DE2721805A1 - CENTRAL CONTROL, IN PARTICULAR FOR REMOTE COMMUNICATION SYSTEMS - Google Patents

Description

Central control, especially for telecommunications switching systems

The invention relates to a central controller, in particular for telecommunications switching systems, with at least two computers with a stored program linked to one another according to the load sharing principle, each of which has at least one a status register and a computer link register having monitoring unit is assigned, wherein the Monitoring units of the two computers by means of two connecting devices are coupled together which correspond to one of the number of signals to be transmitted Number of transmission lines connected to each other.

Such a central control for switching systems, stored program computers on the principle of load or the division of labor interconnected with at least two has been described several times C ^v SSL · FR-PS 1 603 3 ^ 1 (* and magazine "Electrical Communication" 47 (1972) 3, p.157 to 1613. In this multi-computer arrangement, each of the at least two computers simultaneously performs part of the functions of the system to be controlled, for example the coupling network of a telephone exchange. This control technology has a number of advantages. Computer arrangement is much more reliable than single-computer controls. In addition, it reacts more favorably to programming errors, since the various computers each control different peripheral functions and the probability of simultaneous failures is very low. The control is also characterized by high overload capacity If a computer fails, practice the computer that remains in operation takes its functions. To do this without affecting the controlled

(*) S T) T-OS 15 7lf 597

709849/0799

M.A. Longatte et al 3-2-3

To be able to do the installation, the computer in operation must be controlled by the other computer before it failed Workflows are kept up to date. The central control computers therefore exchange messages about the changes made by each of them in the controlled system. It is also necessary that each computer is informed about the status of the other computer or computers in order to be able to make a decision to be able to decide whether he can continue to work in partial load operation or whether he has to take over the entire control work.

Each of the computers mentioned has a monitoring unit assigned, which on the one hand has a status register (or status unit) in which some both the status of the assigned as well as data relating to the state of the other computer are stored, and on the other hand a computer link register (or link unit) into which the messages generated by the assigned computer are storable. The two monitoring units are aware of the aforementioned messages and the status data from a connecting device transmitting to the other computer connected with each other.

Under normal conditions, the two computers are largely operated independently of each other, i.e. they only exchange messages to the extent necessary to avoid conflicts prevent to check the circuits connecting them together and so that every computer is able to do the To take over the control work of the other. As soon as an error occurs, the computers become independent of one another: the faulty computer is operationally separated from the controlled network and from the other computer. He loads the Test programs to be carried out to determine the origin or cause of the error. In most

709849/0799

In cases in which an error occurs, the faulty computer is easy to identify (usually it is the one who detected the error), but this is no longer the case when the error occurs in the computer link devices. In such a case , it is better to first determine which computer is the starting point for the error and then to determine the defective part in this computer.

The invention has for its object to provide a central control system of the type mentioned, which is provided with a computer linking device which makes it possible in conjunction with a monitoring unit of a computer, to check the output from the computer signals.

According to the invention, this object is achieved in that the connecting devices are provided with devices for feedback of the signals they send and with receivers receiving the signals from the respective other connecting device when the first test signal becomes effective , and that, when the two computers are in normal operation, the Output signals of the receiver and, when a second test signal is present, the feedback signals receiving devices .

In the event of an error occurring in the area of the link between the two computers , the defective computer can be determined by a method according to the invention , through which the incorrect data and the identification of the computer are entered in the status registers of the two computers when an error occurs (1.) who has determined this data , and from this computer a first sub-

709B49 / 0799

- 7 MALongatte et al 3-2-3 ^ 'C I 0 Uu

refraction signal is transmitted to the other, second computer, through which, upon receipt of this first signal, the successive transmission of several stored in the computer link register representing a message Words is carried out; that (2.) then a second interrupt signal from the other computer the computer that found the error, and a first test signal that blocks its own receiver as well as generating a second test signal which feeds back the own transmitter outputs; that (3 ·) from the calculator that has detected the error, likewise a first test signal blocking its receiver and one of its transmitter outputs feedback second test signal generated and some test runs from both computers, in which signals via the fed back transmitter and compared the received with the transmitted signals, executed and, if an error is detected during the test runs, a diagnostic flip-flop is transferred to a predetermined state and that (4.) at the end of the test runs and upon receipt of a third interrupt signal sent by the second computer the first and the second test signals are suppressed in both computers, in the case of the specified State of the diagnostic flip-flop of one of the computers of this computer is stopped and the other computer has the Control work is taken over alone and when the diagnostic flip-flops are not in the specified state both computers by the computer that detected the error, which sent the signal for which an error was detected the receiving computer is stopped.

