JP2011065269A - Time series data analysis support program and device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To automate an operation to determine a period when time series data are made almost constant, and to extract the time series data in the period, and to calculate the mean value. <P>SOLUTION: In a client device 100, a parameter input part 151 inputs an allowable width and the shortest period read from a storage part 101, and a determination part 153 determines, based on the allowable width and the shortest period, a period which is equal to or more than the shortest period when the time series data of processing quantity are stored within allowable width as a constant period, and a totaling means 155 calculates the mean value of the time series data of processing quantity and the mean value of the time series data of CPU use rates in each determined constant period, and a result output part 156 performs the display processing of a graph in which the processing quantity is associated with the CPU use rates based on each calculated mean value. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a time series data analysis support program and apparatus, for example, to determine a period in which time series data is substantially constant, extract time series data within the period, and automate an operation of calculating an average value. The present invention relates to a realizable time series data analysis support program and apparatus.

  The load test is a test in which a large amount of load such as an http request is given to a target computer system, and the performance characteristics of the computer system with respect to a processing amount (request occurrence frequency) assumed at the time of operation are confirmed. Here, the target computer system is a computer system on the Internet or an intranet. The confirmation of the performance characteristics includes confirmation that the CPU utilization rate with respect to the processing amount satisfies the performance requirement, and confirmation of a limit processing amount that does not satisfy the performance requirement.

  Also, in the load test, the analyst performs the work of analyzing the time series data. In this work, the analyst calculates an average value of the request occurrence frequency and the CPU utilization rate in a period in which the time series data of the request occurrence frequency and the time series data of the CPU utilization rate of the server device are substantially constant. The calculation result is displayed on the client device as a graph of the request occurrence frequency on the horizontal axis and the CPU utilization rate on the vertical axis according to the operation of the operator.

  The analyst confirms the performance characteristics described above based on the displayed graph. Such a graph may be displayed not only during the load test but also during operation. When a graph is displayed during operation, the limit processing amount cannot be confirmed, but it can be confirmed that the CPU utilization rate with respect to the processing amount satisfies the performance requirement.

JP 2007-241534 A

  In the load test as described above, generally, when a load test tool such as JMeter or WebLOAD is used, the number of virtual users can be adjusted, and the frequency of requests given to the target computer system can be changed. These load test tools generate a request, and after receiving a response from the computer system, generate the next request. Therefore, it takes time until the request occurrence frequency is actually stabilized after the number of virtual users is set. Further, the request occurrence frequency varies with time even after stabilization.

  On the other hand, a load test tool that generates requests at a fixed frequency without waiting for a response from the target computer system can be considered. However, in this method, although the request occurrence frequency is stabilized, it takes time until the CPU utilization rate of the target computer system is stabilized. Also, the CPU utilization rate varies with time even after stabilization.

  In any case, before displaying the above-mentioned graph, the work in which the person visually confirms the change in the request occurrence frequency and the CPU usage rate of the server device in advance, and the period in which the request occurrence frequency and the CPU usage rate are almost constant. Judgment work is required. In addition, it is necessary to cut out the data of the request occurrence frequency and CPU usage rate during the determined period and to calculate and total the average values. Since these operations are complicated and time-consuming, it is desirable to automate them. Although Patent Document 1 discloses an example in which utilization rate data of a CPU or the like is read and summarized in a utilization rate table at each time, for example, a time period in which time-series data is determined to be almost constant, Since the average value of the time series data is not calculated, it does not contribute to the automation of the above-described work.

  The present invention has been made in consideration of the above situation, and it is possible to realize the automation of the operation of determining the period in which the time series data is substantially constant, extracting the time series data within the period, and calculating the average value. It is an object to provide a time series data analysis support program and apparatus.

  One aspect of the present invention is a time-series data analysis support program used in a client device that includes storage means and can receive time-series data of processing amount and time-series data of CPU utilization in a computer system to be analyzed. And the client device writes the input allowable width and the shortest period to the storage means, and based on the allowable width and the shortest period in the storage means, the time-series data of the processing amount is the allowable width A fixed period determination means for determining a period within the shortest period as a fixed period, an average value of the time series data of the processing amount and a time series of the CPU utilization rate for each determined fixed period An average value calculation means for calculating an average value of data, and a graph display process in which the processing amount and the CPU usage rate are associated with each other based on the calculated average values. Is a time-series data analysis support program for causing a graph display processing unit functions as, for.

  Note that one aspect of the present invention is expressed as a program. However, the present invention is not limited thereto, and can be expressed as an apparatus, a method, or a computer-readable storage medium storing a program.

(Function)
In one aspect of the present invention, the fixed period determination unit is configured to set a period equal to or longer than the shortest period in a period in which the time series data of the processing amount is within the allowable range based on the allowable range and the shortest period in the storage unit. Since the average value calculation means calculates the average value of the time-series data of the processing amount and the average value of the time-series data of the CPU usage rate for each fixed period determined by the average value calculation means, the time-series data It is possible to realize the automation of the operation of determining a period in which the current is almost constant, extracting time series data within the period, and calculating the average value.

  As described above, according to the present invention, it is possible to realize the automation of the operation of determining the period in which the time series data is substantially constant, extracting the time series data in the period, and calculating the average value.

