JP7670167B2 - Resource management device and program - Google Patents

Description

The present invention relates to a resource management device and program for managing computational resources.

Technological innovation is progressing in many fields, including machine learning, artificial intelligence (AI), and the Internet of Things (IoT), and by utilizing various data, services are being actively improved and added value is being provided. This type of processing requires large amounts of calculations, and an information processing infrastructure is essential for this.

For example, Non-Patent Document 1 points out that, although attempts are being made to update existing information processing infrastructure, modern computers are unable to keep up with the rapid increase in data. Non-Patent Document 1 also points out that in order to achieve further evolution in the future, "post-Moore technology" that goes beyond Moore's Law must be established.

As a post-Moore technology, for example, Non-Patent Document 2 discloses a technology called flow-centric computing. Flow-centric computing introduces a new concept of moving data to a location where a computing function (computing resource) is located and processing the data there, instead of the conventional computing idea of processing data where the data is located.

“NTT Technology Report for Smart World 2020,” Nippon Telegraph and Telephone Corporation, 2020, https://www.rd.ntt/_assets/pdf/techreport/NTT_TRFSW_2020_EN_W.pdf R. Takano and T. Kudoh, “Flow-centric computing leveraged by photonic circuit switching for the post-moore era,” Tenth IEEE/ACM International Symposium on Networks-on-Chip (NOCS), Nara, 2016, pp. 1-3.

To realize flow-centric computing as described above, it is necessary to properly manage which hardware constitutes the computing resources. For example, if the computing resources are constituted by the hardware of a computer with a high load without proper management, delays may occur in the processing of the computing resources.

The objective of the present invention is to appropriately manage the hardware configuration of multiple computing resources that share the services of processing the data to be processed.

In order to solve the above problems, the resource management device of the present invention is configured using part of the hardware of a computing system consisting of multiple computers, and is a resource management device that manages the hardware configuration of multiple computing resources that share the service of processing data to be processed, and is equipped with a required quality acquisition unit that acquires the quality related to the processing time of the service required for the service as a required quality, an estimation result acquisition unit that acquires the result of estimating the quality of the service performed by the multiple computing resources of a first hardware configuration during operation of the computing system as a quality estimation result, and a configuration control unit that changes the hardware configuration of the multiple computing resources from the first hardware configuration to a second hardware configuration different from the first hardware configuration during operation of the computing system if the quality estimation result does not satisfy the required quality.

In order to solve the above problem, the program of the present invention causes a computer that is configured using part of the hardware of a computing system consisting of multiple computers and manages the hardware configuration of multiple computing resources that share the service of processing the target data to execute the following steps: a required quality acquisition step of acquiring the quality related to the processing time of the service required for the service as a required quality; an estimation result acquisition step of acquiring, as a quality estimation result, a result of estimating the quality of the service performed by the multiple computing resources of a first hardware configuration while the computing system is operating; and a configuration control step of changing the hardware configuration of the multiple computing resources from the first hardware configuration to a second hardware configuration different from the first hardware configuration while the computing system is operating if the quality estimation result does not satisfy the required quality.

According to the present invention, the hardware configuration of multiple computing resources that share the service of processing the data to be processed is appropriately managed.

FIG. 1 is a hardware configuration diagram showing the configuration of a resource management device and a computer according to a first embodiment of the present invention. FIG. 2 is a block diagram showing the configuration of the computer shown in FIG. FIG. 3 is a block diagram showing the configuration of the resource management device of FIG. FIG. 4 is a flowchart of a resource management process according to the first embodiment. FIG. 5 is a block diagram showing the configuration of a resource management device according to the second embodiment. FIG. 6 is a block diagram showing the configuration of a resource management device according to the third embodiment. FIG. 7 is a flowchart of a resource management process according to the third embodiment.

The following describes an embodiment of the present invention with reference to the drawings. In the following description, elements having the same function and elements having different functions but corresponding to each other are appropriately denoted by the same reference numerals. Also, in the drawings, for multiple elements having the same function or corresponding to each other, reference numerals may be denoted only for some of the elements.

