CN115421917A - A load balancing method for embedded system multiprocessor modules - Google Patents

Abstract

The invention relates to a load balancing method for multiprocessor modules of an embedded system, which belongs to the technical fields of embedded system communication technology and redundancy management. The invention does not require manual intervention, and can automatically realize the designation and distribution of the control process of the terminal equipment on the processor module; through real-time dynamic monitoring of the program execution time of each slave processor module, the control of the running time and execution efficiency of each slave processor module is realized. Load balancing improves the resource utilization of the entire system; the efficiency of console task execution and the distributed processing capability of the system are improved through reasonable and optimal configuration of control process resources.

Description

A load balancing method for embedded system multiprocessor modules

technical field

The invention belongs to the field of embedded system communication technology and redundant management technology, and in particular relates to a load balancing method for multiprocessor modules of an embedded system.

Background technique

In a certain control system, the console device needs to realize the control of multiple terminal devices through the processor computing unit. The processor calculation unit receives the control instructions sent by the console device through the Ethernet, analyzes and processes the control instructions, and sends the analyzed and processed control instructions to each terminal device through different interfaces. Each terminal device sends the command feedback data generated by itself to the processor computing unit through different interfaces. After the processor calculation unit receives the command feedback data of each terminal device, the command feedback data is grouped and processed according to the Ethernet message, and the console feedback command is generated, and the generated console feedback command is sent to the console device through the Ethernet .

In the above control system, the processor computing unit is the core component of the control system. In order to ensure the reliability of the core component, it is necessary to perform redundant backup on the processor computing unit. The processor computing unit is composed of multiple processor modules, including a master processor module and multiple slave processor modules. The console device sends control processes to different terminal devices through Ethernet, and the control process of each terminal device is the same. After the main processor module receives the control process of all terminal devices sent by the console device, according to the principle of load balancing, let a certain slave processor module control one or more terminal devices to execute and complete the control process. The core of the above control system is how the main processor module reasonably distributes the terminal device control process sent by the console device to different slave processor modules. The existing terminal equipment control process allocation strategy usually relies on the experience value of the operator, and randomly allocates the terminal equipment control process to different slave processor modules. This approach will lead to two adverse consequences: (1) Different operator experience different, resulting in a part of the slave processor module being in a relatively busy state, while another part of the slave processor module is in a relatively idle state, resulting in a large difference in the execution efficiency of different terminal devices; (2) in some extreme cases, a slave processor module Because the processor executes multiple terminal equipment control processes, the operating load is heavy. The actual operating cycle of the processor program is greater than the execution cycle designed in the program, resulting in the program not running normally and the control process being interrupted abnormally.

Contents of the invention

(1) Technical problems to be solved

The technical problem to be solved by the present invention is: how to solve the problem of "uneven distribution of control process resources of multiple slave processor modules of the processor computing unit" in the current control system, and enhance the communication between the console, the processor computing unit and the terminal equipment. The reliability of communication improves the processing efficiency of terminal equipment control process.

(2) Technical solution

In order to solve the above-mentioned technical problems, the present invention provides a load balancing method for embedded system multiprocessor modules, the method is implemented based on a control system, the control system includes: a console device, a processor computing unit and a terminal device;

The console device is used to send control instructions to the processor computing unit through the Ethernet, and receive the console device feedback instructions sent by the processor computing unit through the Ethernet, and all control commands sent by the console device follow the established process Executed, so as to realize the control flow sent to each terminal device, and the control flow sent to each terminal device is the same;

The processor calculation unit is used to receive the control flow sent by the console device through Ethernet, and analyze and process each control flow; the processor calculation unit is composed of multiple processor modules, including a main processor module and multiple From the processor module; wherein, the main processing module is used to receive and save all the control flow sent by the console device, and send the control flow to the slave processor module; the slave processor module is used to analyze and execute the control sent by the main processor module process, the effective data content of each control command in the control process is sent to the terminal device through different interfaces, and the command feedback data sent by the terminal device is received, and the command feedback data is packaged according to the Ethernet message to generate Console device feedback command, send the generated console device feedback command to the console device through Ethernet; the program running in each slave processor module is the same; the main processing module in the processor computing unit is also used to The execution time of the program in each slave processor module dynamically adjusts the number of control processes executed by each slave processor module;

