GB2100482A - Computer testing device - Google Patents

Abstract

A device (2) for testing a computer (4) is operable to determine the times at which the computer system is unavailable by repeatedly and automatically sending messages to the computer (4), and determining whether a response is received within a predetermined interval. The device (2) preferably decides that the computer system is available in response to an acknowledgement that the device has successfully logged onto the computer. The device (2) then preferably repeatedly sends transaction messages to which a program in the computer (4) responds. If the device (2) fails to receive responses to a predetermined number of transaction messages within a predetermined interval, the device (2) decides that the computer system has become unavailable. The device (2) provides a printed record indicating the times at which the computer system became available and unavailable, and generates an immediate alarm whenever the computer system becomes unavailable or a response to a transaction message is not received within a predetermined interval. <IMAGE>

Description

SPECIFICATION Computer testing device This invention relates to computers, and particularly to devices for monitoring the performance of computers.

The invention is particularly applicable to interactive computer systems which have a main computer linked to a number of userterminals. The performance of such systems depends on a large number of factors, including both the operation of various hardware items such as the devices for controlling communication between the computer and the terminals, and storage devices, etc., and software design.

Because the particular design of each computer installation is unique, generally speaking, it is dif ficu It to ensure that each system operates at its best level for the particular installation where it is used.

The problem is exacerbated by the requirement that the computer activities be shared between a number of userterminals. This not only makes it more complicated to design the system to give good performance, but also means that the performance of the computer for any given user may vary considerably depending upon how the computer is used by the other users.

Much work is therefore necessary' to determine how to adapt andior use any computer system to obtain the best possible performance. Some prior art devices have been proposed for assisting this. These devices generally comprise monitors which are linked either to the main computer or to a user ter- minal. The monitor keeps a record of the activity between the computer and one or more terminals. It is therefore possible to keep a record of how the computer system is in fact operating, and this record provides an indication of what, if any, changes are needed in the system. These monitors are, however, of limited value.

In accordance with the present invention there is provided a computer testing device which is operable repeatedly and automatically to test whether a computer system is available for use by sending messages to the computer and determining whether responses are received within predetermined intervals.

The device would normally take the place of a computerterminal, and instead of merely monitoring the traffic on the communication line between the terminal and the computer, it would automatically initiate responses from the computer to determine whether the computer system is available.

Computers generally have an overall operating system to which users send messages when they first establish communication with (i.e. log onto) the computer. The device can be arranged simply to test this operating system by successively sending logon messages; if a response is received within the predetermined interval the device determines that the system is available, and if not the device determines that the system is not available.

However, more thorough testing can be achieved by actually having a "user program" stored in the computer and repeatedly operating the program in order to get a response from the computer.

In the preferred embodiment, the device determines that the computer system is available when it receives an acknowledgement signal from the computer in response to a logon request. Thereafter, the device repeatedly sends transaction messages each of which causes the program stored in the computer to produce a response from the computer. If the device fails to receive responses to a predetermined number of successive transaction messages within predetermined intervals, the device determines that the computer system has become unavailable.

The computer may have several operating systems, and these may be arranged in a hiererchy so that a user must log onto successive operating systems before gaining access to a program stored in the computer. The device can be arranged to test any of these operating systems, or all of the systems for example by successively operating the stored program.

The device preferably keeps a record (e.g. a printed record) of the availability of the computer system. The device may record the times at which the system becomes available and unavailable, from which the amount of time for which the computer system was unavailable can be determined. The device may be arranged to work out and record the percentage of time, during a predetermined time period, for which the computer system was unavailable.

Unlike the monitors described above, the device of the present invention actively tests the computer and is able to provide a comprehensive record of the availability of the computer throughout a predetermined period, which includes times when the computerwould not normally be in use.

An important feature of the preferred embodiment of the invention is the provision of an alarm which gives an immediate warning whenever the computer system becomes unavailable. The device may also be arranged to give an alarm each time the device fails to receive a response to a transaction message within a predetermined interval. The provision of an alarm enables an operator to take immediate steps to investigate the reason for the computer system becoming unavailable, and/or correct any problems which may have arisen. The warning given by the device can be a visual and/or audible alarm.