[First embodiment]
As shown in FIG. 1, the resource management device 10 according to the present embodiment is provided so as to be able to communicate with each of the multiple computers 30-1 to 30-N (where N is a natural number of 2 or more) constituting the computing system 20 via a network NW such as the Internet or a local area network (LAN). The resource management device 10 is configured using accelerators 35 (to be described later) of the computers 30-1 to 30-N, which are part of the hardware of the computing system 20, and manages the hardware configuration of multiple computing resources R that share and perform services to process data to be processed. More specifically, the resource management device 10 manages which hardware is used to realize the multiple computing resources R for each service. Below, the configuration of the computers 30-1 to 30-N will be described first, and then the configuration of the resource management device 10 will be described. The computers 30-1 to 30-N are also collectively referred to as computers 30.

The computers 30-1 to 30-N can communicate with each other via the network NW, and the computing system 20 as a whole performs multiple types of services using a large number of computing resources R. The services include image processing, etc. One service includes multiple processes, and the multiple processes are shared and executed by the multiple computing resources R. For example, multiple computing resources R performing one service are connected via a virtual network configured in the network NW, etc., and process the data to be processed in series and/or in parallel. For example, as one service, image data as the data to be processed is binarized by parallel processing by the two computing resources R of the computer 30-1, and then image recognition processing is performed on the binarized image data by the computing resource R of the computer 30-2, and the processing result is returned to the provider of the image data (not shown). The provider is a client computer of the user of the service, etc. The data to be processed and the processing result are transmitted and received to the provider via the resource management device 10. A series of processes constituting each service is performed, for example, under the control of the resource management device 10. For example, the storage device 13 of the resource management device 10 stores the addresses of a plurality of computing resources R for each service, and the resource management device 10 specifies the destination of the processing result data output by the computing resources R.

Each of the computers 30-1 to 30-N has a similar configuration, although the processes that it can execute are different. Hereinafter, the computers 30-1 to 30-N are collectively referred to as computers 30. The configuration of each computer 30 will be described below.

The computer 30 is a computer such as a personal computer, a smartphone, or a tablet. The computer 30 includes a processor 31, a main memory 32 of the processor 31, a non-volatile storage device 33 that stores programs and various data, and a network interface card (NIC) 34 connected to the network NW. The computer 30 further includes an accelerator 35 that improves the functions of the computer 30.

The processor 31 is composed of a CPU (Central Processing Unit) and the like, and executes or uses programs and various data stored in the storage device 33 to control the entire computer 30. The main memory 32 is composed of a RAM (Random Access Memory) and the like. The programs and various data are read into the main memory 32 as appropriate. The storage device 33 is composed of an SSD (Solid State Drive) and the like. The NIC 34 transmits and receives data to and from the network NW under the control of the processor 31.

The accelerator 35 is composed of hardware such as a field-programmable gate array (FPGA), an application specific integrated circuit (ASIC), a sub-CPU, and a graphics processing unit (GPU). When the accelerator 35 is an FPGA, the processor 31 can dynamically delete and reconfigure a specific arithmetic circuit as an arithmetic resource R in at least a part of the reconfigurable area of the FPGA while the computer 30 is running. Here, one or more arithmetic resources R are configured by the accelerator 35. For example, when the accelerator 35 is an FPGA, the arithmetic circuit that operates as the arithmetic resource R is configured in the accelerator 35, thereby configuring the arithmetic resource R. The area in which the FPGA arithmetic resource R is configured becomes the hardware that configures the arithmetic resource. The arithmetic resource R may be configured by the processor 31 as, for example, a virtual machine. The arithmetic resource R may be configured by hardware in which a GPU is logically divided into multiple parts. Data input to the computing resource R, transfer of processed data from the computing resource R, etc. are controlled by a processor 31, for example.

By executing the above program, the processor 31 operates as a load monitoring unit 31A, a performance estimation unit 31B, and a resource management unit 31C as shown in Figure 2.

The load monitoring unit 31A monitors the load on the computer 30. Examples of the load include the amount of data input to the computer 30 per unit time, the amount of data processed by the computer 30 per unit time, the amount of data output by the computer 30 per unit time, and the usage rate of the computational resource R of the computer 30. Examples of the load also include the usage rate of memory resources such as the main memory 32 or a memory unit provided in the accelerator 35 for storing intermediate data in the processing process or storing computational parameters. The load is not limited to this, and may be, for example, the input data rate, the output data rate, or the frame rate in the case of video processing. The load is composed of information that affects the presence or absence of a delay in processing time by the computer 30 and the degree of the delay.

