLOOD state; if not, it is an IN frame, then enter the synchronization state.

For the state transition of the waiting state: it is also judged whether _E receives the data frame within the timing TE, if it is received, it still stays in the waiting state, if not, then judges whether the local member value of the node reaches the state threshold If the node's local membership value reaches the status threshold Then enter the synchronization state; if the state threshold is not reached Then enter the asynchronous state.

For the state transition in the FLOOD state: judge whether to receive PCF equally in the timing duration in state management module 30, if receive, then continue to judge the type of the PCF that receives; If not received, then judge whether to satisfy configuration Info conditions (configuration about flood_recv and closed_window values). For the situation of receiving the PCF within the above specified time, continue to judge whether the received PCF is a CS frame; if yes, continue to maintain the FLOOD state; if not, then judge whether it is a CA frame. If it is a CA frame, it enters the waiting state; if not, it judges whether the configuration information condition is satisfied (the configuration of the flood_recv and closed_window values). For the above-mentioned situation that PCF is not received within the timing duration in the state management module 30, if the configuration information condition is satisfied, it will enter the asynchronous state; if not, it will still stay in the FLOOD state. The timing duration and configuration information condition are based on the FLOOD state in the state management module 30 depends on.

For the state transition in the synchronous state: judge whether to detect whether the agglomeration situation and the data receiving queue Q _recv receive the CA frame, if the detection has the agglomeration situation, then enter the asynchronous state; if not, then continue to remain in sync status. If a CA frame is received, it enters a waiting state; if not, it remains in a synchronous state. The clustering situation mentioned is specified in SAEAS6802 (Time-Triggered Ethernet).

In the present invention, after the node is started, the compression controller CM ^B first enters the integration state, and judges whether the size of the new member received in the state management module 30 reaches the state threshold value in the time duration D and time T _D The IN frame, if yes, enters the synchronous state; otherwise, enters the asynchronous state.

For a state transition in an asynchronous state: first determine whether a CS frame is received or the value of the new member is greater than the state threshold CA frame, if it is, enter the enable state; if not, continue to judge whether the size of the received new member reaches the state threshold IN frame. If the size of the new member received reaches the status threshold If the IN frame is not received, it will enter the synchronous state; if it is not received, it will still stay in the asynchronous state.

For the state transition in the enabling state: judge whether the data frame is received within the timing TF of _F , if yes, then stay in the enabling state; if not, then enter the waiting state.

For the state transition in the waiting state: judge whether the size of the new member received within G timing T _G reaches the state threshold The IN frame, if it is, enters the synchronous state; if not, then enters the asynchronous state.

For the state transition in the synchronous state: check whether there is a clumping situation, if there is a clumping situation, enter the integrated state; if not, then stay in the synchronous state.

In the present invention, the synchronization client SC ^B first enters the integrated state after the node starts, and judges whether the size of the received new member reaches the state threshold IN frame, if it is, the conversion enters the synchronous state; otherwise, it remains in the integrated state.

For the state transition in the synchronous state: check whether there is a clumping situation, if there is a clumping situation, enter the integrated state; if not, then stay in the synchronous state.

Data processing module 60

Referring to FIG. 3 , the data processing module 60 mainly includes a data frame generating unit 601 , a data frame sending unit 602 , a data frame receiving unit 603 , a data frame curing unit 604 and a data frame compressing unit 605 .

After data processing module 60 receives data frame, according to node state $V^{\mathrm{User}}=\{Z_{\mathrm{IN}}^{\mathrm{User}},Z_{\mathrm{AS}}^{\mathrm{User}},Z_{\mathrm{FL}}^{\mathrm{User}},Z_{\mathrm{WT}}^{\mathrm{User}},Z_{\mathrm{EA}}^{\mathrm{User}},Z_{\mathrm{Sy}}^{\mathrm{User}}\}$ Any one of the states and call the data frame transmission role MB ^User in the Ethernet protocol instruction database = {SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } any node role to process the data frame, Realize the generation, transmission, reception, curing and compression of data frames.

The data processing module 60 is configured according to the node role To build a data processing queue said Including data sending queue Q _send , data receiving queue Q _recv , VL receiving queue Q _VLrecv , data curing queue Q _perm , data compression queue Q _comp , and data collection queue Q _coll ; it is expressed as ${\mathrm{D.}}_{\mathrm{queue}}^{20}=\{Q_{\mathrm{send}},Q_{\mathrm{recv}},Q_{\mathrm{VLrecv}},Q_{\mathrm{perm}},Q_{\mathrm{comp}},Q_{\mathrm{coll}}\}.$

In the present invention, the data compression queue Q _comp is only used in the compression controller.

Data frame generating unit 601

In the first aspect, the data frame generation unit 601 dynamically generates the protocol control frame FR according to the SSM ^B and SCM ^B in the data frame transmission role MB ^User ={SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B };

The protocol control frame FR contains the address of the source node, the address of the destination node, and the VL (Virtual Link, virtual link) information to be experienced by its transmission

The type of the FR satisfies the node state $V^{\mathrm{User}}=\{Z_{\mathrm{IN}}^{\mathrm{User}},Z_{\mathrm{AS}}^{\mathrm{User}},Z_{\mathrm{FL}}^{\mathrm{User}},Z_{\mathrm{WT}}^{\mathrm{User}},Z_{\mathrm{EA}}^{\mathrm{User}},Z_{\mathrm{Sy}}^{\mathrm{User}}\}$ The provisions of the different states in .

Referring to FIG. 3A , the second aspect dynamically generates a protocol control frame PCF according to state information and state behavior information. The generated PCF is divided into three types: cold start (CS, ColdStart) frame, cold start confirmation (CA, ColdStartAcknowledge) frame and integration (IN, Integrate) frame.

The PCF is a standard Ethernet frame with a minimum payload (46 bytes), and its Ethernet type is 0x891d. In the data frame generating unit 601, first generate 28 bytes of PCF suitable for clock synchronization, and reserve 18 bytes for distribution later. The structure diagram of PCF is shown in Fig. 3A;

Specifically, in the PCF frame structure,

Integration period I _pcf = integration period value for node i

The new member is a 32bit variable, xxxx xxxx xxxx xxxx xxxx xxxx xxxx xxxx; the new member value of the generated PCF S _pcf = the local member value of the sending node This new subscript should change

The maximum number of synchronous controllers that can be carried in the simulated switched Ethernet is 32. At initialization, the local member value of the node configured as a synchronous controller is the same as the serial number of its own synchronous controller. The third Take the initial member value of the synchronization controller as an example. Initially, the local member value of this node is 0000 0000 0000 0000 00000000 0000 0100. Therefore, the new member in the PCF sent for the first time = 0000 0000 0000 0000 0000 00000000 0100;

Synchronization priority = the synchronization priority of the sending node;

Synchronization domain = the synchronization domain where the sending node is located;

The transparent clock domain part is mainly used to record the time elapsed in each process from the time when the PCF is generated. The transparent clock when the PCF enters the current processing unit is denoted as The transparent clock when the PCF leaves the current processing unit is denoted as If the node generating the data frame is the source node, the transparent clock value of the PCF in the data frame generating unit 601 ${\mathrm{TC}}_{\mathrm{out}}^{601}=0.$

Data frame sending unit 602

In the first aspect, the data frame sending unit 602 forwards the data frame received by the data frame receiving unit 603 according to MSC ^B in the data frame transmission role MB ^User ={SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } ;then execute the fourth and fifth aspects;

The second aspect of the data frame sending unit 602 is to receive the output of the data frame generating unit 601 according to the SSM ^B and SCM ^B in the data frame transmission role MB ^User ={SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } the FR;

In the third aspect, record the sending time point of FR In the fourth aspect, add FR to the data sending queue Q _send of the node. Among them, FR enters and exits the team in a first-in first-out manner.

In the fifth aspect, record the data distribution delay time TD ^C and the transmission delay time TS ^C , and change the transparent clock in FR value. Then, according to the data distribution delay time TD ^C , the transmission delay time TS ^C and the formula Obtain the transparent clock leaving the data frame sending unit 602

Data frame receiving unit 603

FIG. 3B is a flowchart of the data frame receiving unit 603 . The data frame receiving unit 603 is used to receive the data frame FR sent by the data frame sending unit 602, and store the received data frame into the data receiving queue Q _VLrecv ;

In the first aspect, it is judged whether the VL receiving queue Q _VLrecv of the node is empty, if it is not empty, it means that the data frame has been received, if it is empty, then return to the initial state, and wait for the arrival of the data frame;

In the second aspect, for the non-empty situation of the above-mentioned VL receiving queue Q _VLrecv , first judge whether the received data frame is valid, if yes, then perform information extraction processing on the data frame; if not, trigger the fault detection and processing module 70 for discarding.