The device preferably has means whereby a user can set the interval within which a response from the computer can be received without producing an alarm.

A preferred embodiment of the invention is arranged to measure the time taken for the computer to respond to the messages from the device.

The device can be set so that it sends messages to the computer, and measures the response time, at short intervals when it is desired to diagnose the performance of a particular piece of software, or can be set to send messages at long intervals so as to monitor the computer performance over a period of, for example, several days, without using any significant amount of computer time.

The device preferably stores and/or displays response time information, such as the time taken for each response from the computer, the number of times the response time threshold has been exceeded, the time of day when response time was recorded, etc. The device may print out a record of the computer performance, or display the information in some other way. It may also be arranged to process the information, for example by determining an average response time.

The device may also be arranged to send the response time information to the main computer so thatthe latter can perform any desired processing of the information.

The program stored in the main computer may be a short, trivial one sufficient merely to send an ack nowledgement signal to the device. Preferably how everthe program enables the computer to send data to the device; the data may for example be the date and time of day. The device is capable of accepting and recording this data.

The program in the computer may be more extensive, in which case the device can be used to monitor the overall operation of the program.

An arrangement embodying the invention will now be described by way of example with reference to the accompanying drawings, in which: Figure 1 schematically illustrates a device in accordance with the invention, Figure 2 shows the front panel of the device, and Figure 3 is a flow chart for describing the operation of the device.

The device 2 shown in Figure 1 is linked to a main computer4 via a cable 6. The main computer is also connected with several user terminals (not shown).

The connection between the main computer4 and the device 2, and the connection between the main computer and anyuserterminal, may be a permanent cable link, or a telecommunication link which is set up only when the device or terminal is in use, such as a telephone line.

In the illustrated embodiment, the device 2 comprises a microcomputer8 incorporating a microprocessor, random access memory, a read-only memory storing a program for controlling the operation of the device, and input/output circuits. A suitable device is the Texas Instruments TM990/1 01 M single-board microcomputer. This is supplied with an operating monitor, and suitably programmed EPROM's can be added to achieve appropriate operation of the microcomputer.

The microcomputer 8 has two serial input/output circuits. One of these supplies messages to, and receives messages from, the cable 6 leading to the main computer 4. The other is connected to a printer 16 so that information stored in the microcomputer8 can be displayed and/or printed out.

The microcomputer8 also has parallel input/output circuits which are connected to an interface circuit 10. The circuit 10 is connected to a number of switches, indicated at 12, to enable the switches to be read by the microcomputer 8. The switches are set by a user of the device to control the detailed operation of the device.

The interface circuit 10 is also connected to display circuitry 14.

The preferred embodiment incorporates a program-control' d microprocessor because it is felt that this is the sin pest and most economical way of achieving the device of the invention. However, the use of a microprocessor is not essential.

The device has a housing the front panel of which is shown in Figure 2.

The panel is provided with an ON/OFF switch 18.

An indicator light 19 is illuminated when the switch is on, i.e., when power is applied to the device. A further switch 20 is in the form of a push button and, when pressed, initialises the device so that it is ready for operation. Another indicator light 21 pulses regularlyto show that the microcomputer in the device is operating.

A warning light30 is provided to give an immediate indication when the response time threshold has been exceeded.

An electrically alterable display 32, in this case incorporating L.E.D.'s capable of showing four alphanumeric characters, can provide any one of seven different displays depending upon the setting of a seven-way rotary switch 36. The different displays which can be obtained are as follows: (a) the current setting for the interval between the sending of messages to the computer (b) the current setting of the response time threshold (c) the average measured response time (d) the time taken for the last response from the computer (e) the day number (f) the current hour (g) the current minute.

When any of the first four settings is chosen, the left-most digit of the display 32 displays the charac ter"A", "B", "C"or"D",to indicate which setting has been selected.

The displays provided at settings (e), (f) and (g) represent the current time as stored in the device.

This time, and the settings for the interval between the sending of messages to the computer and the response time threshold can all be altered by selecting the appropriate position of the switch 36, and then operating push button switches 22 and 24.

