KR102400158B1 - Dynamic Resource Allocation Method and Apparatus for Service Chaining in Cloud-Edge-Radio 5G Network - Google Patents

Abstract

Provided are a method and an apparatus for allocating a dynamic resource for service chaining in a hierarchical 5G network structure. The method for allocating a dynamic resource for service chaining in a hierarchical 5G network structure, which is implemented through a computer device according to an embodiment, may include: processing a function required for a service in at least one of a cloud server and an edge server according to a service request of a user through a user terminal; and transmitting the service processed in at least one of the cloud server and the edge server to the user through at least one of a backhaul, a fronthaul link, and a wireless access network connecting the cloud server, the edge server, and the user terminal. According to the embodiment, it is possible to increase service satisfaction of the user.

Description

Dynamic Resource Allocation Method and Apparatus for Service Chaining in Cloud-Edge-Radio 5G Network}

The following embodiments relate to a method and apparatus for dynamic resource allocation for service chaining in a hierarchical 5G network structure, and more particularly, to a service requested by a user in a hierarchical 5G network structure, and to a user with a small delay A dynamic resource allocation method and apparatus for selecting a service path to provide a high quality service.

With the introduction of virtualization technology, services that cluster networking resources, processing resources, and storage resources and use them together have emerged. Software-defined networking (SDN) and network functions virtualization (NFV) technologies are virtualization technologies, and by using these technologies, networks such as deep packet inspection (DPI), intrusion detection system (IDS), firewall Functional functions can now be virtualized through virtual machines or container technologies in general servers.

Before the introduction of virtualization technology, each of these network function functions was implemented based on hardware, so in the case of a service that requires the corresponding network function, additional hardware supporting the function must be purchased and passed through the hardware. Virtualization technology enables the use of various virtual functions, such as not only network function functions, but also virus scanning and computing function functions corresponding to data processing, and joint management of network, process, and storage resources.

As a result, infrastructure administrators can add network function functions or compute function functions required for a service to a general server at no cost, if necessary. Thanks to this, network administrators can provide various services by efficiently using network, process, and storage resources through flexible resource management. For example, when providing a video streaming service, high-quality services are provided to users by jointly optimizing network and processing resources such as transcoding and DPI.

Korea Patent No. 10-2156439 relates to such a cloud-edge system and a data processing method thereof, and describes a technology for a system and method capable of processing data through an analysis model of a pipeline structure in the cloud-edge, and there is.

Korean Patent No. 10-2156439

The embodiments describe a method and apparatus for dynamic resource allocation for service chaining in a hierarchical 5G network structure, and more specifically, cloud-backhaul, edge, fronthaul, It provides technology to manage dynamic resources of all layers of the network up to the wireless access network.

In a situation where the network and service environment change in real time, the embodiments provide a dynamic network and processing resource allocation in consideration of all layers (cloud, edge, wireless access network), thereby maximizing service satisfaction of all users and lowering service delay. An object of the present invention is to provide a dynamic resource allocation method and apparatus for service chaining in a hierarchical 5G network structure.

A dynamic resource allocation method for service chaining in a hierarchical 5G network structure implemented through a computer device according to an embodiment, according to a service request of a user through a user terminal, in at least one of a cloud server and an edge server processing a function required for the service; and the service processed by at least one of the cloud server and the edge server through at least one of a backhaul, a fronthaul link, and a wireless access network connecting the cloud server, the edge server, and the user terminal. It can be made including the step of delivering to.

The processing of the function required for the service may include: a service chaining step of determining which of the cloud server and the edge server to process each network function or computing function from the cloud server to the service path; and a scheduling step of controlling process or network scheduling by determining the use of process resources or network resources in the cloud server.

The processing of the function required for the service may include: a scheduling step of an edge server in which each of the edge servers is operated according to a decision of the cloud server, and controlling process scheduling in each of the edge servers; and a user scheduling step of making the radio access network into a beam activation state when the base station provides a service to the user in the edge server, and determining which user to provide the service for which the processing has been completed.

In a situation where networks and service environments change in real time, network and processing resources may be dynamically allocated in consideration of all layers of the cloud server, the edge server, and the wireless access network.