The advantages of the invention are, among other things, that it is possible with it after a computer has been found to be faulty has been determined to be operationally separated from the other computer or computers and thereby the

709849/0799 - / -

Linking device to facilitate maintenance work to be carried out.

An embodiment of the invention is explained below with reference to the drawing. Show it:

Fig.l a central control according to the invention with two computers connected to one another by their linking devices, in a block diagram representation,

2 shows the linking devices of the two computers the control according to Fig.l and

3 and 4, the operations of the two computers to be executed in the controller according to Fig.l monitoring method according to the invention in flow chart representation of _i.

The central control of a telecommunications switching system, shown in simplified form in FIG. 1, is provided with two memory-programmed computers A and B that work together according to the load sharing principle. The computers A and B each have a central unit or central processing unit IA, IB with an associated memory 2A, 2B . The central units IA, IB are each assigned a monitoring unit 3A, 3B and they are connected to one another via connecting or linking devices 4A, 4B. As already mentioned, the monitoring units assigned to the central computers each have a status unit or status register storing the computer data and a computer link register in which the messages transmitted to the other computer via the connecting device are stored. These data messages or messages consist of four data words,

- 9 M.A. Longatte et al 3-2-3

which are able to inform the other computer about the progressing functional processes. Thus, each computer not only knows the changes that it has introduced in the system, but also the changes introduced by the other computer. The transmitted messages contain redundant data that enable the receiving computer to check the transmission quality. In the exemplary embodiment described, a message consists of four words transmitted one after the other, the last word representing a test coding of the three preceding words. The data are exchanged between the connection devices A and 4b in question-and-answer mode, over as many connection lines as there are signals to be transmitted. Simplified it can be assumed that each monitoring unit request signals RQA or RQB via the connecting devices ¹ IA, 4B transmits to the other monitoring unit and receives response or data signals DATA or DATB. The request or data signals are transmitted with the aid of a transmission circuit, which is part of the connection device assigned to a computer, and received with the aid of a reception circuit, which is part of the connection device of the other computer.

If an error is found in the signal transmission between the connecting devices, this can Errors can be due to a failure on either the sending or the receiving side. It must now by Test programs are carried out, the faulty computer is determined in order to be able to disconnect it. According to the invention this is carried out with the device shown in FIG. Since the central control is symmetrical corresponds to each 'sent by the computer A and from the computer B received a signal received by the computer B

709849/0799

- 10 M.A. Longatte et al 3-2-3 2 / Z10 Ub

sends and received by the computer A equivalent signal. Accordingly, FIG. 2 shows a request signal RQA and a data signal DATA, which originate from computer A, as well as corresponding signals RQB and DATA originating from computer B. Each signal sent by the connecting device of a computer Hk or HB is transmitted not only to the receiving circuit in the connecting device of the other computer but also to an additional receiving circuit provided in the connecting device of the sending computer. The signal RQA emitted by a transmission circuit 50 is transmitted to a ^{receiving circuit 60 contained in the connecting device 1} IB of the computer B and to a further receiving circuit 51 which is part of the connecting device A of the computer A. The output signal of the receiving circuit 51 is applied to an AND logic element 53, which also receives an activation signal TSTA. An OR gate 5H receives both the output of gate 53 and the output signal of a receiving circuit 52 to which a signal coming from the connecting device HB request signal RQB is transmitted. The output of the logic element 5H corresponds either to the signal RQA or the signal RQB in the event that the AND logic element / is activated. The same circuit arrangement is used to generate the request signal RQB or the data signals DATA and DATB. It should be noted, however, that when the output of the receiving circuits 51 and 56 is activated by the signal TSTA, the receiving circuits 52 and 57 must be blocked. This is accomplished in the connection device Hk by a signal RPDA and in the connection device HB by a signal RPDB. So when the signals TSTA and RPDA are present, the computer A is operationally independent of the computer B; the output of the monitoring unit 3A

709849/0799

- 11 M.A. Longatte et al 3-2-3 ^ I O UO

is fed back in a closed loop.

