JP2001325041A - Method for utilizing computer resource and system for the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To open and resell surplus parts of the information processing capabilities of a computer connected to a network to a third person, and to allow a user to easily use computer resources, as necessary. SOLUTION: A request from a user 3 for calculation processing is accepted (step 104), and the calculation processing is divided into plural calculation unit blocks which can be processed in parallel (step 105), and the assignment of the calculation unit blocks to each provider 4 is decided (step 106), and the calculation unit blocks respectively assigned to the plural providers 4 are distributed (step 107). Each provider 4 calculates the calculation unit blocks (step 108), and transmits the results to a center system 2 (step 109). The results are gathered (step 110), and transmitted to the user 3 (step 111), and charging processing is executed (steps 112-115).

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a system including a large number of computers connected via a network, and more particularly to a computer for reselling and providing a surplus portion of computer resources for others. It relates to a resource utilization method and a system.

[0002]

2. Description of the Related Art As a calculation requiring a large amount of computer time and computer power, for example, various simulations (large-scale simulation mainly using vector operation,
Image processing such as evacuation in the event of a disaster), image processing such as rendering, risk analysis that requires a large number of case studies, optimization problems with a very large search area, and computer-aided design using the finite element method ( CAD; computer assited design). Conventionally, these calculations have been performed using a high-performance computer such as a supercomputer.

[0003] On the other hand, the processing capacity of personal computers (PCs) and server computers in recent years has been remarkably improved, and from the viewpoint of the rate of increase in performance in recent years, it can be said that the performance has greatly exceeded that of supercomputers and the like. Processors used in personal computers and servers are also on the market inexpensively and in large quantities. As a result, even if the number of the computers is of the personal computer class, the information processing capacity of the computer as a whole has surpassed that of the supercomputer.

Therefore, instead of a supercomputer using a single or a small number of processors, there is a type of parallel computer in which a large number of personal computer class processors are arranged and connected to each other, and each processor shares processing. Proposed. Further, there has been proposed a distributed computer in which computers that can operate independently, such as personal computers, are connected to a network, and the computers are distributed to execute processing. In particular, performing parallel / distributed computation using computer resources (such as a personal computer) dispersedly connected to a wide area network is also referred to as global computing (global computing) or cluster computing.

[0005]

Since the processing power that can be executed by a single computer is limited, a case where the above-described large-scale calculation is always performed is aside.
Providing your own computing power that can handle such large-scale calculations when you do large-scale calculations that are not so frequent requires the labor of purchasing expensive computers and maintaining them Therefore, it becomes quite expensive. Even when a large-scale calculation is performed constantly, it is difficult to assume that the required calculation amount does not change with time, so computers and the like are deployed according to the peak request amount. It will be difficult to get results worth the investment.

[0006] Considering the case of outsourcing information processing in a company, at present, only services using limited computers operated by contractors can be obtained. This outsourcing service is not always suitable for applications that require extreme computing power, and at least for a certain period of time, the usage fee is fixed, and for service users, the validity of the service fee is determined in real time. It was difficult to grasp.

As described above, the number of personal computers and servers spread and the processing capacity tend to increase significantly. For example, considering the use of personal computers in offices, it is necessary to perform processing suitable for the improvement in the processing capacity of the personal computers. Is not running, and there is considerable room for processing power. Specifically, at very limited peak times, they may be using their PCs to their full potential, but most of the rest of the time they are running out of PCs. The extra capacity in such processing power can be used to share and perform computations by techniques such as global computing, and potentially to provide it to others and receive compensation for it. However, at present, there is no mechanism for releasing the extra portion for use by others as much as one wants. Also, even for users who perform large-scale calculations, even if there is enough hardware in the world to perform the calculations using global computing, those hardware can be used to perform the large-scale calculations. There is no mechanism to do this.

[0008] Accordingly, an object of the present invention is to allow a computer connected to a network to open a surplus portion of information processing capability to others and resell it, and to perform calculations for users who intend to perform calculations. It is an object of the present invention to provide a method and a system for making it possible to easily use computer resources as needed.

[0009]

SUMMARY OF THE INVENTION The present invention provides a method and system for purchasing and reselling the remaining information processing capacity of a computer accessible to a network such as a wide area network (WAN). The purchase cost of the surplus part may be fixedly set, but it is better to provide a part that changes in conjunction with the supply and demand balance of both the capacity providing side and the service receiving side in order to provide incentives. .

