JP5011174B2 - Information device management method - Google Patents

Description

  The present invention relates to management of information equipment that performs processing by user operation.

  Information devices are becoming more sophisticated year by year, and faster computer environments are being installed in many devices, including home appliances and mobile devices. Accordingly, the scale of software is steadily increasing, and it becomes difficult to completely guarantee the reliability of software implementation at the time of shipment as it becomes more complex.

  Software update mechanisms after product shipment are also standardized in many fields, and at the same time, in order to analyze the cause of defects that occurred after product shipment, methods for collecting information on actual machines have been studied and PCs etc. It is being installed. The example of Patent Document 1 is an example of collecting software faults after shipping as a product, and has an object to collect information when software processing abnormally ends due to a fault.

JP 2003-179541 A

  However, the abnormal state of the software includes not only a case where the abnormal state is detected and terminated, but also a freeze that completely falls into a non-responding state. This cannot be detected.

  It is very difficult for a user to distinguish between a normal waiting time of an application / system and a frozen state when it becomes unresponsive, especially in information equipment with a user interface. , Contributing to user stress.

  If it is in a frozen state, it can only be interrupted by forcibly terminating it, but it cannot be distinguished as described above. There is a danger of end. At that time, there is a risk of data loss. In particular, in the case of a device configuration that has only power-off or hardware reset as a means for forced termination, there is also a risk of device damage.

  Although it is possible in some cases to consider the forced termination by the user as an abnormal termination in the example of Patent Document 1, even in such a case, if the user forcibly terminates without distinguishing between freeze and waiting time, the information that can be collected is also mixed Therefore, it becomes difficult to use for the analysis of the freeze state.

  Therefore, for example, a means for requesting the user of the information device to determine whether the normal non-response state is a waiting state or a frozen state is provided. In response to the request, the information device makes an inquiry to the device management server using the event sequence information up to that state as a search condition. Providing users who have made inquiries about the expected waiting time and the possibility of freezing based on the history data of past non-response results reported in the history database managed by the device management server. . In addition, when there is no record in the history database with the same event sequence or when there is not enough storage, the device management server does not indicate any event sequence that has been queried to the client information device to be managed. Request response status monitoring and results reporting. As a result, a history of results for event sequences that are likely to cause no response is collected, and accurate waiting information can be responded to inquiries from clients regarding similar operations in the future. Furthermore, by narrowing down the event sequence that causes the freeze state, more detailed profile information is also acquired when monitoring the state of the client, which is useful for solving the freeze state that is a defect.

  According to the above means, for example, in an information device having a user interface, user stress is alleviated by supporting determination of a waiting state and a freeze state. In addition, by suppressing easy forced termination by the user in the waiting state, the risk of equipment damage and data loss is avoided. In addition, the device management server statistically manages the discrimination between the device waiting state and the freeze state that occurred in all the information devices to be managed, improves the accuracy of the collected information about the freeze state, and promotes failure analysis.

  In an information device having a user interface, it is a great user stress that the waiting time at the time of executing the application cannot be distinguished from the frozen state at the time of abnormality. Moreover, even in a normal waiting state, there is a risk that the user may be misidentified as a frozen state and forcibly terminated. By providing the user with a means for supporting both judgments, it is possible to reduce user stress and avoid dangers such as device damage and data loss due to forced termination. By requesting from the device management server, it is possible to collect more detailed information to clarify the cause of the freeze state by strengthening the collection of profile information in the information device to be managed when an event sequence that reaches the freeze state occurs .

  An example (example) of an embodiment of the present invention will be described in detail with reference to the drawings.

  FIG. 1 shows a configuration example of a system that realizes the present embodiment. This embodiment is configured by a device management server 101 and information devices 110, 120, and 130 managed by the device management server 101, which are connected through a network 103. In practice, it is not limited to the three units shown in the drawing, and is configured by more managed information devices. In addition, although it demonstrates by the system containing a terminal and a server, it does not prevent the right acquisition and exercise of rights with respect to not a system but a terminal unit or a server unit.

  The management server 101 is connected to the history database 102 for the purpose of managing the connected information devices 110, 120, and 130. The information device 110 is assumed to be any device that performs information processing by user operation, and includes a CPU 111, a communication unit 112 connected to the network 103, a main storage device 113, and a secondary storage device 114 as shown in FIG. And an output device 115 such as a display for outputting, and an input device 116 such as a switch / keyboard / mouse / touch panel. Information devices 120 and 130 and other information devices to be managed by the device management server 101 also have the same configuration and functions as the information device 110.

