JP7350694B2 - Control device, information processing device, information processing control method, and computer program - Google Patents

Description

The present invention relates to a control device, an information processing device, an information processing control method, and a computer program.

2. Description of the Related Art Stream processing is conventionally known in which data processing is sequentially executed by a predetermined application when data is generated. According to stream processing, processing such as aggregation and analysis can be performed on sequentially generated data in real time. As information processing techniques for executing stream processing, for example, the techniques described in Patent Document 1 and Non-Patent Document 1 are known. In the conventional technology described in Patent Document 1, a switch executes primary processing of data and transfers the execution result to a stream cluster. The switch performs stream processing based on the descriptor received from the server.
Non-Patent Document 1 describes a distributed stream processing system that includes a "Worker Node" that executes stream processing by an application, a "Driver Program" that schedules application jobs, and a "Cluster Manager" that manages resources. has been done.

Patent No. 5802215

Apache Spark, “Cluster Mode Overview”, Internet <URL: https://spark.apache.org/docs/latest/cluster-overview.html>

However, in the conventional technology described in Patent Document 1 mentioned above, it is not possible to perform stream processing in consideration of delay requirements requested by an application. In the conventional technology described in Non-Patent Document 2, resources are fixedly allocated to "Worker Node", so if the amount of data to be stream processed per unit time varies depending on the time period, It is necessary to allocate a fixed amount of resources according to the maximum amount of generation. Therefore, surplus resources are generated during a time period when the amount of data to be stream processed is generated in a small amount per unit time, and resource utilization efficiency is reduced.

The present invention has been made in consideration of such circumstances, and its purpose is to perform stream processing while taking into account delay requirements for stream processing, and to improve resource utilization efficiency.

(1) One aspect of the present invention provides an intra-group scheduler that schedules jobs for each application based on the delay requirements of each application for each group composed of one or more applications; an inter-group scheduler that controls the amount of resources in each of the groups that can be used for job execution based on the degree of compliance with application delay requirements, which is the degree of compliance with the application delay requirements , which is the degree of compliance with the application delay requirements; The scheduler calculates the progress schedule achievement level and required resource amount, which are the degrees indicating whether the job is progressing as planned or not, for each job based on the delay requirements of each application, and calculates the progress schedule achievement level of each job. and determining the priority of each job based on the amount of required resources, and the intra-group scheduler determines the priority of the job by weighting the degree of achievement of the scheduled progress of the job and weighting the amount of required resources of the job. It is a control device.
( 2 ) One aspect of the present invention is that the intra-group scheduler controls each weighting coefficient for the progress schedule achievement level and the required resource amount of the job of the application so that the application's compliance with delay requirements is improved. This is the control device described in ( 1 ) above.
( 3 ) One aspect of the present invention provides an intra-group scheduler that schedules jobs for each application based on the delay requirements of each application for each group composed of one or more applications, and an intra-group scheduler that schedules jobs for each application based on delay requirements of each application. an inter-group scheduler that controls the amount of resources of each of the groups that can be used to execute a job based on the degree of compliance of the application with delay requirements, which is the degree of compliance with the application's delay requirements; The control device calculates the degree of compliance with the delay requirements based on the difference between the sum of the job execution time of the application plus the record read delay time and the processing delay time requested by the application.

( 4 ) One aspect of the present invention includes an execution unit that is provided for each group composed of one or more applications and executes a job, and a control device according to any one of (1) to ( 3 ) above. The execution unit is an information processing device that executes a job of an application of its own group according to scheduling by the control device.

( 5 ) One aspect of the present invention includes an intra-group scheduling step in which the intra-group scheduler schedules jobs for each application based on the delay requirements of each application for each group consisting of one or more applications; an interval scheduler controls the amount of resources of each of the groups that can be used to execute the job based on the degree of compliance with application delay requirements, which is the degree of compliance with application delay requirements in each of the groups; an inter-group scheduling step;
including;
The intra-group scheduler calculates the progress schedule achievement level and required resource amount, which are degrees of progress indicating whether the job is progressing as planned, for each job based on the delay requirements of each application, and calculates the calculated progress of each job. Determine the priority of each job based on the degree of target achievement and the amount of required resources,
The intra-group scheduler determines the priority of the job by weighting the progress plan achievement level of the job and weighting the required resource amount of the job.
This is an information processing control method.
(6) One aspect of the present invention includes an intra-group scheduling step in which the intra-group scheduler schedules jobs for each application based on the delay requirements of each application for each group consisting of one or more applications; an interval scheduler controls the amount of resources of each of the groups that can be used to execute the job based on the degree of compliance with application delay requirements, which is the degree of compliance with application delay requirements in each of the groups; an inter-group scheduling step, wherein the degree of compliance with the delay requirement of the application is calculated based on the difference between the sum of the job execution time of the application plus the record loading delay time and the processing delay time requested by the application. , an information processing control method.