Dynamic resource allocation apparatus for service chaining in a hierarchical 5G network structure according to another embodiment, according to a service request of a user through a user terminal, each network function or computing function as a service path to any one of the cloud server and the edge server a cloud server that determines whether to process in the server, and determines the use of process resources or network resources to control process or network scheduling; And it operates according to the decision of the cloud server, controls the process scheduling of the edge server, sets the radio access network to a beam activation state when the base station provides a service to the user, and provides a service that has been processed to a certain user It can be made by including the edge server to determine.

According to embodiments, in a situation where the network and service environment change in real time, by allocating network and processing resources dynamically in consideration of all layers (cloud, edge, wireless access network), it is possible to reduce service delay while maximizing service satisfaction of all users. It is possible to provide a dynamic resource allocation method and apparatus for service chaining in a hierarchical 5G network structure.

According to embodiments, an end-to-end service delay occurring in a process for a user to receive a service may be reduced, and a user's service satisfaction may be increased.

1 is a diagram for explaining a hierarchical 5G network structure according to an embodiment.
2A is a flowchart illustrating a dynamic resource allocation method for service chaining according to an embodiment.
2B is a flowchart illustrating a method of processing a function of a dynamic resource allocation method for service chaining according to an embodiment.
3 is a diagram for explaining a queue and control model of a hierarchical 5G network structure according to an embodiment.
4 is a diagram for explaining a mathematical principle of algorithm derivation for solving a defined problem according to an embodiment.
5 is a diagram illustrating a comparison between a proposed algorithm and a modified algorithm in a uniform user distribution according to an embodiment.
6 is a diagram illustrating a comparison between a proposed algorithm and a modified algorithm in an unequal user distribution according to an embodiment.

Hereinafter, embodiments will be described with reference to the accompanying drawings. However, the described embodiments may be modified in various other forms, and the scope of the present invention is not limited by the embodiments described below. In addition, several embodiments are provided in order to more completely explain the present invention to those of ordinary skill in the art. The shapes and sizes of elements in the drawings may be exaggerated for clearer description.

Recently, technology that provides services including network and computing function functions in a hierarchical network structure that covers all layers from cloud servers, edge servers, and end devices is attracting attention.

The present embodiments consider a hierarchical 5G mobile communication network network structure in which a single cloud server, multiple edge servers, base stations, and even a wireless access network can be considered at once. The hierarchical 5G network structure consists of a single cloud, an edge in charge of a certain area, a backhaul link at the wire end connecting the cloud and the edge, a base station in each area, a fronthaul link at the wire end connecting the edge and the base station, and the mobile in each area. It consists of a user and a wireless access network between each base station and a mobile user. Cloud and edge servers have processing resources (CPU, GPU) that can process services, and network resources (available bandwidth of wired/wireless ends) exist in backhaul/fronthaul/wireless access network links. In this structure, when a mobile user requests a service from a mobile communication network, the functions required for the service are processed in the cloud and edge, and the service is delivered through the backhaul/fronthaul link and wireless access network.

In such a hierarchical 5G network structure, the cloud exists at the top layer, and it is effective in integrating and optimizing the entire system through a large amount of processing resources. However, it may be inefficient to manage resources while receiving feedback from detailed characteristics of each region in the cloud. In this case, it is easy to fine-tune resource control in each region by utilizing the edge resources in the hierarchical network structure. However, since the edge has a weakness of having fewer processing resources than the cloud, high service quality can be provided to users only by efficiently managing the processing resources of the cloud and the edge at the same time. In addition, even if the service requested by the user is processed with abundant processing resources in the cloud or edge, if the network resources of the backhaul/fronthaul/wireless access network link are not efficiently managed, the user will be provided with low-quality services. Therefore, it is necessary to efficiently use network resources and processing resources at the same time among the possible service paths and to achieve high service satisfaction at the same time.

In the previous study, only the service path problem in the cloud-edge hierarchical structure was dealt with, but the service path problem in the cloud-edge-wireless access network hierarchical structure was not dealt with. However, in the current network situation where various services must be provided quickly, it is natural for service processing and resource management to occur at the same time.

This embodiment proposes a technology for processing a service requested by a user in a hierarchical 5G network structure and selecting a service path to provide a high-quality service to the user with a small delay.