[0010] In the present invention, a person (provider) who intends to resell the remaining capacity in computer resources for another person registers the fact in advance in the center. The registered information is stored in the database of the center. Then, the center divides a series of processes requested by a user (a person who wants to enjoy the service) into a plurality of unit processes that can be processed in parallel, and allocates the divided processes to a spare capacity of a plurality of computers on the provider side. This makes it possible to realize an inexpensive and extremely high-speed processing capacity. By dividing a series of processes requested by the user into a plurality of unit processes and executing the processes on a plurality of computers, data security at the time of performing the information processing service for the user is enhanced.

When it is necessary to increase the reliability of the information processing service for the user, the same unit processing is allocated to a plurality of computers in parallel.

To enable the above-described processing, each computer on the capability providing side is equipped with a software module (client software) for that purpose. The software module includes various functions necessary for realizing each of the above-described services while ensuring security, and is preferably expandable in the future.

[0013]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, a preferred embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the overall configuration of a computer resource utilization system according to an embodiment of the present invention. The computer resource utilization system is a network system to which the computer resource utilization method of the present invention is applied. It is assumed that a computer user who can access the network system purchases a surplus part of information processing capability and wants to utilize the surplus part. This is a system for resale to users. The calculation that the user wants to perform is based on the method of global computing,
Distributed / parallelized by a number of computers trying to provide extra power.

The computer resource utilization system shown in FIG. 1 includes a wide area network (WAN) 1, a center system 2 is connected to the wide area network 1, and a user 3 who intends to use computer resources. And a computer 5 owned by a provider 4 who intends to provide a surplus portion of information processing capability or having a right to use. The number of computers 5 to which each provider 4 intends to provide spare capacity may be one or more for each provider 4. On the other hand, it is highly desirable that the number of the providers 4 be plural from the viewpoint of security and the like as described later. The number of computers 5 as a whole is set to a number larger than or equal to a value that enables effective calculation by global computing. In practice, the provider 4 registers the computer 5 owned by the provider 4 in the center system 2 (offers a sale of a surplus part of the information processing capability), so that each computer 5 is connected to the center system 2 via the wide area network 1. It will be logically connected. The wide area network 1 may be a network having a speed sufficient for performing global computing, and may use the Internet or the like. In this case, as described later, the center system 2 side sends the computer of the provider 4 Since data is transmitted and received at any time with respect to 5, it is extremely preferable to be of a constant connection type.

The center system 2 accepts a request for sale of the remaining information processing capacity from the provider 4, receives a calculation processing request from the user 3, and can perform the calculation processing requested by the user 3 in parallel. The processing is divided into a plurality of unit processes, each unit process (calculation unit block) is distributed to the computer 5 of each provider 4, the processing results are collected from these computers 5, and the result is returned to the user 3 by the user 3. To do. Further, the center system 2 executes a billing, billing, and clearing process for the user 3 and a payment process for the provider 3 based on the usage history. Further, in the present invention, the purchase price / sales price of the surplus information processing capacity is determined according to the demand balance between the provision (supply) of the information processing capacity and the use (demand) of such capacity. However, when price setting is performed in accordance with such a supply and demand balance, the center system 2 also executes a process of calculating a purchase price / sales price and presenting it to the provider 4 / user 3. .

Next, the overall processing flow in the computer resource utilization system will be described with reference to FIG.

The provider 4 who intends to provide the surplus computing capacity of the computer 5 owned / managed by the user first registers the provider 4 with the center system 2 via the wide area network 1. Center system 2
(Step 101). Here, it is assumed that a plurality of providers 4 have made a request for provider registration. The center system 2 receives the request from each provider 4 and registers the provider 4 in a surplus processing capacity database described later so that the provider 4 can share and execute the calculation process of the user 3. For example, necessary settings (provider registration) inside the center system 2 are performed (step 102). After completing the provider registration, the center system 2 distributes the client software to each provider 4 via the wide area network 1 (step 103). Provider 4
Installs the client software on its own computer 5. This client software is connected to the center system 2 to receive a calculation unit block, processes the calculation unit block, and transmits the calculation result to the center system 2 in order to share and execute the calculation processing by the method of global computing. , Which is commonly used irrespective of the user 3 requesting the calculation processing or the content of the calculation processing actually requested. That is, once the client software is installed, the computer 5 can be any calculation unit block derived from any user 3 (although of course there are restrictions on the operating system (OS) used and computer architecture). Can be processed. In order to secure data and prevent unauthorized use of the system, the connection between the center system 2 and the computer 5 of the provider 4 preferably uses an encrypted dedicated protocol. It is preferable to support such an encrypted dedicated protocol. The processing of steps 101 and 103 described above can be executed as a known processing of registering on a website on the Internet and receiving distribution of software. Of course, the client software is stored on a recording medium such as a CD-ROM, and the CD-RO
It is also possible to physically distribute M to the provider 4. In this case, the provider 4 stores the CD-ROM in the computer 5
The client software is installed in the computer 5 by executing a predetermined command. On a recording medium such as a CD-ROM, an online sign-up program for provider registration is stored in addition to the client software, and the recording medium is distributed to the provider 4 in advance, and the provider is signed by online sign-up. The registration may be performed, and the client software may be installed in the computer 5.