( 7 ) One aspect of the present invention is to provide a computer with an intra-group scheduling step of scheduling jobs for each application based on the delay requirements of each application for each group consisting of one or more applications; an inter-group scheduling step of controlling the amount of resources of each of said groups available for execution of the job based on the degree of compliance with application delay requirements, which is the degree of compliance with application delay requirements; The intra-group scheduling step calculates and calculates the progress schedule achievement level and the required resource amount, which is a degree indicating whether the job is progressing as planned or not, for each job based on the delay requirements of each application. The priority of each job is determined based on the progress schedule achievement level and required resource amount of each job, and the intra-group scheduling step weights the progress schedule achievement level of the job and the required resource amount of the job. A computer program that determines job priorities .
(8) One aspect of the present invention is to provide a computer with an intra-group scheduling step of scheduling jobs for each application based on the delay requirements of each application for each group consisting of one or more applications; an inter-group scheduling step of controlling the amount of resources of each of said groups available for execution of the job based on the degree of compliance with application delay requirements, which is the degree of compliance with application delay requirements; The computer program is executed, and the degree of compliance with the delay requirements of the application is calculated based on the difference between the sum of the job execution time of the application plus the record read delay time and the processing delay time requested by the application.

According to the present invention, it is possible to perform stream processing in consideration of delay requirements for stream processing, and to improve resource utilization efficiency.

FIG. 1 is a diagram illustrating a configuration example of an information processing device according to an embodiment. 1 is a flowchart illustrating an example of the overall procedure of an information processing control method according to an embodiment. 3 is a flowchart illustrating an example of a procedure for intra-group scheduling according to an embodiment. 3 is a flowchart illustrating an example of a procedure for inter-group scheduling according to an embodiment.

Embodiments of the present invention will be described below with reference to the drawings.
FIG. 1 is a diagram illustrating a configuration example of an information processing apparatus according to an embodiment. In FIG. 1, the information processing device 1 includes a control device 10 and processing nodes wnd1 and wnd2. The control device 10 controls information processing for processing nodes wnd1 and wnd2 that execute stream processing by applications.

The processing nodes wnd1 and wnd2 include execution units exc1 and exc2 that execute jobs issued by applications. Each execution unit exc1, exc2 is provided for each application group. An application group is composed of one or more applications. The execution unit exc1 executes jobs issued by applications app1, app2, and app3 belonging to group A. The execution unit exc2 executes a job issued by an application app4 belonging to group B. The upper limit of information processing resources that can be used by each of the execution units exc1 and exc2 is controlled by the control device 10 for each processing node wnd1 and wnd2. The execution units exc1 and exc2 are constructed using, for example, a container-type virtual execution environment.

Hereinafter, information processing resources may be simply referred to as resources. Furthermore, when the processing nodes wnd1 and wnd2 are not particularly distinguished, they are referred to as processing nodes wnd. Furthermore, when the execution units exc1 and exc2 are not particularly distinguished, they are referred to as execution units exc. Furthermore, when groups A and B are not particularly distinguished, they are referred to as groups. Moreover, when the applications app1, app2, app3, and app4 are not particularly distinguished, they are referred to as applications app.

In this embodiment, each application app issues jobs so that only one job is executed at a time. Therefore, in each application, the next job is not issued until the previous job is completed. As a result, the execution unit exc executes only one job for one application at the same time.

The control device 10 includes an intra-group scheduler 11, an inter-group scheduler 12, an application manager 13, and a metrics collector 14. .

The functions of the control device 10 are realized when the control device 10 includes computer hardware such as a CPU (Central Processing Unit) and a memory, and the CPU executes a computer program stored in the memory. Note that the control device 10 may be configured using a general-purpose computer device, or may be configured as a dedicated hardware device. For example, the control device 10 may be configured using a server computer connected to a communication network such as the Internet. Moreover, each function of the control device 10 may be realized by cloud computing.

The intra-group scheduler 11 schedules jobs for each application within the group for each processing node wnd. The intra-group scheduler 11 schedules the jobs of each application app for each group based on the delay requirements of each application app. The execution unit exc executes the job of each application according to the scheduling of the intra-group scheduler 11. The delay requirements (Service Level Objective: SLO) of each application app are set in advance in the control device 10.