The embodiments can reduce service delay while maximizing service satisfaction for all users by allocating dynamic network and processing resources in consideration of all layers (cloud, edge, wireless access network) in a situation where the network and service environment change in real time. . This technique is based on (1) service chaining (determining whether to process each network/compute function in the cloud/edge as a service path), (2) process/network scheduling (use of process/network resources) in the cloud at every moment to support the service. at each edge, (1) process scheduling, (2) beam activation of the radio access network (activated when the base station provides a service to users) and user scheduling (to which user the service that has been processed is to be provided) decide).

1 is a diagram for explaining a hierarchical 5G network structure according to an embodiment.

As shown in Fig. 1, with a small delay in the hierarchical 5G network environment consisting of the cloud (cloud server, 110), edge (edge server, 130), and wireless communication networks (120, 140, 160), high quality to users You can select a service path and allocate dynamic resources to provide the services of Here, the cloud 110 may include a computing function 111 , a network function 112 , and a process resource 1130 .

A plurality of edges 130 are connected to one cloud 110 by a backhaul link 120 , and several base stations 150 connected to each edge 130 by a fronthaul link 140 exist. The user terminal 170 in the area controlled by each edge 130 receives a service through the base station 150 wireless access network 160 in the corresponding edge 130 .

At this time, the radio access network 160 is a situation in which CoMP is considered, and when the user 170 is located at a distance that can be affected by the base station 150, the service can be received from several base stations 150, When the user 170 receives a service, a plurality of beams may be allocated within one base station 150 .

The user service 180 is defined as a continuous form of a network function function and a computing function function. Each function in the service requested by the user 170 is processed through the processing of the cloud 110 or the edge 130 .

2A is a flowchart illustrating a dynamic resource allocation method for service chaining according to an embodiment.

Referring to FIG. 2A , in a method for dynamic resource allocation for service chaining in a hierarchical 5G network structure implemented through a computer device according to an embodiment, according to a user's service request through a user terminal, one of a cloud server and an edge server Processing a function required for a service in at least one or more servers (S110), and through at least one of a backhaul, a fronthaul link, and a wireless access network connecting the cloud server, the edge server, and the user terminal, the cloud server and the edge A service processed by at least one of the servers may be delivered to the user ( S120 ).

2B is a flowchart illustrating a method of processing a function of a dynamic resource allocation method for service chaining according to an embodiment.

Referring to FIG. 2B , the step S110 is a service chaining step S111 of determining which server of the cloud server and the edge server to process each network function or computing function from the cloud server to the service path, and the process in the cloud server A scheduling step (S112) of controlling a process or network scheduling by determining the use of a resource or network resource may be included.

In addition, each edge server is operated according to the decision of the cloud server, and when the edge server provides a service to the user in the edge server scheduling step (S113), which controls the process scheduling in each edge server, and the edge server provides a service to the user, the wireless access network is beamed. The method may further include a user scheduling step (S114) of determining which user is to be in an activated state and to which a service for which the processing has been completed is to be provided.

According to embodiments, in a situation where networks and service environments change in real time, network and processing resources may be dynamically allocated in consideration of all layers of the cloud server, edge server, and wireless access network.

The dynamic resource allocation method for service chaining according to an embodiment may be described with an example of the dynamic resource allocation apparatus for service chaining according to an embodiment. The dynamic resource allocation apparatus for service chaining according to an embodiment may include a cloud server and an edge server.

Below, each step of the method for dynamic resource allocation for service chaining according to an embodiment will be described in more detail. The dynamic resource allocation method for service chaining according to an embodiment may be performed through a dynamic resource allocation apparatus for service chaining.

In step S110, according to the user's service request through the user terminal, at least one of the cloud server and the edge server may process a function required for the service.

More specifically, in the service chaining step (S111), the cloud server may determine which of the cloud server and the edge server to process each network function or computing function as a service path according to the user's service request through the user terminal. .

In the scheduling step S112 , the cloud server may control process or network scheduling by determining the use of a process resource or a network resource.

In the edge server scheduling step S113, the edge server operates according to the decision of the cloud server, and may control process scheduling of the edge server.

In the user scheduling step ( S114 ), the edge server sets the radio access network to a beam activation state when the base station provides a service to the user, and may determine which user to provide the processed service to.