There are three push button switches 22, each associated with a respective one of the three rightmost digits of the display 32. Each time one of the push button switches 22 is pressed, the corresponding digit in the display is stepped upwardly once.

After the push button switches 22 have been operated to obtain the desired setting on the display, this is entered into the device by operating the push button switch 24.

Preferably, the average response time is determined by taking the times recorded for the last predetermined number (e.g. ten) of messages sent to the computer.

The device also has a four-way rotary switch 38 which controls the external printer 16 which can be connected to the housing of the device. The four settings of the switch correspond to: (a) no printer operation (b) the printing of the measured response time for the last response received from the computer (c) printing of the average of a number of measured response times (d) the printing of all information appearing on the line between the device and the computer.

Preferably, when printing either the average response time or the last response time, the printer also prints a histogram to provide a graphical representation of the measured response of the computer.

Whenever the printer is in operation, irrespective of which of the settings (b), (c) or (d) is selected, the printer also prints the time at which the main computer becomes available and unavailable. Also, after regular, predetermined intervals, the printer prints out the percentage of time for which the computer has been available (or unavailable) during the preceding interval. If desired, a switch could be provided for setting these intervals, but in the illustrated embodiment the intervals are fixed.

The device also includes a switch 40 for selectively enabling or disabling an audible alarm which, if enabled, gives an indication of when a preset response time threshold has been exceeded. The numeral 44 represents a horn forming the audible alarm.

An indicator light 46 is illuminated when traffic is on the line between the device and the computer.

On the rear (not shown) of the housing there are switches for selecting the rate of data transmission to the computer and the printer and the type of parity to be used in communication. There are also ports for communication with the printer and the main computer, and a socket for the connection of an external alarm device, if desired.

The device may also be provided with a test mode selection switch which, when operated, allows further switches to cause the display 32 to show the number of times the response time threshold has been exceeded since the last initialisation of the device, and the number of incorrect characters which have been detected on the line between the device and the computer.

The operation of the device will now be described with reference to the flow chart of Figure 3. The flow chart is intended to provide a broad description of the overall operation of the device. The details of the way in which each step in the operation is carried out are not shown because these details are not necessary for an understanding of the operation of the device, and can be implemented in a variety of ways.

The flow chart describes only the main operation of the device, and does not describe various necessary ancillary operations such as the reading of the switches of the device, the operation of the timers in the device, etc., because these operations can also be carried out in a variety of ways and their detailed operation is not essential to an understanding of the invention.

The first step 100 in the flow chart is the initialisation of the microcomputer8, which is carried out following an operation of the switch 20. During this initialisation, various counters and registers are cleared, and flags set up, in the memory of the microcomputer8 in accordance with both the program stored in the microcomputer and the setting of the various switches of the device.

The program then proceeds to point A, following which, in step 102, the device establ!shes communication with the main computer4 by transmitting a "LOGON" message along the line 6. This message serves to identify the device to the computer4, and the computer transmits an acknowledgement of the message along the line 6.

In step 102 the device also sets up, in a register, an error delay time which corresponds to the maximum expected time for a response from the main computer4.

The operation then proceeds to step 106, in which the device determines whether the error delay time has expired, and if not the operation then proceeds to step 108 to determine whether a character has been received from the computer. If not, the program determining the operation of the device loops back to step 106, in which it is again determined whether the error delay time has expired.

The program continues until either the error delay time has expired, in which case the program proceeds as indicated at B to an error routine to be described later, or a character is received, in which case the program proceeds to step 110.

In step 110, an input buffer in the inputioutput circuit 10 is cleared to permit receipt of responses from the main computer4. The program also arranges for the printer to print-out an indication that the device 2 has established communication with the main computer 4, i.e. that the device has "logged on" to the computer. The printer also prints the time at which this has occurred. Also, the program starts a timer, whose contents represent the time for which the main computer is available.

The program then proceeds, via point C, to step 112. At this stage, a transaction message is sent to the main computer4 via the line 6. The transaction message may be of any desired form so long as it is recognisable by the main computer4, and may for example merely consist of a few characters. The program may arrange for the printer to print the transaction message during step 112.

The program then proceeds to step 114, in which a timer for timing the response of the computer is started.