The inter-group scheduler 12 allocates resources between groups for each processing node wnd. Resources include a CPU, memory, storage, network, etc. used by the processing node wnd. The inter-group scheduler 12 controls the amount of resources in each group that can be used to execute a job, based on the degree of compliance with the delay requirements of the applications in each group. The degree of compliance is a degree indicating whether or not the delay requirements of the application app are satisfied.

The inter-group scheduler 12 constantly monitors to what extent the delay requirements of the application app are satisfied in each group. Based on the monitoring results, the inter-group scheduler 12 reduces the amount of resources that can be used by the group whose delay requirements of the application app are sufficiently satisfied, while reducing the amount of resources that can be used by the group whose delay requirements of the application app are not satisfied. Periodically execute control to increase the amount of resources.

The application management unit 13 has an interface function for executing, stopping, and checking the status of jobs issued by the application app.

The metric collection unit 14 monitors the status of the application APP. The metric collection unit 14 collects information for determining whether the delay requirements of each application app are satisfied from each processing node wnd.

The overall procedure of the information processing control method according to this embodiment will be described with reference to FIG. 2. FIG. 2 is a flowchart illustrating an example of the overall procedure of the information processing control method according to the present embodiment. Here, the overall procedure of the information processing control method will be explained focusing on one processing node wnd, but the same procedure is performed for each of the plurality of processing nodes wnd.

(Step S1) The inter-group scheduler 12 performs the first inter-group schedule. In the first inter-group schedule, resources are distributed between groups according to initial settings for each group. For this initial setting, a default value or an arbitrary setting value set by the user may be used, or a value based on past results of resource allocation between groups may be used. According to the first inter-group schedule, an initial value of the upper limit of usable resource amount is set for the execution unit exc corresponding to each group.

(Step S2) The intra-group scheduler 11 performs intra-group scheduling. In intra-group scheduling, the intra-group scheduler 11 schedules the jobs of each application app for each group based on the delay requirements of each application app. Details of intra-group scheduling will be described later.

(Step S3) The inter-group scheduler 12 calculates the degree of compliance with the delay requirements of the application in each group. The method for calculating the degree of compliance will be described later.

(Step S4) The inter-group scheduler 12 determines whether rescheduling between groups is necessary based on the degree of compliance with the delay requirements of the applications in each group. As a result of this determination, if rescheduling between groups is necessary (step S4, YES), the process advances to step S5, and if not (step S4, NO), the process returns to step S2.

(Step S5) The inter-group scheduler 12 performs inter-group scheduling. In this inter-group scheduling, the inter-group scheduler 12 reallocates resources between groups based on the degree of compliance with delay requirements of applications in each group. Details of inter-group scheduling will be described later.

(Step S6) If the execution of the stream processing in the processing node wnd has ended (step S6, YES), the process in FIG. 2 is ended. On the other hand, if the execution of the stream processing at the processing node wnd is to be continued (step S6, NO), the process returns to step S2.

Next, the procedure for intra-group scheduling according to this embodiment will be explained with reference to FIG. 3. FIG. 3 is a flowchart illustrating an example of a procedure for intra-group scheduling according to this embodiment.

(Step S11) The intra-group scheduler 11 calculates the required resource amount rJ for the job of each application app for each group. The required resource amount rJ is calculated for each application and for each job.

Note that although intra-group scheduling (step S2 in FIG. 2) is executed periodically, the initially calculated value of the required resource amount rJ for a job is used until the job is completed. Therefore, the intra-group scheduler 11 holds the required resource amount rJ initially calculated for each job, and uses the held required resource amount rJ without changing it until the job is completed.

(Required resource amount calculation method)
Here, a method for calculating the amount of required resources according to this embodiment will be explained. In order to satisfy the delay requirements (SLO) of the application app, it is necessary to satisfy the following equation.
SLO ≧ J+Rd
However, SLO is a processing delay time required by the application. J is the time required to execute the job (job execution time). Rd is the time required to read records necessary for the job (record read delay time).

As an example of how to measure the record read delay time Rd, use the timestamp of the storage system where the record is written, calculate the difference between the time the record was written and the time the record was read, and calculate the difference between the record read time and the record read time. Used for reading delay time Rd. Another example of how to measure the record read delay time Rd is to write a probe (dummy data) to the storage system where the record is written, measure the time until the probe is read, and calculate the record read delay time as the measurement result. Used for time Rd. The intra-group scheduler 11 uses the record read delay time Rd measured by the record read delay time Rd measuring method.