  An example of processing in the system configuration of FIG. 1 will be described using the processing flow of FIG. In the flow of FIG. 2, the processing on the management server 101 is expressed as step 20x, the processing on the information device 110 is expressed as step 22x, and the processing on the information devices 120 and 130 other management target information devices is expressed as step 24x. The flow of each process is indicated by a solid line, and the communication between the processes is indicated by a broken line.

  As a premise, in the information device 110, a fixed number of event sequences such as input by the input device 116 and application / system state transition are always accumulated and held in the main storage device 113 or the secondary storage device 114. The data format of the held event sequence is illustrated in FIG. As shown in this example, the event sequence is a sequence of events such as the start of application A that occurred, selection of menu B, and button C pressed, added each time a new event occurs, Discard the buffer count from the overflow event.

  The flow is started when an application is running on the information device 110 and the application falls into a state of no response to the user.

(1) In the information device 110, the user issues a determination request for both states using the input device 116. This request is transmitted to the management server 101 via the network 103 via the CPU 111 / communication unit 112. In accordance with this request, the event sequence is also transmitted to the management server 101. (Step 201)
This request is made by the input device 116. In the multitasking environment, when only the application is not responding, it can be realized by a normal input event.However, when there is no response with the OS, NMI (Non-Maskable The accuracy of request issuance can be improved by realizing this request with interrupts that do not go through the OS, such as Interrupt.

(2) The information device 101 monitors the behavior of an application or system that falls into a no-response state after transmission. (Step 202)
(3) The device management server 101 receives the request from the user in step 201 together with the event sequence. In response to the request in step 201, the history of the non-response state caused by the same event sequence on the history database 102 is searched. (Steps 221 and 222)
The search procedure in step 222 is detailed.

  An example of the record configuration of the history database 102 is shown in FIG. In the history database 102, the result of accumulating the result of the non-response state generated from the same event sequence and the time to the result in association with the event sequence 401, 411 that occurred in the information device 110, 120, 130, etc. that occurred in the past The history tables 402 and 412 are held. The history database 102 can store a large number of pairs of event sequences and result history tables, not limited to the two sets shown in the figure, and can be added as needed in the course of this flow as will be described later.

  At the time of the search for the request in step 201, the elements of the event sequence transmitted together with the request are sequentially compared with the record sets on the history database 102 to find the same event sequence.

When comparing event sequences, it is not an exact match, and application startup and operations from the root menu are regarded as the start of a series of events, and comparisons are made within that range, or comparisons are made after canceling the cancellation operation and the previous event. Comparison accuracy can be improved by a search technique such as
In order to improve the accuracy, the application can explicitly register the process to be performed by itself in the event sequence together with the application information in the form of process name or function name.

  As a result of history search, waiting time information is returned based on a result history table associated with an event sequence that has been successfully compared. When the event sequence 401 is found as a result of the search in FIG. 4, it is understood that the processing by the event sequence that is the target of the search is completed as the past history between 10 seconds and 45 seconds using the search history table 402. .

Also, from the standpoint of distinguishing between the waiting state and the frozen state, in the result history table on the history database 102, the user forced termination within the time that can be statistically regarded as normal termination is ignored from the time range in which the normal recovery has occurred in the past. Only the user forced termination that occurred outside the range can be regarded as having a high possibility of freezing.
This information is notified to the user in the subsequent step.

(4) In the device management server 101, if the result history corresponding to the searched event sequence is found in the history database 102, the prediction result based on the analysis is found, and if not found, the information device 110 that is the request source To notify. (Step 223)
(5) The information device 101 receives the result of step 223 and displays it to the user. (Step 203)
At this time, in the information terminal 110 in a non-response state, there may be a case where there is a delay in displaying this information through the output device 115 or the information terminal 110 cannot display it. At the time of notification in step 223, transmission to a user portable terminal other than the information device 110 that is the notification source can also be performed to provide information to the user more reliably in response to the request in step 201.

(6) The device management server 101 confirms the number of history records in the result history table 402 of FIG. 4 obtained as the history search in step 222. If the number is not sufficient as statistical data, the device management server 101 manages An information collection request for processing by an event sequence that is a search key is issued to the target information devices 120, 130, and the like. (Steps 224 and 225)
If there is enough history as statistical data, nothing is done. Whether the number is sufficient is determined by a threshold value given as a setting value of the device management server 101. The information collection request in step 225 is requested to all the client information devices to be managed by the device management server 101, and the information regarding the non-response state can be efficiently collected by performing distributed monitoring.