Thereafter, at step 116, an error delay is set up; the error delay corresponds to the maximum time within which a response should be received from the computer4.

At the next step,118, the device decides whether the error delay has been exceeded, and if so proceeds to an error routine to be described later as indicated at D. Otherwise, the program proceeds to step 120, where it is determined whether a character has been received from the computer 4. If not, the program loops back to step 118.

On detection of a character, the program proceeds to step 122. At this step, the program looks at the character and determines whether or not it is a correct one. For example, the main computer may be programmed to respond to the device with a first character "$", and thereaftertransmitthe date and the time of day. If, at step 122, it is found that the first character received is not "$", the program loops back to step 116, whereupon the error delay is again set up.

Assuming that the correct character was received, the program proceeds to step 124, whereupon the device stores the time indicated by the timer which was started in step 114.

The program then proceeds to wait for further data from the main computer4. In step 126, another error delay time is set up, and in the following step,128, the program determines whether this error delay time has been exceeded, and if not proceeds to step 130 to determine whether or not a further character has been received from the main computer 4. If not, the program loops back to step 128, and continues until either the error time has been exceeded in which case the program proceeds as indicated at D to the above-mentioned error routine, or a character is detected, in which case the program proceeds to step 132, whereupon the character is stored.

The main computer is arranged to terminate its messages to the device with a carriage return and a line feed. The program therefore includes a step 134 to determine whether these have been received, and if notto loop back to step 128 to waitforthe next character.

Once the message to the device from the computer has finished, the program proceeds to step 136.

At this step, the device sends another message to the main computer4, this message comprising the measured and stored response time. The main computer 4 can then, if desired, process this information.

The device can also, in step 136, display the measured response time, update the average response time, print-out the response time or average response time, and clear any registers, flags, etc. to enable a subsequent response time measurement The program also sets up an interval delay time in accordance with a previous setting by a user operating the switches 22,24 and 36.

The program then proceeds to step 140, whereupon the program continues to loop until the interval delay time has expired, following which the program proceeds to point C, and then step 112, to commence a further response time measurement.

If, after sending the "LOGON" message to the main computer4, the device does not receive an acknowledgement within the set error delay time, the program proceeds as indicated at B to step 150.

Here, the device turns on the alarm light30, and, if enabled, the horn 44. The device can also, if desired, increment a displayable count which represents the number of times the device has failed to communicate properly with the computer 4.

The device also sets up an interval delay time, and then proceeds to step 152 where the program loops until this interval delay time has expired. The program then proceeds to point A, following which the device again sends the "LOGON" message to the main computer 4.

If, either at step 118 or at step 128, the program proceeds as indicated at D due to the failure to receive a correct response from the main computer4 within a predetermined time, the program proceeds to step 153. At this step, the device increments the counter representing the number of times a correct response has not been received within a predetermined interval, tu ns on the alarm light 30, and, if it is enabled, turns un the horn 44.

The fact that the program has reached step 153 could be simply a result of a slow response from the main computer, or itcould be a resu It of the computer system becoming unavailable. To determine this, the program counts the number of successive times that the computer has failed to receive a correct response within the predetermined interval, and if this "fail count" exceeds ten the program decides that the computer system has become unavailable.

Accordingly the "fail count" is incremented at step 153.

The program then decides at step 154 whether the "fail count" is equal to one. If so, the program proceeds to step 155, where the current time is stored. If there are more successive failures to receive a response from the main computer and the program decides that the computer system is unavailable, then this current time will represent the time at which the computer system became unavailable.

The program also stores the contents of the "availability timer" which was started at step 110.

The program then proceeds to step 156, where the device decides whether the fail count exceeds 10, and if not the program proceeds to point C whereupon the device again transmits its transaction message to the computerto try once more to obtain a response within the predetermined interval.

If, however, there have been more than ten unsuccessful attempts to obtain a response within the predetermined interval, the program proceeds from step 156 to step 157. Here, the program prints out the fact that the computer system is unavailable, and prints out the time which was stored at step 155. The value of the availability timer which was stored at step 155 is added to a register which accumulates the total time for which the computer system was available over a predetermined period.