1 is a block diagram illustrating a client device and its peripheral configuration according to a first embodiment of the present invention. It is a block diagram which illustrates the composition of the time series data analysis support part in the embodiment. It is a flowchart for demonstrating the operation | movement in the embodiment. It is a graph for demonstrating the time series data and parameter of the processing amount in the embodiment. It is a schematic diagram for demonstrating the algorithm which judges the fixed period of the processing amount in the embodiment. It is a flowchart for demonstrating operation | movement of the judgment part in the embodiment. It is a graph which shows the time series data of the processing amount in the embodiment. It is the graph which showed the fixed period of the judgment result in the vicinity of the time series data of the processing amount in the same embodiment. It is a schematic diagram which shows the number, start time, and end time which identify the said fixed period for every fixed period of the processing amount in the embodiment. It is a graph which shows the time series data of CPU utilization in the same embodiment. It is a schematic diagram which shows the number which identifies the said fixed period for every fixed period in the same embodiment, start time, end time, request occurrence frequency, and CPU utilization. It is a graph which shows the request generation frequency and CPU utilization rate in the embodiment in association with each other. It is the graph which showed the fixed period judged in the vicinity of the time series data of the processing amount in the embodiment in order of the new time series data. It is a schematic diagram which compares and compares the fixed period judged from the oldest with the fixed period judged from the newest with respect to the same time series data in the embodiment. It is a block diagram which illustrates the composition of the time series data analysis support part applied to the client device concerning a 2nd embodiment of the present invention. It is a flowchart for demonstrating the operation | movement in the embodiment. It is a flowchart for demonstrating operation | movement of the judgment part in the embodiment. It is a block diagram which illustrates the composition of the time series data analysis support part applied to the client device concerning a 3rd embodiment of the present invention. It is a flowchart for demonstrating the operation | movement which concerns on the same embodiment.

  Hereinafter, each embodiment of the present invention will be described with reference to the drawings, but before that, an outline of each embodiment will be described. The outline of each embodiment is to determine a period in which either or both of the request occurrence frequency and the CPU usage rate are substantially constant, and for each determined period, the request occurrence frequency and the CPU usage rate are extracted and averaged. A value is calculated, and the request occurrence frequency and the CPU usage rate are associated with each other and displayed in a graph. Here, in the first embodiment, a period in which the request occurrence frequency is substantially constant is determined, and in the second embodiment, a period in which the CPU utilization rate is constant is determined. In the third embodiment, it is determined that both the request occurrence frequency and the CPU usage rate are substantially constant. The above is the outline of each embodiment.

  In each of such embodiments, for each device, either a hardware configuration or a combination configuration of hardware resources and software can be implemented. As the software of the combined configuration, a program that is installed in advance on a computer of a corresponding device from a network or a storage medium and that realizes the function of the corresponding device is used.

  Subsequently, specific description will be made sequentially from the first embodiment.

(First embodiment)
FIG. 1 is a block diagram illustrating a client device according to the first embodiment of the present invention and its peripheral configuration according to the first embodiment of the present invention, and FIG. 2 is a time-series data analysis support unit in the client device. It is a block diagram which illustrates the composition of. As shown in FIG. 1, the computer system CS to be analyzed includes a server group such as a Web server device 1, an application server device 2, and a database server device 3. Such a computer system CS and the client device 100 are connected to each other via a network 200 so that they can communicate with each other.

  Each of the server devices 1, 2, 3,... In the computer system CS is an arbitrary server device that can communicate via the network 200, and responds to the http request received from the client device 100 after executing the generated processing. A return function and a function of transmitting the time series data of the processing amount and the time series data of the CPU utilization rate to the client device 100.

  Here, as the time-series data of the processing amount, the request occurrence frequency generated from the virtual user is measured by the load test tool unit 140 during the load test. Further, when the computer system CS is operated, a usage history is obtained by a log function of the Web server device 1 or the application server device 2 or the like. As the usage history, for example, the following two types (1) to (2) can be used. In any case, the processing amount of each processing given to the computer system CS within a certain time interval can be known.

  (1) A usage history in which the type of processing started in response to a request from a client and the time of occurrence are recorded as in an access log.

  (2) A usage history in which the amount of generation for each process in the section is recorded at regular intervals.

  The time series data of the CPU usage rate is the CPU usage rate of each of the server devices 1, 2, 3,... In the computer system CS, and is measured at regular time intervals such as 10 seconds. Further, the time series data of the CPU usage rate can be measured and output by a program tool such as a performance monitor (perfmon.exe).

  On the other hand, the client device 100 includes a storage unit 110, an input unit 120, a communication I / F unit 130, a load test tool unit 140, a time series data analysis support unit 150, and a display unit 160.

  Here, the storage unit 110 is a storage device that can be read / written from each of the units 120 to 160, for example, parameters input from the input unit 110, time series data received by the communication I / F unit 130, and each unit 140. , 150, etc., and data of processing results are stored.

  The input unit 110 is an input device such as a keyboard and a mouse operated by an operator.

  The communication I / F unit 130 is a communication interface for communicating with each server 1, 2, 3,... Of the computer system CS via the network 200. For example, the http request generated by the load test tool unit 140 is transmitted to the computer system. It has a function of transmitting to CS and a function of receiving time-series data transmitted from computer system CS. Examples of the time series data include time series data including a measurement value of processing amount and time data, and time series data including a measurement value of CPU utilization and time data.

  The load test tool unit 140 transmits a http request according to the number of virtual users set from the input unit 110 to the computer system CS by the communication I / F unit 130 and performs a load test, and communicates from the computer system CS. It has a function of inputting the time series data of the processing amount and the time series data of the CPU usage rate received by the I / F unit 130 to the processing amount input unit 152 and the performance data input unit 154 in the time series data analysis support unit 150. Note that the load test tool unit 140 can be omitted when there is no load test for the use of the client device 100 (when the use is only in operation).

  As shown in FIG. 2, the time-series data analysis support unit 150 includes a parameter input unit 151, a processing amount input unit 152, a determination unit 153, a performance data input unit 154, a totaling unit 155, and a result output unit 156.

  The parameter input unit 151 has a function of inputting the parameter input from the input unit 110 to the determination unit 153 by the operation of the operator. Here, examples of the parameter include an allowable width W and a shortest period T used for determination of a certain period by the determination unit 153. However, not only these parameters W and T but also the allowable number of detachment V may be used as a parameter. The allowable number of deviations V will be described in the second embodiment.