(7) The information devices 120, 130, etc. that have received the request in step 225 are also accumulating event sequences that have occurred themselves, similar to the information device 110. (Step 241)
(8) The information devices 120, 130, etc. always check whether or not an equivalent event sequence has occurred as compared with the event sequence received in step 225 when accumulating new events. (Steps 242 and 243)
(9) If it is determined that an equivalent event sequence has occurred, the information devices 120, 130, etc. are considered to be in a no-response state as with the information device 110, and monitoring of application processing is started. (Step 224)
(10) In the information devices 110 and 120, 130, etc., the application monitoring is performed in the same manner in steps 202 and 224, and the monitoring server is notified of the result of the non-response state to be monitored. . (Steps 204 and 245)
This result information is composed of the result of whether the user is forcibly terminated or returned to normal with the event sequence information as a target, and the elapsed time to reach the result, and is reported together with the terminal information of the information device itself. . Although it can not be said that a no-response state occurs in response to an event sequence requested by each information device, in this case, it is reflected in the result history as a non-response state with a waiting time of zero. Further, the information collection in the information device may be continued until sufficient information is accumulated in the device management server 101 without ending with one report.

  (11) Upon receiving the reports in steps 204 and 245, the device management server 101 adds the terminal information, time, and result to the result search table 402 corresponding to the corresponding event sequence.

  As described above, in response to the user request in step 201, it becomes possible to present non-response state determination support information by statistical information across a plurality of client information devices. Termination can be suppressed.

  By repeating the cycle of the flow of FIG. 2, the history database 102 accumulates a history of non-response states for many event sequences. As a result, the probability of successful search in step 222 increases, and significant information can be notified by the enhancement of statistical information.

  In many embedded devices, the same product configuration is the same, such as CPU and memory, and the same processing time is expected to be close. If it is assumed that all the management targets of the device management server 101 are devices having the same configuration, it is not necessary to consider the difference in specifications in the search result of step 222 in FIG. When the management targets of the device management server are not the same, the difference in the waiting time / freeze state obtained as a result can be improved by considering the difference in the search result of step 222.

  FIG. 5 shows an example in which the hardware specification information of the terminal is added to the result history table stored in the history database 102. In this example, the CPU clock performance 501 and the memory capacity 502 are combined and stored in the result history.

  At the time of the request in step 201 in FIG. 2, the specification information possessed by the information device 101 is also notified according to the request and used for the history search in step 222. Note that the specification information of each client is registered in advance in the management server 101, and a mechanism of notifying its own registration ID when requesting step 201 is also conceivable. In this case, only the registration ID is left in the result history table on the history database.

The following corrections can be considered to calculate the expected return time of the no-response state according to the specifications of the information device 101 that is the request source.
When the CPU clock of the information device 101 is 100 MHz, which is the same as that of the terminals L and N in FIG. 5, it is conceivable to use the results of L and N preferentially for time prediction. Also, by simply converting the result of 300MHz M as a third of the figure, it can be expected that it will end in 45 to 75 seconds at 100MHz as an example of normal termination. In addition, it is conceivable that the freeze state is caused by a memory shortage. When there is a correlation between the memory loading amount and the result as shown in FIG. 5, the information device 101 has a memory loading amount smaller than 16 MB. It is possible to notify the user or developer that the risk of falling into a frozen state is high.

In addition to such hardware spec information, state parameters such as dynamically changing free memory and CPU usage are also stored in the result history table on the history database 102, and the closer information is weighted to give no response. As a result, it is possible to obtain a more accurate forecast.
Also in the calculation of the normal end time range in the detection of the possibility of the freeze state detailed in the procedure of step 222, the accuracy can be improved by taking the hardware specifications / dynamic state parameters into consideration as described above.

  In step 225, the device management server 101 requests the information devices 120 and 130 to be managed to collect information in a non-response state with respect to the event sequence. In the description of the flow of FIG. 2 above, the information devices 120, 130, etc. have the result and the time as this information. By providing the monitoring profiler on the information devices 120 and 130, it is possible to enable the monitoring simultaneously with the start of the monitoring of application processing, and to monitor the behavior of the application in more detail.