On the other hand, the user 3 who wants to request the calculation processing,
It connects to the center system 2 via the wide area network 1 and requests the center system 2 for calculation processing (step 104). In this embodiment, as described later, since the calculation processing unit price can be changed according to the supply and demand balance of the information processing ability, in such a case, the user 3 is presented by the center system 2. After confirming the current calculation processing unit price, a request for calculation processing is made. A calculation job to be requested to be processed can be transferred from the user 3 to the center system 2 by a protocol such as FTP (file transfer protocol).

The center system 2 that has received the request for the calculation process divides a series of processes (jobs) requested by the user 3 into a plurality of unit processes (calculation unit blocks) that can be processed in parallel (step 105). The allocation of the calculation unit block to each provider 4 is determined so as to be most suitable for executing a series of processing, for example, to obtain the calculation result in the shortest time by the parallel processing by the plurality of providers 4. (Step 106), and distributes each of the allocated calculation unit blocks to the corresponding provider 4 via the wide area network 1 (Step 107). Each provider 4 that has received the distribution of the calculation unit block executes the arithmetic processing of the calculation unit block (step 108), and transmits the calculation result to the center system 2 via the wide area network 1 (step 109).

The center system 2 compiles the calculation results transmitted from each provider 4 (step 11).
0), and transmits to the user 3 as the processing result (step 1)
11). If the requested computation is a search problem with a wide search range and each calculation unit block is a simple division of the search range into smaller blocks, combine the special results. Instead, the search result (calculation result) for each block may be directly returned to the user 3 as the processing result. However, if the method of dividing the processing (job) in step 105 is complicated, Since it is difficult to grasp the result of the requested processing only by the calculation result for each calculation unit block, the calculation result is edited based on the method of dividing the processing and the edited calculation result is used as the processing result Is transmitted to the third party. For example, in a case where a large-scale matrix operation is divided into sub-matrix operations, even if the calculation result of each sub-matrix is transmitted to the user 3 as it is, it is difficult for the user to understand the original result of the matrix operation. As a summary of the results, it is desirable to edit the results so that the results of the original matrix operation can be easily understood.

After that, the center system 2 performs a billing process. That is, the center system 2 determines the amount of execution based on the amount of processing performed by the computer 5
By calculating the product of the execution amount and the unit price, a charge is calculated (step 112), and a fee is charged to the user 3 (step 113). The user 3 pays a fee (step 114), and the center system 2 pays each user 4 according to the amount of execution for each user (step 115).

FIG. 3 shows an example of the configuration of the center system 2 that executes such processing.

The center system 2 is a supervisor control unit 1 for integrating the entire center system 2.
1 and the amount of information processing capability provided by the provider 4 and the user 3
A CPU price setting mechanism 12 for setting a price per unit calculation capacity (CPU) based on the demand-supply balance of the required processing capacity 3 of the user, and a plurality of calculation processes (jobs) requested by the user 3 suitable for global computing. A job division mechanism 13 for dividing the job into unit processes (unit jobs), a surplus processing capacity database 14 for storing information on the provider 4 and the computer 5 thereof, For example, the scheduler 15 and the assignment result of the scheduler 15 determine which unit process should be assigned to the provider 4 (which computer 5) so that the efficiency is the highest, for example. A processing distribution / monitoring mechanism 16 which distributes the results to the computer 5 of the provider 4 and collects the results; . Supervisor control unit 1
1 executes processing such as registration of the provider 4, reception of a processing request from the user 3, compilation of calculation results, transmission of the calculation results to the user 3, and billing processing. The processing distribution / monitoring mechanism 16 also monitors the state of the network with the provider 4 via the wide area network 1.

