KR101996768B1 - Method and apparatus of managing resource - Google Patents

Abstract

A resource management method and an apparatus are disclosed. According to an embodiment of the present invention, the resource management method comprises the steps of: receiving a reservation request for a network bandwidth resource; determining a scheduling possibility for the network bandwidth resource based on the reservation request; reserving the network bandwidth resource based on the scheduling possibility for the network bandwidth resource; determining schedulability for a CPU bandwidth resource based on the reserved network bandwidth resource; and reserving the CPU bandwidth resource based on the schedulability for the CPU bandwidth resource.

Description

[0001] METHOD AND APPARATUS OF MANAGING RESOURCE [0002]

The following embodiments relate to a resource management method and apparatus.

Linux provides a cgroup (Control Group) framework that manages various hardware resources on a task group basis. In addition, container technologies used in Linux environments (eg, Docker, LXC (LinuX Containers)) use cgroups to reserve resources.

However, since cgroups use a resource management method that focuses on resource constraints, the protection against reserved resources is low. In order to solve these problems, technologies aiming at increasing the protection of resources such as CPU and disk have been studied. However, there is still insufficient research to improve the protection of network resources.

Linux has provided a tc (Traffic Controller) framework to manage network resources, and cgroups can use tc to reserve network resources on a task group basis. However, tc does not consider reserved network resources for resources used in different queues, but only for managing the upper bound of network bandwidth through qdiscs in each queue.

Therefore, if the processes using the network resources are scattered in different queues in a situation where the available network bandwidth is insufficient, the network bandwidth may not be ensured due to mutual network resource interference.

In addition, when a process consumes network bandwidth, it can not guarantee provision of network bandwidth resources reserved for the process when resources such as central processing unit (CPU) and memory are interfered with by other processes.

Therefore, it is difficult to support the network-intensive real-time application with only the network resource management function using the existing Linux tc and cgroup, and it is also difficult to operate the billing system precisely based on the network resources used by the container.

Embodiments may provide techniques for managing resources.

A resource management method according to an exemplary embodiment includes receiving a reservation request for a network bandwidth resource, determining schedulability for the network bandwidth resource based on the reservation request, scheduling for the network bandwidth resource, Determining a scheduling probability for a CPU bandwidth resource based on a reserved network bandwidth resource; and determining a scheduling probability for the CPU bandwidth resource based on schedulability for the CPU bandwidth resource, For example.

Wherein determining the schedulability for the network bandwidth resource comprises: calculating a sum of reserved network bandwidth resources; comparing the sum of the maximum available network bandwidth resources and the reserved network bandwidth resources to determine schedulability; . ≪ / RTI >

The resource management method further comprises setting a limit of the network bandwidth resource to a value obtained by subtracting the sum of the reserved network bandwidth resources from the maximum available network bandwidth resource after reserving the network bandwidth resource .

Wherein determining the schedulability for the CPU bandwidth resource comprises: calculating a required CPU bandwidth resource per unit network bandwidth; comparing the sum of the CPU bandwidth resources needed per sum and the unit network bandwidth to the maximum available CPU bandwidth resource; And determining a schedulability possibility.

Wherein the resource management method further comprises: setting a limit of the CPU bandwidth resource to a value obtained by subtracting the product from the maximum available CPU bandwidth resource after reserving the CPU bandwidth resource based on schedulability for the CPU bandwidth resource As shown in FIG.

The resource management method may further include monitoring a usage amount of the network bandwidth resource.

The monitoring may include generating a default queue of the network bandwidth resource and managing the non-reserved bandwidth non-real-time network flow with the default queue.

The monitoring may further include adjusting a network bandwidth of the default queue when a process for reserving the network bandwidth resource occurs or a change occurs.

A resource management apparatus according to an embodiment includes a receiver for receiving a reservation request for a network bandwidth resource, a scheduler for determining schedulability for a network bandwidth resource based on the reservation request, and for determining a scheduling probability for the network bandwidth resource Determining a scheduling probability for a CPU bandwidth resource based on a reserved network bandwidth resource and reserving the CPU bandwidth resource based on schedulability for the CPU bandwidth resource, .