  In general, the monitoring profiler is often problematic in terms of performance because it interrupts normal processing and monitors its behavior. By enabling the monitoring profiler only when a specific event sequence occurs as in steps 242 and 243 of the flow of FIG. 2, it is possible to monitor a non-response state without imposing a performance load on normal processing. it can.

  From the viewpoint of reducing the load, the conditions for enabling the profile can be further narrowed down. An example of validation determination is shown using the flow of FIG.

(1) By limiting the purpose of monitoring to resolution of the frozen state, it is possible to narrow down the event sequence used as the monitoring profile start condition. In accordance with the determination criteria described in step 222, it is possible to suppress the activation of the profile in the event sequence that is unlikely to be frozen. (Step 601)
(2) In the result history in the history database 102, similar hardware among the information devices 110, 120, 130, etc. managed by the device management server 101 as compared with the hardware specifications of the information device that has caused the freeze state Only the target information device having the wear spec is instructed to validate the profile in accordance with the information collection request in step 225. (Step 602)
It is also possible to notify only the hardware specification and determine whether the profile is valid by comparing each of the information devices 110, 120, 130, etc. with their own hardware specifications.

  (3) Similarly, dynamic parameters such as the amount of free memory and the CPU usage rate at the time when the freeze state occurs can be used for profile validation determination as in step 602. In this case, the dynamic parameters at the time when the no-response state occurs are sent together with the step 225 information collection request. When each information device is judged to be similar to its own dynamic parameter compared to the sent one in accordance with the event sequence generation decision steps 242 and 243 for which information collection was requested Only for the purpose of validating the profile.

  With the above procedure, it is possible to narrow down the validation of profiles that affect processing performance according to the conditions necessary for freeze state analysis, and collect necessary fault information while suppressing the load of monitoring profiles on information devices. .

  It can be used for improving the operability of information devices that involve user operations and for analyzing failures (freeze states).

System configuration example Example of processing in the system Data structure example of event sequence with events arranged in time series Data structure example of history database 102 Data configuration example in which hardware specification information of the terminal is added to the result history table stored in the history database 102 Judgment example of whether to enable the monitoring profiler

Explanation of symbols

101 Device management server
102 history database
103 network
110 Information equipment
111 CPU
112 Communication Department
113 Main storage
114 Secondary storage
115 Output device
116 Input device
120 Information equipment
130 Information equipment

Claims (7)

An information device management method for performing processing by user operation,
In the device management server, the event sequence on the information device that has reached the no-response state is used as a key, and a history database is built that stores records whose values are the result history of the state brought about by the event sequence,
The history database is searched using, as a key, an event sequence that causes a non-response state in the device management server by a user request when a non-response state occurs. An information device management method that predicts the end time of a state or the possibility of freezing and presents it to the user. An information device management method according to claim 1,
In the case of the user request, when the number of records in the history database is not sufficient at the time of searching in the device management server, the information collection of the state reached by the event sequence when it occurs for the information device managed by the device management server Information device management method that broadcasts requests and replenishes history database information. An information device management method according to claim 1,
Notify the result predicted from the search result of the history database by the device management server not only to the requesting information device in the non-response state, but also notify other terminals that can be referred to by the requesting user. Information device management method that communicates to the requesting user. An information device management method according to claim 1,
Taking into account the dynamic parameters of the hardware specifications and CPU usage, available memory, etc. of the device management information device resulted in the result history based on statistics of the result history at the server, the result information equipment management method for performing prediction . An information device management method according to any one of claims 1 to 4,
Regarding forced termination that occurred at a time outside the range that can be predicted from the statistical information of the history of the waiting state that has returned to normal among the history of forced termination by the user in the result history for each event sequence stored in the history database An information device management method that determines the possibility of a frozen state and notifies the developer and the requesting user. An information device management method according to any one of claims 2 to 5,
When broadcasting an information collection request from the device management server to the management target information device, more detailed information collection when it is determined that there is a possibility of a freeze state from the result history of the target event sequence An information device management method that enables a profiler on an information device to collect information for resolving a frozen state. An information device management method according to claim 6,
As a condition for enabling the profiler on the information device, the information device hardware specifications and the dynamic parameters such as the CPU usage rate and free memory that resulted in the history of the freeze state, only information devices with closer conditions An information device management method for restricting information devices that give a profile load by enabling a profiler.

Images