JP2012043010A - Load distribution system and load distribution method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a load distribution system which reduces a verification processing time by distributing verification processing by means of distributed processing in a source code verification system using a model inspection technique.SOLUTION: An input information weighting part 23 makes the weight of each process constant from information, which is the input of a process of a model inspection technique, such as program information or modeling information to weight it by model inspection process units and generates a job based on the weighted process. A weight sum value calculation part 41 by devices defines the weight sum value by processes as the weight of the job, and a distribution destination device determination part 42 determines a distribution destination device of a job to be next input based on the weight of each job and the weight sum value possessed by the distribution destination device and input the job to distribute jobs in processing devices, thereby distributing load uniformly to distribution destination devices.

Description

  The present invention relates to a load distribution system, and more particularly to a load distribution system in a source code verification system using a model checking method.

  In a source code verification system using a conventional model checking method, verification processing is performed on the same device and processor regardless of the source code scale to be verified. Therefore, a large amount of verification time is required when the source code scale is large. It was.

  The model is a logical expression after the source code is converted into “Control Flow Graph”.

  The model checking method is a method that uses a model to verify all data patterns and all processing paths during program execution.

  The source code verification system is a system that makes it possible to find a description mistake or a malfunction at the time of operation without moving the program before or during the build of the program.

  Build means to compile source code, link libraries, etc., and create a final executable file.

  As a related technique, Japanese Patent Application Laid-Open No. 2004-287801 discloses an information processing system, an information processing apparatus, a distributed information processing method, and a computer program. In this related technique, the distribution destination device is determined from the availability of the processor. Also, the distribution destination device is determined from the availability of the processor. The management is performed according to the transfer status of the candidate computer and the status of the unloaded status resource. The management is performed according to the transfer status of the candidate computer and the status of the unloaded status resource. Further, the amount of information processing load requested is calculated by measuring the amount of information to be processed per unit time. Further, it is determined based on the load size, the operation status monitoring result, and other device performance information. Also, the calculation is performed based on the data sent from the main computer to the candidate computer and the size of the processing result. In addition, a coefficient value determined in advance according to each of the plurality of task execution conditions is weighted.

  Japanese Patent Application Laid-Open No. 2009-053756 discloses a source code verification processing time calculation apparatus and method. In this related technology, the source code verification processing time calculation apparatus includes a property extraction unit and a property count unit. The property extraction unit extracts property information from model information obtained by modeling the source code. The property count unit calculates a timeout period of source code verification of the source code based on the property information.

JP 2004-287801 A JP 2009-053756 A

  An object of the present invention is to provide a load distribution system that shortens verification processing time by distributing verification processing by distributed processing in a source code verification system using a model checking method.

  Distributed processing is to distribute calculation processing using a plurality of computers or processors.

  The load distribution system of the present invention includes a processing device and a distribution device. The processing device receives the source code and the logical model as input information, and calculates different weights depending on the model checking process based on the input information. The distribution device obtains the job weight from the weight for each process, determines the distribution destination processing device from the job weight, and distributes the job to the determined distribution destination processing device.

  In the load distribution method of the present invention, the processing device receives the source code and the logical model as input information, and performs calculation of different weights depending on the model checking process based on the input information. Determining the job weight from the job weight, determining the distribution destination processing device from the job weight, and distributing the job to the determined distribution destination processing device.

  The program of the present invention is a program for causing a processor to execute any one of the processing apparatus and the distribution apparatus in the load distribution method. Note that the program of the present invention can be stored in a storage device or a storage medium.

  It is possible to distribute the load almost evenly to the distribution destination processing devices.

It is a figure which shows the structural example of the load distribution system of this invention. It is a flowchart which shows the flow of the load distribution process in this invention.

<First Embodiment>
Hereinafter, a first embodiment of the present invention will be described with reference to the accompanying drawings.
As shown in FIG. 1, the load distribution system of the present invention includes an input device 1, a processing device 2, a storage device 3, and a distribution device 4.

[overall structure]
The input device 1 is a device for inputting information (input information) such as a source code and a logical model. The processing device 2 performs weighting in units of processes. The storage device 3 stores information regarding weighting. The distribution device 4 generates a job based on the weighted process, uses the total weight of each process as the job weight, and distributes the job to the corresponding distribution destination processing device according to the weight of each job.

  In addition, the processing apparatus 2 may be plural. The storage device 3 may be integrated with the processing device 2. The processing device 2 and the distribution device 4 are assumed to be independent from each other.

[Relationship between jobs and processes]
A plurality of processes that are successively executed for one process are treated as one job. That is, a plurality of processes gather to form one job. Of course, there may be one process per job.