In step S120, the service processed by at least any one of the cloud server and the edge server through at least any one of a backhaul, a fronthaul link, and a wireless access network connecting the cloud server, the edge server, and the user terminal is performed by the user. can be passed on to

Embodiments provide an algorithm for efficiently allocating dynamic resources for service chaining in a hierarchical 5G network structure consisting of a cloud-edge-radio access network. For this reason, an algorithm is provided that reduces the end-to-end service delay that occurs in the process of receiving the service and increases the user's service satisfaction.

The embodiments propose algorithms that can efficiently use limited processes and network resources as the network structure of the cloud-edge-wireless access network develops, so that all layers (cloud, edge, wireless access network) as well as process and network resources can be efficiently allocated according to the situation.

3 is a diagram for explaining a queue and control model of a hierarchical 5G network structure according to an embodiment.

Referring to FIG. 3 , a queue and control model shown in the hierarchical 5G network structure is shown, and the expressed variables can be summarized as shown in Table 1 below.

[Table 1]

를 넣어준다. 이 때,

와

경로는

에,

경로는

에 트래픽이 쌓인다. 클라우드는

을 결정하고, 이 때 처리된 트래픽은

에 쌓인다.

에 쌓인 트래픽은 클라우드의

결정에 따라서

경로는

에,

와

경로는

에 쌓인다.Variables to be controlled in this embodiment are the same as red variables in Table 1. Looking at the movement of traffic according to the determined variable, the cloud determines the user's service chaining path and queues the traffic on that path.

put in At this time,

Wow

the path is

to,

the path is

traffic accumulates in the cloud

is determined, and the traffic processed at this time is

piled up on

The traffic accumulated in the cloud

according to the decision

the path is

to,

Wow

the path is

piled up on

에 쌓인 트래픽은 엣지의

결정으로 처리되며

에 쌓인다. 마지막으로 엣지별 무선접속망에서 기지국

에서 사용자

에 대해

를 결정해 프로세싱이 완료된 최종 데이터를 사용자에게 전송한다.

The traffic accumulated at the edge

treated as a decision

piled up on Finally, the base station in the wireless access network for each edge

user from

About

to send the final data that has been processed to the user.

The problem definition and algorithm derivation method are described below.

First, the objective function (P1) of the problem can be expressed as the following equation.

[Equation 1]

)(P1):

)

In addition, the constraints (C1), (C2), (C3), (C4), (C5), and (C5) can be expressed as the following equations.

[Equation 2]

(C1):

(C2):

(C3):

(C4):

(C5):

(C6): Mean-rate stability of all queues.

은 유틸리티 함수로, 사용자

의 평균 쓰루풋을 사용해 서비스 만족도의 정도를 나타낼 수 있다. 제약사항 (C1)은 단위시간 t에서의 사용자

의 최대 처리율의 제약, (C2)는 클라우드와 엣지의 프로세싱 자원의 가용 용량 제약사항, (C3)와 (C4)는 각각 백홀과 프론트홀의 네트워크 자원의 가용 용량 제약사항, (C5)는 사용자

의 최소 평균 쓰루풋의 제약, (C6)은 모든 서비스는 유한한 시간 내에 처리되어야 한다는 제약사항이다. 본 문제를 최적화하기 위해 변수 사용자의 서비스 체이닝 옵션별 트래픽 주입량

, 프로세스 스케줄링

, 네트워크 스케줄링

, 빔 활성화 및 사용자 스케줄링

를 제어한다. The purpose is to maximize the service satisfaction of users in the hierarchical 5G network structure composed of cloud-edge-wireless access network. In this case, in the objective function

is a utility function, the user

The average throughput of can be used to indicate the degree of service satisfaction. Constraint (C1) is the user in unit time t

(C2) is the available capacity constraint of cloud and edge processing resources, (C3) and (C4) are the available capacity constraints of the backhaul and fronthaul network resources, respectively, and (C5) is the user

The constraint of the minimum average throughput of (C6) is that all services must be processed within a finite time. To optimize this problem, the amount of traffic injected by service chaining option of variable users

, process scheduling

, network scheduling

, Beam Activation and User Scheduling

to control

에 대한 제약사항을 추가한다. 이후에 보조 변수의 추가와 jensen's inequality를 이용해 사용해 유틸리티 함수의 시간평균을 최대화하는 문제로 변형할 수 있다. 이는 매 단위시간마다 하위 문제를 풀어가는 문제로, 본래의 문제보다 훨씬 다루기 쉽다. 따라서 최종적으로 본래 문제의 최적성을 유지한 채 원래의 장기적 관점 문제를 매 단위시간 마다 문제를 풀어가는 알고리즘을 제공한다.Problem (P1) is the problem of maximizing the utility function of the time average. In this case, since the utility function is a non-linear function, it must be solved for a long time. Therefore, in order to transform the problem so that it can be solved every unit time, first

