JPH04360215A - Accounting controller and method for controlling load - Google Patents

Abstract

PURPOSE:To provide the efficient utilization of resources by smoothing the load of a computer resource to the lapse of time and preventing the resource from the increase of loads. CONSTITUTION:A request having priority order is inputted from a terminal 1 and stored in a queue 3 by the computer resource (CPU) 2. The request whose priority order is more than a reference value calculated from a utilization rate is extracted from the queue 3 and executed. At this time, a load recording means 50 records the utilization rate of the resource in a utilization rate file 6, a utilization recording means 51 stores information such as the utilization time of the resource in an accounting file 7 and an accounting means 51 executes accounting calculation in proportional to the utilization rate and forms an accounting information file 8. Since accounting proportional to a load state is executed, the excess state of loads to the resource or the idle state of the resource without using it can be removed.

Description

[Detailed description of the invention]

0001

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a billing control device for a computer system or the like that charges for the use of resources, and also to a load control method for controlling the load state of resources.

0002

2. Description of the Related Art A conventional method for controlling the execution of requests is to divide requests into a plurality of classes and control them, as shown in FIG. This illustration is by James
L. Peterson, Abraham Silb
erschatz; Translated by Koichi Utsunomiya and Akira Fukuda: "Concepts of Operating Systems", P. 133 1987
This method was published by Baifukan Co., Ltd. on July 10, 2007, and was used in the OS/MFT system. In the figure, 36 is a system request queue, 37 is an interactive request queue, 38 is an interactive edit request queue, 39 is a batch request queue,
40 represents a queue of student requests.

Next, the operation will be explained. Requests in each queue are prioritized for execution in the following order: system requests, interactive requests, interactive editing requests, batch requests, and student requests. Requests in each queue are executed according to priority. Requests in the batch queue will not be executed unless the system request, interactive request, and interactive edit request queues are all empty. If an interactive edit request is queued while a batch request is being executed, the interactive edit request will be executed. It is also possible to allocate more execution time to the queues so that the higher the priority, the more execution time is allocated to the queue.

0004

[Problems to be Solved by the Invention] In such conventional request scheduling methods, when there are many requests, the resource utilization rate is almost 1, and the response of the computer from the user's perspective is very slow. A situation arises where it cannot be used. Conversely, when there are few requests, the resource utilization rate decreases, resulting in a state in which computer resources are idle without being used. As described above, the conventional method has a problem in that the load on computer resources depends only on the usage status of the user and cannot be controlled on the system side.

[0005] This invention was made to solve the above problems, and it smoothes the load on computer resources over time, avoids an extreme increase in the load on resources, and improves the efficiency of resources. The purpose of the present invention is to provide a billing control device and a load control method that can be expected to have efficient operation.

[0006]

[Means for Solving the Problems] A billing control device according to the present invention has the following elements. (a) Load recording means for recording the load state of computer resources; (b) Usage recording means for recording how much computer resources are used by each user; (c) Load state recorded by the load recording means. and charging means for calculating charging information for computer resource usage based on the usage result recorded by the usage recording means.

The load control method according to the present invention includes the following steps. (a) Execution determination step of determining the execution of a computer resource utilization request by comparing it with the load condition of the computer resource; (b) Market price determination step of determining the market price of charging for the utilization request from the load condition of the computer resource. , (c) a billing step for calculating charges for the usage request executed as determined in the execution determining step, in accordance with the billing market price determined in the market price determining step.

[0008]

[Operation] In the charging control device according to the present invention, the charging means calculates a corresponding charge depending on the magnitude of the load on computer resources, so when the load is large, the charge is also large, and when the user has a small load, the charge is large. It is expected that usage requests will be issued.

[0009] In the load control method according to the present invention, the priority of the request is compared with the load condition in the execution determination step, and the request is executed only when the priority is exceeded. The market rate of charges is weighted, and in the billing process, charges are calculated from usage information according to the billing rate determined in the market price determination process. It can be expected that usage requests will be issued when there are few.

0010

【Example】

Example 1. FIG. 1 is a block diagram of a CPU load control system showing an embodiment of the present invention. 1 is the terminal that issues the request, 2 is the CPU that processes and executes the request, 3 is the queue that stores the request, 4 is the memory, 5 is the disk,
6 is a usage rate file that records the load state of the CPU 2 as a usage rate y (0≦y≦1), 7 is an accounting file that records usage information such as request execution start time and end time, and 8 is a file used for request execution. This is a billing information file that records billing information that has been calculated.

50 is a load recording means that constantly monitors the CPU 2 and records the usage rate y of the CPU 2 when a request is executed in a usage rate file; 51 is a usage unit that records the start and end of usage requests in an accounting file 7; A recording unit 52 is a billing unit that calculates billing from the usage rate y and usage information and records it in the billing information file 8.

