JPH01152545A - Performance measurement system for virtual machine - Google Patents

Abstract

PURPOSE:To automatically executing a performance measurement by the benchmark job of a virtual machine to a full virtual machine and an object without worrying about measuring environment and to reduce man hour by using a communicating function between the virtual machine. CONSTITUTION:By a performance measurement starting command from a console device 6, a communication processing part 5 of a virtual machine 4 opens a message log file 7 and indicates the starting of a processing to a virtual machine 15 of a user. The machine 15 outputs a text 18 to indicate the name of an operating system 17. A processing part 5 obtains an applicable benchmark job 14 from a corresponding table 11, and convert-inputs it through a spool file 12 to a benchmark job 14 of the machine 15. The machine 15 executes the job 14, outputs the data to the text 18, send them to a machine 4, stores them in the file 7 and outputs them as editing data 10 when the processing of the system 17 is completed, and the data are analyzed and evaluated.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an electronic computer system, and particularly to a computer performance measurement technique suitable for grasping the performance of a virtual computer system.

[Conventional technology]

Conventional performance measurement of computer systems is described in Japanese Patent Application Laid-open No. 59-20
As described in Japanese Patent No. 6953, the execution time of a program is measured by incorporating a microprocessor that measures instruction execution time into an actual computer. 5. The purpose of the conventional method is to grasp the performance of a program executed on an actual computer. However, when considering a virtual computer system, the performance of the virtual computer actually used is more important to the user than the performance of the actual computer. Therefore, it is necessary to know the execution time of the program on the virtual machine used by the user.

[Problem that the invention seeks to solve]

The above-mentioned conventional technology is aimed at understanding the performance of a real computer, and does not give consideration to understanding performance data required by a user using a virtual computer. For this reason, the following problems occurred.

(1) In order to understand the performance of a virtual machine, it is necessary to prepare a benchmark job in advance and evaluate it by actually executing the job on the virtual machine.

(2) When implementing the above method, accurate information cannot be obtained unless the device is not in a no-load condition and is actually being used by a user. Also, it is desirable to run the same benchmark job on all virtual machines being used, but
It requires a lot of man-hours. Therefore, it is necessary to limit the virtual computers to be measured.

In this way, measuring the performance of a virtual machine requires a difficult measurement environment and a large amount of man-hours.

An object of the present invention is to automatically measure the performance of all virtual machines using a benchmark job, which is a means of understanding the performance of virtual machines, by using a communication function between virtual machines.

[Means for solving problems]

The above purpose is to realize a data communication function between virtual machines in a virtual computer system, and to submit benchmark jobs and receive execution results without human operations. This is achieved by performing data exchange.

[Effect]

Inputting a benchmark job and receiving the results executed on a user's virtual machine sets and controls a virtual machine for center management. In this method, the center management virtual machine recognizes the operating systems on multiple virtual machines used by users and submits benchmark jobs accordingly. In addition, the execution results are sent back from each virtual machine to the center management virtual machine.

This series of operations is performed according to a "sender code" and a "destination party code" that are automatically assigned from information generated by the system.

In this way, the program recognizes this code and automatically sends and receives data, so there is no possibility of malfunction.

〔Example〕

An embodiment of the present invention will be described below with reference to FIGS. 1, 2, 3, and 4.

FIG. 1 shows the overall configuration and processing flow of the present invention. The virtual computer communication control unit 3 operates within the virtual computer management program 2 that manages the real computer 1 and controls communication between virtual computers. The virtual computer 4 is a virtual computer for operating a communication processing unit 5 that transmits benchmark jobs for performance measurement and receives results, and is managed by the center.

The virtual computers 15 and 19 are virtual computers used by users.

The specific processing flow will be explained according to FIG.

101, input a performance measurement start command from the console device 6;

102. The communication processing unit 5 of the virtual machine 4 performs an open process of the message log file 7 for storing the message text 18 output by the user's virtual machine 15.

103, the communication processing unit 5 of the virtual machine 4 sends IPL (In1tial Prog.
ram L.

ading) Instructs to start processing.

104, the virtual machine 15 processes the IPI of the operating system 17 and outputs text 18 indicating the name of the operating system 17.

The output text 18 is automatically transmitted to the virtual computer 4 by the communication control unit B.

105, the communication processing unit 5 confirms the sent daily text and operating system name.

106, the communication processing unit 5 executes benchmark job 1 for the operating system assigned from the correspondence table 11.
Get 3. Since the job control language used to control job execution differs depending on the operating system, it is necessary to prepare the job to be compatible with the operating system.

For this reason, a correspondence table 11 that associates operating system names with corresponding benchmark jobs is created.
Is required.