The program then proceeds to point A, whereupon the device again sends the "LOGON" message to the main computer4.

The error delay times which are set up in steps 102,116 and 126 form response time thresholds which, if exceeded, cause the program to enter an error routine from step 106, 118 or 128. A record is kept of the number of times the computer takes so long to respond that a response time threshold is exceeded, and an alarm is given immediately each time this occurs.

Thus, the device is testing the time taken for the computer to (a) acknowledge the "LOGON" message, (b) initiate a response to the transaction message, and (c) transmit subsequent characters in response to the transaction message. In the illustrated embodiment, the device uses the same response time threshold for all these procedures, and this response time threshold is set up by the operator using the switches 22,24 and 36.

However, it is not necessary for the device to use the same response time threshold for all three testing steps. Some or all of the thresholds may be set to fixed levels, or the device may be arranged so that the operator can individually set different thresholds.

Furthermore, it is not necessary for all three of these computer responses to be tested. For example, the computer's response to the transaction message may be buffered before transmission, so that there should not normally be any significant interval between the characters in the response, unless there are any faults in the system. Therefore it may be desirable to omit the step of testing the intervals between successive characters in the computer's response, and comparing these intervals with a threshold.

If desired, the time taken for all the abovementioned three types of response can be compared with a threshold, as described above, but the device can be arranged to provide an alarm and increment an error count only when the acknowledgement of the "LOGON" message and/or the start of the response to the transaction message exceeds the threshold.

In the described arrangement, every time a response time threshold is exceeded the program loops back either to re-establish communication with the main computer by sending the "LOGON message", orto initiate another response time measurement by sending the transaction message.

This, however, is not essential, and instead the program may proceed with the response time measurement even if a threshold is exceeded. In this case, the program may, for example, determine at step 136 whether any threshold has been exceeded, and if so provide an alarm.

The error routines entered as indicated at B and D may be retained, and arranged so that they are entered only if the delay in receiving a correct response from the main computer4 is substantially greater than the response time threshold.

The program is arranged so that each time it transmits a character or receives a character via the line 6, the light 46 is illuminated, to provide an indication that the device is working.

If desired, while the device is waiting for a response to the transaction message, the program can be arranged to increment a "bad character" count each time the device receives a character which is not the character "$" indicating the beginning of the computer's response. The count can also be incremented if the device, when waiting for the end of the response, receives a carriage return which is not followed by a line feed.

The "bad character" count can be displayed using the test mode selection switch mentioned above.

Therefore if, for example, the device repeatedly gives alarms to indicate that a response from the computer has not been received within a sufficiently short time, the operator can look at the "bad character" count, and if this is high it will provide a strong indication that the computer is in fact responding sufficiently quickly, but that the responses are being corrupted, possibly by transmission faults.

In the preferred embodiment of the invention the microcomputer8 is programmed to carry out a plurality of programs in accordance with predetermined priorities. For example, the operation described above would form a main program. There would also be another program, having a lower priority, which would be used to print out, at the end of a predetermined period, the percentage of time during that period forwhich the computer was unavailable, using the contents of the register which is updated at step 157.

Another lower priority program would read the switches of the device and store the settings of the switches for use by the main program. These lower-priority programs would be entered when the main program is at a stage where no action needs to be taken, for example in a delay loop.

There would also be a further program for incrementing a timer at predetermined intervals (of for example .1 seconds). This further program would have the highest priority, and thereby would be able to interrupt any other programs, to ensure that the time is kept accurately.

There may also be a separate program for installation purposes, which runs only once when the device is installed, another separate program for initialising the device, and a further program having the lowest priority which, when no other program is running, checks the read-only memory of the microcomputer 8.

The design of several programs for the device, as well as the necessary hardware, is well within the scope of anyone familiar with the art.

The device can print out each response from the computer, and thus if the response contains the date and time the printed record will indicate when each response measurement occurred. A similar result could be achieved by arranging for the device to print out the day and time stored in the device, rather than the data from the computer.