[Configuration of each device]
The processing device 2 includes an input information storage unit 21, a process search unit 22, and an input information weighting unit 23.

  The storage device 3 includes an input information storage unit 31, a process-specific weighting rule storage unit 32, and a weight storage unit 33.

  The distribution device 4 includes a device-specific weight total value calculation unit 41, a distribution destination device determination unit 42, and a job distribution unit 43.

  The input information storage unit 21 stores information such as a source code and a logical model given from the input device 1 in the input information storage unit 31.

  The input information storage unit 21 extracts information such as a source code and a logical model from all information given from the input device 1 as necessary, and stores it in the input information storage unit 31. May be. For example, the input information storage unit 21 may extract only information such as a source code or a logical model described in a specific program language according to a preset condition and store it in the input information storage unit 31. The type of source code or logical model can be determined by the document structure or extension.

  Further, the input information storage unit 21 may notify the process search unit 22 that the input information has been stored in the input information storage unit 31 or the input information itself.

  The process search unit 22 periodically monitors the input information storage unit 31 or acquires information stored in the input information storage unit 31 based on a notification from the input information storage unit 21.

  Based on the information stored in the input information storage unit 31, the process search unit 22 selects a corresponding weight calculation formula from the weight calculation formula (weighting rule) information group stored in the process-specific weight rule storage unit 32. And the information of the found weight calculation formula is notified to the input information weighting unit 23.

  The input information weighting unit 23 calculates the weight for each process from the information stored in the input information storage unit 31 and the weight calculation formula information stored in the process-specific weighting rule storage unit 32 notified from the process search unit 22. And the calculated weight is stored in the weight storage unit 33.

  Here, the input information weighting unit 23 converts the weight of each process into a constant from information that becomes an input of the process of the model checking method, such as program information and modeling information, based on the weight calculation formula.

  Further, the input information weighting unit 23 may notify the device-specific weight total value measurement unit 41 that the calculated weight is stored in the weight storage unit 33 or the calculated weight itself.

  The input information storage unit 31 stores information such as a source code and a logical model given from the input device 1.

  The process-specific weighting rule storage unit 32 stores weight calculation formula information prepared for each model checking process.

  The weight storage unit 33 stores the weight calculated by the input information weighting unit 23.

  The device-by-device weight total value measuring unit 41 periodically monitors the weight storage unit 33 or acquires the weight stored in the weight storage unit 33 based on the notification from the input information weighting unit 23.

  In practice, the distribution device 4 may include the weight storage unit 33. That is, it is possible for the distribution device 4 to centrally manage the weight of each process possessed by a plurality of processing devices. In this case, the weight storage unit 33 stores the identification information of the processing device and the weight of the processing device in association with each other.

  The device-by-device weight total value measurement unit 41 calculates the sum of weights for each process of each processing device that is a candidate for the distribution destination device. Here, the device-specific weight total value measuring unit 41 calculates the weight total value from the weights for each process stored in the weight storage unit 33 for each processing device in units of jobs. The calculated weight total value is a weight of a job configured by these processes.

  The distribution destination device determination unit 42 determines a distribution destination device from the weight total value (job weight) calculated by the device-specific weight total value measurement unit 41. Here, the distribution destination device determination unit 42 determines the processing device having the smallest calculated weight total value as the distribution destination device.

  Actually, the distribution destination device determination unit 42 determines the priority of the distribution destination processing device and the job distributed to the processing device according to the calculated weight total value (in order of the weight total value). The amount may be determined.

  The job distribution unit 43 submits a job to the distribution destination device determined by the distribution destination device determination unit 42.

  Note that the job distribution unit 43 may generate a job according to the weight for each process for the job to be submitted. For example, the job distribution unit 43 may generate jobs by combining the processes so that the weights of the jobs are equal. Or, a low-weight job can be generated by combining a low-weight process for a low-performance processing device, or a high-load job can be generated by combining a high-weight process for a high-performance processing device. good.

[Hardware example]
Examples of the input device 1 include a keyboard, a keypad, a keypad on the screen, a touch panel, a tablet, or a reading device that reads an IC chip or a storage medium. Alternatively, the input device 1 may be an interface (I / F) for acquiring information from an external input device or a storage device.

  As an example of the processing device 2, the storage device 3, and the distribution device 4, a computer such as a PC (personal computer), a workstation, a main frame, or a supercomputer is assumed. The processing device 2 and the distribution device 4 may be processors. The storage device 3 may be a memory. Further, the processing device 2, the storage device 3, and the distribution device 4 are not limited to terminals and servers, but may be relay devices or peripheral devices. The processing device 2, the storage device 3, and the distribution device 4 may be an expansion board mounted on a computer or a virtual machine (Virtual Machine (VM)) built on a physical machine.