The intra-group scheduler 11 uses the known record read delay time Rd and determines the required resource amount rJ of the job so as to satisfy "SLO-Rd≧J". Here, the required resource amount rJ is expressed as a ratio (the following equation) to the total resource amount of the group (the upper limit of resources that can be used by the execution unit exc).
rJ = (amount of resources used by the job) ÷ (total amount of resources in the group)

Note that when there are multiple types of resources (CPU, memory, storage, network, etc.), the required resource amount rJ is represented by an N-dimensional vector corresponding to the resource types (N types). Note that in the following description, a case will be described in which the required resource amount rJ is a one-dimensional value.

There exists a relationship expressed by the following relational expression between the job execution time J, the required resource amount rJ, and the job input data amount IJ.
rJ=f(J,IJ)
However, the job input data amount IJ is the amount of data used by the job.

As for the method for obtaining the job input data amount IJ, if the job input data amount IJ can be measured, the intra-group scheduler 11 uses the measured value of the job input data amount IJ. On the other hand, if it is difficult to measure the job input data amount IJ, the intra-group scheduler 11 estimates the job input data amount IJ from the most recent past data amount of data used by the job. The intra-group scheduler 11 uses the job input data amount IJ (measured value or estimated value) acquired by the job input data amount IJ acquisition method.

The method for deriving f(J, IJ) is, for example, by linear regression analysis to derive f(J, IJ) based on the information of job execution time J, required resource amount rJ, and job input data amount IJ collected by the metric collection unit 14. seek. Generally, there is a proportional relationship between the job input data amount IJ and the job execution time J. Further, there is an inversely proportional relationship between the job execution time J and the required resource amount rJ. From this, linear regression analysis can be cited as a method for deriving f(J, IJ). Note that f(J, IJ) may be obtained using a machine learning algorithm such as a support vector machine (SVM) or deep learning. The intra-group scheduler 11 uses f(J, IJ) obtained by the f(J, IJ) derivation method.

The intra-group scheduler 11 uses the known record read delay time Rd, the job input data amount IJ, and the relational expression "rJ=f(J,IJ)" to schedule the job so that "SLO-Rd ≧ J" is satisfied. The required resource amount rJ is determined.

(Step S12) The intra-group scheduler 11 calculates the scheduled progress achievement level "Pr/Ptr" of the job of each application app for each group. The progress schedule achievement level "Pr/Ptr" is a degree indicating whether or not the job is progressing as planned. The progress plan achievement level "Pr/Ptr" is calculated by the following formula.
"Pr/Ptr" = "Job progress Pr" ÷ "elapsed time percentage Ptr"
Elapsed time ratio Ptr = (t-Ts) ÷ (Tdj-Ts)
However, t is the current time. Ts is the time when the job was issued. Tdj is the scheduled end time (deadline) of the job.

The job progress Pr is measured by the application management unit 13. "0≦Pr≦1". The time Ts at which the job was issued is acquired by the application management unit 13. The scheduled end time Tdj of the job is calculated by the following equation based on the time Ts when the job was issued, the processing delay time SLO requested by the application, and the record reading delay time Rd.
Tdj=Ts+SLO-Rd

When the progress plan achievement level "Pr/Ptr" is 1 or more, the job is progressing as planned or ahead of schedule. On the other hand, when the progress schedule achievement level "Pr/Ptr" is less than 1, the job is behind schedule.

(Step S13) The intra-group scheduler 11 calculates the priority P of the job of each application app for each group. The priority P is calculated using the following formula.
P=α(1÷(Pr/Ptr))+(1-α)×rJ
However, α is a weighting coefficient. "0≦α≦1". The weighting coefficient α of each group can be set arbitrarily.

The priority P is calculated based on the job progress schedule achievement level "Pr/Ptr" and the required resource amount rJ (in the case of 0<α<1). The priority level P becomes higher as the job progress schedule achievement level "Pr/Ptr" is smaller (in the case of 0<α). The priority P becomes higher as the required resource amount rJ becomes larger (in the case of α<1).

(Method of determining weighting coefficient α)
Regarding the method of determining the weighting coefficient α, a static determination method and a dynamic determination method will be explained.

(1) Static determination method Although the weighting coefficient α of each group can be set arbitrarily, the weighting coefficient α of each group can be set statically by using the static determination method of Example 1-1 or Example 1-2 shown below. may be determined.

(Example 1-1) If it is known that the required resource amount rJ of all applications in the group will be approximately the same value, the weight (α) of the planned progress achievement rate "Pr/Ptr" is increased.