[0012] The terminal 1 assigns a priority x (0<x<1) to a request using the CPU 2 and submits the request to the system. The CPU 2 has a queue 3 on the memory 4 that stores requests waiting to be used by the CPU. In the utilization rate file 6 on the disk 5, the utilization rate y when the CPU 2 is assigned to the request is stored by the load recording means 50, and in the accounting file 7, the utilization rate y when the CPU 2 is assigned to the request is stored.
The time when U2 is assigned is stored by the usage recording means 51. The billing information file 8 contains billing information determined by the billing means 52 from the billing market rate f(y)=y, which is a function of the usage rate y in the usage rate file 6, and the allocation time of the CPU 2 in the accounting file 7, for each use. It is added and stored for each person. Based on this billing information, the system administrator periodically charges the system user.

In a system having such a configuration, when a request is submitted, processing based on the flowchart shown in FIG. 2 is started, and the request is taken out from the queue 3 and executed. (1) Execution decision step First, in step 9, it is checked whether there is a request in the queue of the CPU 2. If there are no requests, the process ends. If there is a request, queue 3 in step 10.
Sort the requests in descending order of priority. In step 11, it is checked whether the priority x of the request at the head of the queue 3 is greater than or equal to the utilization rate y. x<y
In this case, repeat steps 10 and 11. (2) In the case of execution step x≧y, in step 12, the utilization rate y is stored in the utilization rate file 6, and the time t at this time is stored as t1 in the accounting file 7, and in step 13, the CPU 2 is assigned to the request. After the CPU 2 is used, the current time t2 is stored in the accounting file 7 in step 14. (3) Market price determination step Next, in step 15, the market price is determined from the usage rate y of the usage rate file 6 stored in step 12. In this example, f(y) = y, so the market price is the utilization rate itself, and 0≦
f(y)≦1. (4) Billing process Next, in step 16, the time T (T = t2 - t1) which is the difference between times t1 and t2 in the accounting file 7 is calculated.
and f(y) obtained from y in the usage rate file 6,
f(y)·T is calculated and added to each user's billing information in the billing information file 8. After this, return to step 7.

[0014] Therefore, since the billing rate changes in proportion to the priority, requests with higher priorities are processed with higher charges, and requests with lower priorities are processed with lower charges.

Example 2. In Example 1, priority x
was 0<x<1, but in the second embodiment, in the configuration diagram of the CPU load control system in FIG.
The case where x=1) will be explained as an example. In the case of the second embodiment, as in the first embodiment, the CPU 2 has a queue 3 on the memory 4 in which requests waiting to be used by the CPU are stored in a FIFO format. The utilization rate file 6 on disk 5 contains the utilization rate y when CPU2 is assigned to a request.
is stored, and the accounting file 7 stores the time when the CPU 2 is assigned to the request. In the billing information file 8, the billing market rate f(y)=y, which is a function of the usage rate y in the usage rate file 4, and the billing information determined from the allocation time of the CPU 2 in the accounting file 5 are added for each user. things are stored. Based on this billing information, the system administrator periodically charges the system user.

In a system having such a configuration, when a request is submitted, processing based on the flowchart shown in FIG. 3 is started, and the request is taken out from the queue 3 and executed. First, in step 17, it is checked whether there is a request in the queue 3 of the CPU 2. If there are no requests, the process ends. If there is a request, in step 18 the utilization rate y is stored in the utilization rate file 6, and this time is stored as t1 in the accounting file 7, and in step 19 the CPU 2 is assigned to the request. After the use of CPU2 is finished, step 20
Then, the time t2 at this time is stored in the accounting file 7. Next, in step 21, the market price f(y) is determined from the usage rate y stored in the usage rate file. Then, in step 22, the time t in the accounting file 7 is
From f(y) obtained from the time T (T = t2 - t1) obtained by difference between 1 and t2 and y in the utilization rate file 6, f(y)・T
is added to the calculated billing information of each user in the billing information file 8. After this, the process returns to step 17.

Therefore, in this second embodiment as well, the billing rate fluctuates in proportion to the CPU usage rate, so that when the CPU usage rate is high, requests are processed at high charges.
When the CPU utilization rate is low, processing will be performed at a low charge, but especially in Example 2, when the priority order x is constant (
x = 1) (or even if there is no priority order), depending only on the usage rate y, as in Example 1, when the load is heavy, the charge will be high, and when the load is low, the user will make a usage request. You can expect it to come out.