  The processing amount input unit 152 uses the load test tool unit 140 (during the load test) or the communication I as the time series data of the processing amount representing the amount actually processed by the computer system CS, such as the frequency of requests to the target computer system CS. / F unit 130 (during operation), and has a function of inputting time-series data of this processing amount to determination unit 153 and totalization unit 155.

  Note that many load test tools applicable to the load test tool unit 140 transmit a next request after receiving a response from the target computer system CS. For this reason, the request occurrence frequency matches the processing amount per unit time of the computer system CS, unless the computer system CS returns an error without being overloaded. Therefore, in the present specification, the request occurrence frequency and the processing amount of the computer system CS are assumed to be the same.

  Based on the function of writing the parameter input from the parameter input unit 151 to the storage unit 110 and the parameter in the storage unit 110, the determination unit 153 has almost the time series data of the processing amount input from the processing amount input unit 152. It has a function of determining a certain period and a function of sending a number for identifying the certain period, a start time, and an end time for each determined period to the counting unit 155.

  Here, the determination function is specifically based on the allowable range and the shortest period in the storage unit 110, and the period of the time series data of the processing amount within the allowable range is set as a fixed period. to decide.

  The performance data input unit 154 uses the load test tool unit 140 (during the load test) or the communication I / F unit 130 as the time series data of the CPU utilization rate (performance data) of the server devices 1, 2, 3,. It has a function of receiving the time series data of the CPU utilization rate from the (during operation) and inputting the time series data of the CPU utilization rate to the totaling unit 155.

  The totaling unit 155 has a function of receiving a number, a start time, and an end time every predetermined period from the determination unit 153, a function of receiving time-series data of the processing amount from the processing amount input unit 152, and a CPU from the performance data input unit 154 to the CPU. Based on the function that receives the time series data of the utilization rate, the number of each fixed period, the start time and the end time, the average value of the time series data of the processing amount and the time series data of the CPU usage rate for the fixed period It has a function of calculating an average value and a function of sending the calculated average values to the result output unit 156.

  The result output unit 156 has a function of receiving the average value of the processing amount and the average value of the CPU usage rate for each fixed period from the totaling unit 155, and the processing amount and the CPU usage rate based on each average value for each fixed period. And a function of executing a graph display process in which The graph display process includes a process of creating display data of the graph and a process of sending the created display data to the display unit 160.

  The display unit 160 is a display device that displays data sent from the units 140 and 150, and a liquid crystal display device can be used as appropriate.

  Next, the operation of the client apparatus configured as described above will be described with reference to the flowchart of FIG.

  The parameter input unit 151 inputs a parameter P for determining a period during which the request occurrence frequency to the target computer system CS is substantially constant to the determination unit 153 by the operation of the input unit 120 by the operator (STEP 100). Here, the parameter P is the allowable width W and the shortest period T.

When the processing amount input unit 152 receives time series data S _i (i = 1, 2,..., N) of processing amounts such as the frequency of requests to the target computer system from the load test tool unit 140, for example, the processing amount when entering the series data _{S i} to the determination unit 153 and the aggregation section 155 of the (STEP200). S _i is a request occurrence frequency from time (i−1) Δt to iΔt. Δt is an observation interval, and description will be made assuming that data from time 0 to time NΔt is observed.

The determination unit 153 writes the parameter P (allowable width W and shortest period T) received from the parameter input unit 151 to the storage unit 110, and based on the parameter P (allowable width W and shortest period T) in the storage unit 110, The period of time series data S _i (i = 1, 2,..., N) of the processing amount received from the processing amount input unit 152 is within the allowable width W, and a period longer than the shortest period T is set to a certain period B _j (j = 1, 2,..., M) (STEP 300). M is the number of periods in which the processing amount is determined to be constant.

Thereafter, the determination unit 153 determines, for each determined fixed period B _j (j = 1, 2,..., M), a number j that identifies the fixed period, a start time Ts _j (j = 1, 2,. M), and the end time Te _j (j = 1, 2,..., M) is sent to the counting unit 155.

The performance data input unit 154 converts the time series data U _i (i = 1, 2,..., N) such as the CPU usage rate of the server device into the load test tool unit 140 (during the load test) or the communication I / F unit 130 ( When receiving from the time of operation), the time series data U _i of the CPU utilization rate is input to the totaling unit 155 (STEP 400).

Based on the number j, the start time Ts _j , and the end time Te _j for each fixed period B _j received from the determination unit 153, the counting unit 155 calculates the average value S _j ave (j = 1) of the request occurrence frequency during that period. , 2,..., M) and CPU utilization average value U _j ave (j = 1, 2,..., M) are calculated (STEP 500), and the calculated average values S _j ave and U _j ave are output as results. To the unit 156.

In the result output unit 156, the average value S _j ave (j = 1, 2,..., M) of the request occurrence frequency and the average value U _j ave (j = 1, 2,. ) Is received, a graph display process in which the average values S _j ave and U _j ave are associated with each other is executed (STEP 600), and the graph is displayed on the display unit 160.

  As described above, since the time series data analysis support unit 150 operates, it is not necessary for a person to look at the time series data of the request occurrence frequency to determine a substantially constant period, and the start time and end of the period. It is not necessary to perform complicated calculations such as cutting out the CPU utilization during that time and calculating the average value.

In addition, since a graph in which the average values S _j ave and U _j ave are associated with each other can be displayed even during the load test, it can be confirmed that the load test is normally performed, and the number of virtual users to be loaded next is determined. You can decide whether to set.

  Next, an example of a method by which the determination unit 153 determines a certain period will be described in detail.

The parameter P is composed of an allowable width W of processing amount and a shortest period T (P = (W, T)). The fixed period means a period in which the time-series data S _i (i = 1, 2,..., N) of the processing amount continuously falls within the allowable width W and is equal to or longer than the shortest period T. FIG. 4 is a graph showing the time-series data of the processing amount, and the height of the square frame in the figure represents the allowable width W. A range that falls within the allowable width W is surrounded by a square frame. If the period is equal to or longer than the shortest period T, the period is determined to be a period in which the processing amount is constant.