The terminal device is used to receive the effective data content of the control command sent by the processor computing unit through different interfaces, and send the command feedback data generated by itself to the processor computing unit through different interfaces;

The load balancing method realized based on the control system includes the following steps:

S ₁ : The main processor module in the processor computing unit receives and saves all the control processes sent by the console device through Ethernet;

The console device sends the preset control process to the processor computing unit through the Ethernet, and receives the console device feedback instructions sent by the processor computing unit through the Ethernet.

S ₂ : The main processor module in the processor computing unit is based on the running time value of the current program in each slave processor module, and applies the load balancing principle to send the control process sent by the console device to different slave processors through Ethernet The module realizes the load balancing management of all slave processor modules, parses and processes the control instructions in the control process from the processor module, and sends the effective data content of the parsed and processed control instructions to the terminal device through different interfaces;

S ₃ : The terminal device sends the command feedback data generated by itself to the processor computing unit through different interfaces;

S ₄ : After the processor calculation unit receives the command feedback data of the terminal equipment, it packs the command feedback data according to the Ethernet message, generates a console device feedback command, and sends the generated console device feedback command through Ethernet to the console device.

Preferably, 1 main processor module and N slave processor modules in the processor computing unit set up a message-centric model based on CORBA middleware; each slave processor module calculates the actual running time value of the current entire program in real time , and save the maximum value of the current program running time; each slave processor module will issue its own number, the number of control processes it is executing, and the maximum value of the current program running time, and the main processor module will pass the CORBA middleware The synchronous update mechanism can access the number of each slave processor module, the number of control processes being executed and the maximum value of the current program running time.

Preferably, in step _S2 , the main processor module in the processor computing unit implements load balancing management for all slave processor modules according to the following steps for the received control flow:

S ₂₁ : Since the number of terminal devices is M, and M≥1, the maximum number of control processes sent by the console device is M; assuming that there is only one main processor module and one slave processor in the processor computing unit Module, through testing, calculate the maximum value T _i of the actual running time of the slave processor module program when the console device sends 1 to M control processes, 1≤i≤M;

S ₂₂ : Assume that the execution cycle of the program in each slave processor module is T _execution ;

When T _M < T is _executed , it indicates that M control processes can be correctly executed on one slave processor module; each time the master processor module assigns a control process to a slave processor module, it reads the current The maximum value of the program running time, find out the number of the slave processor module with the smallest running time among these maximum values, and assign this control flow to the slave processor module; when the slave processor module executes a new control flow , add 1 to the number of control processes being executed, update and save the maximum value of the current program running time of the slave processor module and the number of control processes being executed by the slave processor module in real time; if each slave processor module The _execution period T of the program is equal to the running time of the program, and the main processor module also executes the distribution of the control flow in the same way under the _execution condition of T _M <T;

When T _M ≥ T is _executed , it indicates that M control processes cannot be correctly executed on a slave processor module; according to the T _i value obtained in step _S21 , the minimum value of i is calculated when T _i ≥ T is _executed , the The value is the maximum value that a slave processor module can correctly execute the number of control processes at the same time, which is denoted as C _max , and then step _S23 is executed;

S ₂₃ : Every time the master processor module assigns a control flow to the slave processor module, it reads the number of control flows being executed by each slave processor module, which is recorded as C _i , and if C _i <C _max , then execute the step S ₂₄ ; if C _i is equal to C _max , then compare the size of C _i+1 and C _max , if C _i+1 < C _max , then execute step S ₂₄ ; if all the values of C _i are equal to C _max , it means that there is no A slave processor module can execute the control process sent by the current console device, and the master processor module sends the "no slave processor module available" message to the console device through Ethernet;

S ₂₄ : Assign this control flow to the slave processor module; when the slave processor module executes a new control flow, it adds 1 to the number of the currently executing control flow, and updates and saves the current state of the slave processor module in real time. The maximum value of program running time and the number of control processes being executed by the slave processor module.