The performance estimation unit 31B estimates the performance of the processing time of the computer 30 when one or more computing resources R that execute a part of the processing of one service are operated. The performance may be the processing time itself, or may be an amount that increases or decreases the processing time, such as the processing speed. The processing time is, for example, the time from when data is input to the computer 30 to when the data is processed by the one or more computing resources R and the processing result is output to the outside of the computer 30. The storage device 33 stores a relational expression or table showing the relationship between the circuit scale of the computing resource R and the load on the computer 30, and the performance of the computer 30 when the computing resource R is used. The performance estimation unit 32B estimates the performance using the relational expression or table based on the circuit scale of the computing resource R and the load monitored by the load monitoring unit 31A. When estimating, the performance estimation unit 32B may have the computer 30 use test data to measure the processing time, etc., and estimate the measured processing time, etc. as the above-mentioned performance.

The resource management unit 32C adds or deletes any computation resource R to the accelerator 35 based on an instruction from the resource management device 10 described below. If the accelerator 35 is an FPGA, the resource management unit 32C writes or deletes a computation circuit that operates as the computation resource R to the FPGA.

The resource management device 10 shown in Figure 1 is a computer such as a server computer. The resource management device 10 includes a processor 11, a main memory 12 for the processor 11, a non-volatile storage device 13 for storing programs and various data, and a NIC 14 connected to a network NW.

The processor 11 is composed of a CPU etc., and executes or uses the programs and various data stored in the storage device 13 to control the entire resource management device 10. The main memory 12 is composed of a RAM etc. The programs and various data are read into the main memory 12 as appropriate. The storage device 13 is composed of an SSD etc. The NIC 14 transmits and receives data to the network NW under the control of the processor 11.

By executing the above programs, the processor 11 operates as a required quality acquisition unit 11A, an estimate result acquisition unit 11B, a configuration control unit 11C, a load information acquisition unit 11D, and an input data restriction unit 11E, as shown in FIG. 3. These units 11A to 11E work together to perform the resource management processing shown in FIG. 4. The resource management processing is performed periodically for each service. The resource management processing is performed, for example, every few minutes for one service. The resource management processing is performed dynamically while the computing system 20 is operating. In other words, the resource management processing is performed while the above services are being provided. When performing the resource management processing, the resource management device 10 may instruct multiple computers 30-1 to 30-N to stop execution of the service to be processed.

The operation of each of the units 11A to 11E will be described below with reference to Figures 3 and 4. In the following, the service to be processed is referred to as service X. Each of the multiple computational resources R that provide service X is also referred to as computational resource RX. Each of the multiple computers 30 in which the multiple computational resources RX are configured is also referred to as computer 30X. The storage device 13 of the resource management device 10 stores the addresses of the multiple computational resources RX that perform service X. The addresses also make it possible to identify the computers 30X in which the multiple computational resources RX are each provided.

In the resource management process, first, the requested quality acquisition unit 11A acquires the requested quality, which is the quality related to the processing time of service X that is required for service X (step S101). The requested quality functions as a preset threshold value, and is recorded in the storage device 13 in association with each service. The requested quality acquisition unit 11A acquires the requested quality corresponding to service X from the storage device 13. The requested quality is compared with the above performance estimation result, and has the same dimension as performance. If the performance is processing time, the requested quality is also the processing time required for the service, for example, the upper limit time allowed as processing time. If the performance is processing speed, the requested quality is also processing speed.

Note that performance and quality may be amounts that increase or decrease processing time, such as input data rate, output data rate, or, if service X is video processing, frame rate during processing. The content of performance and quality is determined by the content of each service.

The estimation result acquisition unit 11B also acquires the result of estimating the quality of service X performed by multiple computing resources RX of the current hardware configuration as a quality estimation result (step S102). The estimation result acquisition unit 11B communicates with each computer 30X provided with multiple computing resources RX via the NIC14, the network NW, and the NIC34 (hereinafter, the same applies to communication), and instructs the computer 30X to estimate the performance of the computer 30X when using the computing resources RX. The performance estimation unit 31B (FIG. 2) of the computer 30X transmits the estimation result estimated by the instruction to the estimation result acquisition unit 11B. The estimation result acquisition unit 11B estimates the quality of service X based on the estimation results transmitted from each computer 30X and acquires the quality estimation result. If the performance that is the estimation result is the processing time, the processing time obtained by adding up the estimation results from each computer 30X becomes the quality estimation result. If the performance that is the estimation result is the processing speed, the average speed of the estimation results from each computer 30X becomes the quality estimation result.