The information extraction process is to extract the source node and destination node address information, record the receiving time point T _r , record the online transmission delay time TW ^C between nodes in the process of sending from the previous node to the current node, and receive the delay time TR ^C , changing the transparent clock value

The method of changing the transparent clock value is based on the online transmission delay time TW ^C , the reception delay time TR ^C and the formula Get the transparent clock value leaving the data frame receiving unit 603

The third aspect performs the following operations. Based on the source node and destination node address information extracted in the second aspect, and the DSM ^B and DCM ^B in the data frame transmission role MB ^User ={SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } Whether the current node is the destination node consistent with the address information of the destination node, if yes, it means that the data frame transmission has been completed, and the data frame FR and the receiving time point T _r of the received FR are stored in the data receiving queue Q _recv of the node , trigger the data frame solidification unit 604, and start its solidification operation; if it is not the destination node, then pass the data frame FR to the data frame sending unit 602 of the current node, continue the data sending process, and send the data frame FR to the destination node.

Data frame curing unit 604

FIG. 3C is a flowchart of the data frame curing unit 604 . The data frame curing unit 604 implements the function of rearranging the data frames FR received by the destination node. Through the data frame curing unit 604, the destination node can obtain a group of data frames arranged in the order of sending time of the source node.

In the first aspect, the data receiving queue Q _recv of the node is judged to determine whether the node has received the data frame. If it is received, that is, the data receiving queue Q _recv is not empty, the time point T _r of the received data frame is started. Processing, take out the head member (including time and data frame information) in the data receiving queue, and perform curing operation on it to obtain the curing time point T _p ; if it is empty, return to the initial state and continue to wait for the data receiving queue Q _recv is not empty.

In the second aspect, for the case where the data receiving queue Q _recv is not empty, first judge whether it is an IN frame according to the frame type part in the data frame FR, and if so, the cured IN frame and its curing time point T _p Store the data into the node's data curing queue Q _perm in the order of curing time points; if not, it will not be processed.

In the third aspect, when the type of the data frame in the data receiving queue Q _recv is an IN frame, then select according to the role type B ^User of the node, if the node role is a compression controller, then enter the data frame compression unit 605, and start data The process of the compression operation; otherwise, that is, the role of the node is a synchronization controller or a synchronization client, then enter the clock correction module 80, and start to correct the local clock of the node.

The manner of the curing operation is shown in Figure 3E. First, according to the maximum delay of data transmission And the value TC _n of the transparent clock of the data frame FR (assuming that the destination node is the nth node, TC _n is the total time experienced by the data frame FR from the source node to the destination node) to calculate the curing delay time. The curing delay time is

Then, the formula can be used Calculate the curing time point T _p . Then, continue to _judge whether the cured data frame type is an IN frame according to the frame type information in the data frame FR structure. The order of the time point T _p (the smaller the time, the front of the queue) is inserted into the data curing queue Q _perm of the node, which is convenient for the next operation. Then judge whether the destination node role configuration is a compression controller, if yes, then enter the data frame compression unit 605, and start to perform subsequent operations; if not, then explain that the configuration is a synchronization controller or a synchronization client, then enter the clock correction module 80, The data frames in the curing queue can be used for the correction operation of the local clock.

Data frame compression unit 605

FIG. 3D is a flowchart of the data frame compression unit 605 . The data frame compression unit 605 mainly realizes the function of compressing the curing time point of the data frame, and is a very important unit in establishing a global time reference for the entire network.

In the first aspect, the data frame compression unit 605 performs the collection operation according to the cured data frame FR obtained by the data frame curing unit 604 . The method of the collection operation is in a not greater than (Where, OW, observedwindow, is the observation window, by the formula Calculated) time window to collect data, each node maintains A counter is used to record the number of data frames collected in each OW, and the specific process is as shown in Figure 3D (in the figure, with example).

Start the data frame compression unit 605 according to the curing time point T _p-0 contained in the first element Q _perm [0] in the data curing queue Q _perm obtained by the data frame curing unit 604. This earliest curing time point is the data The starting point of the time window opened by the frame compression unit 605 is recorded as T _start (where T _start =T _p-0 ), and it is stored in the compression queue Q _comp of the node, and the value of the counter 1 is set to 1, Then, judge whether the data curing queue Q _perm is empty (that is, whether there are data frames to be taken out), if it is empty, judge whether the current time has exceeded T _start + 3×OW, and if it has exceeded, do not continue to collect data frame, directly enter the data calculation stage, and calculate the compression time point; if it is not empty, continue to take out the head element of the current solidification queue, obtain its solidification time point T _pi , and judge the obtained time value to determine whether it satisfies T _pi ≤ T _start +OW. If it satisfies T _pi ≤ T _start +OW, it means that the data frame meets the requirements, store it in the data compression queue Q _comp , and continue to judge whether there is a data frame in the data curing queue Q _perm ; if it does not satisfy T _pi ≤ T _start +OW, then judge whether the value of counter 1 is greater than 1 at this time. If the value of counter 1 is not greater than 1, it means that the collection phase is over, the currently received data frame is an incorrect data frame, discard it, and enter the data calculation phase, using the solidification time of the data frame in the data compression queue Q _comp point to calculate the compression time point; if the value of counter 1 is greater than 1, open the second receiving window to collect the data frames whose curing time point is within the range of (T _start +OW,T _start +2×OW]. If curing When the time point is within the range of (T _start +OW,T _start +2×OW], use counter 2 to record the number of data frames within this range at the time point of solidification, and store it in the data compression queue Q _comp , continue Judge whether there is a data frame in the data curing queue Q _perm ; if it is not within this range, then judge whether the value of counter 2 is greater than or equal to 1 at this time. If the value of counter 2 is less than 1, it means that the collection phase is over and the current received The data frame is the wrong data frame, discard it, and enter the data calculation stage, and use the solidification time point of the data frame in the data compression queue Q _comp to calculate the compression time point; if the value of the counter 2 is greater than or equal to 1 at this time, then Open the third receiving window to collect data frames whose curing time point is in the range of (T _start +2×OW,T _start +3×OW]. If the curing time point is in (T _start +2×OW,T _start +3×OW], store it in the data compression queue Q _comp , and continue to judge whether there is a data frame in the data curing queue Q _perm ; if it is not within this range, it means that the wrong data frame is received , discard it, and enter the data calculation stage, and use the curing time point of the data frame in the data compression queue Q _comp to calculate the compression time point. Repeat the above process until the time exceeds (In this example, the For example, until the time exceeds (T _start +3×OW)). In the above process, the data frames collected in the time window will be stored in the data compression queue Q _comp of the node in the order of curing time points, and each time a data frame is put into the data compression queue Q _comp , the node will be corrected once The local member variable value of = the current local member variable value + new member (PCF internal parameter), so as to obtain the new local member variable value.

In the second aspect, when the collection process is over, it begins to enter the calculation phase. Calculate the compression correction value based on the curing time point T _pi collected during the collection process (T _pi represents the value of the i-th (i≥0) curing time point)

The calculation method of the compression correction value is to calculate according to the following calculation formula:

For the case of only one input, the compression correction value

Two inputs: ${\mathrm{D.}}_{\mathrm{comp}\_\mathrm{corr}}^{605}=(T_{p-0}+T_{p-1})/2-T_{p-0}$

Three inputs: ${\mathrm{D.}}_{\mathrm{comp}\_\mathrm{corr}}^{605}=T_{p-1}-T_{p-0}$

Four inputs: ${\mathrm{D.}}_{\mathrm{comp}\_\mathrm{corr}}^{605}=(T_{p-1}+T_{p-2})/2-T_{p-0}$

Five inputs: ${\mathrm{D.}}_{\mathrm{comp}\_\mathrm{corr}}^{605}=(T_{p-1}+T_{p-2}+T_{p-3})/2-T_{p-0}$

More than five inputs: ${\mathrm{D.}}_{\mathrm{comp}\_\mathrm{corr}}^{605}=\underset{k=1}{\overset{\mathrm{no}-2}{\mathrm{\Σ}}}{T}_{p-k}/(\mathrm{no}-2)-T_{p-0}$ ( $f_0^{\mathrm{User}}=2$ in the case of)

In the third aspect, the compression correction value calculated according to the second aspect Calculate the compression time point

As shown in Figure 3G, when all available data frames are collected, the delay of the compression process is calculated according to the calculated compression correction value ${\mathrm{D.}}_{\mathrm{comp}\_\mathrm{func}}^{605}=(f_0^{\mathrm{User}}+1)\×\mathrm{ow}+{\mathrm{BD}}_0^{\mathrm{User}}+{\mathrm{D.}}_{\mathrm{comp}\_\mathrm{corr}}^{605},$ Further, the compression time point can be calculated Then, the data frame compression unit 605 will calculate the compression time point The data is transmitted to the clock correction module 80, and the clock correction module 80 is triggered to perform subsequent processing related to correcting the local clock.