The processor may calculate the sum of the reserved network bandwidth resources and determine the schedulability by comparing the sum of the maximum available network bandwidth resources and the reserved network bandwidth resources.

The processor may set the limitation of the network bandwidth resource to a value obtained by subtracting the sum of the reserved network bandwidth resources from the maximum available network bandwidth resource after reserving the network bandwidth resource.

The processor may calculate a required CPU bandwidth resource per unit network bandwidth and compare the maximum available CPU bandwidth resource with a product of the sum and the required CPU bandwidth resource per unit network bandwidth to determine schedulability.

The processor may set a limit of the CPU bandwidth resource to a value obtained by subtracting the product from the maximum available CPU bandwidth resource after reserving the CPU bandwidth resource based on schedulability of the CPU bandwidth resource.

The processor may monitor usage of the network bandwidth resource.

The processor may generate a default queue for the network bandwidth resources and manage non-real-time network flows that are not reserved for bandwidth with the default queue.

The processor may adjust the network bandwidth of the default queue when an end or change of a process that reserves the network bandwidth resource occurs.

1 shows a schematic block diagram of a resource management apparatus according to an embodiment.
Fig. 2 shows an example of the operation of the resource management apparatus shown in Fig.
3 shows an example of the result of the resource protection verification test for the network load.
FIG. 4 shows an example of the results of the resource protection verification test for the CPU load.
Fig. 5 shows an example of bandwidth-specific bandwidth configuration of the load with the largest load.
FIG. 6 shows a flowchart of the operation of the resource management apparatus shown in FIG. 1. FIG.

In the following, embodiments will be described in detail with reference to the accompanying drawings. However, various modifications may be made in the embodiments, and the scope of the patent application is not limited or limited by these embodiments. It is to be understood that all changes, equivalents, and alternatives to the embodiments are included in the scope of the right.

The terms used in the examples are used for descriptive purposes only and are not to be construed as limiting. The singular expressions include plural expressions unless the context clearly dictates otherwise. In this specification, the terms "comprises" or "having" and the like refer to the presence of stated features, integers, steps, operations, elements, components, or combinations thereof, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

The terms first, second, or the like may be used to describe various elements, but the elements should not be limited by terms. The terms may be named for the purpose of distinguishing one element from another, for example, without departing from the scope of the right according to the concept of the embodiment, the first element being referred to as the second element, The second component may also be referred to as a first component.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this embodiment belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Do not.

In the following description of the present invention with reference to the accompanying drawings, the same components are denoted by the same reference numerals regardless of the reference numerals, and redundant explanations thereof will be omitted. In the following description of the embodiments, a detailed description of related arts will be omitted if it is determined that the gist of the embodiments may be unnecessarily blurred.

A module in this specification may mean hardware capable of performing the functions and operations according to the respective names described in this specification and may mean computer program codes capable of performing specific functions and operations , Or an electronic recording medium, e.g., a processor or a microprocessor, equipped with computer program code capable of performing certain functions and operations.

In other words, a module may mean a functional and / or structural combination of hardware for carrying out the technical idea of the present invention and / or software for driving the hardware.

1 shows a schematic block diagram of a resource management apparatus according to an embodiment.

Referring to FIG. 1, the resource management apparatus 10 can manage resources according to a resource reservation request of a user. The resources managed by the resource management apparatus 10 may include network bandwidth resources and CPU bandwidth resources.

The network bandwidth resources may include the transmission rate. For example, the unit of the network bandwidth resource may be bps (bits per second). The CPU bandwidth resources may include the amount of data that the CPU can exchange, or the number of CPU cores.

The resource management apparatus 10 can manage network bandwidth resources and CPU bandwidth resources. The resource management apparatus 10 may analyze the schedulability in the network resource reservation step.

In addition, the resource management apparatus 10 may limit the usage of the network bandwidth resources of the non-real-time network flows to reduce resource interference caused by the non-real-time network flows, and to guarantee the CPU resources of the applications that simultaneously process the reserved real- Thereby guaranteeing the reserved network bandwidth.