After the required quality and the quality estimation result are acquired (after steps S101 and S102), the configuration control unit 11C compares the acquired required quality with the quality estimation result and determines whether the quality estimation result satisfies the required quality (step S103). If the required quality and the quality estimation result are processing times, if the processing time of the quality estimation result is less than or equal to the processing time of the required quality, it is determined that the quality estimation result satisfies the required quality. If both are processing speeds, if the processing speed of the quality estimation result is greater than or equal to the processing speed of the required quality, it is determined that the quality estimation result satisfies the required quality.

If the quality of the service to be processed meets the required quality (step S103; Yes), there is no need to change the hardware configuration of the computational resource RX, and the resource management process in Figure 4 is terminated.

If the quality of the service to be processed does not satisfy the required quality (step S103; No), it is necessary to change the hardware configuration of the multiple computing resources RX. In this case, the load information acquisition unit 11D acquires load information indicating the load on each of the computers 30-1 to 30-N of the computing system 20 (step S104). The load information acquisition unit 11D communicates with the computers 30-1 to 30-N and requests their current loads. In response to the load request, the load monitoring unit 31A in the computers 30-1 to 30-N returns the load being monitored to the load information acquisition unit 11D. In this way, the load information acquisition unit 11D acquires the load on each of the computers 30-1 to 30-N as load information.

After acquiring the load information, the configuration control unit 11C dynamically changes the hardware configuration of multiple computational resources RX from the hardware configuration at the time the estimate result was acquired to a second hardware configuration different from this hardware configuration based on the acquired load information (step S105).

In step S105, the configuration control unit 11C changes, for example, the hardware constituting at least one of the multiple computational resources RX from the hardware of the computer 30X having a high load among the computers 30-1 to 30-N to the hardware of one or more computers 30 having a low load.

As an example, when the hardware is an FPGA, the configuration control unit 11C deletes the computation resource RX written to the accelerator 35 of the first computer 30X, whose load is higher than the first standard. On the other hand, the configuration control unit 11C writes a computation resource R having the same circuit configuration as the deleted computation resource RX to the accelerator 35 of the second computer 30, whose load is lower than the second standard. The computation resource R may be divided and written to multiple second computers 30. The configuration control unit 11C communicates with each of the first and second computers 30 and causes the resource management unit 32C to perform the deletion and writing. After the writing and deletion, the address of the computation resource RX stored in the storage device 33 is also changed. Furthermore, when the computing resource R that is the same as the computing resource RX provided in the computer 30X with a high load is configured in the computer 30 with a low load, the hardware configuration may be changed by switching the computing resources, for example, by rewriting the address of the computing resource RX stored in the storage device 33 to the computing resource R of the computer 30 with a low load. The change in the hardware configuration may include the addition of a new computing resource RX. Such an addition includes the configuration control unit 11C increasing the circuit scale of the computing resource RX and increasing the parallelism of the circuit configuration in order to improve the processing performance. When increasing the circuit scale, the configuration control unit 11C attempts to secure a necessary area in the accelerator 35.

The configuration control unit 11C may change the hardware configuration of the computation resource RX while referring to the quality estimation results and the load information so as to satisfy the quality requirements of service X. The configuration control unit 11C may change the hardware configuration using any algorithm, such as an algorithm that solves an optimization problem regarding the hardware configuration of the computation resource RX.

The configuration control unit 11C may dynamically change the hardware configuration of the multiple computational resources RX without using load information. For example, the storage device 13 of the resource management device 10 stores multiple combinations of addresses of the multiple computational resources RX. The configuration control unit 11C uses one of the multiple combinations for service X, and may change the combination of addresses of the multiple computational resources RX used for service X when changing the hardware configuration.