Preferably, in step S2, after the execution of a control flow of the slave processor module is completed, the number of control flows being executed is decremented by ₁ ; The number of control processes and the maximum value of the program running time are cleared; then, the maximum value of the current program running time of the module and the number of control processes being executed by the module are updated and saved in real time.

Preferably, in step _S2 , the processor computing unit uses the following method to analyze and process the control flow sent by the console device:

The message format of the control process sent by the console device consists of "command identification" + "source identification" + "sink identification" + "command serial number" + "command length" + "command number" + "command 1 type" + "Instruction 1 interval time" + "Instruction 1 effective field content" + "Instruction 2 type" + "Instruction 2 interval time" + "Instruction 2 effective field content" + ... + "Instruction k type" + "Instruction k interval time "+" effective field content of instruction k"+"checksum"; k is an integer greater than 2;

The processor module in the processor calculation unit judges whether the length and the checksum of the current control flow are correct according to the message format; wherein, the content of the last word of the current control flow is a checksum, and the current control flow The content of the other fields except the "checksum" field is calculated according to the CRC16 checksum algorithm; if the length of the control flow and the checksum are both correct, extract each command in the control flow from the current control flow according to the number of instructions. The effective field content of each control instruction is recorded as the effective data content of the console device control instruction, and the instruction interval time of each control instruction is saved; the processor calculation unit divides the console device control process according to the instruction interval time of each control instruction. The effective data content of each control command in the command is sent to the terminal device; after the processor computing unit receives the command feedback data of the terminal device through different interfaces, the processor module in the processor computing unit extracts the "data field" in the command feedback data The content is used as the content of the message, and the length of the valid data in the content and the checksum are calculated. The calculation method of the checksum is the same as that of the control command checksum of the console device, and then according to "command identification" + "source ID" + "sink ID" + "command number" + "command length" + "command effective field content" + "checksum" format to generate a console device feedback command, and the generated console device feedback command will be sent via Ethernet sent to the console device.

Preferably, in step S3, the terminal device sends the command feedback data generated by itself to the processor calculation unit through one of the RS _- 422A interface, the GJB289A interface, the switch interface, and the CAN bus interface.

Preferably, in step _S4 , the processor computing unit sends the generated console device feedback instruction to the console device via Ethernet using the following method:

The slave processor module in the processor calculation unit extracts the content of the "data field" in the terminal device command feedback data as the message content, and calculates the length and checksum of the valid data in the content. The calculation method of the checksum is the same as The verification and calculation method of console device control commands is the same, and then follow the "command ID" + "source ID" + "sink ID" + "command serial number" + "command length" + "command effective field content" + "verify and" format to generate a console device feedback command, and send the generated console device feedback command to the console device via Ethernet.

Preferably, the effective field content of the instruction is a positive integer multiple of 4.

The invention also provides an application of the method in the technical field of embedded system communication.

The invention also provides an application of the method in the technical field of redundancy management.

(3) Beneficial effects

The advantage of the present invention compared with prior art is:

1. Without manual intervention, the designation and distribution of the terminal equipment control process can be automatically realized on the processor module;

2. Through real-time dynamic monitoring of the program execution time of each slave processor module, the load balance of the running time and execution efficiency of each slave processor module is realized, and the resource utilization rate of the entire system is improved;

3. The efficiency of console task execution and the distributed processing capability of the system are improved through the rational optimization and allocation of control process resources.

Description of drawings

Fig. 1 is the realization schematic diagram of the present invention;

Fig. 2 is a flow chart of the method of the present invention.

detailed description

In order to make the purpose, content and advantages of the present invention clearer, the specific implementation manners of the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments.