Hereinafter, the CPU price setting mechanism 12, the job division mechanism 13, the surplus processing capacity database 14, the scheduler 15, and the processing distribution / monitoring mechanism 16 will be described in more detail.

As described above, the CPU price setting mechanism 12 is a mechanism for determining a price suitable for the supply and demand balance. When the CPU 4 purchases the remaining information processing capacity from the provider 4 and resells it to the user 3 For each price of
Since supply and demand can always change, the latest value is always calculated as the unit calculation capacity price (CPU price).
And the user 3.

Here, the unit calculation capacity price will be described. The information processing capability of the computer 5 possessed by each provider 4 varies widely, and the information processing capability that can be provided as a spare portion varies from provider to provider and can change every moment. Since the data transfer speed and the like between the center system 2 and the provider 4 in the wide area network 1 can also change, here, the relative processing capacity including the network throughput is calculated and set as a unit calculation capacity value. . Specifically, as described later,
The monitoring mechanism 16 periodically transmits a test program (probe) to each computer 5, and each computer 5 processes the test program and transmits a processing result to the center system 2 (processing distribution / monitoring mechanism 16). I do.
The center system 2 calculates the processing capability of each computer 5 using the response time required to transmit the test program to each computer 5 and receive the correct answer to the test program. At this time, for example, the processing capacity of the reference personal computer provided in the center system 2 is set to “1”, and the relative processing capacity value is set as the unit calculation capacity. The measurement of the unit calculation capability of each computer 5 is periodically performed, and is stored in the surplus processing capability database 14 as the processing capability time-series data of the provider 4.

The CPU price setting mechanism 12 refers to the surplus processing capacity database 14 and, for example, according to the demand-supply curve model in economics, the total of the unit calculation capacities registered and currently available and the request from the user 3 The unit calculation capacity unit price (CPU unit price) is set so that the total of the unit calculation capacity assumed to be necessary for the performed processing becomes equal. That is, if the total provision capacity (supply) <the total required capacity (demand), the CPU price is increased, and the total provision capacity (supply)>
In the case of the total required capacity (demand), the CPU price is decreased. In this case, the demand and supply change according to the price presented to the user 3 and the provider 4, so that the CPU price setting mechanism 12 adjusts the CPU price at regular intervals. Thereby, the time series data of the CPU price is created.

When the CPU price is adjusted as described above, the CPU price is considered to behave like a stock price. Therefore, if excessive price fluctuations in a short period of time are undesirable, it is possible to limit the price fluctuation rate (or price fluctuation range) per hour and to stabilize the price fluctuations to some extent. It is. This corresponds to price limits in the stock market. Alternatively, a reference price may be set on a weekly or monthly basis, and the price may be varied within a predetermined range according to supply and demand. You can also change the price depending on when the information processing capacity is provided (for example, make it cheaper at night), set a volume discount for doing a lot of calculations,
It is possible to allow a user to reserve a desired information processing capability in advance for the same time period every day over a certain period.

In any case, in this embodiment, the provided information processing capability is calculated by (processing time of actual result) ×
The price determined by the CPU price is paid to the provider 4 and the user 3 who commissions the calculation is charged (processing time of actual result) × CPU price + commission.

Next, the job dividing mechanism 13 will be described.

As described above, the job dividing mechanism 13
A process for dividing a calculation process (job) requested by the user 3 into a plurality of unit processes (unit jobs) suitable for global computing is executed. Specifically, the job dividing mechanism 13 divides the entire calculation process requested by the user 3 into minimum unit operations (the minimum processing unit that does not produce any economic merit even if it is further divided). By classifying the minimum unit operation into a sequence procedure and a parallel procedure, integrating multiple minimum unit operations, and dividing it into calculation unit blocks that have the shortest total response time (the time required to answer the result), The calculated calculation process is divided into a plurality of unit processes. Here, the time-series procedure is a process executed using the result of the preceding process, and the parallel procedure is a process that can be executed in parallel because there is no dependency between the processes.

When dividing a job, if the number of calculation unit blocks is large, processing required for communication and data transfer increases. Further, since the calculation unit blocks that can be processed in parallel can be simultaneously processed by the plurality of computers 5 of the plurality of providers 4, the total response time can be reduced.