  Other examples of the storage device 3 include DAS (Direct Attached Storage), FC-SAN (Fibre Channel-Storage Area Network), NAS (Network Attached Storage), and IP-SAN (IP-Storage Center). The storage device that was used can be considered.

  Examples of networks connecting the input device 1, the processing device 2, the storage device 3, and the distribution device 4 include the Internet, a LAN (Local Area Network), a wireless LAN (Wireless LAN), a WAN (Wide Area Network), and a backbone. (Backbone), cable TV (CATV) line, fixed telephone network, mobile phone network, WiMAX (IEEE 802.16a), 3G (3rd Generation), leased line, IrDA (Infrared Data Association), Bluetooth (registration) Trademark), serial communication line, data bus, and the like.

  Examples of the input information storage unit 21, the process search unit 22, the input information weighting unit 23, the device-specific weight total value calculation unit 41, the distribution destination device determination unit 42, and the job distribution unit 43 are a CPU (Central Processing Unit), a micro A processor, a microcontroller, a semiconductor integrated circuit (Integrated Circuit (IC)) having a dedicated function, or the like can be considered.

  Examples of the input information storage unit 31, the process-specific weighting rule storage unit 32, and the weight storage unit 33 include a RAM (Random Access Memory), a ROM (Read Only Memory), an EEPROM (Electrically Erasable and Programmable Read Only Memory), and a flash memory. Semiconductor storage devices such as HDD (Hard Disk Drive) and SSD (Solid State Drive), or other removable storage devices such as DVD (Digital Versatile Disk), SD memory card (Secure Digital memory), etc. A storage medium (media) can be considered.

  However, actually, it is not limited to these examples.

[Load distribution processing]
Next, the load distribution process of the present invention will be described in detail with reference to the flowchart of FIG.

(1) Step A1
First, the input information storage unit 21 receives information provided from the input device 1 and stores the received information in the input information storage unit 31.

(2) Step A2
Next, the process search unit 22 searches the weight calculation formula for each process from the process-specific weight rule storage unit 32 based on the information stored in the input information storage unit 31.

(3) Step A3
Next, the input information weighting unit 23 calculates the weight for each process based on the information stored in the input information storage unit 31 and the weight calculation formula extracted from the process-specific weighting rule storage unit 32. The calculated weight is stored in the weight storage unit 33.

(4) Step A4
Next, the device-specific weight total value calculation unit 41 calculates the weight total value for each process held by each processing device at the distribution destination. Specifically, the device-specific weight total value measurement unit 41 calculates, for each processing device, a total value of weights for each process stored in the weight storage unit 33 for each job, and uses this weight total value for the job. Use weights.

(5) Step A5
Next, the distribution destination device determination unit 42 determines the distribution destination device from the total weight value (job weight) calculated by the device-specific weight total value calculation unit 41. At this time, the device having the smallest job weight is determined as the distribution destination device.

(6) Step A6
Next, the job distribution unit 43 submits the job to the distribution destination processing device determined by the distribution destination device determination unit 42.

[Example of weight calculation formula (weighting rule)]
In the load distribution system of the present invention, the coefficient value is calculated and determined from the input scale and execution time of the first executed process as a weight calculation formula (weighting rule).

Here, the process of the model checking method is as follows.
"Modeling source code" → "Model checking"

Input information of each process is as follows.
Source code modeling: Source code Model checking: Model

The input scale of each process is as follows.
Process 1: Source code = 3, Model = 5
Process 2: source code = 4, model = 7

The execution time of each process is as follows.
Process 1: Source code modeling = 5 minutes, Model check = 10 minutes

Calculate the input scale × execution time of each process in the process 1.
Source code modeling: 3 × 5 = 15
Model check: 5 × 10 = 50

The ratio of each process in the process 1 is calculated.
The numerical value calculated here is a coefficient value.
“Modeling of source code”: “Model checking” = 15: 50 = 3: 10

The weight of each process is calculated by multiplying the input scale of process 2 by the coefficient value.
Source code modeling: 4 × 3 = 12
Model check: 7 × 10 = 70

  Similarly for the subsequent processing, the weight of each process is calculated.

[Result]
According to the present invention, it becomes possible to distribute the load almost evenly to the processing apparatuses of the distribution destination.

  In addition, weights that could not be given by the conventional method can be given to each model checking process and made constant. As a result, the load is evenly distributed to the distribution destination processing devices, and the overall processing time can be shortened.

[Features of the present invention]
The present invention relates to a load distribution method for a distribution apparatus in model checking.