A load balancing method for embedded system multiprocessor modules proposed by the present invention is realized based on a control system, and the control system includes: a console device, a processor computing unit and a terminal device; the method can realize console devices, processor computing The communication between the unit and the terminal equipment has high requirements on the real-time and reliability of the communication.

The console device is used to send control instructions to the processor computing unit through Ethernet, and receive the console device feedback instructions sent by the processor computing unit through Ethernet. All control instructions sent by the console device are executed according to the established process , so that the control flow sent to each terminal device is realized, and the control flow sent to each terminal device is the same;

The processor computing unit is used to receive the terminal device control process sent by the console device through Ethernet, and analyze and process each control process; the processor computing unit is the core component of the control system. In order to ensure the reliability of the core component, Redundant backup is required for the processor computing unit; the processor computing unit is composed of multiple processor modules, including a main processor module and multiple slave processor modules; among them, the main processing module is used to receive and save the console equipment All the terminal device control processes sent, send the terminal device control process to the slave processor module; the slave processor module is used to parse and execute the terminal device control process sent by the main processor module, and send the control process of each control command in the control process The effective data content of the control command is sent to the terminal device through different interfaces, and the command feedback data sent by the terminal device is received, the command feedback data is packaged according to the Ethernet message, and the console device feedback command is generated, and the generated console device Feedback commands are sent to the console device via Ethernet. The programs run in each slave processor module are identical. In order to improve the processing efficiency of the control process of each terminal device by the console device, the main processing module in the processor computing unit also dynamically adjusts the terminal device executed by each slave processor module according to the execution time of the program in each slave processor module The number of control processes, so as to realize the load balancing of the terminal device control process resources of each slave processor module.

The terminal device is used to receive the effective data content of the control instruction sent by the processor computing unit through different interfaces, and send the command feedback data generated by itself to the processor computing unit through different interfaces.

As shown in Fig. 1 and Fig. 2, a kind of embedded system multiprocessor module load balancing method proposed by the present invention, the steps are as follows:

S ₁ : The main processor module in the processor computing unit receives and saves all the control processes sent by the console device through Ethernet;

The control process for each terminal device is the same. The console device sends the preset terminal device control process to the processor computing unit through Ethernet, and receives the console device feedback command sent by the processor computing unit through Ethernet. .

S ₂ : The main processor module in the processor computing unit is based on the running time value of the current program in each slave processor module, and applies the load balancing principle to send the control process sent by the console device to different slave processors through Ethernet The module realizes the load balancing management of all slave processor modules, parses and processes the control instructions in the control process from the processor module, and sends the effective data content of the parsed and processed control instructions to the terminal device through different interfaces;

One main processor module and N slave processor modules in the processor computing unit establish a message-centric publish-subscribe service model based on CORBA middleware; each slave processor module calculates the actual running time of the current entire program in real time value, and save the maximum value of the current program running time value. Each slave processor module will publish its own number, the number of terminal device control processes it is executing, and the maximum value of the current program running time. The main processor module can access each slave through the synchronous update mechanism of CORBA middleware. The number of the processor module, the number of terminal device control processes being executed, and the maximum value of the current program running time.

_In step S2, the main processor module in the processor computing unit implements load balancing management for all slave processor modules according to the following steps for the received terminal device control process:

S ₂₁ : Since the number of terminal devices is M (M≥1), the maximum number of terminal device control processes sent by the console device is M; assuming that there is only one main processor module and one From the processor module, calculate the actual running time of the slave processor module program when the console device sends 1 terminal device control process, 2 terminal device control processes, ..., M terminal device control processes. Maximum value T _i (1≤i≤M, i is a positive integer);

S ₂₂ : Assume that the execution cycle of the program in each slave processor module is T _execution (T _execution >0);