The determination algorithm will be described with reference to FIG. In particular, a method for positioning a square frame will be described. FIG. 5 is a graph in which the horizontal axis represents time and the vertical axis represents the processing amount, and time-series data S _i (i = 1, 2,..., N) of the processing amount is plotted. In this determination algorithm, the processing time series data S _i (i = 1, 2,..., N) is sequentially read. First, when time-series data S _i first data from S ₁ is read, the value of the first data S ₁ is substituted into the variable min representing the variable max and the minimum value that represents the maximum value of time-series data S _i. The range that the square frame (allowable width W) can take is a range from max-W to min + W (arrow range in the figure).

Then, the second data _{S 2} of the time-series data _{S i} is read. If the value of the second data S ₂ is within a range of max-W of min + W will be a certain period of time followed. Further, compared variable max of the second data S ₂ and the maximum value, the value of the second data S ₂ is substituted for the variable max maximum value when the second data S ₂ is greater. Compared variable min the second data S ₂ and the minimum value, the value of S ₂ is substituted for the variable min of the minimum value when the second data S ₂ is small. 5 In S ₂ is assigned the second value data S ₂ to the variable max of the maximum value because greater than the variable max of the maximum value. This shows that the range (from max-W to min + W) that the square frame (allowable width W) can take is narrowed. Similarly, when the third data S ₃ is read, the third data S ₃ is the value of the third data S ₃ is substituted for the variable max of the maximum value greater it was so than the variable max of the maximum value. Fourth data S ₄ is because min ≦ S ₄ ≦ max, variables min, the change in max no.

Next, when the fifth data S ₅ is read, it is excluded from the rectangular frame of S _5> max + W since tolerance W. If the period of the _first to fourth data S _{1 to} S ₄ is equal to or longer than the shortest period T, it is determined that the processing amount is constant.

  The above determination method will be described in detail with reference to the flowchart of FIG.

  The determination unit 153 initializes the minimum value variable min, the maximum value variable max, and the time series data number i (STEP 301). That is, min = 0, max = 0, and i = 1.

Subsequently, determination unit 153, the time-series data _{S i,} performs a loop process between STEP302~STEP316 (STEP302).

The determination unit 153 determines whether or not the minimum variable min is 0 (STEP 303). If not (STEP 303; No), the i-th data S _i in the time-series data can take the allowable width W. It is determined whether or not (max−W ≦ S _i ≦ min + W) is satisfied (STEP 304).

If the result of the determination in STEP 304 is within the range (STEP 304; Yes), the determination unit 153 compares the i-th data S _i with the minimum value variable min to determine whether S _i <min. If the determination is made (STEP 305) and S _i <min (STEP 305; Yes), the i-th data is substituted for the minimum value min (STEP 306). If not, the value of the minimum value min is maintained as it is. .

Subsequently, the determination unit 153 compares the variable max i-th data _{S i} and the maximum value determines whether _{S i>} max (STEP 307), if _{S i>} max (STEP 307; Yes), the i-th data S _i is substituted into the maximum value variable max (STEP 308), and if not, the maximum value variable max is maintained as it is.

  Thereafter, the determination unit 153 updates the length n, which continues for a certain period, by +1 (STEP 309).

On the other hand, if the result of the determination in STEP 304 is NO (STEP 304; No), the determination unit 153 determines whether or not the length n that continues for a certain period is equal to or longer than the shortest period T (STEP 310). If it is equal to or longer than the shortest period T (STEP 310; Yes), the time series data S _{s to} S _{i-1 up} to the previous (i-1) within a certain period are grouped (STEP 311), but if not (STEP 310) No), do not group time-series data.

  Thereafter, the determination unit 153 initializes the minimum value variable min and the maximum value variable max to 0 (STEP 312).

On the other hand, if the minimum value variable min is 0 as a result of the determination in STEP 303 (STEP 303; Yes), the determination unit 153 sets the value of the i-th data S _i to the minimum value variable min and the maximum value variable max. substituting (STEP313), the start data _{S s} for a certain period of time and _{S i,} and 1 of length n is followed by a period of time (STEP314).

Next, STEP309, after the execution of STEP312 or STEP314, determination unit 153, when the number i of the series data _{S i} by +1 updated (STEP315), repeatedly executes the loop processing from STEP302 (STEP316).

After completion of the loop processing of STEP 302 to STEP 316, the determination unit 153 determines whether or not the length n that continues for a certain period is equal to or longer than the shortest period T (STEP 317), and if it is equal to or longer than the shortest period T (STEP 317). Yes), the time series data S _{s to} S _{i-1 up} to the previous (i-1) within a certain period are grouped (STEP 318). If not (STEP 317; No), the time series data is grouped. Instead, the processing of STEP 300 is terminated.

Figure 7 is a graph showing the time-series data S _i of throughput. The horizontal axis represents time, and the vertical axis represents the request occurrence frequency. FIG. 8 shows the time series data S _i of this processing amount with a certain period determined by the procedure of FIGS. 5 and 6 added. In Figure 8, on a graph of time-series data S _i of the processing amount, the period during which the processing amount is determined to constant are plotted. The parameter value is a result of calculation assuming that the allowable width W is 3.0 [times / second] and the shortest period T is 100 [seconds].

  FIG. 9 shows the number, start time, and end time for each fixed period determined by the determination unit 153. The determination unit 153 sends information on the number, start time, and end time for each determined fixed period to the totaling unit 155.

  Next, the processing of the totaling unit 155 will be described. In the totaling unit 155, the start time and end time of the period in which the processing amount is determined to be constant by the determination unit 153, the time series data of the processing amount from the processing amount input unit 152, and the CPU of the server device from the performance data input unit 154 When performance data such as usage rate is received, the average value of the request occurrence frequency and the average value of the CPU usage rate are calculated based on the start time and end time of the period when the processing amount is determined to be constant. .