The resource management apparatus 10 includes a receiver 100 and a processor 200. [ The resource management device 10 may include a memory 300.

Receiver 100 may receive a reservation request for a network bandwidth resource. The receiver 100 may output the received reservation request to the processor 200.

로 나타낼 수 있다.The processor 200 may determine the schedulability of the network bandwidth resource based on the received reservation request. The processor 200 may calculate the sum of the reserved network bandwidth resources. The sum of the reserved network bandwidth resources

.

와 같이 나타낼 수 있다.The processor 200 may determine the schedulability by comparing the sum of the maximum available network bandwidth resource and the reserved network bandwidth resource. The maximum available network bandwidth resources

As shown in Fig.

The processor 200 may determine that the sum of the reserved network bandwidth resources does not exceed the maximum available network bandwidth resource. This can be expressed by Equation (1).

The processor 200 may determine that scheduling is possible if the sum of the reserved network bandwidth resources does not exceed the maximum available network bandwidth resource. The processor 200 may reserve network bandwidth resources based on schedulability for network bandwidth resources.

At this time, the processor 200 may limit reservations of network bandwidth resources such that the sum of the reserved network bandwidth resources does not exceed the maximum available network bandwidth resource.

을 계산하는 과정은 수학식 2과 같이 나타낼 수 있다.After reserving the network bandwidth resource, the processor 200 may set a limit of the network bandwidth resource by subtracting the sum of the reserved network bandwidth resources from the maximum available network bandwidth resource. Limitations of network bandwidth resources

Can be expressed by Equation (2). &Quot; (2) "

을 계산할 수 있다.The processor 200 may determine the schedulability for the CPU bandwidth resources based on the reserved network bandwidth resources. The processor 200 determines the required CPU bandwidth per unit network bandwidth

Can be calculated.

을 비교하여 스케줄링 가능성을 판단할 수 있다. 이는 수학식 3과 같이 나타낼 수 있다.The processor 200 calculates the product of the sum of the reserved network bandwidth resources and the required CPU bandwidth per unit network bandwidth and the maximum available CPU bandwidth resource

To determine the schedulability. This can be expressed by Equation (3).

That is, the processor 200 may determine that the CPU bandwidth resource is reservable when the product of the sum of the reserved network bandwidth resources and the required CPU bandwidth resource per unit network bandwidth is smaller than the maximum available CPU bandwidth resource. The processor 200 may reserve the CPU bandwidth resources based on the schedulability of the CPU bandwidth resources.

을 설정할 수 있다.After reserving the CPU bandwidth resource based on the schedulability of the CPU bandwidth resource, the processor 200 calculates the sum of the sum of the reserved network bandwidth resources in the maximum available CPU bandwidth resource minus the product of the CPU bandwidth resource required per unit network bandwidth Limit CPU bandwidth resources to

Can be set.

은 CPU 코어의 수를 포함할 수 있다. 예를 들어,

일 수 있다. 프로세서(200)가 제한

을 설정하는 동작은 수학식 4와 같이 나타낼 수 있다.Limitations of CPU bandwidth resources

May include the number of CPU cores. E.g,

Lt; / RTI > Processor 200 is limited

Can be expressed by Equation (4). &Quot; (4) "

The processor 200 may monitor the usage of network bandwidth resources. The processor 200 may generate a default queue of network bandwidth resources and manage the non-reserved non-real-time network flows as default queues. The processor 200 may adjust the network bandwidth of the default queue when an end or change of a process that reserves network bandwidth resources occurs.

The memory 300 may store an operation result processed by the processor 200, a reservation request received by the receiver 100, and the like. In addition, the memory 300 may store information about resources managed by the processor 200. [

Fig. 2 shows an example of the operation of the resource management apparatus shown in Fig.

Referring to FIG. 2, the resource management apparatus 10 can simultaneously manage a network bandwidth resource and a CPU bandwidth resource. The resource management apparatus 10 may be implemented using a daemon process.