After step S105, the input data limiting unit 11E, while referring to the load information, limits the amount of data to be processed input to the computer 30 whose load is greater than a predetermined standard (step S106). This control includes reducing the amount of data to be processed input per unit time, as well as transferring the data to be processed to another computer 30. This process is particularly effective when the load on the entire computing system 20 is large, or when the quality of service X does not improve even after changing the hardware configuration. The input data limiting unit 11E may also perform the above control by further referring to the quality estimation result.

As described above, in this embodiment, when the quality estimation result for a certain service does not satisfy the required quality, the hardware configuration of the multiple computing resources R that share and process this service is changed from the first hardware configuration at the time of obtaining the quality estimation result to a second hardware configuration different from the first hardware configuration during the operation of the computing system 20. Since the quality estimation result is affected by the load of the computer to which the computing resource is provided, according to this embodiment, the hardware configuration of the computing resource R is changed while following the load on the computer 30 that changes from moment to moment. Therefore, even when the load on the computer 30 is high, the deterioration of the quality of the service being provided can be reduced, and the hardware configuration of the multiple computing resources R is appropriately managed. If the above control is not performed, it is necessary to excessively allocate the computing resource R due to the quality requirements of the service, and the power efficiency and the utilization efficiency of the computer 30 will deteriorate. In this embodiment, there is an effect that the power efficiency and the utilization rate of the computer can be kept at an optimal state while maintaining the quality of each service at a good level. In addition, since the computing resource can be added in this embodiment, there is an effect that the quality requirements of each service can be satisfied while maintaining the power efficiency and the utilization rate of the computer at an optimal state.

Furthermore, in this embodiment, based on load information indicating the load on each of computers 30-1 to 30-2, the hardware constituting at least one of the multiple computational resources R is changed from the hardware of the computer 30 with the larger load among computers 30-1 to 30-N to the hardware of the computer 30 with the smaller load. This changes the hardware configuration of the multiple computational resources R from the first hardware configuration to the second hardware configuration. In this way, it is possible to reduce the degradation of service quality due to the load on computer 30, and the hardware configuration of the multiple computational resources R is appropriately managed.

Furthermore, in this embodiment, the amount of input data to a computer among computers 30-1 to 30-N whose load is greater than a predetermined standard is limited based on the load information. This makes it possible to suppress degradation of service quality due to the load on computer 30.

When instructing the computer 30 to estimate performance in step S102, the estimation result acquisition unit 11B may specify conditions for the performance estimation. For example, in order to estimate the performance of the computer 30 when the computational resource R is increased or decreased by the above hardware configuration, the estimation result acquisition unit 11B may specify the increase and decrease amounts of the computational resource R as parameters for the estimation. Similarly, in order to estimate the processing performance when the amount of input data increases or decreases, the increase and decrease amounts of the input data may be specified as parameters for the estimation. The quality estimation results based on these estimation results may be used in step S106, etc.

[Second embodiment]
The configuration of the resource management device 110 according to this embodiment is shown in FIG. 5. The estimate result acquisition unit 11B in FIG. 5 acquires a performance estimation result that estimates the performance of the processing time when using one or more computing resources R configured on one or more computing resources R of one or more computers 30 (one or more computers 30X) configured with multiple computing resources R (e.g., multiple computing resources RX) that share and execute one service (e.g., service X) among the computers 30-1 to 30-N. This performance estimation result may be, for example, an estimation result of the performance of each computer 30 acquired by the estimate result acquisition unit 11B in the first embodiment. The estimate result acquisition unit 11B estimates the quality during operation of the computing system 20 of the service based on a value obtained by adding one or more fluctuation amounts to the acquired performance estimation result. The one or more fluctuation amounts may include, for example, at least the former of a predetermined first fluctuation amount and a second fluctuation amount according to the trend of the load indicated by the past load information.

The estimation result acquisition unit 11B, for example, adds or subtracts 5% of the value of the performance estimation result as the first variation amount to the value of the performance estimation result. The 5% is set as a fixed value. The first variation amount may be a value other than 5%. The estimation result acquisition unit 11B also sequentially records the load indicated by the load information from the load information acquisition unit 11D in the storage device 33 for each computer 30. The estimation result acquisition unit 11B reads out the past load for the computer 30 that is the target of the performance estimation result from the storage device 33, and derives a second variation amount based on the trend of the past load, for example, the average value of the load in the period from the present to a predetermined timing in the past or the load in the same time period as the present. For example, if the average value or load is equal to or greater than a predetermined threshold, an addition of 5% is derived as the second variation amount. The estimation result acquisition unit 11B calculates a value that further takes into account the second variation amount (for example, a value obtained by adding 5% of the value to the value taking into account the first variation amount) as the quality estimation result. The estimate result acquisition unit 11B may calculate the second fluctuation amount by using a neural network model, or may calculate it by a machine learning model based on statistical data.