  FIG. 10 shows time series data of the CPU utilization rate when the processing amount of FIG. 7 is measured. It is assumed that the time series data of the CPU usage rate is data received from the performance data input unit 154. Based on these data, FIG. 11 shows the total results of calculating the average value of the request occurrence frequency and the average value of the CPU utilization rate for each period in which the processing amount is determined to be constant. The calculation result of the request occurrence frequency and the average value of the CPU usage rate is sent to the result output unit 156.

  The result output unit 156 creates display data of a request occurrence frequency on the horizontal axis and a CPU utilization rate graph on the vertical axis, and sends the display data to the display unit 160. A graph displayed on the display unit 160 is shown in FIG. It can be seen that the CPU utilization rate increases as the request generation frequency increases.

As described above, according to the present embodiment, the determination unit 153 determines the processing amount time-series data within the allowable width W based on the allowable width and the shortest period in the storage unit 110 and is equal to or longer than the shortest period T. Is determined as the fixed period B _j , and the average value of the time series data S _i of the processing amount and the average value of the time series data U _i of the CPU usage rate for each fixed period B _j determined by the totaling unit 155 Therefore, it is possible to realize the automation of the operation of determining the period in which the time series data is substantially constant, extracting the time series data within the period, and calculating the average value.

  That is, unlike the prior art, it is possible to automatically determine that a person has visually identified a period in which the request occurrence frequency and the CPU utilization rate of the server device are stable at a substantially constant level. Further, it is not necessary to perform complicated calculations such as cutting out the CPU usage rate between the start time and end time of the period and calculating the average value. Further, for each period, the average value of the request occurrence frequency and the CPU utilization rate of the server device can be calculated and the result can be automatically displayed in a graph.

  Furthermore, in the middle of the load test, the results can be displayed in a graph to confirm that the test has been performed correctly, and how to set the number of virtual users to be loaded next can be determined.

  In the present embodiment, the fixed period is determined in the oldest order of the time-series data of the processing amount. However, the present invention is not limited to this, and the fixed period may be determined in the newest order of the time-series data.

  In general, the processing amount data greatly fluctuates at the moment when the number of virtual users is changed, and then tends to be stable. Therefore, it is possible to specify a period of a processing amount with less variation by processing the time series data of the processing amount in order from the newest.

  FIG. 13 shows a result of performing processing for specifying a fixed processing amount period in the order of data from the time series data of processing amount in FIG.

  Since the difference between FIG. 8 and FIG. 13 is difficult to understand, the time-series data of the processing amount and the oldest data in the time portion (2008/5/27 20:14 to 2008/5/27 20:19) FIG. 14 shows a result of arranging and displaying the period when the fixed amount of processing period is specified and the periods when the fixed amount of processing period is specified in the order of data. It can be seen that the periods specified by the two types of methods are different, and the period when the processing amount fixed period is specified in the order of the newest data is a longer period. In addition, the variance of the period when the fixed amount of processing period is specified in the oldest order of the data is 0.456, and the variance of the period when the fixed amount of processing period is specified in the newest order of the data is 0.295. It can be seen that the fixed amount period is closer to the fixed value.

(Second Embodiment)
Next, a second embodiment of the present invention will be described.

  This embodiment is a modification of the first embodiment, and unlike the first embodiment in which the target for determining a certain period is time-series data of the processing amount, the object for determining the certain period is the CPU. It is time-series data of usage rate.

  FIG. 15 is a block diagram illustrating the configuration of a time-series data analysis support unit applied to a client device according to the second embodiment of the present invention. Similar to the first embodiment, the time-series data analysis support unit 150 includes a parameter input unit 151, a processing amount input unit 152, a determination unit 153, a performance data input unit 154, a totaling unit 155, and a result output unit 156. The However, the functions of the processing amount input unit 152, the determination unit 153, and the performance data input unit 154 are slightly different. Note that the functions of the totaling unit 155 and the result output unit 156 are the same as those in the first embodiment, and a description thereof will be omitted.

  The parameter input unit 151 has a function of inputting the allowable detachment frequency V input from the input unit 120 to the determination unit 153 in addition to the input function of the allowable width W and the shortest period T described above.

  When the processing amount input unit 152 receives time series data of the processing amount such as a request occurrence frequency to the target computer system from the load test tool unit 140 (at the time of the load test) or the communication I / F unit 130 (at the time of operation), It has a function of inputting the time series data of the processing amount to the totaling unit 155.

  The determination unit 153 has a function of determining a period in which the time series data of the CPU utilization rate input from the performance data input unit 154 is substantially constant based on the allowable width W and the shortest period T in the storage unit 110. And a function for sending the number indicating the fixed period, the start time, and the end time to the counting unit 155 for each fixed period.

  Supplementally, in the function described in the first embodiment, the determination unit 153 replaces the time series data of the processing amount with respect to the time series data of the CPU usage rate, and the minimum period T within the period within the allowable width W. A function for determining the above period as a certain period, and a function for writing the input allowable allowable number of times V to the storage unit 110, and the allowable width when the time-series data of the CPU utilization rate deviates from the allowable width W. If the deviation from W is not continuous, the function of ignoring the deviation from the allowable width W is included up to the allowable deviation frequency V.

  When the performance data input unit 154 receives the time series data of the CPU utilization rate from the load test tool unit 140 (during the load test) or the communication I / F unit 130 (during operation), the performance data input unit 154 determines the time series data of the CPU utilization rate. A function of inputting to the section 153 and the totaling section 155.

  Next, the operation of the time-series data analysis support unit applied to the client device configured as described above will be described with reference to the flowchart of FIG.

  Since STEP 500 and STEP 600 are the same as those in FIG. 3 (first embodiment), description thereof is omitted.

  The parameter input unit 151 inputs the allowable width W, the shortest period T, and the allowable number of detachments V as parameters P to the determination unit 153 by the operation of the input unit 120 by the operator (STEP 110).

