RU2738730C1 - Method for fault-tolerant functioning of prospective complex for automation of military command posts and device implementing thereof - Google Patents

Abstract

FIELD: computer equipment.

SUBSTANCE: invention relates to computer engineering. Technical result is achieved by sorting tasks according to importance, monitoring the number of free computers at the time of new information and computing tasks, distribution of tasks on computers, according to degree of importance, as well as taking into account the possibility of reserving calculations at the current moment in time; by means of each computing machine the decision of the accepted design task of the computer system; by means of device for majority selection of product of majority comparison of results of solving information-calculation tasks coming from computers; issuing correct answers to a device for outputting and transmitting an error record when calculating failures.

EFFECT: technical result consists in providing fault tolerance in the operation of a complex of means of automating military command posts and improving reliability of calculations in general when processing critical information-calculation tasks using a complex of automation means.

2 cl, 1 dwg

Description

The invention relates to the means of: ensuring the fault-tolerant functioning of the complex of automation tools for military command posts, detecting failures, correcting errors and monitoring, and can be used to modernize the complex of automation tools at tactical, operational and operational-tactical levels.

The known method and computing system for improving reliability and reliability _:. processing of critical functions of a fault-tolerant computing system (Patent for invention No. 2413975 dated November 17, 2008), when the effect is achieved due to multiple redundancy of paths for the implementation of critical functions, in which the redundancy management processes and the computing process are separated in a real-time loop and partially by hardware. In the capacity of redundant units of the calculator, the so-called computational paths are allocated, consisting of inextricably linked chains of the input interface, processor and output interface. In each of the computational paths, a majority comparison of the signals passed through each path is performed. The ambiguity of the majority comparison of signals in different paths is eliminated by the majority comparison of the results of the comparison of signals. The reliability of the majority control of signals is achieved by statistical processing of a random process of changing the critical parameter, calculating the residual variance and its majority comparison.

The disadvantages of this system include:

1) Deliberate overestimation of the amount of required redundancy due to the combination of several hardware components into one indissoluble unit;

2) Too high amount of computation associated with multilevel majority control combined with statistical processing of path signals;

3) The complexity of the device itself, which, together with the lack of built-in self-control, reduces the reliability of achieving the corresponding technical effect;

4) The impossibility of differentiating the types of failures, failures and errors.

5) Impossibility of unambiguous identification of failed and low-reliable computing modules from the system.

Also known is a device for recording fault signals, comprising: an asynchronous input unit, OR elements, counters, a decoder, a memory unit, triggers, a fault signal input, a "Reset 1" input, a clock input; start input, input input, fault outputs, number of faults outputs, "Ready" output, "Reset 2" input, device blocking output, AND elements, NOT element, AND element (see RF Patent No. 2010313 dated 1990 G06F 11 / 00).

The disadvantage of this device is the low contribution of the registration of signals to the overall reliability of the device, in comparison with redundant systems, as well as the impossibility of organizing a fault-tolerant operation of the system in the event of failure of its computing modules.

The prototype of the invention are: as a method and a computing system for increasing the reliability and reliability of processing critical functions of a fault-tolerant computing system (Patent for invention No. 2413975 dated 11/17/2008), the computing system includes a computing node module, a multifunctional interface module and a module for combining outputs, a method for increasing reliability and reliability of processing includes the principle of independent, but interrelated functioning of systems in the software and hardware space, the principle of independent but interconnected functioning of systems in the Real Time space, and a device for recording fault signals (Patent for invention No. 2010313 dated 06.08.1990), containing block for selecting the last pulse of a burst, a pulse shaper, two flip-flops, two AND elements, a delay element, a first counter, a second counter, a second delay element, an OR element, the third and fourth AND elements, a register block.

The objective of the invention is to develop a method that takes into account all the disadvantages of the known methods, and a device that implements it, providing fault-tolerant operation of the complex of automation tools for military command posts in solving critical problems.