(Example 1-2) If one of the progress schedule achievement level "Pr/Ptr" and the required resource amount rJ is inaccurate, the weight of the other is increased.
For example, if the job execution time J is a long time exceeding several minutes, there is a high possibility that the required resource amount rJ is incorrect, so the weight (α) of the scheduled progress achievement rate "Pr/Ptr" is increased.
For example, if there is a job composed of multiple task sets (TaskSets) among the jobs of the application app in the group (for example, in the case of a multi-stage configuration where the calculation result of the first task set is used in the second task set) ), the progress rate Pr of the job is not proportional to the elapsed time ratio Ptr, so the scheduled progress achievement rate "Pr/Ptr" is not accurate. In this case, the weight (1-α) of the required resource amount rJ is increased.

(2) Dynamic determination method Although the weighting coefficient α of each group can be set arbitrarily, the weighting coefficient α of each group may be dynamically determined by the dynamic determination method described below.

The dynamic determination method dynamically determines a weighting coefficient α suitable for each group. An example of a dynamic decision method will be explained. In advance, the weighting coefficient α is defined as a discrete value, for example, “α∈[0.0, 0.1, 0.2, . . . , 0.9, 1.0]”. Next, in actual scheduling within the group, when each weighting coefficient α is used, the degree of compliance Ga with the delay requirements of the applications within the group is calculated. Thereby, the degree of compliance Ga at each weighting coefficient α is obtained for each group. Note that the initial value of the weighting coefficient α of each group is set in advance using a default value or the above-described static determination method.

The method for calculating the degree of compliance with delay requirements Ga for applications in a group is to calculate the average value of "J+Rd-SLO" for each application using a sliding window, and then calculate the "J+Rd-SLO" of each application. The average value is further averaged within the group, and the average value is set as the compliance level Ga. In this example of the compliance level, the larger the positive value of the compliance level Ga, the more the delay requirements of the application are not satisfied. In step S3 of FIG. 2 described above, the compliance level Ga of the delay requirements of the application app is calculated for each group by this compliance level Ga calculation method.

When the compliance degree Ga of a group is a positive value for a certain weighting coefficient α, the group is in a state where the delay requirements of the application cannot be satisfied with the weighting coefficient α. Therefore, by searching for a weighting coefficient α that tends to reduce the value of the group's compliance level Ga (easily improves the compliance level), the most suitable weighting coefficient α for the group is searched. For example, a Markov decision process can be used as the algorithm for determining the optimal weighting coefficient α.

(Step S14) The intra-group scheduler 11 schedules each job for each group based on the job priority P of each application. In this job scheduling, the intra-group scheduler 11 calculates the scheduling weight W of each scheduled job for each group. The scheduled job is a job waiting to be executed that is registered in the waiting queue. The scheduling weight W is calculated by the following equation.
W=P_w÷(ΣP)
However, P_w is the priority P of the scheduling target job for which the scheduling weight is to be calculated. ΣP is the sum of the priorities P of all jobs waiting to be executed that are registered in the waiting queue.

The intra-group scheduler 11 selects jobs to be executed for each group, based on the scheduling weight W of each scheduling target job, in order from the scheduling target job with the largest scheduling weight W. As a result, among all the jobs waiting for execution in the group, the jobs with the largest scheduling weight W are executed in order.

In this embodiment, intra-group scheduling (step S2 in FIG. 2) is performed periodically. Therefore, the scheduling weight W of the job waiting to be executed within the group is updated every time intra-group scheduling (step S2 in FIG. 2) is executed. This makes it possible to schedule jobs according to the latest job execution status within the group. For example, the required resource amount rJ for a job is fixed until the job ends, but even if the required resource amount rJ for the job is calculated properly, the job input data amount IJ may suddenly become excessive. Therefore, it may become impossible to meet the delay requirements of the application. According to this embodiment, by periodically executing intra-group scheduling, job scheduling can be performed to meet the delay requirements of the application even in the event of such unexpected changes in the job execution environment. It can be carried out.

In addition, in the above-mentioned intra-group scheduling procedure, intra-group scheduling is performed on a job-by-job basis, but when a job of an application is divided into multiple tasks and processed, intra-group scheduling is performed on tasks of the same processing. It may be performed in units of task sets (TaskSet).

Furthermore, in the above description, the case where the target resource of the required resource amount rJ is one type and the required resource amount rJ is a one-dimensional value has been explained, but the following describes the case where the target resource of the required resource amount rJ is multiple types. Explain. Here, method 1 in which multiple types of resource amounts are handled as multidimensional amounts, and method 2 in which multiple types of resource amounts are handled independently.