The interval between response time measurements can be set to O to 999 seconds and the response time threshold can be set to 0 to 99 seconds. This, however, can of course be varied if desired.

m The device is of considerable value both for immediate checking and modification of computer hardware and/or software, and for long-term monitoring of a computer system. The ability to set the response time measurement rate as desired and the provision of an immediate alarm when the response time threshold has been exceeded is of particular value for the first type of testing, whereas the ability of the device automatically to initiate response time measurements whenever desired is of advantage for the second type of testing. Unlike known, passive monitors, the device does not merely present a record of computer performance under existing conditions, but actively tests the computer at times determined by the device itself.

Installation of the device is very simple. The device may simply be connected temporarily in place of a user terminal until monitoring of the main computer has been completed, at which time the user terminal is re-connected.

Claims (28)

1. A computer testing device which is operable repeatedly and automatically to test whether a computer system is available for use by sending messages to the computer and determining whether responses are received within predetermined intervals.

2. A device as claimed in claim 1, operable to keep a record indicating the amount of time for which the computer system has been unavailable.

3. A device as claimed in claim 2, wherein the record indicates the times between which the computer system was unavailable.

4. A device as claimed in claim 2 or claim 3, wherein the record indicates the amount of time for which the computer system has been unavailable during a predetermined time period.

5. A device as claimed in claim 4, wherein the record indicates the percentage of time for which the computer system has been unavailable during said predetermined time period.

6. A device as claimed in any one of claims 2 to 5, wherein the record is a printed record.

7. A device as claimed in any preceding claim, operable to send to the computer a message for logging onto the computer system, and to determine, in response to an acknowledgement of the logon request from the computer within a first predetermined interval, that the computer system is available.

8. A device as claimed in claim 7, wherein the computer stores a program enabling the computer to recognise and respond to a transaction message from the device, the device being operable repeatedly to send transaction messages to the computer and to determine that the computer system is not available in response to a predetermined number of successive failures to receive responses to transaction messages within a second predetermined interval.

9. A device as claimed in claim 8 having means permitting a user to set said first and second predetermined intervals.

10. A device as claimed in claim 8 or 9, wherein said first and second predetermined intervals are equal.

11. A device as claimed in claim 7, operable repeatedly to send messages for logging onto the computer system at predetermined intervals, and to determine that the computer system is available if a response to a logon message is received within said first predetermined interval, and that the computer system is not available if a response is not received within said first predetermined interval.

12. A device as claimed in any preceding claim, operable to provide an immediate alarm upon determining that the computer system is not available.

13. Apparatus as claimed in claim 12 when dependent upon any one of claims 8 to 10, operable to provide an alarm in response to a failure of the computerto respond to any said transaction mes sage within said second predetermined interval.

14. A device as claimed in any one of claims 8 to 10 or 13, operable to measure the time taken for the computerto respond to atransaction message.

15. A device as claimed in any preceding claim, having means whereby a user can vary the intervals at which messages are sent to the computer.

16. A device for testing a computer, substantially as herein described with reference to the accompanying drawing

17. A device t r testing the operation of a computer, the device being automatically operable to send to the computer a message for initiating a response from the computer, and being operable to determine the time taken for the computer to respond.

18. A device as claimed in claim 17, automatically operable successively to send messages to the computer, and operable to measure the time taken for the computerto respond to each message.

19. A device as claimed in claim 18, having means whereby a user can vary the intervals at which the messages are sent to the computer.

20. A device as claimed in any one of claims 17 to 19, operable to determine whether the time taken to receive a response to a message exceeds a predetermined response time threshold.

21. A device as claimed in claim 20, having means whereby a user can vary the response time threshold.

22. A device as claimed in claim 20 or claim 21, having means for providing an immediate alarm when the response time threshold is exceeded.

23. A device as claimed in any one of claims 20 to 22, operable to count the number of times the response time threshold is exceeded.

24. A device as claimed in any one of claims 17 to 23, operable to transmit to the computer information relating to the measured response time.

25. A device as claimed in any one of claims 17 to 24, operable to compute an average of a number of measured response times.

26. A device as claimed in any one of claims 17 to 25, operable to provide a record of information relat ingto measured response times.

27. A device as claimed in claim 26, including means for printing the record.

28. A device as claimed in any one of claims 17 to 27, including electrically alterable display means for displaying information relating to measured response times.