107, the benchmark job 13 is converted into a benchmark job 14 and input to the virtual machine 15- via the virtual machine spool file 123-.

108, The benchmark job 14 is executed on the virtual machine 15, and the start time and end time of the job are output to the message text 18.0109, The message text 1st is sent to the virtual machine 4 through the same process as in 104. Ru.

110, the transmitted message text 18 is stored in the message log file 7 by the communication processing unit 5.

111. The processes from 107 to 110 are repeated until the operating system 17 is terminated.

112, the message log file 7 is closed upon completion.

115. When the operation of the virtual machine 4 ends, the data editing processing unit 9 edits the contents of the message log file 7 and outputs it as edited data 10.

The center administrator analyzes the edited data 10 and performs all performance evaluations.

Referring to FIG. 3, virtual computing machine communication control for actually communicating data will be explained.

To transmit the transmission data 21 from the virtual machine 4 to the virtual machine 15, first, the communication processing unit 5 of the virtual machine 4 issues a specific interrupt to the virtual machine communication control unit 3 to notify the virtual machine communication control unit 3 of data transmission. Next, the data reception processing unit 3-1 of the virtual computing institution communication control unit 3 issues the l'-RECEIVEJ macro and converts the transmission data 21 into reception data 3-2 on the memory, thereby receiving the data. . When the receiver completes the transaction, the receiver notifies the communication processing section 5 of the completion of the transaction using an interrupt. Next, when the data part 21-3 of the transmission data 21 is "0", the transmission destination code 21-2 is checked and the virtual computer 15 is notified of the start of measurement. When the data section 21-3 is "1", execution of IPL is instructed.

On the other hand, the message text output by the virtual machine 15 is changed into the format of the transmission data 22, and is queued by the queuing process 3-5.

The queuing process 3-5 issues a transmission request 5-6 to the data transmission process S-7 when the queuing buffer becomes full. The data transmission process 3-7 is performed by the communication processing unit 54 of the virtual machine 4, rREcEIVEJ
Mac a1 is issued, transmit data 3-8 are converted into receive data 24 on memory, and the reception is received.

The above processing controls data communication between virtual machines.

FIG. 4 shows the method of feeding the benchmark jikob. First, virtual calculation Ig! A4 is the data part 2 of the transmission data 21
1-3 transmits data of "1" to the virtual machine 15 through the communication path 30. At the same time, the benchmark file 13 of the corresponding operating system is output from the benchmark file 8 to the virtual machine spool file 12 using the command WRITEJ. The virtual machine management program 2 changes the owner of the benchmark Jikob to the virtual machine 15. When the virtual machine 15 receives the transmission data 21'5 notifying job submission, it issues the command [READ
J reads the benchmark job 14 and executes the entire process.

In this way, benchmark jobs can be submitted to the other virtual machine by changing the attributes in the virtual machine spool file 12 managed by the virtual machine management program 2, so that even large amounts of data can be performed. It is possible. This allows the communication data size to be small.

By using the above method, the benchmark Jikob is automatically and periodically injected into multiple virtual machines used by the user.
It becomes possible to receive and receive the execution results of jobs executed on that virtual machine.

The center administrator extracts and edits the data regarding the elapsed time of the job from the results of the execution.
This has the effect that the performance of the virtual machine used by the user can be grasped.

〔Effect of the invention〕

According to the present invention, the communication function between virtual machines makes it possible to automatically submit a job to a virtual machine used by a user and receive the results.

By using this function, it is possible to automatically measure the performance of the virtual machine used by the user, which is effective in evaluating the performance of the virtual machine.

[Brief explanation of the drawing]

Figure 1 is an overall configuration diagram of an embodiment of the present invention, Figures 2 and 3.
FIG. 4 is a configuration diagram of an embodiment of the present invention. DESCRIPTION OF SYMBOLS 1... Actual computer, 2... Virtual computer management program, 3... Virtual computing institution communication control unit, 4... Virtual computer for center management, 5... Communication processing unit, 7... Message Log file, 8... Benchmark job storage file, 12... Virtual needle xm spool file, 15
...Virtual computer used by the user.

Claims (1)

[Claims] 1. In order to submit benchmark jobs for performance measurement and edit the execution results, one virtual computer is provided that is managed by the center along with the user's virtual computer, and then This communication function automatically submits benchmark jobs from the virtual machines managed by the center to the virtual machines used by users, executes the benchmark jobs on the users' virtual machines, and exchanges the execution results. The function automatically sends back the execution results to the virtual machines managed by the center, and the returned execution results are edited as performance evaluation data.In this way, the communication function between virtual machines is realized, and job submission and results are A method for measuring the performance of a virtual machine, which is characterized by automatically receiving the information.