When T _M < T is _executed , it indicates that M terminal device control processes can be correctly executed on one slave processor module; each time the master processor module assigns a terminal device control process to a slave processor module, it reads each slave The maximum value of the current program running time of the processor module, find out the number of the slave processor module with the smallest running time among these maximum values, and assign this terminal device control process to the slave processor module; the slave processor module is in When executing a new terminal device control process, add 1 to the number of terminal device control processes being executed, update and save the maximum value of the current program running time of the slave processor module and the terminal device being executed by the slave processor module in real time The number of control processes; if the _execution period T of each slave processor module program is equal to the running time of the program, the master processor module also executes the distribution of the terminal device control process in the same way as under the _execution condition of T _M <T;

When T _M ≥ T is _executed , it indicates that M terminal device control processes cannot be correctly executed on a slave processor module; according to the T _i value obtained in step S ₂₁ , the minimum value of i when T _i ≥ T is _executed is calculated , this value is exactly the maximum value of a slave processor module that can correctly execute the number of terminal device control processes at the same time, denoted as C _max (C _max >0), and then perform step _S23 ;

S ₂₃ : The master processor module reads the number of terminal device control processes being executed by each slave processor module each time it assigns the terminal device control process to the slave processor module, and records it as C _i , if C _i <C _max , then perform step S ₂₄ ; if C _i is equal to C _max , then compare the size of C _i+1 and C _max , if C _i+1 <C _max , then perform step S ₂₄ ; if all C _i values are equal to C _max , indicating that none of the slave processor modules can execute the terminal device control process sent by the current console device, and the master processor module sends a "no slave processor module available" message to the console device through Ethernet;

S ₂₄ : assign this terminal device control process to the slave processor module; when the slave processor module executes a new terminal device control process, add 1 to the number of terminal device control processes being executed, update and save in real time The maximum value of the current program running time of the slave processor module and the number of terminal device control processes being executed by the slave processor module;

Among them, when the execution of a terminal device control process of the slave processor module is completed, the number of terminal device control processes being executed is reduced by 1; The number of terminal equipment control processes and the maximum value of the program running time are cleared; after the above operations are completed, the maximum value of the current program running time of the module and the number of terminal equipment control processes being executed by the module are updated and saved in real time.

Among them, the processor computing unit uses the following method to analyze and process the terminal device control process sent by the console device:

Among them, the message format of the terminal device control process sent by the console device is composed of "command identification (2 bytes)" + "source identification (4 bytes)" + "sink identification (4 bytes)" + "command serial number (4 bytes)" + "instruction length (2 bytes)" + "number of instructions (the total number is K, K>0, 2 bytes)" + "instruction 1 type (2 bytes)" + " Instruction 1 interval time (2 bytes)"+"Instruction 1 effective field content (the total length is W ₁ byte, W ₁ >0 and W ₁ is a positive integer multiple of 4)"+"Instruction 2 type (2 bytes )" + "command 2 interval time (2 bytes)" + "command 2 effective field content (the total length is W ₂ bytes, W ₂ >0 and W ₂ is a positive integer multiple of 4)" + ... + " Command k type (2 bytes)" + "command k interval time (2 bytes)" + "command k effective field content (the total length is W _k bytes, W _k >0 and W _k is a positive integer multiple of 4 )" + "checksum (2 bytes)".

The processor module in the processor calculation unit judges whether the length and the checksum of the current control flow are correct according to the message format; wherein, the content of the last word of the current control flow is a checksum, and the current control flow The content of the other fields except the "checksum" field is calculated according to the CRC16 checksum algorithm; if the length of the control flow and the checksum are both correct, extract each command in the control flow from the current control flow according to the number of instructions. The effective field content of each control instruction is recorded as the effective data content of the console device control instruction, and the instruction interval time of each control instruction is saved; the processor calculation unit divides the console device control process according to the instruction interval time of each control instruction. The effective data content of each control command in the command is sent to the terminal device; after the processor computing unit receives the command feedback data of the terminal device through different interfaces, the processor module in the processor computing unit extracts the "data field" in the command feedback data The content is used as the content of the message, and the length of the valid data in the content (the length of the message) and the checksum are calculated. The calculation method of the checksum is the same as that of the console device control command checksum, and then according to the "command identification (2 bytes)" + "source identification (4 bytes)" + "sink identification (4 bytes)" + "command serial number (2 bytes)" + "command length (2 bytes)" + "command Valid field content (W byte, W>0 and W is a positive integer multiple of 4)" + "checksum (2 bytes)" format to generate a console device feedback command, and pass the generated console device feedback command through Ethernet sent to the console device.