Fault detection and processing module 70

The fault detection and processing module 70 implements the functions of verifying data frames and performing error processing, detecting and processing clumping, detecting the running status of the simulation system software, and processing errors. The fault detection and processing module 70 detects the simulation running process and performs processing corresponding to the error type according to different error types, and transmits the error situation that occurs to the state management module 40, the data processing module 60, and the clock correction module 80 , the result output module 90 . The clustering situation mentioned is specified in SAEAS6802 (Time-Triggered Ethernet).

Referring to FIG. 1 , the fault detection and processing module 70 mainly includes a data frame verification error processing unit 701 , a synchronous clumping detection and processing unit 702 , an asynchronous clumping detection and processing unit 703 and a software exception processing unit 704 . The fault detection and processing module 70 runs during the whole process after the node is started, and is responsible for detecting errors that may occur during the entire process, and processing the errors.

Data frame check error processing unit 701: When the destination node receives the data frame, it judges whether the format of the data frame is correct, if it is wrong, it enters this unit, and directly discards the wrong data frame.

Synchronization clumping detection and processing unit 702: detect possible clumping in the switched Ethernet, so as to prevent synchronization of some nodes in the switched Ethernet. In the first aspect, the synchronous clumping detection and processing unit 702 is based on the value pair of new members in the optimal data frame selected by the data frame selection unit 801 after the end of the data frame selection unit 801 (that is, the moment when the receiving window is closed). Agglomeration status is checked. Wherein, each node stores a local variable M _sync , which is used to record the value of the new member in the optimal data frame selected by the data frame selection unit 801 . In the second aspect, the synchronous clumping detection and processing unit 702 compares the result of the synchronous clumping detection, that is, the size M _sync of the above-mentioned new member, with the state threshold (when the role is configured as a synchronous controller) or status threshold (when the role is configured as a compression controller) or status threshold (when the role is configured as a synchronous client) for comparison, so as to transfer the processing decision made to the state management module 30 . The state management module 30 acts on the state conversion module 40 to convert the node state. Generally, if the synchronous clumping detection is successful (that is, the size of the new member is less than the threshold value), then it indicates that there is a clumping situation, that is, there is a problem in the synchronization, and it starts to report to the state management module 30, and the state management module 30 triggers a state transition The module 40 returns the node state to an asynchronous state, and waits for an appropriate opportunity to re-initiate synchronization.

Asynchronous clumping detection and processing unit 703: Detect possible clumping in the network, so as to prevent synchronization of some nodes in the network. In the first aspect, an array M _async [2] with a size of 2 is used to record the number of unexpected data frames in different periods. The unexpected data frame refers to a data frame that cannot be received in the data selection window and a data frame that is received in the data selection window but the number of integration cycles in the data frame is inconsistent with the integration cycle of the current node. In the present invention, M _async [even number] is used to record the unexpected number of data frames in the even-numbered integration period, and M _async [odd number] is used to record the unexpected number of data frames in the odd-numbered integration period. Asynchronous clique detection is always performed throughout the integration cycle. In the second aspect, for nodes whose role is configured as a synchronization controller and a synchronization client, at the time point of data frame distribution According to the number of unexpected data frames recorded M _async [even number] (or M _async [odd number]) and the state threshold For comparison; for the node whose role is configured as a compression controller, after the end of the data frame selection unit 801 (that is, the moment when the receiving window is closed), according to the recorded unexpected number of data frames M _async [even number] (or M _async [odd number]) and state threshold For comparison, if the number of unexpected data frames M _async [even number] (or M _async [odd number]) is greater than the state threshold value, it means that the asynchronous cluster detection is successful, that is, there is a problem with the synchronization, and it starts to report to the state management module 30. The state management module 30 triggers the state transition module 50 to return the current state of the node to an asynchronous state, and wait for an appropriate opportunity to re-initiate synchronization.

Software exception processing unit 704: This unit mainly deals with various situations when the software is abnormal. When the software is abnormal, this unit will be used to alarm, report the cause of the error, and display it to the user, so that the user can know the error situation , at the same time, it will also switch to the result output module 50, output the corresponding result, and automatically terminate the software running process.

Clock correction module 80

Referring to Fig. 1, shown in Fig. 4, described clock correction module 80 realizes the function that the local clock of node is corrected, and the data frame information obtained by data processing module 60 is processed, thereby corrects local clock; After correction finishes Pass the correction result and the time point when the correction is completed to the state management module 30, and the state management module 30 will act on other modules according to the obtained information to trigger the behavior corresponding to the current node state; in the case of user demand, the clock correction result Pass it to the result output module 50 to output it; pass the data frame and time information of the clock correction process to the fault detection and processing module 70 for processing.

After entering the clock correction module 80, first the node uses the data frame selection unit 801 to select the best data frame FR _best (that is, the data frame with the largest new member value in the receiving window); and transmits the data frame FR _best and its time point For the clock correction calculation unit 802, the trigger clock correction calculation unit 802 uses the selected optimal data frame solidification time point (when the node is configured as a synchronization controller or a synchronization client) or compression time point (when the node is configured as When compressing the controller), calculate the clock correction value, obtain the size of the clock correction value, and pass it to the clock correction execution unit 803, and the clock correction execution unit 803 uses the obtained clock correction value to execute the clock correction execution time point Correction of the local clock. After the clock correction module 80 corrects the local clocks of the nodes, the local clock errors among the nodes in the network can be within a sufficiently small range, so that global clock synchronization can be achieved.

Refer to the flow chart of selecting the best data frame from the data frame shown in FIG. 4 . For the case where the role of the destination node is configured as a synchronization controller or a synchronization client, after the execution of the data frame curing unit 604 is completed, it starts to enter the data frame selection unit 801 in the clock correction module 80 .

In the first aspect, for each channel connected to the node (the channel number is denoted as j, and the total number of channels is denoted as J), first judge whether the data curing queue Q _perm of the node is empty, and if it is empty, return to the data frame curing unit 604 , waiting to receive the cured IN frame; if it is not empty, take out the first element Q _perm [0] of the data curing queue Q _perm , and record its curing time point T _p .

In the second aspect, if the data curing queue Q _perm is not empty, it is judged whether the curing time point T _p of the obtained IN frame after curing is within the receiving window range of the node and the integration of the IN frame after the curing is further judged. Whether the period is consistent with the current integration period of the node, if yes, it means that it is a valid IN frame that can be used, continue to judge whether it is the first IN frame received in the receiving window; if not, then trigger the fault detection and processing module 703 .

In the third aspect, the best data frame FR _best-j of the channel is selected according to the IN frame within the receiving window range of the second aspect and the best data frame selection method. The mode of described optimal data frame selection method is: if the current received IN frame is the first valid IN frame received in the receiving window, it is stored in the node buffer, and it is used as a reference frame with A later received IN frame on the same channel is compared; if not, it can be compared directly with the last obtained reference frame. If the value of the new member of the current IN frame is smaller than the reference frame, the current IN frame is discarded; otherwise, that is, the value of the new member of the current frame is greater than or equal to the value of the new member of the reference frame, the reference frame is discarded, and the current data frame is saved. and use it as the new frame of reference.

In the fourth aspect, the selected best IN frame FR _best-j and the curing time point T _p-best- j of FR _best-j are stored in the data collection queue Q _coll (the above operations are performed for each channel), and Reassigns the node's current integration cycle. If the obtained solidified IN frame is not within the range of the receiving window or is within the receiving window but whether the integration period of the solidified IN frame is consistent with the current integration period of the node, the fault detection and processing module 703 is triggered.