(Method 1: Method of treating multiple types of resource quantities as multidimensional quantities)
In method 1, the required resource amount rJ is represented by an N-dimensional vector (N is an integer of 2 or more). Each of the N elements of the vector of the required resource amount rJ (N-dimensional vector rJ) is the required resource amount of each of the N types of target resources. For example, if the target resources of the required resource amount rJ are two types, CPU and memory, the required resource amount rJ is expressed as a two-dimensional vector "rJ=(CPU_rJ, Memory_rJ)". CPU_rJ is the required resource amount of the CPU. Memory_rJ is the required resource amount of memory. Further, the required resource amount of each target resource is expressed as a percentage of the total amount of each target resource in the group. For example, the required resource amount CPU_rJ of the CPU is expressed by the following equation.
CPU_rJ = (Amount of CPU resources used by the job) ÷ (Total amount of CPU resources for the group)

Then, when calculating the priority P, the required resource amount rJ (N-dimensional vector rJ) is mapped into one dimension. In this mapping, if the mapping result is "rJ1", then "rJ1=|rJ|. |rJ| is the size of the N-dimensional vector rJ. Here, "rJ1" is the size of the N-dimensional vector rJ. The reason for this is that each of the N elements (required resource amount of each target resource of N types) of the vector of the required resource amount rJ (N-dimensional vector rJ) is expressed as a percentage of the total amount of each target resource in the group. This is because it has the same scale regardless of the type of target resource.

Note that as another method of mapping the required resource amount rJ (N-dimensional vector rJ) into one dimension, mapping may be performed by weighting each of the N types of target resources. For example, when "rJ=(a, b, c)", a function that returns "rJ1=√(2a+3b+1c)^2" as a result may be used.

The priority P is calculated by the following formula, which is obtained by applying "rJ1", a result of one-dimensional mapping of the required resource amount rJ (N-dimensional vector rJ), to "rJ" using the above-mentioned formula for calculating the priority P.
P=α(1÷(Pr/Ptr))+(1-α)×rJ1
Furthermore, the scheduling weight W is calculated using the same calculation formula as described above.

(Method 2: Method of handling multiple types of resource quantities independently)
In method 2, the required resource amount rJ is calculated separately for each of the N types of target resources. Then, for each of the N types of target resources, the priority P and scheduling weight W are calculated using the required resource amount of each target resource.

For example, if the target resources for the required resource amount rJ are two types, CPU and memory, the required resource amount CPU_rJ for the CPU and the required resource amount Memory_rJ for the memory are calculated separately using the following equations.
CPU_rJ = (Amount of CPU resources used by the job) ÷ (Total amount of CPU resources for the group)
Memory_rJ = (amount of memory resources used by the job) ÷ (total amount of memory resources for the group)
Then, the CPU priority P_CPU and the memory priority P_Memory are calculated separately using the following equations.
P_CPU=α(1÷(Pr/Ptr))+(1-α)×CPU_rJ
P_Memory=α(1÷(Pr/Ptr))+(1-α)×Memory_rJ

Then, the CPU and the memory calculate the scheduling weight W separately using the same calculation formula as described above. As a result, a scheduling weight W_CPU for scheduling the CPU resource amount and a scheduling weight W_Memory for scheduling the memory resource amount are obtained. The amount of CPU resources is allocated to each job based on the scheduling weight W_CPU. The amount of memory resources is allocated to each job based on the scheduling weight W_Memory.

Next, a procedure for inter-group scheduling according to this embodiment will be described with reference to FIG. 4. FIG. 4 is a flowchart illustrating an example of a procedure for inter-group scheduling according to this embodiment.

In step S3 of FIG. 2 described above, the inter-group scheduler 12 calculates the compliance degree Ga of the delay requirements of the application in each group. Thereby, the degree of compliance Ga is obtained for each group. The compliance level Ga of a certain group is a degree indicating whether the group satisfies the delay requirements of the application app.

(Step S21) The inter-group scheduler 12 determines whether the compliance level Ga is good or bad for each group based on the compliance level Ga of each group.

When the compliance level Ga of a group is a positive value (“Ga>0”), the group is in a state where it cannot satisfy the delay requirements of the application. Therefore, the inter-group scheduler 12 sets groups whose compliance level Ga has a positive value as resource allocation target groups.

On the other hand, if the compliance degree Ga of the group is "Ga<r" (r is a predetermined negative value), the delay requirements of the application are sufficiently satisfied and a sufficient amount of resources has already been allocated to the group. It is thought that there are. Therefore, the inter-group scheduler 12 sets a group whose compliance degree Ga is "Ga<r" (r is a predetermined negative value) as a resource release target group.

(Step S22) The inter-group scheduler 12 redistributes resources to each group. In this resource reallocation, the resource release target group releases a certain amount of resources. Then, the resources released by the resource release target group are distributed to the resource allocation target groups. In this resource allocation method, resources are allocated based on the ratio of "compliance level Ga x allocation priority level Pg" of the resource allocation target group. The distribution priority Pg is set in advance for each group.