These processes are generally performed when performing global computing.
It can be done automatically. However,
In order to further increase the degree of parallel processing and further reduce the total response time, the processing itself requested by the user 3 is designed or coded so as to be suitable for such parallel processing. It is important that Rewriting a certain process into a form particularly suitable for parallel processing requires knowledge of parallel programming, and it is expected that the general user 3 will be left behind.

Therefore, it is conceivable that a program library is built in the center system 2 by previously accumulating various simulation programs and calculation programs that are likely to be used, such as optimization problems. in this case,
The user 3 can execute the selected calculation program by a global computing method only by selecting a desired calculation program in the library and transmitting parameters and data used in the calculation program to the center system 2. This eliminates the problem of rewriting programs for parallel processing. Of course, each calculation program in the program library is assumed to be optimized for execution by global computing.
Another advantage of providing a program library in this way is that the provider 4 has a reliable center system 2 instead of a program executed on its own computer 5 being a calculation program created by an unknown user 3. This is a calculation program that has increased security confidence.

Next, the surplus processing capacity database 14 will be described.

The surplus processing capacity database 14 stores the provider 4 who has registered to provide a surplus part of the information processing capacity.
Is a database that stores information about More specifically, the surplus processing capacity database 14 includes, for each provider 4, information such as the contact information of the provider 4 and the payment procedure / actuals as provider information, and computer resources provided by the provider 4 as resource information. CPU (Central Processing Unit) / OS
Information, available time zone, load priority, etc., information such as network protocol and connection procedure used to connect to the computer 5 of the provider 4 as connection information, and measurement of processing capacity as a performance log Information such as a result and a processing result of the calculation unit block is stored, and a log such as a payment and a report is stored as a report log.

Next, the scheduler 15 will be described.

The process requested by the user 3 is divided into a plurality of calculation unit blocks by the job dividing mechanism 13 as described above.
It is determined which computer 5 of which provider 4 allocates and executes each of these divided calculation unit blocks. At this time, the scheduler 15 determines each provider 4 (so that the calculation result can be obtained as soon as possible based on the provider information and the resource information stored in the surplus processing capacity database 14 and the information on the current processing capacity. Each calculation unit block is allocated to the computer 5). The criterion of the assignment is not only that the calculation result can be obtained earlier, but also that a more accurate result is obtained, or
It is also possible to adopt a method in which a result can be obtained at a lower cost (because a time zone price is set, etc.). A plurality of allocation standards may be prepared in advance, and one of the standards may be selected according to the request of the user 3 or the nature of the calculation process, and the allocation may be performed based on the selected standard. Good. In addition, the provider 4 may also specify a time zone for providing a spare portion of the information processing capacity. In such a case, the allocation of the calculation unit block is performed in consideration of the specified time zone. Do. An allocation method that considers a time zone is also included in the category of the allocation standard here. Further, by setting the wide area network 1 itself on a global scale, the time difference is used to select only the provider 4 that is always in the nighttime (a time zone in which the margin of information processing capability is considered to be large). It is also possible to perform the assignment, and thereby, it is possible to process the calculation of the user 3 at higher speed and at lower cost.

When allocating the calculation unit block, it is most efficient to execute a series of time series processing by a single provider 4 because the overhead related to the network is omitted. In order to prevent the provider 4 from guessing the contents and data of the calculation processing 3 and to increase security, it is preferable to intentionally assign a calculation unit block to a plurality of providers 4. In addition, it is preferable to divide and assign as much as possible the providers 4 in parallel processing in order to increase the security of the contents of calculation processing and data.

In this embodiment, generally, one calculation unit block is allocated to one computer 5 of the user 4 and transmitted to the computer 5, and the processing of the allocated calculation unit block is performed. When the computer 5 is completed, the next calculation unit block is transmitted to the computer 5 again. At that time, provider 4
Because the information processing capabilities provided by may vary,
Instead of first allocating all the calculation unit blocks corresponding to the processing requested by the user 3, dynamically allocate the calculation unit blocks according to the currently provided information processing capability and the like. Is desirable. That is, 1
Each time one calculation unit block is completed and the next calculation unit block is allocated, it is desirable to review which provider 4 to allocate. In addition, when high calculation accuracy is required, or when it is desired to obtain a calculation result anyway, the same calculation unit block may be assigned to a plurality of providers 4 (or a plurality of computers 5).