For the case where the role of the destination node is configured as a compression controller, after the execution of the data frame compression unit 605 is completed, it starts to enter the data frame selection unit 801 in the clock correction module 80 .

In the first aspect, it is judged whether the data compression queue Q _comp of the node is empty, if it is empty, then return to the data frame compression unit 605, and wait to receive the compressed data frame; if it is not empty, then take out the header of the data compression queue Q _comp Element Q _comp [0], record its solidification time point T _q .

In the second aspect, if the data compression queue Q _comp is not empty, judge whether the compression time point T _q of the obtained data frame is within the receiving window range of the node and further judge the integration cycle of the cured IN frame and the node Whether the current integration cycle is consistent, if yes, it indicates that it is a valid IN frame that can be used, and judges whether it is the first IN frame received in the receiving window; if not, then trigger the fault detection and processing module 703.

In the third aspect, the best data frame FR _best of the channel is selected according to the IN frame within the receiving window range of the second aspect and the best data frame selection method. The mode of described optimal data frame selection method is: if the current received IN frame is the first valid IN frame received in the receiving window, it is stored in the node buffer, and it is used as a reference frame with A later received IN frame on the same channel is compared; if not, it can be compared directly with the last obtained reference frame. If the value of the new member of the current IN frame is smaller than that of the last obtained reference frame, then discard the current IN frame; otherwise, that is, the value of the new member is greater than or equal to the value of the new member of the reference frame, then discard the previous data frame and save The current data frame and use it as the new reference frame.

In the fourth aspect, the selected best IN frame FR _best and the compression time point T _q-best of FR _best are stored in the data collection queue Q _coll (the above operations are performed for each channel), and the current integration of nodes Periodic reassignment. If the obtained compressed IN frame is not within the range of the receiving window or within the receiving window but whether the integration period of the solidified IN frame is consistent with the current integration period of the node, trigger the fault detection and processing module 703 . , the fault detection and processing module 703 is triggered.

For the case where the role of the destination node is configured as a synchronization controller or a synchronization client: first, use the following formula to calculate the expected receiving time point T _schd (also can be understood as the time point of the local clock at this time): $T_{\mathrm{schd}}=T_{\mathrm{dp}}^{602}+2\×{\mathrm{E.}}_0^{\mathrm{User}}+{\mathrm{D.}}_{\mathrm{comp}\_\mathrm{func}}^{605}+{\mathrm{TD}}^C;$ Each node opens a _network with a length of The receiving window of , that is, the receiving window range is After receiving the IN frame solidified by the data frame solidification unit 604, start to judge whether the solidification time point T _p of the IN frame is within the scope of the receiving window, if so, continue to operate it, if not within the scope of the receiving window, Then the IN frame is discarded. Indicates the sending time point of FR, Indicates the maximum delay time of data transmission, Indicates the compression correction value calculated based on the curing time point, and TD ^C indicates the dispatch delay time.

For the case where the destination node is configured as a compression controller, after the execution of the data frame compression unit 605 is completed, it starts to enter the data frame selection unit 801 in the clock correction module 80 . First, use the following formula to calculate the time point at which the node expects to receive the IN frame, T _schd (it can also be understood as the time point of the local clock at this time): $T_{\mathrm{schd}}=T_{\mathrm{dp}}+{\mathrm{E.}}_0^{\mathrm{User}}+{\mathrm{D.}}_{\mathrm{comp}\_\mathrm{func}}^{605}+{\mathrm{TD}}^C,$ (Where, T _dp is the starting time point of any received IN frame) Each node opens a length of The receiving window of , that is, the receiving window range is After receiving the IN frame compressed by the data frame compression unit 605, start to judge whether the time of the IN frame is within the scope of the receiving window, if so, continue to operate it, if not within the scope of the receiving window, then discard the IN frame .

For the case where the destination node is configured as a sync controller or a sync client:

In the first aspect, the judgment is made according to the data collection queue Q _coll acquired in the data frame selection unit 801 . Determine whether the data collection queue Q _coll is empty, if it is empty, it means that the clock deviation between each node is too large, do not execute the operation temporarily, and enter the waiting state; if it is not empty, use the data collected in the data collection queue Q _coll The curing time point T _p-best-j of the best frame is used to calculate the clock correction value. Wherein, the data collection queue Q _coll stores the best IN frame (with the largest new member value) in each channel, and the best curing time point of the jth channel is T _p-best-j ).

In the second aspect, using the formula Calculating Clock Correction Values and correct the clock value Pass it to the clock correction execution unit 803.

For the situation that the purpose role is configured as a compression controller, since only one compression controller is configured by default in each channel in the present invention, there is no problem of channel selection existing in the synchronous controller. Only one data element will be collected at most, so the best data element (that is, the IN frame with the largest new member value) can be directly used to obtain the time T of the compression time point contained in the best data element _q-best , after which clock correction values are calculated and correct the clock value Pass it to the clock correction execution unit 803.

In the first aspect, the clock correction execution unit 803 calculates the delay value according to the T _schd delivered by the data frame selection unit The clock correction time point T _clock-corr is obtained, and the local clock t _local is corrected from the clock correction time point T _clock-corr .

The calculation method of the clock correction time point is:

In the second aspect, at the clock correction time point T _clock-corr , the node makes a size of the local clock t _local , and the revised local clock is recorded as Dt _local , then the clock value is

In the third aspect, for the case where the role of the destination node is configured as a compression controller, the clock correction execution unit 803 re-enters the data frame generation unit 601 at time T _q-best after the clock correction is completed, and generates data frame.

Result output module 50

Referring to Fig. 1, the result output module 50: mainly responsible for outputting the current node status DV ^User and clock correction value of the node and the corrected local clock Dt _local information result.

In the present invention, the result output module 50 can output the above information in an xml format file or an excel format file. The output format file is used to directly pass to the program interface of the upper layer, so that the upper layer program can obtain information such as whether the clock is synchronized or not.

Claims (11)