Note that if there is no group in which the application delay requirement is not satisfied (a group whose compliance level Ga is a positive value), step S22 described above is not executed. In this case, there is no need to reschedule between groups, and the result of determination in step S4 in FIG. 2 described above corresponds to "NO". On the other hand, if there is a group that does not satisfy the application delay requirements (a group whose compliance degree Ga is a positive value), it is necessary to reschedule between the groups, and the decision made in step S4 of FIG. The result corresponds to "YES".

In addition, when there is a group in which the application delay requirement is not satisfied (a group whose compliance degree Ga is a positive value), there is a group in which the application delay requirement is sufficiently satisfied (a group whose compliance degree Ga is “Ga< r (where r is a predetermined negative value) does not exist, rescheduling between the groups is required, but in step S22, there is no resource release target group, so the resource allocation is Since resources to be allocated to the target group are not released, resources are not reallocated to the resource allocation target group.

According to the embodiment described above, the intra-group scheduler 11 schedules the jobs of each application app for each group based on the delay requirements of each application app, thereby performing stream processing in consideration of the delay requirements for stream processing by the application apps. can be executed. Furthermore, the inter-group scheduler 12 controls the amount of resources in each group that can be used to execute a job based on the compliance level Ga of the delay requirements of the application in each group, thereby improving resource utilization efficiency. can.

Further, by grouping the applications, it is possible to adjust the trade-off between independence and efficiency of execution of the applications. By grouping applications, for example, even if a failure occurs in the execution unit EXC of one group and the application becomes unable to run, this will not affect the execution of applications in other groups. , independence of execution of applications between groups is obtained. Furthermore, by dividing into a plurality of groups and scheduling for each group, even if the number of applications in a certain group increases, it can be handled efficiently by scheduling within the group. Furthermore, by dividing into a plurality of groups and reducing the number of applications in one group, it is possible to suppress interference between applications in a group and perform efficient scheduling.

Explain how to group applications. Although the grouping of applications can be set arbitrarily, the grouping of applications may be determined by the grouping method described below.

(1) Grouping method that takes failure into account Applications belonging to the same group may all become unexecutable at the same time if a failure occurs in the execution unit of the group. Now, consider implementing a service that provides value by using multiple applications. If the same group is made up of multiple applications that provide multiple services that are unrelated to each other, if a failure occurs in the execution unit of the group, multiple services may go down at the same time. For this reason, applications that provide the same service are grouped together. This makes it possible to limit the impact of a failure in the execution unit on services to specific services. When applications that provide the same service are grouped together, the number of groups is equal to the number of services.

(2) Grouping method considering performance If the execution cycle of inter-group scheduling is sufficiently longer than the execution cycle of intra-group scheduling, the amount of job input data IJ suddenly increases and the amount of resources within the group becomes insufficient. However, it takes time for resources to be replenished by the next inter-group scheduling. This is undesirable from the perspective of the delay requirements of the application.

In order to deal with this problem, applications in which the amount of job input data IJ repeatedly increases and decreases in short cycles and whose timing trends of increase and decrease in the amount of job input data IJ are the same are not placed in the same group. As a result, even if there are applications in the same group whose job input data amount IJ repeatedly increases or decreases, the timing of increase or decrease in the job input data amount IJ will be different, so resource shortages will be suppressed for the group as a whole. Ru.

It is also preferable to create groups so that the number of applications belonging to each group is as equal as possible. This is because when the number of applications belonging to a group is small, there is generally less margin for scheduling adjustments between jobs in intra-group scheduling, so groups with a small number of applications are particularly vulnerable to fluctuations in the amount of job input data IJ. This is because

Although the embodiment of the present invention has been described above in detail with reference to the drawings, the specific configuration is not limited to this embodiment, and design changes and the like may be made without departing from the gist of the present invention.

Further, a computer program for realizing the functions of each device described above may be recorded on a computer-readable recording medium, and the program recorded on the recording medium may be read into a computer system and executed. Note that the "computer system" here may include hardware such as an OS and peripheral devices.
Furthermore, "computer-readable recording media" refers to flexible disks, magneto-optical disks, ROMs, writable non-volatile memories such as flash memory, portable media such as DVDs (Digital Versatile Discs), and media built into computer systems. A storage device such as a hard disk.

Furthermore, "computer-readable recording medium" refers to volatile memory (for example, DRAM (Dynamic It also includes those that retain programs for a certain period of time, such as Random Access Memory).
Further, the program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in a transmission medium. Here, the "transmission medium" that transmits the program refers to a medium that has a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line.
Moreover, the above-mentioned program may be for realizing a part of the above-mentioned functions. Furthermore, it may be a so-called difference file (difference program) that can realize the above-described functions in combination with a program already recorded in the computer system.