A probe for calculating the current processing capacity of the provider 4 (computer 5) is also periodically allocated to each provider 4 and each computer 5 as a calculation unit block.

Next, the processing distribution / monitoring mechanism 16 will be described.

The basic function of the processing distribution / monitoring mechanism 16 is to distribute each calculation unit block to the computer 5 of the corresponding provider 4 via the wide area network 1 based on the assignment determined by the scheduler 15. This is to receive a response such as a processing result or a calculation progress status for the delivered calculation unit block. When receiving the response, the processing distribution / monitoring mechanism 16 sends the response to the supervisor control unit 11 and notifies the scheduler 1 of the response.
Notify 5 At this time, if no response is obtained within the assumed time, it is considered that a failure has occurred in the computer 5 of the provider 4 or a failure has occurred in the communication line configuring the wide area network 1 and the provider 4 Is instructed to cancel the calculation, and the fact that no response was obtained
Is returned to the scheduler 15 and “calculation failure” is returned to the scheduler 15 to cause the scheduler 15 to determine a provider to distribute the same calculation unit block again.

When a response to the probe is returned, the processing distribution / monitoring mechanism 16 determines the processing capability of (the computer 5 of) the provider 4 according to the time until the response to the probe is returned. Then, the determined processing capacity is recorded in the surplus processing capacity database 14. It should be noted that the provider 4 does not know which is the calculation unit block of the probe and which is the actual calculation unit block of the user 3.

Here, the calculation unit block will be described. FIG. 4 shows an example of the configuration of a calculation unit block. The calculation unit block 21 includes an authentication information block 22 for performing authentication of the calculation unit block itself, a data import block 23 that describes a process of importing data used for calculation from the center system 2, and an actual calculation unit block. And a data export block 25 describing a process of exporting the calculation result to the center system 2 side. As described above, except for the authentication information block 22, the calculation unit block 21 is composed of executable codes. The executable code is preferably a binary code if the calculation speed is prioritized. If the executable code is executable regardless of the architecture or OS of the computer 5 of the provider 4 and the security is emphasized, the programming is preferably performed. Language J
It is preferable to use a code described in AVA (registered trademark) or the like.

Next, processing on the computer 5 side of the provider 4 will be described.

As described above, the provider 4 registered in the center system 2 downloads client software from, for example, the website of the center system 2 and installs it on the computer 5 owned / managed by the provider. FIG. 5 is a flowchart showing processing in the computer 5 when the calculation unit block is sent from the center system 2 to the computer 5 of the provider 4 in a state where the client software is installed in this way.

When the calculation unit block is sent, the client software installed in the computer 5 receives the calculation unit block (step 121). First, the client software executes the calculation unit block to maintain security. Using the authentication information block, it is confirmed whether the calculation unit block is really from the center system 2 (step 122). Thereafter, the client software imports data necessary for the calculation process from the center system 2 based on the data import block (step 1).
23), a series of processing is executed based on the processing block (step 124). When the execution of a series of processing is completed,
The client software exports the calculation result to the center system 2 based on the data export block (step 125). As described above, since the processing for the received calculation unit block is completed, the client software deletes from the computer 5 of the provider 4 all the data received from the center system 2 in relation to the calculation unit block and the calculation unit block. (Step 126).

When the client software receives the processing interruption instruction from the center system 2, the client software terminates the processing at that time, and receives from the center system 2 the calculation unit block and the calculation unit block. All of the data which has been deleted from the computer 5 of the provider 4 is deleted.

The processing by the client software has been described above. Further, the client software may display information such as the processing capacity of resources provided by the computer 5 and the CPU unit price on the screen of the computer 5. .

In this computer resource utilization system, the provider 4 does not need to guarantee the surplus information processing capacity to be provided or the certainty of the provision. And a price calculated based on the CPU price at the time. Basically,
The resources provided by the provider 4 for implementing the computer resource utilization method are a CPU, a memory, a network, and some work disk space in the computer 5, and the data from the center system 2 is transmitted to the environment of the provider 4. It does not save to.