Add restrictions on Afterwards, it can be transformed into a problem of maximizing the time average of the utility function by adding auxiliary variables and using jensen's inequality. This is a problem in which a subproblem is solved every unit time, and it is much easier to deal with than the original problem. Therefore, we provide an algorithm that solves the original long-term problem at every unit time while maintaining the optimality of the original problem.

4 is a diagram for explaining a mathematical principle of algorithm derivation for solving a defined problem according to an embodiment.

As a result, the algorithm for optimally pooling (P3) 430 is not only summarized in FIG. 4 but also for optimally pooling (P1) 410, so the RACER algorithm to satisfy problem (P3) 430 will be described below. provides

The proposed algorithm is described below.

와

를 추가한다. 각 계층에서의 변수 제어는 다음 식과 같이 나타낼 수 있다.Finally, an algorithm was developed as a problem to solve a sub-problem every unit time, and this sub-problem is divided into problems for each physical layer (cloud, edge, wireless access network) and solved. The algorithm independently controls each layer using only information from the user's queue every hour. Virtual queue for control of auxiliary variables added when deriving the algorithm

Wow

add Variable control in each layer can be expressed as the following equation.

의 제어를 다음 식과 같이 표현할 수 있다. auxiliary variable

can be expressed as the following equation.

[Equation 3]

의 안정화에 따라 위의 식을 통해 제어한다. Here, the increment of the utility and the virtual queue

is controlled through the above equation according to the stabilization of

The control of the traffic injection amount for each service chaining option of the user can be expressed as the following equation.

[Equation 4]

의 가능한 세가지 경로 중, 첫 번째 큐에 해당하는 큐의

값이 가장 적은 큐에 트래픽을 주입한다. 큐 길이가 가상 큐의 한계점보다 작을 때 트래픽을 주입할 수 있다. Here, the user

of the three possible paths of the queue corresponding to the first

Inject traffic to the queue with the lowest value. Traffic can be injected when the queue length is less than the threshold of the virtual queue.

을 다음 식과 같이 표현할 수 있다. Process Scheduling in the Cloud

can be expressed as the following formula.

[Equation 5]

의 값이 음수인 값 중 가장 작은 값부터 프로세싱 자원을 한 개씩 할당 받아서 처리하게 된다. 이 때, 스케줄링될 수 있는 경로는 해당 단위시간의 프로세싱 자원 내에서만 처리될 수 있다. 또한, 백홀의 네트워킹 큐와 클라우드의 프로세싱 큐 길이의 차가 같을 경우에,

가 작을수록 프로세싱 자원을 먼저 할당 받게 된다.here,

Processing resources are allocated and processed one by one starting from the smallest value among the negative values of . In this case, the scheduled path may be processed only within the processing resource of the corresponding unit time. In addition, when the difference between the length of the networking queue of the backhaul and the processing queue of the cloud is the same,

As is smaller, processing resources are allocated first.

을 다음 식과 같이 표현할 수 있다. Network Scheduling in Backhaul

can be expressed as the following formula.

[Equation 6]

의 값이 음수인 값 중 가장 작은 값부터 네트워크 자원을 한 개씩 할당 받아 처리하게 된다. 이 때, 스케줄링될 수 있는 경로는 해당 단위시간의 네트워크 자원 내에서만 처리될 수 있다. 또한,

의 계산하는 분자의 값이 같을 경우에,

가 작을수록 네트워크 자원을 먼저 할당 받게 된다.here,

Network resources are allocated and processed one by one, starting with the smallest value among negative values of . In this case, the scheduled path may be processed only within the network resource of the corresponding unit time. In addition,

If the values of the numerators to be calculated of are the same,

The smaller is, the more network resources are allocated first.

을 다음 식과 같이 표현할 수 있다. Process Scheduling at the Edge

can be expressed as the following formula.