In the case of the conventional tc network bandwidth resource management method, since there is no network bandwidth restriction for the basic task group, it may be vulnerable to interference caused by an unmanaged network bandwidth resource use factor such as excessive network bandwidth resource usage of the basic task group.

The resource management apparatus 10 can prevent network resource interference by setting a network bandwidth limitation for the basic task group. In addition, the resource management device 10 can set a limitation of the CPU bandwidth resource so that the use of the CPU bandwidth resource of the task group that reserved the network bandwidth resource due to the use of the CPU bandwidth resource of the task that does not use the network bandwidth resource is prevented Can be prevented.

The resource management apparatus 10 may include a command line interface (CLI) 210, a network resource monitor 220, and a resource management module 230.

The resource management module 230 includes a central module 231, a network schedulability analyzer 233, a required CPU bandwidth calculation module 235, and a CPU bandwidth schedulability analyzer bandwidth schedulability anyalyer, 237).

The resource management apparatus 10 can guarantee the network I / O bandwidth by using the CLI 210, the network resource monitor 220, and the resource management module 230. [

The user interface provided by the CLI 210 can be represented as shown in Table 1.

index
(index) program
(program) Command
(command) Argument
(argument) One ./netbw_demon dev [dev_name] link [link_speed] netpercpu [core_bandwidth] 2 ./netbw_cli add_d pid [pid] core [core_no] 3 ./netbw_cli add, change pid [pid], core [core_no], lower [net_bandwidth] delete pid [pid], core [core_no] 4 ./netbw_cli quit

Here, ./netbw_demon may mean an API for the system administrator to set the maximum link speed of the initial system and the CPU resource requirement (netpercpu) per unit network time.

An API corresponding to indexes 2, 3, and 4 may refer to an API for the user to add network bandwidth for each connection.

When the user executes the resource management device 10, the receiver 100 receives an input for information of the non-real-time network flow process through the CLI 210, and registers the network bandwidth of the real-time network flow and the CPU bandwidth of the process , ≪ / RTI > modification, and deletion. When the use is terminated, the user can terminate the resource management apparatus 10.

The resource management apparatus 10 can register a device to manage a resource, manage a CPU resource requirement of a process corresponding to a real-time flow, and guarantee a network bandwidth.

Also, the resource management apparatus 10 can manage resources by controlling a cgroup (control group) 240, a tc (traffic controller) 250, a task group 260, and a queue disk (qdisc) .

The cgroup 240 may refer to a framework for allocating resources such as CPU time, memory 300, and network to each task group. The cgroup 240 can manage system resources for each task group through a sub-interface called a subsystem.

The resource management apparatus 10 can utilize the tc 250 and the cgroup 240 framework to reserve network bandwidth. tc 250 may refer to a traffic shaping tool provided by the kernel. tc 250 may include a queuing discipline, a class, and a filter. .

tc 250 may create a queue disk 270 in the link layer and queue packets arriving at the device for transmission and determine what action to perform according to the settings. In addition, a class having a hierarchical structure may be created in the cue disk 270 to classify the packets according to the filter rule defined by the user.

tc 250 can prevent excessive resource use by designating a bandwidth upper limit for each device and class. The cgroup 240 may include a net_cls subsystem as a subsystem for managing network bandwidth resources. The cgroup 240 may support the operation of the tc 250 by marking the class ID in the packet generated in the specific task group through the net_cls subsystem.

CPU resources of Linux are managed through the cpu subsystem, and can be classified into real-time task groups and non-real-time task groups according to the scheduler used by the tasks in the task group. Linux supports the non-real-time scheduler, the Completely Fair Scheduler (CFS), as the default scheduler, and the CFS scheduler can support the resource-reservation features of the shared-rate and bandwidth-based schemes.

Among these, the resource reservation of the sharing ratio scheme can reserve resource shares among the task groups. The reserved occupancy rate value is managed with a weight value corresponding to the nice value of the CFS, and the scheduling algorithm in Linux can distribute as much CPU resources as the reserved resource rate for the available resources.