1. A simulation system for analog switched Ethernet clock synchronization, the simulation system includes an Ethernet protocol command database and a timer; the Ethernet protocol command database and the timer are based on the IEEE 802.3 ISO/IEC 8802-3 standard specification It is constructed with the instruction and state association information pointed out in SAEAS 6802; it is characterized in that: it also includes a node and link attribute configuration module (10), a synchronous topology generation module (20), a state management module (30), a state transition module (40), data processing module (60), fault detection and processing module (70), clock correction module (80) and result output module (50); The node and link attribute configuration module ( ₁₀ ) configures the required simulation configuration information D10 according to the end system in the actual Ethernet, the switch, and the physical connection between the end system and the switch; Synchronous topology generating module (20) is used to generate synchronous topology, and output synchronous topology information D ₂₀ ; A state management module (30), configured to manage the execution behavior of node roles and node states; The state conversion module (40) is used to switch the node state; the node states are respectively integrated state, asynchronous state, waiting state, FLOOD state, enabling state and synchronous state; The data processing module (60) completes the generation, sending, receiving, curing, and compression of the data frame according to the current node state; The type information of the frame; after the data is solidified or compressed, it acts on the clock correction module, and passes the solidified or compressed data frame to the clock correction module; in the process of data processing, it acts on the fault detection and processing module, and the The information is transmitted to the fault detection and processing module; it is composed of a data frame generation unit (601), a data frame sending unit (602), a data frame receiving unit (603), a data frame curing unit (604), and a data frame compression unit (605) ; Fault detection and processing module (70), used for data frame verification and error handling; The situation of agglomeration is detected and processed; The detection and processing of the simulation operation status; The data frame verification error processing unit (701) , a synchronous clumping detection and processing unit (702), an asynchronous clumping detection and processing unit (703) and a software exception processing unit (704); A clock correction module (80), used to correct the local clock of the node role; consists of a data frame selection unit (801), a clock correction calculation unit (802) and a clock correction execution unit (803); Result output module (50), for the result of current state information, clock correction value size and the current local clock value of node with the format file output of xml format file or excel; Described node and link attribute configuration module (10) call in the Ethernet protocol command database in the first aspect to configure connection information The second aspect calls the Ethernet protocol instruction database TD ^C ,TS ^C ,TW ^C ,TR ^C , Perform initial assignment to obtain the simulation configuration information after assignment is the physical connection between nodes, and the identification number of any node is marked as node; The clock precision set when running the simulation system; is the integration cycle length set when running the simulation system; is the maximum number of integration cycles set when running the simulation system; It is the number of network fault tolerance set when running the simulation system; TD ^C is the distribution delay time; TS ^C is the transmission delay time; TW ^C is the online transmission delay time; TR ^C is the receiving delay time; is the maximum delay time of data transmission; is the clock drift set when running the simulation system; The first aspect of the synchronous topology generating module (20) is used for receiving The second aspect calls the node role B ^User in the Ethernet protocol command database and Configure node roles in a synchronized manner Get a synchronous topology. 2. the emulation system of analog switched Ethernet clock synchronization according to claim 1 is characterized in that: after the state transition module (40) receives the information that the state management module (30) transmits: After the node is started, the synchronous controller SM ^B first enters the integration state, and judges whether A receives a data frame within the timing T _A , and if it receives it, continues to judge its type; if it does not receive it, it enters the asynchronous state; if it receives If the data frame type is an IN frame, continue to judge whether the local member value of the node reaches the state threshold If it is not an IN frame, then judge whether it is a CA frame; for the above-mentioned situation where the received PCF is an IN frame, if the local member value of the node reaches the state threshold value Then enter the synchronous state; if not, then enter the waiting state; for the above-mentioned situation that the received data frame is not an IN frame, if the received PCF is a CA frame, then enter the waiting state; if not, then enter the asynchronous state; For the state transition in the asynchronous state: also judge whether the data frame is received at _B timing TB, if received, then further judge the type of the received data frame; if not received, then still stay in the asynchronous state; if it is CA frame, enter the waiting state; if not, further judge whether it is a CS frame; if it is a CS frame, enter the FLOOD state; if not, it is an IN frame, then enter the synchronization state; For the state transition of the waiting state: it is also judged whether _E receives the data frame within the timing TE, if it is received, it still stays in the waiting state, if not, then judges whether the local member value of the node reaches the state threshold If the node's local membership value reaches the status threshold Then enter the synchronization state; if the state threshold is not reached Then enter the asynchronous state; For the state transition in the FLOOD state: judge whether to receive PCF equally in the timing duration in state management module (30), if receive, then continue to judge the type of the PCF that receives; If not received, then judge whether Satisfy the configuration information conditions; for the PCF received within the specified time, continue to judge whether the received PCF is a CS frame; if yes, continue to maintain the FLOOD state; if not, judge whether it is a CA frame; Frame, then enter waiting state; If not, then judge whether to satisfy configuration information condition; For above-mentioned situation that does not receive PCF in the timing duration in state management module (30), if satisfy configuration information condition, then enter asynchronous state, If not satisfied, then still stay in the FLOOD state; said timing duration and configuration information condition are based on the FLOOD state in the state management module (30) depends on; For the state transition in the synchronous state: judge whether the clumping situation is detected and whether the data receiving queue Q _recv receives the CA frame, if the clumping situation is detected, enter the asynchronous state; if not, continue to keep in the synchronous state state; if a CA frame is received, it enters a waiting state; if not, it remains in a synchronous state; is the execution behavior of the synchronization controller SM ^B in the integrated state; After the node starts, the compression controller CM ^B first enters the integration state, and judges whether the size of the new member received in the state processing module (30) reaches the state threshold value in the time duration D and time T _D If it is an IN frame, enter the synchronous state; otherwise, enter the asynchronous state; for the state transition in the asynchronous state: first determine whether the CS frame is received or the value of the new member is greater than the state threshold CA frame, if it is, enter the enable state; if not, continue to judge whether the size of the received new member reaches the state threshold IN frame; if the size of the received new member reaches the status threshold If it does not receive the IN frame, it will enter the synchronous state; if it is not received, it will still stay in the asynchronous state; for the state transition in the enabled state: judge whether the data frame is received in the _F timing TF, if it is, it will still stay in the enabled state. If not, then enter the waiting state; for the state transition in the waiting state: judge whether the size of the new member received in G timing T _G reaches the state threshold value IN frame, if it is, then enter the synchronous state; if not, then enter the asynchronous state; for the state transition in the synchronous state: detect whether there is agglomeration, if there is agglomeration, then enter the integrated state; if not , it still stays in the synchronous state; After the node is started, the synchronization client SC ^B first enters the integration state, and judges whether the size of the new member received reaches the state threshold IN frame of IN frame, if it is, that is, the conversion enters the synchronous state; otherwise, it still stays in the integrated state; for the state transition in the synchronous state: detect whether there is agglomeration situation, if there is an agglomeration situation, then enter the integrated state; if If not, it remains in sync. 3. the emulation system of analog switched Ethernet clock synchronization according to claim 1 is characterized in that: the first aspect of data frame generation unit (601) is based on data frame transmission role MB ^User ={SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } in SSM ^B , SCM ^B to dynamically generate the protocol control frame FR; the second aspect is to dynamically generate the protocol control frame PCF according to the state information and state behavior information; SSM ^B is a synchronization controller used to transmit data frames as a source node; SCM ^B is a compression controller used to transmit data frames as a source node; DSM ^B is a synchronization controller used to transmit data frames as a destination node; DCM ^B is a compression controller used to transmit data frames as a destination node; DSC ^B is a synchronization client for transmitting data frames as a destination node; MSC ^B is a synchronization client for transmitting data frames as an intermediate node; The first aspect of the data frame sending unit (602) forwards the data frame receiving unit (603) according to MSC ^B in the data frame transmission role MB ^User ={SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } Received data frame; then perform the fourth and fifth aspects; the second aspect is based on the data frame transmission role MB ^User = SSM ^B , SCM in {SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } ^B to receive the FR output by the data frame generating unit (601); in the third aspect, record the sending time point of FR In the fourth aspect, add FR to the node’s data sending queue Q _send ; in the fifth aspect, record the data dispatch delay time TD ^C and send delay time TS ^C , and change the transparent clock in FR value; then, according to the data distribution delay time TD ^C , sending delay time TS ^C and the formula Obtain the transparent clock leaving the data frame sending unit (602) The data frame receiving unit (603) is used to receive the data frame FR sent by the data frame sending unit (602), and deposits the received data frame into the data receiving queue Q _VLrecv ; in the first aspect, judge the VL receiving queue Q _VLrecv of the node Is it empty? If it is not empty, it means that the data frame has been received. If it is empty, it will return to the initial state and wait for the arrival of the data frame; in the second aspect, for the above-mentioned VL receiving queue Q _{VLrecv is} not empty, first judge the reception Arrived whether it is an effective data frame, if so, then carry out information extraction processing to data frame; If not, then trigger fault detection and processing module (70) to do discarding processing; Described information extraction processing is to extract source node and destination node Address information, record the receiving time point T _r , record the online transmission delay time TW ^C between nodes in