[Equation 7]

의 값이 음수인 값 중 가장 작은 값부터 프로세싱 자원을 한 개씩 할당 받아서 처리하게 된다. 이 때, 스케줄링될 수 있는 경로는 해당 단위시간의 프로세싱 자원 내에서만 처리될 수 있다. 또한, 프론트홀 링크/무선접속망의 네트워킹 큐와 엣지의 프로세싱 큐 길이의 차가 같을 경우에,

가 작을수록 프로세싱 자원을 먼저 할당 받게 된다.here,

Processing resources are allocated and processed one by one starting from the smallest value among the negative values of . In this case, the scheduled path may be processed only within the processing resource of the corresponding unit time. In addition, when the difference between the networking queue of the fronthaul link/wireless access network and the processing queue length of the edge is the same,

As is smaller, processing resources are allocated first.

을 다음 식과 같이 표현할 수 있다. Beam activation and user scheduling in wireless access network

can be expressed as the following formula.

[Equation 8]

To solve this problem, we propose a greedy-based algorithm. The algorithm is as follows.

[Table 2]

In this embodiment, simulation is performed using MATLAB to verify the performance of the proposed algorithm. (a) When there is the same user at all edges, (b) When users are distributed differently, the simulation is performed in two situations.

In order to check how important the control of each layer is, only one layer among all the layers was modified to operate in a different way than the proposed algorithm and compared with the proposed algorithm.

The modified algorithm maintains the behavior of other layers as the proposed algorithm, (1) RACER-SO: preset to have only one possible service path for users, (2) RACER-CRR: Process scheduling in the cloud is in a round-robin format (3) RACER-ERR: Edge process scheduling operates in round-robin format, (4) RACER-BRR: Backhaul network scheduling operates in round-robin format, (5) RACER-RRS: Radio access network beam Activation and user scheduling operate randomly, (6) RACER-RNCP: operates in the same way as the proposed algorithm in a situation where CoMP is not considered in the radio access network, (7) RACER-RLQ: the user's network queue length in the radio access network is long Operation so that the user is scheduled, (8) RACER-RDR: Operation so that the user with a large user signal in the radio access network is scheduled. There are a total of 8 types. Here, the algorithm proposed in this embodiment is RACER.

5 and 6 show the average throughput for each average queue length of the proposed algorithm and the modified algorithm. More specifically, FIG. 5 is a diagram illustrating a comparison between a proposed algorithm and a modified algorithm in a uniform user distribution according to an embodiment. Also, FIG. 6 is a diagram illustrating a comparison between a proposed algorithm and a modified algorithm in an unequal user distribution according to an embodiment.

Referring to FIG. 5 , a result is shown when the same number of users are randomly distributed at all edges, and FIG. 6 is shown when a different number of users are randomly distributed to each edge.

Two analyzes can be made through the following results. (1) The proposed algorithm can greatly increase user throughput by allowing a small amount of delay (average queue length). (2) The proposed algorithm can handle more traffic than static algorithms under the same average queue length. This performance difference is more pronounced when the users are unevenly distributed than when the users are evenly distributed. This is because the proposed algorithm can adaptively allocate resources to the changing environment.

Through the simulation, it was confirmed that if any one layer does not operate efficiently in the algorithm proposed in this embodiment, the performance does not come out well. This means that the proposed algorithm efficiently integrates and controls the dynamic resource allocation of the cloud-edge-wireless access network.

As described above, embodiments may provide a dynamic resource management technology in a 5G network infrastructure consisting of a cloud-edge-wireless access network. In particular, it is possible to provide a dynamic resource management method capable of reducing user service delay and increasing service satisfaction, and providing a method for managing network and processing resources together. In addition, it is possible to provide a resource management method supporting both the network function function and the computing function function.

The device described above may be implemented as a hardware component, a software component, and/or a combination of the hardware component and the software component. For example, devices and components described in the embodiments may include, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA), It may be implemented using one or more general purpose or special purpose computers, such as a programmable logic unit (PLU), microprocessor, or any other device capable of executing and responding to instructions. The processing device may execute an operating system (OS) and one or more software applications running on the operating system. A processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For convenience of understanding, although one processing device is sometimes described as being used, one of ordinary skill in the art will recognize that the processing device includes a plurality of processing elements and/or a plurality of types of processing elements. It can be seen that can include For example, the processing device may include a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as parallel processors.