When the performance data input unit 154 receives the time series data U _i (i = 1, 2,..., N) such as the CPU usage rate of the server device from the load test tool unit 140, for example, the time series data U of the CPU usage rate. _i is input to the determination unit 153 and the totaling unit 155 (STEP 210).

When the determination unit 153 writes the allowable width W, the shortest period T, and the allowable number of deviations V received from the parameter input unit 151 in the storage unit 110, the determination unit 153 receives the time series data U _i of the CPU utilization rate from the performance data input unit 154. Based on the allowable width W in the storage unit 110, the shortest period T, and the allowable number of deviations V, a period B _j (j = 1, 2,..., M) in which the time series data of the CPU usage rate is constant is determined. (STEP 330).

Thereafter, the determination unit 153 determines, for each determined fixed period B _j (j = 1, 2,..., M), a number j that identifies the fixed period, a start time Ts _j (j = 1, 2,. M), and the end time Te _j (j = 1, 2,..., M) is sent to the counting unit 155.

When the processing amount input unit 152 receives the time series data S _i (i = 1, 2,..., M) of the processing amount such as the request occurrence frequency to the target computer system from the load test tool unit 140, the processing amount input unit 152 The time-series data S _i is sent to the totaling unit 155 (STEP 410).

As can be seen by comparing FIG. 7 and FIG. 10, it can be seen that the time series data U _i of the CPU utilization rate has a larger fluctuation than the time series data S _i of the request occurrence frequency. Therefore, it becomes difficult to determine a period when the CPU usage rate is constant.

  Therefore, when the number of times that can deviate from the allowable width W (allowable number of deviations) V is added to the parameter P, and data that deviates from the allowable width W appears when determining a period during which the CPU usage rate is constant, If the deviation from the allowable width W does not continue, an algorithm is used that determines a period in which the CPU utilization rate is constant while allowing the allowable deviation number V up to V times.

FIG. 17 is a flowchart embodying this algorithm, and a part of STEP 331 indicated by a broken line and STEP 340-1 to STEP 340-4 are greatly different from the flowchart of FIG. 6 described above. 6 is different from FIG. 6 in that the target is not the time series data S _i of the processing amount but the time series data U _i of the CPU usage rate. Hereinafter, the operation shown in FIG. 17 will be described in detail.

  The determination unit 153 initializes the minimum value variable min, the maximum value variable max, the time series data number i, and the number k of deviations out of the allowable width W (STEP 331). That is, min = 0, max = 0, i = 1, k = 0. Further, the determination unit 153 writes the allowable number V of detachment in the storage unit 110 by the operation of the input unit 120 by the operator.

Subsequently, the determination unit 153 performs a loop process between STEP 332 and STEP 346 for the time series data U _i (STEP 332).

The determination unit 153 determines whether or not the minimum value variable min is 0 (STEP 333). If not (STEP 333; No), the i-th data U _i in the time-series data is within the allowable range W. It is determined whether or not (max−W ≦ U _i ≦ min + W) is satisfied (STEP 334).

If it is within the range as a result of the determination in STEP 334 (STEP 334; Yes), the determination unit 153 compares the i-th data U _i and the minimum value variable min to determine whether or not U _i <min. If the determination is made (STEP 335) and U _i <min (STEP 335; Yes), the i-th data is substituted for the minimum variable min (STEP 336). If not, the value of the minimum variable min is maintained as it is. .

Subsequently, the determination unit 153 compares the variable max i-th data _{U i} and the maximum value it is determined whether or not _{U i>} max (STEP337), if _{U i>} max (STEP337; Yes), the i-th data U _i is substituted for the variable max of the maximum value (STEP 338). If not, the variable max of the maximum value is maintained as it is.

  Thereafter, the determination unit 153 updates the length n, which continues for a certain period, by +1 (STEP 339).

On the other hand, when the result of the determination in STEP 334 is NO (STEP 334; No), the determination unit 153 determines whether or not the length n for which the predetermined period continues is equal to or longer than the shortest period T (STEP 340). If it is not less than the shortest period T (STEP 340; Yes), the time series data U _{s to} U _{i-1 up} to the previous (i-1) within a certain period are grouped (STEP 341), but if not (STEP 340) No), the number i of the time series data U _i out of the range is stored in the storage unit 110, and the number k of outliers is updated by +1 (STEP 340-1).

  Subsequently, the determination unit 153 determines whether or not the stored number i is continuous (STEP 340-1), and if it is continuous, proceeds to STEP 342.

On the other hand, if it is not continuous (STEP 340-1; No), the determination unit 153 determines whether or not the number of detachment k is equal to or less than the allowable number of detachment V (STEP 340-3). If so, go to STEP345. In other words, when the time series data U _i of the CPU utilization rate deviates from the allowable width W and the deviation from the allowable width W does not continue, the determining unit 153 determines the allowable width W up to the allowable detachment frequency V. Ignore being off.

  Further, when the number of detachment k exceeds the number of allowed detachment V (STEP 340-3; No), the determination unit 153 erases the number i that has been removed from the storage unit 110 and initializes the number of detachment k to 0. Proceed to STEP 342.

  Thereafter, the determination unit 153 initializes the minimum value variable min and the maximum value variable max to 0 (STEP 342).

On the other hand, as a result of the determination in STEP 333, if the minimum value variable min is 0 (STEP 333; Yes), the determination unit 153 sets the value of the i-th data U _i to the minimum value variable min and the maximum value variable max. Is substituted (STEP 343), the start data U _s for a fixed period is set to U _i, and the length n for which the fixed period continues is set to 1 (STEP 344).

Next, STEP339, after the execution of STEP342 or STEP344, determination unit 153, when the number i of the series data _{U i} by +1 updated (STEP345), repeatedly executes the loop processing from STEP332 (STEP346).