  In the present invention, for a distributed processing method of a source code verification system using a model checking method, weighting is performed for each process of model checking, a job is generated based on the weighted process, and a total weight value for each process is calculated. A method of distributing the load evenly to the distribution destination devices by using the job weight and distributing the job to each processing device according to the weight of each job is used.

  Specifically, the weight of each process is made constant from information that is input to the process of model checking, such as program information and modeling information, and the distribution of the job to be submitted next is calculated from the total weight value of the processing device at the distribution destination Determine the destination device. This makes it possible to equalize the load distribution.

  In the present invention, the input source code scale and model size are used.

  In the present invention, each model check process is distributedly executed as one processing unit.

  In the present invention, the total weight of each process held by the distribution destination device is used as the job weight, the distribution destination device is determined based on the job weight, and the distribution destination device is managed.

  In the present invention, the magnitude of the load is calculated from each process of model checking and input information of each process. For example, the magnitude of the load is calculated from the model size and a default (initial setting) coefficient assigned to the model checking process.

  In the present invention, the weight measured from the source code scale and model size is multiplied by the coefficient value given to each model checking process.

  In the present invention, the coefficient value is determined by taking the ratio of each process to the product of the size of the input information of each process and the execution time.

  In the present invention, load distribution is evenly distributed to the distribution destination devices by performing distributed processing by making the weight of each process constant from the input information of the process of the model checking method.

  The load distribution apparatus according to the present invention is installed in a processing apparatus and calculates different weights depending on the model checking process.

  Or the load distribution apparatus which concerns on this invention is installed in a processing apparatus, and provides the information of the calculation formula of a weight which changes with processes of a model check.

  Alternatively, the load distribution apparatus according to the present invention is installed in the distribution apparatus and calculates the weight of a job that exists for each apparatus.

  Alternatively, the load distribution apparatus according to the present invention is installed in the distribution apparatus and determines the distribution destination apparatus from the job weight for each apparatus.

  As mentioned above, although embodiment of this invention was explained in full detail, actually, it is not restricted to said embodiment, Even if there is a change of the range which does not deviate from the summary of this invention, it is included in this invention.

DESCRIPTION OF SYMBOLS 1 ... Input device 2 ... Processing apparatus 21 ... Input information storage part 22 ... Process search part 23 ... Input information weighting part 3 ... Memory | storage device 31 ... Input information storage part 32 ... Weighting rule storage part classified by process 33 ... Weight storage part 4 ... Distributing apparatus 41 ... Device-specific weight total value calculating section 42 ... Distributing destination apparatus determining section 43 ... Job distributing section

Claims (10)

A processing device that receives source code and a logical model as input information, and calculates different weights depending on the model checking process based on the input information;
A load distribution system comprising: a distribution device that obtains a job weight from a weight for each process, determines a distribution processing device from the job weight, and distributes the job to the determined distribution processing device. The load distribution system according to claim 1,
The processor is
An input information storage unit for receiving a source code and a logical model as the input information;
Based on the input information, a process search unit for searching for a weight calculation formula for each process;
A load distribution system comprising: the input information; and an input information weighting unit that calculates a weight for each process based on the discovered weight calculation formula. The load distribution system according to claim 1 or 2,
The dispersing device includes:
A weight total value measuring unit for each device that calculates the weight of the job by summing the weights for each process of each processing device that is a candidate for the distribution destination device;
A distribution destination device determination unit that determines a distribution destination device based on the weight of the job;
A load distribution system comprising: a job distribution unit that inputs a job to the determined distribution destination device.   The processing apparatus used with the load distribution system as described in any one of Claims 1 thru | or 3.   The distribution apparatus used with the load distribution system as described in any one of Claims 1 thru | or 3. In the processing device, receiving source code and a logical model as input information, and performing calculation of different weights depending on the model checking process based on the input information;
A load distribution method, comprising: obtaining a job weight from a process-by-process weight in a distribution apparatus; determining a distribution destination processing apparatus from the job weight; and distributing the job to the determined distribution destination processing apparatus. The load distribution method according to claim 6,
In the processing apparatus,
Retrieving a weight calculation formula for each process based on the input information;
A load distribution method further comprising: calculating a weight for each process based on the input information and the found weight calculation formula. The load distribution method according to claim 6 or 7,
In the dispersing device,
Calculating the job weight by summing the weights for each process of each processing device that is a candidate for the distribution destination device;
Determining the destination device based on the weight of the job;
A load distribution method further comprising submitting a job to the determined distribution destination device.   The program for making a processor perform operation | movement as a processing apparatus in the load distribution method as described in any one of Claims 6 thru | or 8.   The program for making a processor perform the operation | movement as a distribution apparatus in the load distribution method as described in any one of Claims 6 thru | or 8.

Images