The input data limiting unit 11E may set the load information to a value taking into account one or more of the above fluctuations as new load information, and limit the amount of data to be processed that is input to a computer 30 where the load indicated by the load information is greater than a specified standard.

The rest of the configuration is the same as in the first embodiment, so its description will be omitted.

According to this embodiment, a predetermined amount of fluctuation is taken into account in the quality estimation result and the load information, so that it is possible to secure computing resources R within a range that can be predicted in advance for the load on the computer 30, which changes from moment to moment. As a result, even if the load on the computer 30 fluctuates, degradation in the quality of the service can be reduced, and the hardware configuration of multiple computing resources R is appropriately managed. In addition, when securing computing resources R within a range that can be predicted in advance, redundancy against failure of computing resources R is also achieved by arranging physically different computing resources R.

[Third embodiment]
6 shows the configuration of a resource management device 210 according to this embodiment. The resource management device 210 includes a priority control unit 211G in addition to the units 11A to 11E included in the resource management device 10. The priority control unit 211G is a unit that the processor 11 controls to control the load information acquisition unit 11D, which acquires from the computing system 20 not only the load on each of the computers 30-1 to 30-N, but also total load information indicating the total load on the entire computing system 20. The load information acquisition unit 11D may calculate the total load based on the load on each of the computers 30-1 to 30-N.

Each unit 11A to 11E and 211G of the resource management device 210 cooperates to periodically execute the resource management process shown in Fig. 7. The resource management process will be described below with reference to Figs. 6 and 7. Note that the computing system 20 is capable of providing multiple services, and it is assumed that the multiple services are assigned priorities that specify which service is to be executed first in the event of a conflict in execution timing.

In the resource management process according to this embodiment, first, the priority control unit 211G acquires total load information from the load information acquisition unit 11D (step S301). After that, the priority control unit 211G determines whether the total load on the computing system 20 indicated by the total load information is higher than a predetermined threshold (step S302). If the total load is not higher than the predetermined threshold (step S302; No), the resource management process ends. If the total load is higher than the predetermined threshold (step S302; Yes), there is a possibility that a delay has occurred in the provision of services by the computing system 20. In this case, the priority control unit 211G executes the process of step S303. Step S303 may be executed multiple times. In the Pth step S303, where P is a natural number, the priority control unit 211G selects the service with the Pth priority. After that, steps S102 to S106 are executed with the Pth service as the service to be processed. The explanation of these processes conforms to that of the first embodiment. However, if the determination result of step S103 is positive, the process of step S303 is executed again. If the number of times P of the process of step S303 reaches the number of services that can be provided by the computing system 20, the process of step S301 may be executed again. In addition, the process of step S301 may be executed after the process of step S106.

In this embodiment, similar to the first embodiment, the computing system 20 is configured with the multiple computing resources for each of the multiple services provided by the computing system 20. The resource management device 210 of this embodiment includes a load information acquisition unit 11D that acquires load information indicating the overall load on the entire computing system 20, and a priority control unit 211G that determines which of the multiple services should be prioritized as the target of processing (steps S101 to S106) by the required quality acquisition unit, the estimate result acquisition unit, and the configuration control unit when the overall load indicated by the load information exceeds a predetermined threshold. The priority control unit 211G determines which of the services should be prioritized as the target of the processing based on the priority order of processing execution when the multiple services compete.

Instead of or in addition to the priority order, the priority control unit 211G may determine which service should be prioritized for the processing based on the degree of deviation between the quality estimation result and the required quality for each service. For example, the priority control unit 211G accumulates the difference between the required quality and the quality estimation result acquired in steps S101 and S102 for each service in the storage device 33, and the larger the difference, the higher the priority of the service to be processed. The priority control unit 211G may add a weight corresponding to the difference to the priority order for each service in step S303, for example, and set the priority order of the service to be higher the smaller the value after the weighting.