S ₃ : The terminal device sends the command feedback data generated by itself to the processor computing unit through different interfaces;

Among them, the terminal device sends the command feedback data generated by itself to the processor calculation unit through different interfaces such as RS-422A interface, GJB289A interface, switch interface, CAN bus interface and so on.

S ₄ : After the processor calculation unit receives the command feedback data of the terminal equipment, it packs the command feedback data according to the Ethernet message, generates a console device feedback command, and sends the generated console device feedback command through Ethernet to the console device.

Among them, in order to reduce the communication load of the main processor module and improve the execution efficiency of the main processor module, each terminal device sends the command feedback data to the corresponding slave processor module, and the slave processor module sends it to the console device.

Wherein, the processor computing unit sends the generated console device feedback command to the console device via Ethernet in the following way:

The processor module in the processor calculation unit extracts the content of the "data field" in the terminal device command feedback data as the message content, and calculates the length (message length) and checksum of the valid data in the content, and checks The calculation method of the sum is the same as the calculation method of the control command checksum of the console device, and then according to "command identification (2 bytes)" + "source identification (4 bytes)" + "sink identification (4 bytes)" + "Instruction serial number (2 bytes)" + "instruction length (2 bytes)" + "instruction effective field content (N bytes, N>0 and N is a positive integer multiple of 4)" + "checksum (2 byte)” format to generate a console device feedback command, and send the generated console device feedback command to the console device via Ethernet.

It can be seen that a load balancing method for embedded system multiprocessor modules provided by the present invention does not require manual intervention, and can automatically realize the designation and distribution of the control process of the terminal equipment on the processor module; through real-time dynamic monitoring of each The program execution time of each slave processor module realizes the load balance of the running time and execution efficiency of each slave processor module, which improves the resource utilization of the entire system; through the reasonable optimal configuration of control process resources, the execution efficiency of console tasks is improved. Efficiency and system distributed processing capabilities.

Adopting the scheme of the present invention can perfectly solve the problem of "uneven distribution of control flow resources of multiple slave processor modules of the processor computing unit" in the current control system, without manually specifying and allocating the control flow of the terminal equipment, and avoiding the occurrence of slave processor modules The resource allocation of terminal equipment control processes is uneven, and even individual slave processor modules cannot complete all task processes within the specified time due to the execution of more terminal equipment control processes.

The above is only a preferred embodiment of the present invention, it should be pointed out that for those of ordinary skill in the art, without departing from the technical principle of the present invention, some improvements and modifications can also be made. It should also be regarded as the protection scope of the present invention.

Claims (10)

1. A load balancing method for multiprocessor modules of an embedded system is characterized in that the method is realized based on a control system, and the control system comprises: the system comprises console equipment, a processor computing unit and terminal equipment;

the control console device is used for sending a control instruction to the processor computing unit through the Ethernet and receiving a control console device feedback instruction sent by the processor computing unit through the Ethernet, all the control instructions sent by the control console device are executed according to a set flow, so that the control flow is sent to each terminal device, and the control flow sent to each terminal device is the same;

the processor computing unit is used for receiving the control flows sent by the console equipment through the Ethernet and analyzing each control flow; the processor computing unit consists of a plurality of processor modules and comprises a main processor module and a plurality of slave processor modules; the main processing module is used for receiving and storing all control flows sent by the console equipment and sending the control flows to the auxiliary processing module; the slave processor module is used for analyzing and executing the control flow sent by the master processor module, sending the effective data content of the control instruction of each control instruction in the control flow to the terminal equipment through different interfaces, receiving the command feedback data sent by the terminal equipment, packaging the command feedback data according to the Ethernet message to generate a console equipment feedback instruction, and sending the generated console equipment feedback instruction to the console equipment through the Ethernet; the programs running in each slave processor module are identical; the main processing module in the processor computing unit is also used for dynamically adjusting the number of the control flows executed by each slave processing module according to the execution time of the program in each slave processing module;