After completion of the loop processing of STEP332 to STEP346, the determination unit 153 determines whether or not the length n for which the predetermined period continues is equal to or longer than the shortest period T (STEP347), and if it is equal to or longer than the shortest period T (STEP347). Yes), the time series data U _{s to} U _{i-1 up} to the previous (i-1) within a certain period are grouped (STEP 348). If not (STEP 347; No), the time series data is grouped. Without processing, the processing of STEP 330 is terminated.

As described above, the determination unit 153 can determine a certain period even if the CPU utilization rate time-series data U _i varies greatly.

As described above, according to the present embodiment, the determination unit 153 uses the shortest period in which the time-series data U _i of the CPU usage rate is within the allowable width W based on the allowable width and the shortest period in the storage unit 110. A period equal to or longer than the period T is determined as a certain period B _j , and the aggregation unit 155 determines the average value of the time series data S _i of the processing amount and the time series data U _i of the CPU usage rate for each determined period B _j. Since the average value is calculated, it is possible to determine the period in which the time series data is substantially constant, extract the time series data within the period, and automate the operation of calculating the average value.

  For example, even when the load test tool unit 140 generates a request at a fixed frequency without waiting for a response from the target computer system CS, the CPUs of the server devices 1, 2, 3,. It is possible to automatically determine the period during which the utilization rate is stable. In addition, even for a CPU usage rate that has a large variation compared to the processing amount, it is possible to automatically determine a certain period based on the allowable number V of deviations.

Further, as described above, it is also possible to automatically calculate the average value of the request occurrence frequency and the average value of the CPU usage rate of the server device for each fixed period B _j and display the result in a graph.

(Third embodiment)
Next, a third embodiment of the present invention will be described.

  The present embodiment is a modification of the first and second embodiments. The target for determining a certain period is both the time-series data of the processing amount and the time-series data of the CPU utilization rate. A period determined to be a certain period is finally determined to be a certain period.

  FIG. 18 is a block diagram illustrating the configuration of a time-series data analysis support unit applied to a client device according to the third embodiment of the present invention. Similar to the first embodiment, the time-series data analysis support unit 150 includes a parameter input unit 151, a processing amount input unit 152, a determination unit 153, a performance data input unit 154, a totaling unit 155, and a result output unit 156. The However, the functions of the determination unit 153 and the performance data input unit 154 are slightly different. Further, the function of the parameter input unit 151 is the same as that of the second embodiment, and the functions of the processing amount input unit 152, the totaling unit 155, and the result output unit 156 are the same as those of the first embodiment, and thus description thereof is omitted. To do.

  The determination unit 153 writes the parameters (the allowable width W, the shortest period T, and the allowable number of misses V) input from the parameter input unit 151 to the storage unit 110, and the processing amount such as the request occurrence frequency from the processing amount input unit 152 A function of receiving a series of data and receiving a time series data of a CPU utilization rate from the performance data input unit 154, based on a parameter in the storage unit 110, and determining a certain period in which the processing amount and the performance data are substantially constant; It has a function of sending a number, a start time, and an end time for each determined period to the counting unit 155.

  Supplementally, the function of determining the predetermined period of the determination unit 153 includes the following functions (f153-1) to (f153-4).

(f153-1) based on the allowable width W and the shortest period T in the storage unit 110, determines the duration of more than the shortest period T as a period of time in a period in which the time-series data S _i is within the allowed range W of the processing amount Function to do.

  (f153-2) A function of writing the input allowable number of detachment times V in the storage unit 110.

(f153-3) Based on the allowable width W in the storage unit 110, the shortest period T, and the allowable number of deviations V, the time period in which the CPU utilization time series data U _i is within the allowable width W is equal to or greater than the shortest period T. This is a function for determining the period as a fixed period. In this determination, when the time series data U _i of the CPU usage rate deviates from the allowable width W, it is continuously determined that the time is not within the allowable width W. Otherwise, the function of ignoring the deviation from the allowable width W up to the allowable deviation frequency V.

(f153-4) A period in which the period determined as the fixed period for the time series data S _i of the processing amount and the period determined as the fixed period for the time series data U _i of the CPU usage rate are finally fixed Function to determine the period.

  As in the second embodiment, the performance data input unit 154 receives time series data such as a CPU usage rate from the load test tool unit 140 (during load test) or the communication I / F unit 130 (during operation). It has a function of sending time-series data of the CPU usage rate to the determination unit 153 and the totaling unit 155.

  Next, the operation of the time-series data analysis support unit applied to the client device configured as described above will be described with reference to the flowchart of FIG. However, STEP 110, STEP 210, and STEP 330 are the same as those in FIG. 16 (second embodiment) and the like, and STEP 200, STEP 300, STEP 500, and STEP 600 are the same as those in FIG. .

As described above, the determination unit 153 inputs the parameters W, T, and V of STEP 110, the input of the time series data S _i of the request occurrence frequency of STEP 200, and the time series data U _i of the CPU utilization rate of STEP 210. Is input.

  As described above, the determination unit 153 determines the fixed period for the request occurrence frequency of STEP 300 illustrated in FIGS. 3 and 6 and determines the fixed period for the CPU usage rate of STEP 330 illustrated in FIGS. 16 and 17. Execute.

Thereafter, the determination unit 153 finally sets a period in which both the period determined as the fixed period for the time series data S _i of the processing amount and the period determined as the fixed period for the time series data U _i of the CPU usage rate match. Therefore, it is determined that the predetermined period B _j (j = 1, 2,..., M) (STEP 360).

Thereafter, the determination unit 153 determines, for each determined fixed period B _j (j = 1, 2,..., M), a number j that identifies the fixed period, a start time Ts _j (j = 1, 2,. M), and the end time Te _j (j = 1, 2,..., M) is sent to the counting unit 155.

Hereinafter, similarly to the above, the calculation of the mean value based on a predetermined period _{B j} of Step 500, the display of the graph in STEP600 is performed.