In addition, the bandwidth scheme reserves the maximum CPU bandwidth available to each task group, allowing the task scheduler in Linux to limit excessive CPU resource usage of task groups that have reserved CPU bandwidth resources.

The receiver 100 may include a CLI 210. The resource management apparatus 10 may receive a reservation request of the user's network resources through the CLI 210. [ The resource management apparatus 10 can calculate the required CPU bandwidth resource for the requested network I / O bandwidth resource in response to the received reservation request.

The processor 200 may include a network resource monitor 220 and a resource management module 230. The resource management module 230 may perform the schedulability analysis on the reservation-requested network bandwidth resource and the CPU bandwidth resource obtained through the calculation. Also, the resource management module 230 can perform resource reservation through the cgroup 240 for the resource reservation that has passed the schedulability analysis.

The network resource monitor 220 of the resource management device 10 can generate a default queue at an initialization time and manage non-real-time network flows for which bandwidth resources are not reserved through the default queue.

The schedulability analysis for the network bandwidth resource can be performed at the time when the reservation and change instruction of the network bandwidth resource is executed.

과 예약된 네트워크 대역폭 자원의 총합

에 대해 수학식 1을 만족하는지 판단할 수 있다.The resource management module 230 determines the maximum available network bandwidth resource of the system

And the sum of reserved network bandwidth resources

It is possible to determine whether or not the expression (1) is satisfied.

The resource management module 230 may set the network bandwidth resource limit of the default queue as shown in Equation 2 if the network bandwidth resource reservation request passes the schedulability analysis.

The network resource monitor 220 manages a list of processes for reserving network bandwidth resources through the resource management apparatus 10 and periodically searches the process list to monitor the network bandwidth resource usage of the non-real-time network flow.

In addition, since the bandwidth resource usage of the non-real-time network flow does not exceed the limit of the network bandwidth resource of the default queue, the network resource monitor 220 continuously monitors the event, When this happens, you can control the use of network bandwidth resources in non-real-time network flows through adjustment of the bandwidth resources of the default queue.

The resource management module 230 registers the core affinity of the reserved real-time flow process and can guarantee CPU resources for each core based on the core affinity information. Core affinity can be classified into a default core (default_core) that can operate on a CPU and a per_core group that is specified to operate on a specific CPU.

The initial CPU bandwidth resource cap for the classified task group may be 1 for a percore group and may be set to n (total number of cores in the system) for a default core.

을 나타낼 수 있다.Each group consists of a net subgroup in which real-time flows operate and a default subgroup in which processes except real-time flows operate. At this time, the CPU bandwidth of the default sub group can be set to the resource maximum value (CPU bandwidth of the parent group). The setting is the maximum available CPU bandwidth resource

Lt; / RTI >

The CPU bandwidth schedulability analyzer 237 can analyze the schedulability of CPU bandwidth resources when the real-time flow process reserves network bandwidth resources. At this time, the schedulability analysis can be limited to satisfy the equation (3) of the real time network flow so as not to exceed the maximum available CPU bandwidth resource.

The CPU bandwidth resource reservation requests that have passed the schedulability analysis can be applied to the net subgroups of the task groups of each CPU of the cgroup 240. In addition, the CPU bandwidth schedulability analyzer 237 can set the limitation of the CPU bandwidth resource as shown in Equation (4).

The processes of the default sub group may be limited in resources usage over bandwidth depending on the bandwidth resource management method of the cgroup 240. With this restriction, the resource management device 10 reserves resources within the maximum available bandwidth resource and can guarantee smooth provision of reserved CPU resources.

3 shows an example of the result of the resource protection verification test for the network load.

Referring to FIG. 3, the performance of the resource management apparatus 10 can be verified through experiments. The experimental environment assumes a situation where a plurality of clients try to upload to one server at the same time, and resource protection can be confirmed by measuring network bandwidth of a real-time flow that reserves a network bandwidth resource.

The composition of the detailed experiment is shown in Table 2.