the terminal equipment is used for receiving effective data content of the control instruction sent by the processor computing unit through different interfaces and sending command feedback data generated by the terminal equipment to the processor computing unit through different interfaces;

the load balancing method based on the control system comprises the following steps:

S ₁ : a main processor module in the processor computing unit receives and stores all control flows sent by the console equipment through the Ethernet;

the console device sends a preset control flow to the processor computing unit through the Ethernet, and receives a console device feedback instruction sent by the processor computing unit through the Ethernet.

S ₂ : a master processor module in a processor computing unit sends a control flow sent by console equipment to different slave processor modules through Ethernet by applying a load balancing principle based on the running time value of a current program in each slave processor module, so as to realize load balancing management of all the slave processor modules, the slave processor modules analyze control instructions in the control flow and send effective data contents of the analyzed control instructions to terminal equipment through different interfaces;

S ₃ : the terminal equipment sends the command feedback data generated by the terminal equipment to the processor computing unit through different interfaces;

S ₄ : and after receiving the command feedback data of the terminal equipment, the processor computing unit performs packet packaging processing on the command feedback data according to the Ethernet message to generate a console equipment feedback instruction, and sends the generated console equipment feedback instruction to the console equipment through the Ethernet.

2. The method of claim 1 wherein 1 master processor module and N slave processor modules in the processor computing unit build a message-centric model based on CORBA middleware; each slave processor module calculates the actual running time value of the current whole program in real time and stores the maximum value of the current program running time value; each slave processor module can issue the number of the slave processor module, the number of the control processes being executed by the slave processor module and the maximum value of the current program running time value, and the master processor module can access the number of each slave processor module, the number of the control processes being executed and the maximum value of the current program running time value through a synchronous updating mechanism of CORBA middleware.

3. The method of claim 2, wherein step S ₂ In the method, a main processor module in a processor computing unit implements load balancing management on all slave processor modules according to the following steps for a received control flow:

S ₂₁ : the number of the terminal devices is M which is more than or equal to 1, so thatThe maximum value of the number of the control processes sent by the console equipment is M; assuming that only 1 master processor module and 1 slave processor module are arranged in the processor computing unit, the maximum value T of the actual running time of the slave processor module program when the console device sends 1 to M control flows is respectively computed through testing _i ，1≤i≤M；

S ₂₂ : assume that the execution cycle of the program in each slave processor module is T _Execute ；

When T is _M <T _Execute When the flow is in the first control flow, the flow indicates that M control flows can be correctly executed on 1 slave processor module; when the master processor module allocates a control flow to the slave processor module each time, reading the maximum value of the current program running time of each slave processor module, finding out the serial number of the slave processor module with the minimum running time in the maximum values, and allocating the current control flow to the slave processor module; when the slave processor module executes a new control flow, adding 1 to the number of the control flows which are executed, and updating and storing the maximum value of the current program running time value of the slave processor module and the number of the control flows which are executed by the slave processor module in real time; if the execution period T of each slave processor module program _Execute Equal to the running time of the program, the main processor module also follows T _M <T _Execute The same method is used to execute the distribution of the control flow under the same condition;

when T is _M ≥T _Execute When the M control flows are executed, the M control flows cannot be correctly executed on one slave processor module; according to step S ₂₁ T obtained in _i Value, calculate when T _i ≥T _Execute The minimum value of time i is the maximum value of the number of control flows which can be simultaneously and correctly executed by one slave processor module and is marked as C _max Then executing step S ₂₃ ；