Here, the hardware configuration of the center system 2 in the computer resource utilization system described above will be described. As shown in FIG. 3, the supervisor control unit 11, the CPU price determining mechanism 12, the job dividing mechanism 13, the surplus processing capacity database 14, the scheduler 15, and the processing distribution / monitoring mechanism 16 By providing corresponding hardware mechanisms, the center system 2 can be configured. However, in general, the center system 2 can also be realized by causing a computer such as a server computer to read a computer program for realizing the functions of the above-described components and executing the program. A program for implementing the center system is read into a computer by a recording medium such as a magnetic tape or a CD-ROM. FIG. 6 is a block diagram illustrating a configuration of a computer system used to implement the center system 2.

This computer system includes a central processing unit (C
PU) 31, a hard disk device 32 for storing programs and data, a main memory 33, an input device 34 such as a keyboard and a mouse, a display device 35 such as a CRT, and a recording device such as a magnetic tape or CD-ROM. It comprises a reading device 36 for reading the medium 37 and a communication interface 38 for connecting to the wide area network 1, and these are connected by an internal bus shown by a double line in the figure.
Interconnected. In this computer system, a recording medium 37 storing a program for realizing the center system 2 is mounted on a reading device 36, and the recording medium 37
The central processing unit 31 reads out the program from the hard disk device 32 and executes the program stored in the hard disk device 32 to function as the above-described center system 2.

The computer 5 of the user 4 has the same basic configuration as that of the computer system shown in FIG. 6, although the type of central processing unit (CPU) and the memory capacity may be different. However, as described above, since the client software may be distributed via the wide area network 1, in such a case, a program that can receive the program via the network (a so-called Internet browser program) is used. What is necessary is just to make it read from recording media, such as CD-ROM.

Next, as a field in which the merit can be obtained by applying the processing by the computer resource utilization method described above, that is, a processing service (global computing service) by connecting a wide area network with computer capabilities, for example, vector arithmetic is mainly used. Large-scale simulations, image processing such as rendering, risk analysis that requires a large number of case studies, and optimization problems with a very large search area.
Since the calculation processing in these fields includes many parallel processings, by applying the global computing service, the processing time can be significantly reduced as compared with the case where a conventional computer is used.

In a risk analysis or the like that requires a large number of case studies, even if the existing style program is used as it is, the case studies under different conditions can be performed in parallel to sufficiently obtain the advantages of the global computing service. Can be.

Since the global computing service is suitable for a series of processes composed of a plurality of processes having small relevance to each other, it is considered that the applicable field is further increased by changing the method of constructing a program.

By providing library programs suitable for global computing on the center system side, services that can reduce the burden on the user and receive the advantages of global computing become effective.

[0059]

As described above, the present invention enables a widely used computer to be connected to a wide area by a network so that a margin of its processing capacity can be widely opened to others and related to it. Providing the billing process procedure has the effects of reducing outsourcing costs for information processing and increasing the utilization rate of information processing capabilities of each computer.

That is, in the past, those who had a computer could not effectively use the computer even if the computer had enough processing power. However, according to the present invention, the computer sold the remaining processing power. By doing so, it is possible to effectively use the surplus portion. Conventionally, when a process requiring a large information processing capacity is executed, a very expensive special computer had to be used. By applying the present invention, the large information processing capacity can be realized at a low cost. It can be used. Further, according to the present invention, it is possible to use the processing capacity at a reasonable cost (resale price) presented in real time.

[Brief description of the drawings]

FIG. 1 is a block diagram showing a configuration of a computer resource utilization system according to an embodiment of the present invention.

FIG. 2 is a flowchart showing an example of processing based on the computer resource utilization method of the present invention, which is executed using the system shown in FIG.

FIG. 3 is a block diagram illustrating an example of a configuration of a center system.

FIG. 4 is a diagram illustrating an example of a configuration of a calculation unit block.

FIG. 5 is a flowchart showing processing on a provider's computer.

FIG. 6 is a block diagram showing an example of a configuration of a computer system that can be used as a center system or a provider computer.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Wide area network 2 Center system 3 User 4 Provider 5 Computer 11 Supervisor control unit 12 CPU price determination mechanism 13 Job division mechanism 14 Surplus processing capacity database 15 Scheduler 16 Processing distribution / monitoring mechanism 21 Calculation unit block 22 Authentication information block 23 Data Import block 24 Processing block 25 Data export block 101-115, 121-126 steps

Claims (13)