Experiment Experiment 1 Increase Network Load Experiment 2 Increased CPU load component Server, Client, Switch Server Specifications OS: Ubuntu 16.04, Kernel v4.4.0 CPU: Intel XEON HASWELL 2-way Deca-core 3.1Ghz Client specification OS: CentOS 7.4, Kernel v3.10.0 CPU: Intel XEON HASWELL 2-way Deca-core 3.1Ghz Switch Specifications Mellanox SX1012 12-port 40 / 56GibE Switch Experimental measurement Micro-benchmark: iperfMeasuring factors: The sum of the network bandwidth usage of the real-time flows that reserve resources (10G / 5G / 2.5G / 2.5G), respectively. How to manage resources baseline: Default Linux environment without reserved bandwidth resources.w / tc: An environment with Traffic controller's network bandwidth resource management.
w / guaranteed net: An environment with Network Bandwidth Guarantee Technique.
w / guarantted net, cpu: An environment with network bandwidth guarantee method considering CPU and network load. Background process iperf benchmarks running on anycore (to minimize cpu load increase) iperf benchmarks running on the same core (to maximize cpu load increase) Experiment Detail Network bandwidth resource reservation: 20 Gbit / sec
Manipulation variable: Increase background workload by one (0 to 9).

Experimental results show that the network I / O bandwidth usage in real - time flow is measured by network and CPU load in the same hardware environment, and the performance of four different resource management methods can be compared.

The resource management method includes: 1) baseline of Linux; 2) reservation of network resources (w / tc) through tc (Traffic Controller); 3) network I / O bandwidth guarantee technique of resource management device 10 (W / guaranteed net), 4) guarantee of network I / O bandwidth (w / guaranteed net, cpu) considering CPU resources and network resources of the resource management apparatus 10.

Experimental environment configuration can be used iperf, a micro benchmark tool, and the result of multiple clients performing unidirectional communication with one server can be measured.

The first experiment can be an experiment to confirm network bandwidth guarantee of real-time flow as the load of network bandwidth increases gradually. Three graphs, excluding the baseline, are used to reserve bandwidths of 10 Gbps, 5 Gbps, 2.5 Gbps, and 2.5 Gbps, respectively, for clients in four real-time flows, and the sum of the network bandwidth used by four clients using the iperf benchmark It may be measured.

To minimize the impact of CPU resources on background client growth, all clients can be managed in the default_core group. The experiment can be performed by measuring the network bandwidth of the client that reserved the network bandwidth by incrementing the background client from 0 to 10 one by one, measuring the average value for the measurement result of 6 minutes each, and the respective average values can be summed .

In the example of FIG. 3, the x-axis of the graph represents the number of background clients and the y-axis of the graph may represent the sum of the network bandwidth of the iperf benchmark. In the case of a baseline graph without bandwidth resource reservation, it can be seen that the bandwidth decreases in inverse proportion as the number of background clients increases.

In the w / tc graph, as the network load increases, it is possible to confirm that the bandwidth is not as high as the reserved resource, although the bandwidth is used in comparison with the baseline. In addition, the resource guarantee for the reserved network bandwidth can be confirmed by using the resource management device 10 through the w / guaranteed net graph and the w / guaranteed net and cpu graphs.

If you use only tc in the graph to manage network resources, you can see a performance decrease of up to 42.1%. In addition, this performance degradation may show larger differences as the network bandwidth load increases. Experimental results show that the network I / O bandwidth management tool is able to confirm resource protection against network load.

FIG. 4 shows an example of the results of the resource protection verification test for the CPU load.

Referring to FIG. 4, in order to maximize the CPU resource competition between a client reserving a bandwidth resource and a background client in comparison with an existing experiment in which all clients are managed in the default core group, Core. ≪ / RTI >

We also confirmed that 36.2% of the CPU resources are used to guarantee 20Gbps of bandwidth in the experimental environment by using an experimental method for CPU resource management. In order to guarantee maximum 40% CPU resources in the experimental environment, CPU resources Can be reserved. The other configuration of the environment may be the same as the experiment of Fig. The example of FIG. 4 can show the results of the resource protection verification test for the CPU load.