DESCRIPTION OF SYMBOLS 1... Information processing device, 10... Control device, 11... Intra-group scheduler, 12... Inter-group scheduler, 13... Application management unit, 14... Metric collection unit, app1, app2, app3, app4... Application, exc1, exc2... Execution part, wnd1, wnd2...processing node

Claims (8)

an intra-group scheduler that schedules jobs for each application based on the delay requirements of each application for each group consisting of one or more applications;
an inter-group scheduler that controls the amount of resources of each of the groups that can be used for job execution based on the degree of compliance with application delay requirements, which is the degree of compliance with application delay requirements in each of the groups; ,
Equipped with
The intra-group scheduler calculates the progress schedule achievement level and required resource amount, which are degrees of progress indicating whether the job is progressing as planned, for each job based on the delay requirements of each application, and calculates the calculated progress of each job. Determine the priority of each job based on the degree of target achievement and the amount of required resources,
The intra-group scheduler determines the priority of the job by weighting the progress plan achievement level of the job and weighting the required resource amount of the job.
Control device. The intra-group scheduler controls each weighting coefficient for the progress schedule achievement level and the required resource amount of the job of the application so that the application's compliance with delay requirements is improved.
The control device according to claim 1 . an intra-group scheduler that schedules jobs for each application based on the delay requirements of each application for each group consisting of one or more applications;
an inter-group scheduler that controls the amount of resources of each of the groups that can be used for job execution based on the degree of compliance with application delay requirements, which is the degree of compliance with application delay requirements in each of the groups; ,
Equipped with
The degree of compliance with the delay requirements of the application is calculated based on the difference between the sum of the job execution time of the application plus the record read delay time and the processing delay time requested by the application.
Control device. an execution unit that is provided for each group consisting of one or more applications and executes a job;
A control device according to any one of claims 1 to 3 ,
The execution unit executes the job of the application of its own group according to the scheduling by the control device.
Information processing device. an intra-group scheduling step in which the intra-group scheduler schedules jobs for each application based on the delay requirements of each application for each group consisting of one or more applications;
An inter-group scheduler controls the amount of resources available for each said group for job execution based on the degree of compliance with application delay requirements, which is the degree to which application delay requirements in each said group are met. an intergroup scheduling step to
including;
The intra-group scheduler calculates the progress schedule achievement level and required resource amount, which are degrees of progress indicating whether the job is progressing as planned, for each job based on the delay requirements of each application, and calculates the calculated progress of each job. Determine the priority of each job based on the degree of target achievement and the amount of required resources,
The intra-group scheduler determines the priority of the job by weighting the progress plan achievement level of the job and weighting the required resource amount of the job.
Information processing control method. an intra-group scheduling step in which the intra-group scheduler schedules jobs for each application based on the delay requirements of each application for each group consisting of one or more applications;
An inter-group scheduler controls the amount of resources available for each said group for job execution based on the degree of compliance with application delay requirements, which is the degree to which application delay requirements in each said group are met. an intergroup scheduling step to
including;
The degree of compliance with the delay requirements of the application is calculated based on the difference between the sum of the job execution time of the application plus the record read delay time and the processing delay time requested by the application.
Information processing control method. to the computer,
an intra-group scheduling step of scheduling jobs for each application based on the delay requirements of each application for each group consisting of one or more applications;
an inter-group scheduling step for controlling the amount of resources of each of the groups available for execution of the job based on the degree of compliance with application delay requirements, which is the degree of compliance with the application delay requirements in each of the groups; and,
run the
The intra-group scheduling step calculates the progress schedule achievement level and the required resource amount, which are the degrees indicating whether the job is progressing as planned, for each job based on the delay requirements of each application, and calculates the amount of required resources for each job. Determine the priority of each job based on the progress plan achievement level and the amount of required resources,
The intra-group scheduling step determines the priority of the job by weighting the degree of achievement of the scheduled progress of the job and weighting the required resource amount of the job.
computer program. to the computer,
an intra-group scheduling step of scheduling jobs for each application based on the delay requirements of each application for each group consisting of one or more applications;
an inter-group scheduling step for controlling the amount of resources of each of the groups available for execution of the job based on the degree of compliance with application delay requirements, which is the degree of compliance with the application delay requirements in each of the groups; and,
run the
The degree of compliance with the delay requirements of the application is calculated based on the difference between the sum of the job execution time of the application plus the record read delay time and the processing delay time requested by the application.
computer program.

Images