According to this embodiment, when the load on the entire computing system 20 exceeds a predetermined threshold, the hardware configuration of the computing resources of the first embodiment is controlled preferentially for services with high priority. Therefore, even when the load on the entire system is high, the hardware configuration of the computing resources R can be controlled for services with high priority, thereby reducing degradation in quality for services with high priority.

[Scope of the present invention]
The present invention is not limited to the above-mentioned embodiment and modified examples. For example, the present invention includes various modifications to the above-mentioned embodiment and modified examples that can be understood by a person skilled in the art within the scope of the technical idea of the present invention. The configurations listed in the above-mentioned embodiment and modified examples can be appropriately combined within a range without contradiction. In addition, it is also possible to delete any of the above-mentioned configurations. The above-mentioned various programs may be stored in a non-transient computer-readable storage medium, not limited to the non-volatile storage device 33.

10...resource management device, 11...processor, 11A...required quality acquisition unit, 11B...estimate result acquisition unit, 11C...configuration control unit, 11D...load information acquisition unit, 11E...input data restriction unit, 12...main memory, 13...storage device, 20...computing system, 30, 30-1 to 30-N...computers, 31...processor, 31A...load monitoring unit, 31B...performance estimation unit, 31C...resource management unit, 32...main memory, 32B...performance estimation unit, 32C...resource management unit, 33...storage device, 35...accelerator, 110...resource management device, 210...resource management device, 211G...priority control unit, R...computational resource.

Claims (7)

A resource management device that is configured using a part of the hardware of a computing system consisting of a plurality of computers and manages the hardware configuration of a plurality of computing resources that share a service of processing data to be processed, comprising:
a required quality acquisition unit that acquires a quality related to a processing time of the service that is required for the service as a required quality;
an estimation result acquisition unit that acquires a result of estimating the quality of the service performed by the plurality of computing resources of a first hardware configuration during operation of the computing system as a quality estimation result;
a configuration control unit that changes a hardware configuration of the plurality of computing resources from the first hardware configuration to a second hardware configuration different from the first hardware configuration during operation of the computing system when the quality estimation result does not satisfy the required quality;
The computing system includes a plurality of computing resources for each of a plurality of services provided by the computing system,
a load information acquisition unit that acquires first load information indicating a load on the entire computing system;
and a priority control unit that determines, when the load indicated by the first load information exceeds a predetermined threshold, which of the plurality of services should be prioritized as a target for processing by the required quality acquisition unit, the estimate result acquisition unit, and the configuration control unit.
Resource management device. the load information acquisition unit acquires second load information indicating a load on each of the plurality of computers;
the configuration control unit changes hardware constituting at least one of the plurality of computing resources from hardware of a first computer having a large load to hardware of a second computer having a small load, based on the second load information, thereby changing the hardware configuration of the plurality of computing resources from the first hardware configuration to the second hardware configuration;
The resource management device according to claim 1 . the load information acquisition unit acquires second load information indicating a load on each of the plurality of computers;
an input data limiting unit that limits an amount of input data to a computer among the plurality of computers whose load is greater than a predetermined standard based on the second load information;
The resource management device according to claim 1 or 2, comprising: the estimation result acquisition unit acquires a performance estimation result that estimates performance related to a processing time of each of the one or more computers configured with the plurality of computing resources among the plurality of computers when using one or more computing resources configured on the computer among the plurality of computing resources, and estimates the quality of the service during operation of the computing system based on a value obtained by adding one or more fluctuation amounts to the acquired performance estimation result;
The resource management device according to any one of claims 1 to 3. the load information acquisition unit acquires second load information indicating a load on each of the plurality of computers;
the estimation result acquisition unit estimates the quality of the service during operation of the computing system based on a value obtained by adding a predetermined first fluctuation amount and a second fluctuation amount according to a trend of the load indicated by the past second load information to the acquired performance estimation result;
The resource management device according to claim 4. the priority control unit determines which of the services is to be given priority for the processing based on at least one of a priority order of processing execution when the plurality of services compete with each other and a degree of deviation between the quality estimation result and the required quality for each service.
The resource management device according to claim 1 . A program that causes a computer to function as the resource management device according to claim 1 .

Images