Technical. The result of the invention is to ensure fault tolerance in the operation of a set of automation tools for military command posts and to increase the reliability of calculations in general when processing critical information and computational tasks with a complex of automation tools.

Achievement of the declared technical result and, as a consequence, the solution of the task is ensured: both by majority reservation of computing paths, and by means of a built-in control system implemented in a majority choice and control device, by the fact that a device that ensures fault-tolerant operation of the command post automation complex military purpose, contains a task distribution device, computers, a majority choice and control device, an output device, a failure storage device, so that the request distribution device is connected to computers, computers are connected to a majority choice and control device, a majority choice device and control is connected to the output device and to the device for storing failures, which makes it possible to increase the reliability of calculations in general when processing critical information and computational tasks by a set of automatic tization of the command post.

A diagram of a device that implements a method for ensuring the fault-tolerant operation of a promising complex of automation equipment for military command posts is shown in Fig. one.

The diagram uses the following designations:

1 - task distribution device;

2 - n-th computer;

3 - device of majority choice and control;

4 - output device;

5 - device for storing failures;

A - the line of the flow of tasks, along which we are transmitted, information about the importance of each task, the logic of majoring, the maximum time for calculating each information and calculation task;

B - data transmission line, through which requests are sent from the task distribution device to determine the state of the computer and computational tasks are transmitted to it;

B - feedback line for transferring the state (ready, not ready) of the computer to the task distribution device;

Г - control line, along it from the task distribution device to the majority choice and control device, information is transmitted about which tasks will be solved by which machines. If it is necessary to majorize by multiple computations on one machine - the machine number and the number of recalculations are transmitted along this line;

D - data transmission line, through which the results of calculations are fed from the computer to the device of majority choice and control;

E - data transmission line, through which reliable calculation results are fed to the output device from the device of majority choice and control;

F - line of transmission of error messages, through it from the device of majority choice and control, the number of the computer and the computational task in which the error occurred is fed to the device for storing failures.

The task distribution device 1 is intended for sorting tasks by degree of importance, controlling the number of free machines at the time of the emergence of new information and computational tasks, distributing tasks among machines, according to their degree of importance, and also taking into account the possibility of backing up computations at the current time. The task distribution device 1 is connected to n computers 2, as well as to the majority choice and control device 3.

n-th computer 2. is a computing path capable of independently solving the accepted information-computational problem of the N-th computer system.

The device for majority choice and control 3 is intended for the majority comparison of the results of solving information and computational problems coming from computers 2, issuing correct answers to output device 4 and transferring error records during calculations to the device for storing failures 5.

Units 1, 3, 5 can be made in the form of software and hardware modules on the expansion board of personal electronic computers with the architecture of the X86 type, installed at the corresponding command posts for military purposes.

The operation of the device for ensuring the fault-tolerant functioning of a promising complex of automation equipment for command posts is as follows.

The task distribution device 1 is connected to the majority choice and control device 3, computers 2. In the course of work, the initial data is loaded into the task distribution device 1 along line A: the type of information and calculation problem and its importance, in accordance with which the required degree of majorization is determined, as well as the logic of majoring and the maximum computation time for each information and computational task. The higher the importance of the information and computational problem, the higher the level of majorization is set, the maximum level of majorization corresponds to the number of free computers at the moment in time. Further, the task distribution device 1 on line B sends a request to computers 2, determining which of them are currently busy. Computing machine 2 sends its state along line B. Based on the answers, the type of computation reservation is selected.

If there are less than three computers 2 free, and the importance of the task requires its majorization, then one of the free computers 2 is sent by the task distribution device 1 along line B to one task many times. In this case, information about the number of the computer and the number of backup passes is fed to the device of majority choice and control 3 via line D. Computing machine 2 sends a signal about the end of calculations, as well as the results of calculations on line D to the device of majority choice and control 3. The device of majority choice and control 3 stores in memory the results of calculations received on line D for their subsequent majority comparison according to a predetermined logic “Majority”, for example, “two out of three”, “three out of four”.