In FIG. 4, the baseline graph and the w / tc graph show a trend similar to that of FIG. Also, it can be seen that the performance of the w / guaranteed net using the resource management device 10 is greatly reduced to 62.5% as compared with other experimental results. On the other hand, the experimental results of w / guaranteed net and cpu using the resource management device 10 confirm that the bandwidth of the reserved network is ensured.

Fig. 5 shows an example of bandwidth-specific bandwidth configuration of the load with the largest load.

5, the example of FIG. 5 is a result of the experiment results of an environment in which 10 background clients that operate with the greatest CPU load are operated in a flow-by-flow basis. For the three resource management methods except for the base line, It can indicate bandwidth usage.

In the 10Gbps flow of FIG. 5, the largest difference is seen in comparison with the ideal case, and it can be confirmed that the client that reserves a large network bandwidth is greatly affected by the CPU load.

Also, it can be confirmed that the network bandwidth of each flow is guaranteed through the measurement result of w / guaranteed net and cpu graph. Therefore, in order to guarantee the correct network I / O bandwidth, it is necessary to consider not only the guarantee on the network bandwidth but also the guarantee of the CPU resource of the process of performing the network operation.

FIG. 6 shows a flowchart of the operation of the resource management apparatus shown in FIG. 1. FIG.

Referring to FIG. 6, the resource management apparatus 10 may receive a reservation request for a network bandwidth resource (610).

The resource management apparatus 10 may determine the schedulability of the network bandwidth resource based on the received reservation request (630). Specifically, the resource management apparatus 10 can calculate the sum of the reserved network bandwidth resources and determine the schedulability by comparing the sum of the maximum available network bandwidth resource and the reserved network bandwidth resource.

The resource management device 10 may reserve network bandwidth resources based on schedulability for network bandwidth resources (650).

The resource management device 10 may determine 670 schedulability for CPU bandwidth resources based on the reserved network bandwidth resources. Specifically, the resource management device 10 can calculate the required CPU bandwidth per unit network bandwidth, and compare the maximum available CPU bandwidth resource with the product of the sum of the reserved network bandwidth resources and the required CPU bandwidth per unit network bandwidth, Can be determined.

The resource management device 10 may reserve 690 CPU bandwidth resources based on schedulability for CPU bandwidth resources.

Through the above operation, the resource management apparatus 10 can prevent interference with network and CPU bandwidth resources from other tasks in order to guarantee the network I / O bandwidth unlike the conventional resource management method. Therefore, the resource management apparatus 10 can stably provide the network I / O service to the reserved resource even in the change of the system resource situation.

The method according to an embodiment may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer readable medium may include program instructions, data files, data structures, and the like, alone or in combination. Program instructions to be recorded on the medium may be those specially designed and constructed for the embodiments or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Magneto-optical media, and hardware devices specifically configured to store and execute program instructions such as ROM, RAM, flash memory, and the like. Examples of program instructions include machine language code such as those produced by a compiler, as well as high-level language code that can be executed by a computer using an interpreter or the like. The hardware device may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

The software may include a computer program, code, instructions, or a combination of one or more of the foregoing, and may be configured to configure the processing device to operate as desired or to process it collectively or collectively Device can be commanded. The software and / or data may be in the form of any type of machine, component, physical device, virtual equipment, computer storage media, or device , Or may be permanently or temporarily embodied in a transmitted signal wave. The software may be distributed over a networked computer system and stored or executed in a distributed manner. The software and data may be stored on one or more computer readable recording media.

Although the embodiments have been described with reference to the drawings, various technical modifications and variations may be applied to those skilled in the art. For example, it is to be understood that the techniques described may be performed in a different order than the described methods, and / or that components of the described systems, structures, devices, circuits, Lt; / RTI > or equivalents, even if it is replaced or replaced.

Therefore, other implementations, other embodiments, and equivalents to the claims are also within the scope of the following claims.