S ₂₃ : when the master processor module distributes the control flow to the slave processor module each time, the number of the control flows which are executed by each slave processor module is read and recorded as C _i If C is _i <C _max Then execute step S ₂₄ (ii) a If C _i Is equal to C _max Then, thenComparison C _i+1 And C _max Size of (2), if C _i+1 <C _max Then execute step S ₂₄ (ii) a If all C are _i Value equal to C _max If the slave processor module is not capable of executing the control flow sent by the current console device, the master processor module sends 'no available slave processor module' information to the console device through the Ethernet;

S ₂₄ : distributing the control flow to the slave processor module; when the slave processor module executes a new control flow, the number of the control flows which are executed is added with 1, and the maximum value of the current program running time value of the slave processor module and the number of the control flows which are executed by the slave processor module are updated and stored in real time.

4. The method of claim 3, wherein step S ₂ When the execution of one control flow of the slave processor module is finished, subtracting 1 from the number of the control flows which are executed; if all the control flows responsible by the slave processor module are executed, clearing the number of the control flows being executed and the maximum value of the program running time; then, the maximum value of the current program running time value of the module and the number of the control flows being executed by the module are updated and saved in real time.

5. The method of claim 4, wherein step S ₂ The processor computing unit analyzes the control flow sent by the console device by adopting the following method:

the message format of the control flow sent by the console equipment consists of an instruction identifier, a source identifier, a sink identifier, a plus instruction sequence number, an instruction length, an instruction number, an instruction 1 type, an instruction 1 interval time, an instruction 1 effective field content, a plus instruction 2 type, an instruction 2 interval time, an instruction 2 effective field content, a plus 8230, an instruction k type, an instruction k interval time, an instruction k effective field content, a plus checksum; k is an integer greater than 2;

a processor module in the processor computing unit judges whether the length and the checksum of the current control flow are correct or not according to the message format; the content of the last word of the current control flow is a checksum, and the checksum is calculated according to a CRC16 check algorithm from the content of other fields except the 'checksum' field in the current control flow; if the length of the control flow and the checksum are correct, respectively extracting the effective field content of each control instruction in the control flow from the current control flow according to the number of the instructions, recording the effective field content as the effective data content of the control instruction of the console equipment, and storing the instruction interval time of each control instruction; the processor computing unit sends the effective data content of each control instruction in the control flow of the console equipment to the terminal equipment according to the instruction interval time of each control instruction; after the processor computing unit receives command feedback data of the terminal equipment through different interfaces, a processor module in the processor computing unit extracts content of a data field in the command feedback data as message content, computes the length of effective data in the content and a checksum, wherein the computation method of the checksum is the same as that of the checksum of the console equipment control command, then generates a console equipment feedback command according to the format of command identification + information source identification + information sink identification + instruction sequence + instruction length + instruction effective field content + checksum, and sends the generated console equipment feedback command to the console equipment through Ethernet.

6. The method of claim 1, wherein step S ₃ And the terminal equipment sends the self-generated command feedback data to the processor computing unit through one of an RS-422A interface, a GJB289A interface, a switching value interface and a CAN bus interface.

7. The method of claim 5, wherein step S ₄ In the method, the processor computing unit sends the generated console device feedback instruction to the console device through the ethernet by adopting the following method:

the processor module in the processor computing unit extracts the content of a data field in the command feedback data of the terminal equipment as message content, computes the length and the checksum of effective data in the content, the computation method of the checksum is the same as the computation method of the control instruction checksum of the console equipment, generates a console equipment feedback instruction according to the format of instruction identification plus information source identification plus information sink identification plus instruction sequence number plus instruction length plus instruction effective field content plus checksum, and sends the generated console equipment feedback instruction to the console equipment through the Ethernet.

8. The method of claim 1, wherein the valid field content of the instruction is a positive integer multiple of 4.

9. Use of a method according to any of claims 1 to 8 in the field of embedded system communication technology.

10. Use of a method according to any one of claims 1 to 8 in the field of redundancy management.

Images