Example 3. FIG. 4 is a block diagram of a disk device load control system showing another embodiment of the present invention. Terminal 1 assigns priority x (
0<x≦1) and input it into the system. disk 5
The priority is set to x=1 only in the case of a request to release . The disk 5 has a queue 3 in the memory 4 that stores requests waiting to use the disk. The utilization rate y of the disk is stored in the utilization rate file 6 on the disk 5. The accounting file 7 stores the number of blocks on the disk 5 being used and the time when the allocation of the disk 5 changes, and in the initial state, they are all 0. In the billing information file 8, the billing market rate f(y)=1/|y-1|, which is a function of the usage rate y in the usage rate file 6, and the billing information determined from the accounting file 7 are added for each user. things are stored. Furthermore, the priority table 26 on the memory 4 records the priority x of the currently used disk block specified by the user at the time of request. 27 is a CPU.

In a disk device having such a configuration, upon input of a request, processing based on the flowchart shown in FIG. 5 is started, the request is taken out from the queue 3, and disk blocks are allocated and released. First, in step 28, it is checked whether there is a request in the queue 3 of the memory 4. If there are no requests, the process ends. If there are requests, the requests in the queue 3 are sorted in descending order of priority in step 29. In step 30, it is checked whether the priority x of the request at the head of the queue 3 is greater than or equal to the utilization rate y. If x<y, in step 31 it is reported to the user of the request in the queue that the usage rate y is greater than the priority x specified by the user, and steps 29 and 30 are repeated. If x≧y, in step 32 it is determined whether x=1. When x≠1, since this is an allocation request, the requested disk 5 is allocated in step 33. In step 34, the changed utilization rate y is stored in the utilization rate file 6, and the changed allocation block number b
The time t at this time is stored in the accounting file 7 as t1, and the priority table 26 on the memory 4 is updated. In step 35, x<y in the priority table 26
It is reported to the user of the disk that satisfies the above that the usage rate y is greater than the priority x specified by the user, and then the process returns to step 28. Next, when x=1, it is a release request, so in step 36, the blocks that have been allocated so far on the disk 5 are released. Then, in step 37, the charging market rate f(y) is calculated from the usage rate y stored in the usage rate file 6 by executing step 35. Next, in step 38, the allocated block number b for each request stored in the accounting file 7, the difference T (T=t-t1) between the time t1 in the accounting file 7 and the current time t, and the utilization rate file 6 are stored. Product f of charging market f(y) found from stored usage rate y
(y)・b・T is added to each user's billing information in the billing information file 8.

[0020] According to the above method, when the usage rate of the disk device increases and the billing rate becomes larger than the specified value, this can be reported to the user in step 35 and the user can be encouraged to release the disk. Therefore, the usage status of disk devices can be smoothed out. In other words, by reporting to users that the billing market has increased and encouraging them to release resources, it is possible to smooth out the resource load and allow users to keep the billing amount within a certain range. do.

Example 4. In the above embodiment, the load control system is a CPU or a disk, but other computer resources such as a memory or a printer may be used. Further, the charging market price may also be determined using other predetermined calculation formulas.

[0022]

[Effects of the Invention] As explained above, according to the present invention,
When the utilization rate of computer resources is high, the charge is automatically increased to suppress the use of resources, and when the utilization rate of the computer is low, the charge is automatically reduced to encourage the use of resources. As a result, the load on computer resources is smoothed over time, avoiding an extreme increase in the load on computer resources,
Efficient use of computer resources can be expected.

[Brief explanation of drawings]

FIG. 1 is a schematic block diagram of a CPU load control system according to an embodiment of the present invention.

FIG. 2 is a flow diagram of a method for executing a request in FIG. 1;

FIG. 3 is a flow diagram of a method for executing a request in FIG. 1;

FIG. 4 is a schematic block diagram of a disk device load control system according to an embodiment of the present invention.

FIG. 5 is a flow diagram of a method for executing a request in FIG. 4;

FIG. 6 is a block diagram showing a conventional method.

[Explanation of symbols]

2 CPU (an example of computer resources) 5 Disk (an example of computer resources) 6 Utilization rate file 7 Accounting file 8 Billing information file 50 Load recording means 51 Usage recording means 52 Billing means 60 Execution decision process 61　Execution process 62 Market price determination process 63 Billing process

Claims (2)

[Claims] Claim 1: A billing control device (a
) Load recording means for recording the load state of computer resources; (b)
) a usage recording means for recording how much computer resources are used by each user; (c) a usage recording means for recording how much computer resources are used by each user; Billing means that calculates billing information for resource usage. Claim 2: A load control method (a) comprising the following steps:
) an execution determination step of determining the execution of the computer resource usage request by comparing it with the load state of the computer resource; (b) a price determination step of determining the market price of charging for the usage request from the load state of the computer resource; c) A charging step of calculating charges for the usage request executed as determined in the execution determining step, in accordance with the charging market price determined in the market price determining step.