If, in response to a request on line B of the task distribution device 1, more than two readiness messages from computers 2 arrive on line B, and the importance of the task requires its majorization, then the task is sent to several computers 2, and their number is determined as the number of free computers. machines at the moment in time, and the importance of the problem. Information about the numbers of computers on which the problem is solved in parallel is fed through line G to the majority choice and control device 3.

If, in response to a request on line B, the task distribution device 1 does not receive a single message about readiness from computers 2 on line C, then the task is written to the queue and waits for the readiness of at least one of the computers 2. When, in response to a request, on line B of the task distribution device 1 arrives via line C a message about readiness from the first released computer 2, then the task distribution device 1 repeatedly submits the task from the queue to this computer 2, taking into account the level of its majorization. Information about the number of the computer on which the problem is being solved and the number of backup passes is fed through line D to the majority choice and control device 3. The majority choice and control device 3 stores in memory the results of calculations received through line D for their subsequent majority comparison according to a predetermined "majority" logic, for example, "two out of three", "three out of four."

The device of majority choice and control 3 in the course of operation receives a signal via line D. about the end of the calculations and the results. computations from computers 2, as well as control signals on line D from the task distribution device 1. After receiving signals about the end of computations from all computers, the results of computations for one task are compared with each other. Majority selection and monitoring device 3 feeds the correct values to output device 4 on line E.

The correct values are chosen by majority according to the majority logic preset in the system, for example, “two out of three”, “three out of four”.

If not all calculation results are equal, then an error is recorded, the number of the computer in which the erroneous results were obtained is remembered, and the problem is recalculated. The recalculation process is carried out until the correct result is obtained, or the task recalculation cycle time does not exceed the specified time. If the specified time for recalculation of the task is exceeded, a signal is sent to the engineering and technical staff for the occurrence of a software error and the failure of the computer 2, along the line Ж this information is written into the failure memory device 5 and then is used to identify faulty blocks.

The device for storing failures 5 receives information about the flow of failures of computers, as well as errors during the execution of specific information and computational tasks from the device of majority choice and control 3. The information includes the number of the failed machine, the time and nature of the failure.

If, after determining the type of information-computational task and its submission by the task distribution device 1 to several computers 2, the computation time of the task on one of the computers exceeds the maximum allowable time, information on which must be recorded in advance in the task distribution device 1 for each information- a computational problem, then this computer is considered to be faulty, the task distribution device 1 sends a signal to the engineering and technical personnel to restore it through the majority choice and control device 3, and the computer itself is excluded from the calculations.

Claims (2)

1. A complex of automation tools for military command posts, containing: a task distribution device, computers, a majority choice and control device, an output device and a failure memory device, while the task distribution device is configured to sort tasks by degree of importance, control the number of free computing machines at the time of the emergence of new information and computational problems, the distribution of tasks among computers, according to the degree of their importance, as well as taking into account the possibility of redundancy of computations at the current time, and is connected to computers, and is also connected by its output to the input of the majority choice device and control, where each computer is connected to the majority voting device and is a computing path capable of independently solving the accepted computational problem of the computing system, and the majority voting and control device is connected to the device output, with a storage device and is made with the possibility of a majority comparison of the results of solving information and computational tasks coming from computers, issuing correct answers to the output device and transferring error records during calculations to the failure memory device. 2. A method of automating military command posts, including the stages at which, by means of a task distribution device: sort tasks by degree of importance, control the number of free computers at the time of the emergence of new information and computational tasks, distribute tasks among computers, according to their degree of importance, and also taking into account the possibility of backing up calculations at the current time; by means of each computer solve the accepted computational problem of the computer system; by means of a majority choice device, a majority comparison is made of the results of solving information and computational problems coming from computers; give correct answers to the output device; and transferring the calculation error records to the fault memory device.

Images