the process of sending from the previous node to the current node, receive delay time TR ^C , change the transparent clock value The method of changing the transparent clock value is based on the online transmission delay time TW ^C , the reception delay time TR ^C and the formula Get the transparent clock value leaving the data frame receiving unit (603) The third aspect performs the following operations; according to the source node and destination node address information extracted in the second aspect, and the data frame transmission role MB ^User ={SSM ^B , SCM ^B , DSM ^B , DCM ^B , DSC ^B , MSC ^B } DSM ^B and DCM ^B in DSM B judge whether the current node is the destination node consistent with the address information of the destination node, if yes, it means that the data frame transmission has been completed, and store the data frame FR and the receiving time point T _r of the received FR into In the data receiving queue Q _recv of the node, trigger the data frame curing unit (604), start to carry out its curing operation; if it is not the destination node, then pass the data frame FR to the data frame sending unit (602) of the current node, continue Data sending process, sending data frame FR to the destination node; The data frame curing unit (604) realizes the function of rearranging the data frame FR received by the destination node; through the data frame curing unit (604), the destination node can obtain a group of data frames arranged according to the sending time order of the source node; the second On the one hand, judge the data receiving queue Q _recv of the node to judge whether the node has received the data frame, if it is received, that is, the data receiving queue Q _recv is not empty, start processing the time point T _r of the received data frame , take out the head member in the data receiving queue, and perform curing operations on it to obtain the curing time point T _p ; if it is empty, return to the initial state and continue to wait for the data receiving queue Q _recv to be not empty; the second aspect , for the case that the data receiving queue Q _recv is not empty, first judge whether it is an IN frame according to the frame type part in the data frame FR, and if it is, the cured IN frame and its curing time point T _p are determined according to the curing time The order of the points is stored in the data curing queue Q _perm of the node; if not, it will not be processed; in the third aspect, if the type of the data frame in the data receiving queue Q _recv is an IN frame, then according to the role type B ^User of the node Select, if the node role is a compression controller, then enter the data frame compression unit (605), and start the process of data compression operation; otherwise, that is, the node role is a synchronization controller or a synchronization client, then enter the clock correction module (80) , start to correct the local clock of the node; The data frame compression unit (605) mainly realizes the function of compressing the curing time point of the data frame, and is a very important unit in the whole network to establish a global time reference; in the first aspect, the data frame compression unit (605) is based on data The frame curing unit (604) collects the cured data frame FR obtained by the frame curing unit (604); in the second aspect, when the collection process ends, it starts to enter the calculation stage; calculate the compression correction value according to the curing time point T _pi collected during the collection process In the third aspect, the compression correction value calculated according to the second aspect Calculate the compression time point 4. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that: described fault detection and processing module (70) realizes data frame verification and carries out error handling, to the situation of clumping The functions of detection and processing, detection of the operating status of the simulation system software and processing of errors; Data frame verification error processing unit (701): when the destination node receives the data frame, it is judged whether the format of the data frame is correct, if it is wrong, it enters this unit, and directly discards the erroneous data frame; A synchronous clustering detection and processing unit (702): detects a possible clustering condition in the switched Ethernet, so as to prevent the synchronization of some nodes in the switched Ethernet; In the first aspect, after the end of the data frame selection unit (801), the synchronous clumping detection and processing unit (702) performs the clumping condition according to the value of the new member in the optimal data frame selected by the data frame selection unit (801). detection; Wherein, a local variable M _sync is stored in each node, which is used to record the value of the new member inside the optimal data frame selected by the data frame selection unit (801); In the second aspect, the synchronous clumping detection and processing unit (702) combines the result of the synchronous clumping detection, that is, the size M _sync of the above-mentioned new member, with the state threshold or state threshold or state threshold Compare, thereby the processing decision that will make is converted to state management module (30); Act on state conversion module (40) by state management module (30) and node state is converted; Generally, if synchronous clumping detects success, then Explain that the state of clustering occurs, that is, there is a problem in the synchronization, and it starts to report to the state management module (30), and the state management module (30) triggers the state transition module (40) to return the node state to an asynchronous state, waiting for a suitable Time to re-initiate synchronization; Asynchronous clustering detection and processing unit (703): detect possible clustering conditions in the network, so as to prevent synchronization of some nodes in the network; In the first aspect, use a size to be (2) array M _async [2] to record the number of unexpected data frames in different cycles; the unexpected data frames refer to those that fail to receive in the data selection window The data frame and the data frame received in the data selection window but the number of integration cycles in the data frame is inconsistent with the integration cycle of the current node; use M _async [even number] to record the unexpected number of data frames in the even integration cycle, use M _async [Odd number] Record the number of unexpected data frames in the odd-numbered integration cycle; throughout the entire integration cycle, asynchronous cluster detection is always performed; In the second aspect, for nodes whose role is configured as a synchronization controller and a synchronization client, at the time point of data frame distribution According to the number of unexpected data frames recorded M _async [even number] or M _async [odd number] and the state threshold Compare; For the node that role is configured as a compression controller, after the data frame selection unit (801) ends, according to the recorded unexpected number of data frames M _async [even number] or M _async [odd number] and the state threshold value For comparison, if the unexpected number of data frames M _async [even number] or M _async [odd number] is greater than the state threshold value, it means that the asynchronous cluster detection is successful, that is, there is a problem in synchronization, and it starts to report to the state management module (30 ), the state management module (30) triggers the state transition module (40) to return the node state to an asynchronous state, waiting for an appropriate opportunity to re-initiate synchronization; Software exception processing unit (704): This unit mainly deals with various situations when the software is abnormal. When the software is abnormal, it will use this unit to alarm, report the cause of the error, and display it to the user, so that the user can know In the case of an error, at the same time, it will also switch to the result output module (50), output the corresponding result, and automatically terminate the software running process. 5. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that: clock correction module (80) realizes the function that the local clock of node is revised, to obtained by data processing module (60) Process the data frame information of the local clock, thereby correcting the local clock; after the correction is completed, the correction result and the time point of completing the correction are passed to the state management module (30); For the situation that the destination node role is configured as a synchronous controller or a synchronous client, after the execution of the data frame solidification unit (604), start to enter the data frame selection unit (801) in the clock correction module (80); in the first aspect, for Each channel connected to the node first judges whether the data solidification queue Q _perm of the node is empty, if empty, then returns to the data frame solidification unit (604), and waits to receive the solidified IN frame; if not empty, then take out The first element Q _perm [0] of the data curing queue Q _perm , record its curing time point T _p ; second aspect, for the situation that the data curing queue Q _perm is not empty, judge the curing time of the obtained IN frame after curing Whether the point T _p is within the receiving window range of the node and further judge whether the integration period of the solidified IN frame is consistent with the current integration period of the node, if yes, it means that it is a valid IN frame that can be used, and continue to judge whether it is a receiving window In the first IN frame that receives in; If not, then trigger fault detection and processing module (703); The third aspect, according to the IN frame and the best data frame selection method in the receiving window range of the second aspect, select out The best data frame FR _best-j of this channel; In the fourth aspect, the selected best IN frame FR _best-j and the curing time point T _p-best- j of FR _best- j are stored in the data collection queue Q _coll , and re-assign the current integration period of the node; if the obtained solidified IN frame is not within the receiving window range or within the receiving window but the integration period of the solidified IN frame is consistent with the current integration period of the node, trigger A fault detection and processing module (703). 6. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that: clock correction module (80) realizes the function that the local clock of node is revised, to obtained by data processing module (60) Process the data frame information of the local clock, thereby correcting the local clock; after the correction is completed, the correction result and the time point of completing the correction are passed to the state management module (30); For the situation that the destination node role is configured as a compression controller, after the execution of the data frame compression unit (605), start to enter the data frame selection unit (801) in the clock correction module (80); Whether the compressed queue Q _comp is empty, if empty, then return to the data frame compression unit (605), waiting to receive the compressed data frame; if not empty, then take out the first element Q _comp [0] of the data compressed queue Q _comp , record its curing time point T _q ; secondly, for the data compression queue Q _comp is not empty, judge whether the compression time point T _q of the obtained data frame is within the receiving window range of the node and further judge the curing time point T q Whether the integration period of the last IN frame is consistent with the current integration period of the node, if yes, it means that it is a valid IN frame that can be used, and judge whether it is the first IN frame received in the receiving window; if not, trigger a fault Detection and processing module (703); In the third aspect, according to the IN frame and the best data frame selection method in the receiving window range of the second aspect, select the best data frame FR _best of the channel; in the fourth aspect, the selected The best IN frame FR _best and the compression time point T _q-best of FR _best are stored in the data collection queue Q _coll , and the current integration cycle of the node is reassigned; if the obtained compressed IN frame is not within the receiving window range Or within the receiving window but whether the integration period of the solidified IN frame is consistent with the current integration period of the node, trigger the fault detection and processing module (703). 7. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that: clock correction module (80) realizes the function that the local clock of node is revised, to obtained by data processing module (60) Process the data frame information of the local clock, thereby correcting the local clock; after the correction is completed, the correction result and the time point of completing the correction are passed to the state management module (30); For the case where the role of the destination node is configured as a synchronous controller or a synchronous client: first, use the following formula to calculate the expected receiving time point T _schd : Each node opens a _network with a length of The receiving window of , that is, the receiving window range is After receiving the IN frame solidified by the data frame solidification unit (604), start to judge whether the solidification time point T _p of the IN frame is within the scope of the receiving window, if so, continue to operate it, if not within the scope of the receiving window , the IN frame is discarded; Indicates the sending time point of FR; Indicates the maximum delay time of data transmission; Indicates the compression correction value calculated based on the curing time point; TD ^C represents the dispatch delay time. 8. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that: clock correction module (80) realizes the function that the local clock of node is revised, to obtained by data processing module (60) Process the data frame information of the local clock, thereby correcting the local clock; after the correction is completed, the correction result and the time point of completing the correction are passed to the state management module (30); For the situation that the destination node is configured as a compression controller, after the data frame compression unit (605) executes, it starts to enter the data frame selection unit (801) in the clock correction module (80); at first, use the following formula to calculate the expected reception of the node To the point in time of the IN frame, T _schd : Each node opens a network with a length of The receiving window of , that is, the receiving window range is After receiving the IN frame compressed by the data frame compression unit (605), start to judge whether the time of the IN frame is within the scope of the receiving window, if so, continue to operate it, if not within the scope of the receiving window, then discard IN frame. 9. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that: clock correction module (80) realizes the function that the local clock of node is revised, to obtained by data processing module (60) Process the data frame information of the local clock, thereby correcting the local clock; after the correction is completed, the correction result and the time point of completing the correction are passed to the state management module (30); For the situation that the destination node is configured as a synchronous controller or a synchronous client, in the first aspect, judge according to the data collection queue Q _coll obtained in the data frame selection unit (801); judge whether the data collection queue Q _coll is empty, if empty , it means that the clock deviation between each node is too large, do not execute the operation temporarily, and enter the waiting state; if it is not empty, use the curing time point T _p-best- of the best frame collected in the data collection queue Q _{coll j} to calculate the clock correction value; second aspect, using the formula Calculating Clock Correction Values and correct the clock value Pass it to the clock correction execution unit (803). 10. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that: clock correction module (80) realizes the function that the local clock of node is revised, to obtained by data processing module (60) Process the data frame information of the local clock, thereby correcting the local clock; after the correction is completed, the correction result and the time point of completing the correction are passed to the state management module (30); For the case where the destination role is configured as a compression controller, since only one compression controller is configured in each channel by default, there is no problem of channel selection in the synchronization controller, and the data collection queue will only collect at most A data element, so the best data element can be directly used to obtain the time T _q-best of the compression time point contained in the best data element, and then calculate the clock correction value and correct the clock value Delivered to the clock correction execution unit (803); in the first aspect, the clock correction execution unit (803) transmits the T _schd and the calculation delay value according to the data frame selection unit Obtain the clock correction time point T _clock-corr , and start to correct the local clock t _local from the clock correction time point T _clock-corr ; secondly, at the clock correction time point T _clock-corr , the node makes a size for the local clock t _local , and the revised local clock is recorded as Dt _local , then the clock value is In the third aspect, for the case where the role of the destination node is configured as a compression controller, the clock correction execution unit (803) re-enters the data frame generation unit (601) at T _q-best time after the clock correction is completed, and generates device as the data frame of the source node. 11. the emulation system of analog switched Ethernet clock synchronization according to claim 1, is characterized in that state management module (30) is to the management of state: In the first aspect, according to the information in the D ₂₀ of the synchronous topology generation module (20), the state transition module (40), the data processing module (60), the fault detection and processing module (70) and the clock correction module (80), the integrated Analyze and generate execution behaviors that match the node states in the state transition module (40); In the second aspect, passing the execution behavior obtained by the first aspect to the matching execution behavior refers to that when the node state transferred by the state transition module (40) is an integrated state: the execution of the synchronous controller SM ^B in the integrated state Behavior Execution Behavior of Compression Controller CM ^B in Integrated State Execution behavior of synchronization client SC ^B in integrated state When the node state delivered by the state transition module (40) is a synchronous state: the execution behavior of the synchronous controller SM ^B in the synchronous state Execution Behavior of Compression Controller CM ^B in Synchronous State Execution behavior of sync client SC ^B in sync state When the node state delivered by the state transition module (40) is an asynchronous state: the execution behavior of the synchronous controller SM ^B in the asynchronous state Execution Behavior of Compression Controller CM ^B in Asynchronous State When the node state delivered by the state transition module (40) is a waiting state: the execution behavior of the synchronization controller SM ^B in the waiting state Execution Behavior of Compression Controller CM ^B in Waiting State When the node state delivered by the state transition module (40) is the FLOOD state: the execution behavior of the synchronization controller SM ^B in the FLOOD state When the node state delivered by the state transition module (40) is the enable state: the execution behavior of the compression controller CM ^B in the enable state First, for a node whose role is configured as a sync controller: (A) The execution behavior of the synchronous controller SM ^B in the integrated state Means: in A timing _TA , the state management module (30) judges the elements in the data solidification queue Q _perm delivered by the data processing module (60), judges whether there are data elements therein, and if there is, then further judges the receipt The type of the data frame FR: if it is a CA frame, then the trigger state transition module (40) is converted to a waiting state with the node state, and triggers the synchronous cluster detection and processing unit (702) to set M _sync to 1; if it is IN frame, continue to judge whether the synchronization member variable reaches the state threshold If reach, then trigger state conversion module (40) node state is converted into synchronous state, and start to trigger data processing module (60) to record above-mentioned IN frame and utilize data processing module (60) to carry out the processing to IN frame; Do not reach, then trigger state transition module (40) node state is converted into waiting state; If after A timing T _A finishes, data curing queue Q _perm is still empty, then trigger state transition module (40) node state is converted into non- Synchronous state, the timing duration is set to _TB , triggers the data frame generation unit (601) in the data processing module (60), starts to generate CS frame, then operates according to each unit sequence in the data processing module (60), initiates Handshake process; (B) Execution behavior of the synchronous controller SM ^B in an asynchronous state Refer to: in _B timing TB, the element in the data solidification formation Q _perm that data processing module (60) transmits is judged, judges whether there is data element wherein, if then further judges the data frame type that receives: if If it is a CS frame, then the trigger state conversion module (40) converts the node state into a FLOOD state; if it is a CA frame, then the trigger state conversion module (40) converts the node state into a waiting state, and triggers a synchronous clumping detection and processing unit (702) M _sync is set to 1; if it is an IN frame, then continue to judge whether the synchronization member variable reaches If the node synchronization member variable reaches Then the trigger state conversion module (40) converts the node state into a synchronous state, and starts to trigger the data processing module (60) to record it and perform the processing to the IN frame in the data processing module (60); if not reached, trigger State conversion module (40) converts the node state into a waiting state; if after the timing TB of _B ends, the data solidification queue Q _perm of the node is still empty, then the trigger state conversion module (40) converts the node state into an asynchronous state, and this , trigger the data frame generation unit (601) in the data processing module (60), start to generate the CS frame, then operate according to the order of each unit in the data processing module (60), and initiate the handshake process; (C) Execution behavior of the synchronous controller SM ^B in the FLOOD state Refer to: the state management module 30 sets the timing duration according to the current variable configuration information of the node; The variable configuration information includes flood_recv and closed_window; In the timing TC of _C , the element added in the data solidification queue Q _perm delivered by the data processing module (60) is judged to judge whether there is a data element in it, if there is, then further judge the PCF frame type received: if it is CS frame, then triggering state transition module (40) node state is set to FLOOD state again; If it is a CA frame, and under the configuration situation that the flood_recv value is 1 and the closed_window value is 0, then the state transition module (40) is triggered to convert the node state to a waiting state; After C timing T _C ends, the data solidification queue Q _perm of node is still empty, then judge flood_recv value and closed_window value, if when all being 0, then triggering state conversion module (40) node state is put into FLOOD state again, And trigger data processing module (60) to send CA frame, meanwhile, flood_recv is assigned a value of 1; If the flood_recv value is 1, and the closed_window value is 0, then within _I timing T1, continue to judge the elements in the data curing queue Q _perm delivered by the data processing module (60), judge whether there is a data element in it, if not , then the trigger state transition module (40) is set to the FLOOD state again with the node state, and does not carry out state transition; If the flood_recv value and the closed_window value are all 1, then in J timing T _J , continue to judge the elements in the data curing queue Q _perm delivered by the data processing module (60), judge whether there is a data element therein, if not, Then the trigger state conversion module (40) converts the node state into an asynchronous state, and assigns flood_recv value and closed_window value to 0; (D) Execution behavior of the synchronous controller SM ^B in the waiting state Refer to: in _E timing TE, judge the element in the data solidification queue Q _perm that data processing module (60) transmits, judge whether there is data element wherein, if have, then further judge the data frame type that receives: If it is a CS frame, then the trigger state transition module (40) converts the node state to the FLOOD state; if it is a CA frame, then the trigger state transition module (40) continues to stay in the waiting state with the node state, and triggers synchronous clumping detection and The processing unit (702) sets M _sync to 1; if the data receiving queue Q _recv is empty, and the local member value of the node reaches the state threshold value Then the trigger state conversion module (40) converts the node state into a synchronous state, and triggers the data frame generation unit (601) in the data processing module (60) simultaneously, and starts to generate an IN frame, that is, the local member value of new member=node in the frame , to initiate the synchronization process; if the node's local membership value does not reach the state threshold Then trigger the state transition module (40) to convert the node state into an asynchronous state; (E) The execution behavior of the synchronous controller SM ^B in the synchronous state Means: in this state, the cold start handshake has been successful, and the clock difference of each node remains within the scope of the set clock accuracy; the state management module (30) triggers the data processing module (60) at the starting point of each integration cycle ) generates an IN frame, and performs operations such as sending, receiving, and solidifying it; the trigger clock correction module (80) performs clock correction after receiving the solidified IN frame; the trigger fault detection and processing module (70) performs the operation process Monitoring, utilizing the synchronous clumping detection and processing unit (702) and the asynchronous clumping detection and processing unit (703) in the fault detection and processing module (70) to detect whether there is clumping in the synchronous process, and make Processing; in the synchronous state, the data processing module (60) feeds back the elements added in the data solidification queue Q _perm delivered to the state management module (30) to judge, if it is an IN frame, then continue to remain in the synchronous state, and continue the follow-up operation If it is a CA frame, it indicates that there is an asynchronous situation at present, and the trigger state transition module (40) converts the node state to a waiting state, and begins to wait for re-initiation of the synchronization process; fault detection and processing module (70) will detect report to the state management module (30), when the synchronous clumping detection and processing unit (702) or the asynchronous clumping detection and processing unit (703) detects the clumping situation, the situation will be reported to the status management module (30), the state management module (30) triggers the state transition module (40) to convert the current state into an asynchronous state; For a node configured as a compression controller: discuss according to different state information delivered by the state transition module (40); (A) Execution behavior of the compression controller CM ^B in the integrated state Refer to: in D timing T _D , the element in the data solidification queue Q _perm that data processing module (60) transmits is judged, judges whether there is data element wherein, if then further judges the data frame type that receives: if If it is a CS or CA frame, it will be regarded as something that should not happen, and the data frame will be discarded directly; if it is an IN frame, continue to judge whether the synchronization member variable reaches the state threshold If the synchronization member variable reaches the state threshold Then the trigger state conversion module (40) converts the node state into a synchronous state, and starts to trigger the data processing module (60) to record it and utilize the data processing module (60) and the clock correction module (80) to carry out corresponding operations; if not Reach, then data frame is discarded; If after D timing T _D ends, the data solidification queue Q _perm of node is still empty, then enter triggering state conversion module (40) and node state is converted into asynchronous state; (B) Execution behavior of the compression controller CM ^B in an asynchronous state Refer to: judge the elements in the data solidification queue Q _perm of the node delivered by the data processing module (60), judge whether there are data elements in it, if there is, then further judge the received data frame type: if it is a CS frame, or It is a CA frame and the member value information of the received CA frame is greater than the status threshold Then the trigger state conversion module (40) converts the node state into an enabled state, and the trigger data processing module (60) utilizes the data frame generation unit (601) to generate CS or CA frames, and according to the steps in the data processing module (60) Generate and send the data frame; if it is an IN frame, and the value of the new member of the IN frame is greater than the state threshold Then the trigger state conversion module (40) converts the node state into a synchronous state, and triggers the data processing module (60) subsequent processing; if the data solidification queue Q _perm is always empty, then the node stays in the asynchronous state; is the threshold value of the compression controller CM ^B from the asynchronous state to the synchronous state; The steps in the described data processing module (60) include: because the generated data frame is substantially consistent with the received data frame format, therefore, when generating a new data frame, the received PCF internal structure is changed The method for modifying the transparent clock domain of the frame is 0, and adding new VL information in the data processing process, and sending it back to the corresponding destination node according to the subsequent steps in the data processing module (60); (C) Execution behavior of the compression controller CM ^B in an enabled state Refer to: in _F timing TF, the element in the data solidification formation Q _perm that data processing module (60) transmits is judged, judges whether there is data element wherein, if then further judges the data frame type that receives: if If it is a CS or IN frame, it is regarded as an error, and the data frame is directly discarded without operation; if a CA frame is received, the CA frame will be forwarded, and the data processing module (60) is triggered to generate a CA frame , and send the data frame according to the steps in the data processing module (60); in this state, whenever a CA frame is received, the forwarding behavior is performed, because the generated data frame is basically consistent with the received data frame format , therefore, when generating a new data frame, take the method of changing the received PCF content, modify the transparent clock domain of the frame, and add a new VL in the data processing process, according to the data processing module (60) The follow-up steps send it back to the corresponding destination node. In the case of receiving the data frame, the node continues to stay in the current state without state transition; if the node data receiving queue Q _recv is still empty after the end of _F timing TF , that is, no new data is received, then the state transition module (40) is triggered to convert the node state to a waiting state; (D) Execution behavior of the compression controller CM ^B in the waiting state Refer to: in G timing T _G , the element in the data solidification queue Q _perm that data processing module (60) transmits is judged, judges whether there is data element wherein, if then further judges the data frame type that receives: if If it is a CS or CA frame, it is regarded as an error, and the data frame is discarded directly, and no operation is performed; if it is an IN frame, and the value state threshold of the new member of the IN frame Then the trigger state conversion module (40) converts the node state into a synchronous state, and triggers the data processing module (60) and the clock correction module (80) to operate on the received data frame; if G timing T _G ends, the node It is always empty in the data solidification queue Q _perm , that is, no new data is received, then the state transition module (40) is triggered to convert the node state to an asynchronous state; (E) Execution behavior of the compression controller CM ^B in a synchronous state Refers to: when in this state, it means that the cold start handshake is successful, and the clock difference of each node remains within the range of the clock accuracy set; the state management module (30) according to the data frame fed back by the data processing module (60) Carry out corresponding operation: if receive the CS after curing or CA frame, then regard as the situation of error, directly discard data frame, do not give operation; If receive the IN frame after curing, then trigger data processing module (60) The data frame compression unit (605) in the IN frame is collected, compressed and other operations, and the clock correction module (80) is used to correct the clock, and the synchronous cluster detection and processing unit in the fault detection and processing module (70) is used (702) and asynchronous clumping detection and processing unit (703) detect whether clumping occurs in the synchronous process, and make processing; Fault detection and processing module (70) reports the detected problem to the state management module (30), when the synchronous clumping detection and processing unit (702) or the asynchronous clumping detection and processing unit (703) detect the situation of clumping, the situation that occurs will be reported to the status management module (30), the status management module (30) trigger state transition module (40) node state is converted into integrated state, thereby wait again to start the process of handshake synchronization; If the situation of clustering is not detected, then continue to remain in the synchronous state, also according to the data processing module ( 60) when processing data in the corresponding situation updates the local configuration information of the node; For the node configured as a synchronous client: discuss according to the different state information delivered by the state transition module (40); (A) Execution behavior of the synchronization client SC ^B in the integrated state Refer to: first judge the elements in the data curing queue Q _perm delivered by the data processing module (60), judge whether there are data elements in it, if there is, then further judge the received data frame type: if it is an IN frame, then further Determine whether its synchronization member variable reaches the state threshold If yes, then the trigger state conversion module (40) converts the node state into a synchronous state, and triggers the data processing module (60), processes it according to the method when receiving the IN frame in the data processing module (60), and to The local information of the node must be updated; if not, the state threshold cannot be reached Then explain the state that can not reach synchronization yet, continue to trigger data processing module (60), carry out corresponding data forwarding operation; If the PCF type that receives is CS or CA frame, also continue to trigger data processing module (60), carry out corresponding data forwarding operation; (B) The execution behavior of the synchronization client SC ^B in the synchronization state Refers to: the state management module (30) triggers the data processing module (60), the clock correction module (80), etc. to process the received IN frame, and corrects the current local clock of the node, and at the same time, triggers fault detection and processing Module (70) detects whether there is agglomeration, if it occurs, then triggers the state transition module (40) to convert the current state into an integrated state; otherwise, it remains in the synchronous state; at the same time, the current configuration information of the node is correspondingly Modifications.