[Claims] 1. A computer resource utilization method that utilizes information processing capabilities of a plurality of computers connected to a network, the method including receiving an offer to provide a spare portion of information processing capability, and storing information on a computer corresponding to the offer in a database. And receives a request for a calculation process from a user, divides the requested calculation process into a plurality of unit processes that can be processed in parallel, and refers to the database to execute the unit processes for the plurality of computers. Determine the allocation, distribute the allocated unit processing to the corresponding computer, receive the calculation result corresponding to the distributed unit processing from each of the computers, and transmit the processing result to the user based on the calculation result A method of utilizing computer resources that performs billing according to the calculation results. 2. A program for measuring the processing capacity of each computer is delivered to each of the computers, a response to the program is received from each of the computers, and a processing capacity of each of the computers is determined based on the response. 2. The computer resource utilization method according to claim 1, wherein the determined processing capacity is stored in a database. 3. The method according to claim 1, wherein the unit price used in the charging process is changed such that the offer of the surplus portion is supplied and the request for the calculation process is the demand, and the demand and the supply are balanced. How to use computer resources described. 4. When transmitting a processing result to a user, a calculation result from each computer is collected as the processing result according to a division method when the requested calculation processing is divided into unit processing. Item 4. The computer resource utilization method according to any one of Items 1 to 3. 5. A unit processing includes a recognition information block including recognition information, a data import block describing processing for importing data required for executing the unit processing from a center system, and a substantial unit of the unit processing. Is delivered to each computer as a calculation unit block consisting of a processing block describing the various processes and a data export block describing a process for exporting a result obtained by the processing block to the center system. The computer resource utilization method according to claim 1. 6. Upon receiving an offer to provide a spare portion of information processing capability, a software for receiving unit processing from the center system and executing the unit processing is provided to a person who made the offer via a network. The computer resource utilization method according to claim 1, wherein the computer resource is distributed. 7. A computer resource utilization method that utilizes a surplus portion of information processing capacity of a computer owned or managed by a self, and provides an offer of the surplus portion to a center system. A computer resource utilization method for receiving a process, executing the process by the computer, transmitting a result obtained by the execution to the center system, and receiving a fee according to a calculation result. 8. A computer resource utilization system that utilizes the information processing capabilities of a plurality of computers connected to a network, comprising: a database that stores information relating to computers that are to provide extra information processing capability; A job dividing unit that divides the processing into a plurality of unit processes that can be processed in parallel; a scheduler that determines the allocation of the unit processes to the plurality of computers by referring to the database; A processing distribution / monitoring mechanism for distributing the unit processes to the corresponding computers via a network and collecting calculation results from the computers; and accepting a request for a calculation process from a user, and executing the requested calculation process. The processing result is passed to the job division mechanism, and based on the calculation result collected by the processing distribution / monitoring mechanism. Computer resources utilization system having a supervisor control unit transmits to the user, executes an accounting process corresponding to the calculation results. 9. The processing distribution / monitoring mechanism distributes a program for measuring the processing capacity of each computer to each computer, receives a response to the program from each computer, and executes each computer based on the response. 9. The computer resource utilization system according to claim 8, wherein a processing capacity is determined for each computer, and the determined processing capacity is stored in a database. 10. A CP for determining a resale price of information processing capability of a computer based on the demand and the supply, with an offer to provide a spare portion as a supply and a request for a calculation process as a demand.
10. The computer resource utilization system according to claim 8, further comprising a U price determining mechanism. 11. A CPU price determination mechanism determines a resale price at any time based on supply and demand, and requests the determined resale price to a person who intends to offer an extra portion and a request for calculation processing. Claim 10 to present to the person trying to
Computer resource utilization system described in 1. 12. A recording medium readable by a computer system, wherein the computer system accepts an offer to provide a surplus portion of information processing capacity, registers information on a computer corresponding to the offer in a database, and performs computation processing from a user. Receiving the request, dividing the requested calculation process into a plurality of unit processes capable of parallel processing, determining the allocation of the unit processes to the plurality of computers with reference to the database, and determining the allocated unit processes. Distribute to the corresponding computer, receive the calculation result corresponding to the distributed unit process from each of the computers, transmit the processing result to the user based on the calculation result, and execute the charging process according to the calculation result A storage medium storing a program for causing the computer system to execute the processing to be performed. 13. A recording medium readable by a computer system, comprising: receiving a calculation unit block from a center system; executing a calculation process based on the calculation unit block; A recording medium storing a program for causing the computer system to execute a process of deleting data received from the center system from the computer system based on the block and the calculation unit block.