The advantages of such a feedback device for the output signals of the connecting device of a computer are obvious. For one, the computer can have its own circuits, especially those that touch the computer link, check out what makes a simplification troubleshooting and checking circuit boards after manufacture. A related one The advantage lies in the simpler maintenance of a computer belonging to a multi-computer control. Finally, the high speed with which the diagnosis runs in a multi-computer control is particularly advantageous can be carried out as soon as an error has been detected in the computer link device is. The following are the main features of the invention Determination of a defective computer explained. If an error affects both computers A and B, i.e. theirs Linking or connecting devices, this can either be done by the faulty computer or by the faultless one working, but with an error afflicted data receiving computer can be determined. This is for example, the case when an error is found when decoding the data message. Once a Computer detects an error, either its own central unit or the central unit of the other begins Computer running a special program that identifies the computer that is causing the error and enables the necessary measurements to be taken. Assuming computer A detects an error, while he is executing a command relating to computer B (for example, when executing a data transfer command from computer B to computer A), then this is indicated by the states ERA = I and ERB = O. Of the

709849/0799 - / -

M.A. Longatte et al 3-2-3

Computer A then begins a test run, ie it carries out a corresponding program and transmits a signal to computer B in order to also cause it to perform a test run. This signal is called the first test and interrupt signal. When this first interruption signal is received, computer B begins transmitting a data message. This means that the computer B writes four words forming this data message in the registers provided for this purpose in the connection unit of this computer. After the registers are loaded, a special interrupt signal is sent to computer A and then the four words are read out by computer A one after the other. After the four words have been transmitted, a new interrupt signal to the computer A is transmitted from the host B. The computer B then executes a feedback command, ie it activates the outputs of the receiving circuits 61 and 66 by generating the signal TSTB and blocks the receiving circuits 60 and 65 by applying the blocking signal RPDB to them. At the same time , a second test and interrupt signal is transmitted to the computer A, which thus learns that the computer B is carrying out a test run and that it can then execute a feedback command itself. The two computers then carry out test commands and enter the results of these tests in a diagnostic flip-flop DIAG ( not shown in the drawing). These tests consist in the transmission of data via the transmitter circuits 50 and 55 and in checking whether the output signals of the OR ^{gates 5 1} * and 59 correspond to the transmitted data. Let it now be established that DIAGA = I if computer A detects an error on the basis of these tests, and DIAGB = I if computer B does this. The two computers interrupt the test operation as soon as they each come from a third computer

709849/0799 "'"

M.A. Longatte et al 3-2-3

Interrupt and test signal received. After that, each calculator analyzes the accumulated data in the following Way. First of all, computer A is examined:

(1) If the diagnostic flip-flop DIAGA is in state 1, computer A stops itself.

(2) If the diagnostic flip-flop DIAGB is in state 1, computer A alone controls the Switching system continues.

(3) If none of the flip-flops DIAGA and DIAGB are in state 1, computer A examines the state the variables ERA and ERB.

(4) If ERA = I and is the error in one of the computer A signal transmitted to computer B has been detected, computer A stops computer B, otherwise the computer A stops itself.

(5) In the cases in which ERA = O, the following applies:

a) if ERB = I, the computer B hits the final one Decision;

b) if ERB = O, this means that there is no error.

Now let us examine the case of computer B:

I ¹ ) If the diagnostic flip-flop DIAGA is in state 1, computer B takes over the entire control work of the switching system.

709849/0799 "^;"

M.A. Longatte et al 3-2-3

2 ") If the diagnostic flip-flop DIAGB is in the state 1, the computer B is stopped.

3 ¹ ) If none of the flip-flops DIAGA and DIAGB are in state 1, computer B examines the state of the variables ERA and ERB.

4 ') If ERA = I, computer A makes the final decision.

5 ') If ERB = I and is the error in one of the computer B signal transmitted to the computer has been detected, the computer B stops the computer A, otherwise the computer B stops itself.