Claims (16)

Receiving a reservation request for a network bandwidth resource;
Determining schedulability for the network bandwidth resource based on the reservation request;
Reserving the network bandwidth resource based on schedulability for the network bandwidth resource;
Determining schedulability for a central processing unit (CPU) bandwidth resource based on a reserved network bandwidth resource; And
Reserving the CPU bandwidth resource based on schedulability for the CPU bandwidth resource
The resource management method comprising:
The method according to claim 1,
Wherein determining the schedulability for the network bandwidth resource comprises:
Calculating a sum of reserved network bandwidth resources; And
Comparing the maximum available network bandwidth resource with the sum of the reserved network bandwidth resources to determine schedulability
The resource management method comprising:
3. The method of claim 2,
After the step of reserving the network bandwidth resource,
Setting a limit of the network bandwidth resource to a value obtained by subtracting the sum of the reserved network bandwidth resources from the maximum available network bandwidth resource
Further comprising the steps of:
3. The method of claim 2,
Wherein determining the schedulability for the CPU bandwidth resource comprises:
Calculating a required CPU bandwidth resource per unit network bandwidth; And
Determining the schedulability by comparing the sum of the sum of the required CPU bandwidth resources per unit network bandwidth and the maximum available CPU bandwidth resource
The resource management method comprising:
5. The method of claim 4,
After the step of reserving the CPU bandwidth resource based on schedulability for the CPU bandwidth resource,
Setting a limit of the CPU bandwidth resource to a value obtained by subtracting the product from the maximum available CPU bandwidth resource
Further comprising the steps of:
The method according to claim 1,
Monitoring the usage of the network bandwidth resource
Further comprising the steps of:
The method according to claim 6,
Wherein the monitoring comprises:
Generating a default queue for the network bandwidth resource; And
Managing non-real-time network flows that are not reserved for bandwidth with the default queue
The resource management method comprising:
8. The method of claim 7,
Wherein the monitoring comprises:
Adjusting a network bandwidth of the default queue when an end or a change of a process for reserving the network bandwidth resource occurs,
Further comprising the steps of:
A receiver for receiving a reservation request for a network bandwidth resource; And
Determining a schedulability for a network bandwidth resource based on the reservation request, reserving the network bandwidth resource based on schedulability for the network bandwidth resource, and determining a central processing unit (CPU) based on the reserved network bandwidth resource, A processor for determining the schedulability for a bandwidth resource and reserving the CPU bandwidth resource based on schedulability for the CPU bandwidth resource,
The resource management apparatus comprising:
10. The method of claim 9,
The processor comprising:
Calculating a sum of the reserved network bandwidth resources, and comparing the sum of the maximum available network bandwidth resource and the reserved network bandwidth resource to determine schedulability
Resource management device.
11. The method of claim 10,
The processor comprising:
After reserving the network bandwidth resource,
Setting a limit of the network bandwidth resource to a value obtained by subtracting the sum of the reserved network bandwidth resources from the maximum available network bandwidth resource
Resource management device.
11. The method of claim 10,
The processor comprising:
Calculates a necessary CPU bandwidth resource per unit network bandwidth, and determines schedulability by comparing the sum of the sum of necessary CPU bandwidth resources per unit network bandwidth and the maximum available CPU bandwidth resource
Resource management device.
13. The method of claim 12,
The processor comprising:
After reserving the CPU bandwidth resources based on schedulability for the CPU bandwidth resources,
Setting a limit of the CPU bandwidth resource to a value obtained by subtracting the product from the maximum available CPU bandwidth resource
Resource management device.
10. The method of claim 9,
The processor comprising:
Monitoring the usage of the network bandwidth resource
Resource management device.
15. The method of claim 14,
The processor comprising:
Generates a default queue of the network bandwidth resource, and manages a non-reserved non-real-time network flow with the bandwidth as the default queue
Resource management device.
16. The method of claim 15,
The processor comprising:
The network bandwidth of the default queue is adjusted when an end or a change of the process of reserving the network bandwidth resource occurs
Resource management device.

Images