As described above, according to the present embodiment, in addition to the effects of the first and second embodiments, the time-series data S _i (i = 1, 2,..., N) of the processing amount, the CPU usage rate, etc. Since the period when both of the performance time series data U _i (i = 1, 2,..., N) are determined to be constant is selected, it is possible to aggregate the results more stably at a constant value.

  Supplementally, in this embodiment, since it is possible to automatically determine a period in which both the processing amount such as the request occurrence frequency and the CPU utilization rate of the target computer system are stable, it is possible to specify a more stable period. Can do. Further, for each period, the average value of the request occurrence frequency and the CPU utilization rate of the server device can be calculated and the result can be automatically displayed in a graph.

  The above embodiments have been described by taking the case of analyzing time series data at the time of a load test as an example. However, the present invention is not limited to this, and each embodiment is applicable even when analyzing time series data at the time of operation. The same effect can be obtained by carrying out similarly.

  In each of the above-described embodiments, an example of a determination method for a certain period by the determination unit 153 has been described. However, the present invention is not limited to this. You may deform | transform so that it may produce | generate.

  Note that the method described in the above embodiment includes a magnetic disk (floppy (registered trademark) disk, hard disk, etc.), an optical disk (CD-ROM, DVD, etc.), a magneto-optical disk (MO) as programs that can be executed by a computer. ), And can be distributed in a storage medium such as a semiconductor memory.

  In addition, as long as the storage medium can store a program and can be read by a computer, the storage format may be any form.

  In addition, an OS (operating system) running on a computer based on an instruction of a program installed in the computer from a storage medium, MW (middleware) such as database management software, network software, and the like realize the above-described embodiment. A part of each process may be executed.

  Further, the storage medium in the present invention is not limited to a medium independent of a computer, but also includes a storage medium in which a program transmitted via a LAN, the Internet, or the like is downloaded and stored or temporarily stored.

  Further, the number of storage media is not limited to one, and the case where the processing in the above embodiment is executed from a plurality of media is also included in the storage media in the present invention, and the media configuration may be any configuration.

  The computer according to the present invention executes each process in the above-described embodiment based on a program stored in a storage medium, and is a single device such as a personal computer or a system in which a plurality of devices are connected to a network. Any configuration may be used.

  In addition, the computer in the present invention is not limited to a personal computer, but includes an arithmetic processing device, a microcomputer, and the like included in an information processing device, and is a generic term for devices and devices that can realize the functions of the present invention by a program. .

  Note that the present invention is not limited to the above-described embodiment as it is, and can be embodied by modifying the constituent elements without departing from the scope of the invention in the implementation stage. Moreover, various inventions can be formed by appropriately combining a plurality of constituent elements disclosed in the embodiment. For example, some components may be deleted from all the components shown in the embodiment. Furthermore, constituent elements over different embodiments may be appropriately combined.

  DESCRIPTION OF SYMBOLS 1 ... Web server apparatus, 2 ... Application server apparatus, 3 ... Database server apparatus, 100 ... Client apparatus, 110 ... Memory | storage part, 120 ... Input part, 130 ... Communication I / F part, 140 ... Load test tool part, 150 ... Time series data analysis support unit 151... Parameter input unit 152 152 Processing amount input unit 153. Judgment unit 154. Performance data input unit 155 ... Totaling unit 156 ... Result output unit 160 ... Display unit 200 Network, CS ... computer system.

Claims (4)

A time series data analysis support program for use in a client device comprising a storage means and capable of receiving processing time series data and CPU utilization time series data in a computer system to be analyzed,
The client device;
Means for writing the inputted allowable width and the shortest period in the storage means;
Based on the allowable range in the storage unit and the shortest period, a fixed period determination unit that determines a period equal to or greater than the shortest period in a period in which the time-series data of the processing amount is within the allowable range,
An average value calculating means for calculating an average value of the time-series data of the processing amount and an average value of the time-series data of the CPU utilization rate for each determined period;
Graph display processing means for executing a graph display process in which a processing amount and a CPU usage rate are associated with each other based on the calculated average values;
Time series data analysis support program to function as In the time series data analysis support program according to claim 1,
The fixed period determination means includes
In place of the processing amount time-series data, the CPU utilization rate time-series data is a means for determining a period that is within the allowable range and a period equal to or longer than the shortest period as a fixed period, and
Means for writing the inputted allowable number of erasures in the storage means;
Means for ignoring the deviation from the allowable range up to the allowable number of detachments if the time series data of the CPU utilization rate deviates from the allowable range and the deviation from the allowable range does not continue. ,
A time-series data analysis support program characterized by including: In the time series data analysis support program according to claim 1,
The fixed period determination means includes
Means for writing the inputted allowable number of erasures in the storage means;
Based on the allowable width in the storage means, the shortest period, and the allowable number of detachments, a time period in which the time series data of the CPU usage rate is within the allowable width is determined as a fixed period. In this determination, when the time series data of the CPU utilization rate deviates from the allowable range, if it is not continuously deviated from the allowable range, up to the allowable number of deviations. Means for ignoring the deviation from the allowable width;
Means for finally determining a period in which both the period determined as the fixed period for the time-series data of the processing amount and the period determined as the fixed period for the time-series data of the CPU utilization rate are the fixed period;
A time-series data analysis support program characterized by including: A client device capable of receiving time-series data of processing amount and time-series data of CPU utilization in a computer system to be analyzed,
Storage means for storing the input tolerance and the shortest period;
Based on the allowable range in the storage unit and the shortest period, a fixed period determination unit that determines a period equal to or greater than the shortest period in a period in which the time-series data of the processing amount is within the allowable range;
Average value calculating means for calculating an average value of the time-series data of the processing amount and an average value of the time-series data of the CPU usage rate for each determined fixed period;
Graph display processing means for executing a graph display process in which a processing amount and a CPU usage rate are associated with each other based on the calculated average values;
A client device comprising:

Images