6 ') If ERB = 0, this indicates that there is no error.

In the analysis carried out by computers A and B. a state is considered impossible in which the two diagnostic flip-flops DIAGA and DIAGB assume the value 1. In fact, as soon as a computer detects an error, it carries out test runs and a diagnosis, that too Include the other computer to determine the cause. So the probability is that on both Computers an error occurs at the same time in the area of the computer link is extremely unlikely.

In the evident from the Figures 3 and ¹ J flowcharts is the order in which the function sequences described above are carried out, is shown. These functional processes are carried out in both computers at the same time and they take very little time. Other side

709849/0798 - / -

M.A. Longatte et al 3-2-3

after completion of these functional processes, the central control is able to control the switching system - either in the Two-computer operation, if no permanent error has been found, or in simple operation, if one of the computers has been found to be faulty.

It can be seen from FIG. 2 that if the feedback loop is closed in one of the connecting devices ΊΑ or B, the receiving circuits (for example) 52 and 57 are not checked. Therefore, the transmission of a data message is carried out before a feedback command is executed in each of the computers. As a result of such a data transmission, a decision can be made as to whether the detected error concerns a signal transmitted from computer A to computer B or a signal transmitted in the opposite direction. In conjunction with the results of the test runs carried out after the feedback, this information makes it possible, if necessary, to determine whether the error relates to a receiving circuit of the connection device UA or the connection device 4b. This fact can also Fig. ¹ "be seen.

709849/0799

Blank page - Ak-

Claims (1)

Patent attorney Stuttgart M.A. Longatte-J.-P.Nicolle-X.Y.C. Penet 3-2-3 INTERNATIONAL STANDARD ELECTRIC CORPORATION, NEW YORK Claims 1. Central control, in particular for telecommunications switching systems, with at least two computers linked to one another according to the load sharing principle with a stored program, each of which is assigned a monitoring unit having at least one status register and a computer link register, the monitoring units of the two computers being coupled together by means of two connecting devices, which are connected to one another with a number of transmission lines corresponding to the number of signals to be transmitted, characterized in that the connecting ^{devices (1} IA, ¹ IB) have devices (51 »56, 6l, 66) for feeding back the signals they send as well as with a first test signal effective blocking devices for the signals from the respective other connecting device (ΊΒ, ¹ IA) receiving receivers (52, 57, 60, 65) are provided, and that in normal operation of the two computers (IA, IB) the output signals All of the receivers (52, 57, 60, 65) and, if a second test signal is present, the feedback signals May 10, 1977
Bk / Mr 709849/0799 ORIGINAL INSPECTED - 2 M.A. Longatte et al 3-2-3 IÖUÖ receiving devices (5 ¹ I, 59, 6 * 1, 69). Method for monitoring the central controller according to claim 1, by which the faulty computer is determined when an error occurs in the area of the link between the two computers, characterized in that (1.) the faulty data and the identification of the in the status registers of the two computers Computer that has determined this data is stored and a first interrupt signal is transmitted from this computer to the other, second computer, by which the successive transmission of a plurality of words representing a message stored in the computer link register is carried out on receipt of this first signal; that (2.) thereafter a second interrupt signal is transmitted from the other computer to the computer which has detected the error and a first test signal blocking its own receiver and a second test signal which feeds back its own transmitter outputs are generated; that (3.) the computer that has detected the error also generates a first test signal blocking its receiver and a second test signal that feeds back its transmitter outputs, and some test runs from both computers in which signals are passed through the feedback transmitter and the received signals are transmitted to the transmitted signals are compared, executed and, if an error is detected in the test runs, a diagnostic flip-flop is transferred to a predetermined state, and that ( ¹ I.) at the end of the test runs and upon receipt of a third sent by the second computer Interrupt signal in both computers, the first and the second test signals are suppressed, if the diagnosis-FlJp-Flop is in the predetermined state, one of the computers of this computer is stopped and the other computer takes over the control work alone 709849/0739 M.A. Longatte et al 3-2-3 and if the diagnostic flip-flops of the two computers are not in the specified state by the Computer that detected the error receiving the signal with which an error was detected Computer is stopped. 709849/0799