Software may comprise a computer program, code, instructions, or a combination of one or more thereof, which configures a processing device to operate as desired or is independently or collectively processed You can command the device. The software and/or data may be any kind of machine, component, physical device, virtual equipment, computer storage medium or apparatus, to be interpreted by or to provide instructions or data to the processing device. may be embodied in The software may be distributed over networked computer systems and stored or executed in a distributed manner. Software and data may be stored in one or more computer-readable recording media.

The method according to the embodiment may be implemented in the form of program instructions 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, etc. alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the embodiment, or may be known and available to those skilled in the art of computer software. Examples of the computer-readable recording medium include magnetic media such as hard disks, floppy disks and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic such as floppy disks. - includes magneto-optical media, and hardware devices specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those generated by a compiler, but also high-level language codes that can be executed by a computer using an interpreter or the like.

As described above, although the embodiments have been described with reference to the limited embodiments and drawings, various modifications and variations are possible from the above description by those skilled in the art. For example, the described techniques are performed in an order different from the described method, and/or the described components of the system, structure, apparatus, circuit, etc. are combined or combined in a different form than the described method, or other components Or substituted or substituted by equivalents may achieve an appropriate result.

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

Claims (5)

A dynamic resource allocation method for service chaining in a hierarchical 5G network structure implemented through a computer device, the method comprising:
processing a function required for the service in at least one of a cloud server and an edge server according to a user's service request through a user terminal; and
The service processed by at least one of the cloud server and the edge server is provided to the user through at least one of a backhaul, a fronthaul link, and a wireless access network connecting the cloud server, the edge server, and the user terminal. stage to be passed
including,
The step of processing the function required for the service is,
a service chaining step of determining which of the cloud server and the edge server to process each network function or computing function from the cloud server to the service path;
a scheduling step of controlling process or network scheduling by determining the use of a process resource or a network resource in the cloud server;
a scheduling step of an edge server in which each of the edge servers is operated according to the determination of the cloud server, and controlling process scheduling in each of the edge servers; and
In the edge server, when the base station provides the service to the user, the wireless access network is in the beam activation state, and the user scheduling step of determining which user to provide the service for which the processing has been completed
including,
The cloud server determines process scheduling, processed traffic is accumulated in a networking queue of the backhaul path from the cloud server to the edge server, and the accumulated traffic is determined by the network scheduling of the cloud server according to the network scheduling decision of the cloud server. The path through which the service is processed in the cloud server is accumulated in the networking queue of the wireless access network and the fronthaul link from the edge server to the user. The path through which the service is processed is accumulated in the processing queue of the path of the edge server, and the traffic accumulated in the processing queue of the path of the edge server is processed by the process scheduling decision of the edge server, and is connected from the edge server to the user. accumulating in the networking queue of the fronthaul link and the radio access network
characterized in that, a dynamic resource allocation method. delete delete According to claim 1,
Allocating network and processing resources dynamically in consideration of all layers of the cloud server, the edge server, and the wireless access network in a situation where the network and service environment change in real time
characterized in that, a dynamic resource allocation method. In the device for dynamic resource allocation for service chaining in a hierarchical 5G network structure,
According to the user's service request through the user terminal, each network function or computing function is processed by the cloud server or the edge server in the service path, and the process or network scheduling is performed by determining the use of the process resource or network resource. Cloud server to control; and
It operates according to the decision of the cloud server, controls the process scheduling of the edge server, sets the radio access network to the beam activation state when the base station provides the service to the user, and decides which user to provide the service for which the processing has been completed edge server
including,
The cloud server,
After service chaining that determines which server of the cloud server and edge server to process each network function or computing function from the cloud server to the service path, the cloud server determines the use of a process resource or network resource to schedule a process or network to control,
The edge server is
Each of the edge servers is operated according to the decision of the cloud server, and each edge server controls process scheduling, and when the edge server provides a service to the user from the base station, the wireless access network is in a beam activation state, and processing decides which user to give the completed service to,
The cloud server determines process scheduling, processed traffic is accumulated in a networking queue of a backhaul path from the cloud server to the edge server, and the accumulated traffic is determined by the network scheduling of the cloud server according to the network scheduling decision of the cloud server. The path where the service is processed in the server is accumulated in the fronthaul link from the edge server to the user and the networking queue of the wireless access network, the path where the service is processed in the edge server, and the service is provided in both the cloud server and the edge server The processed path is accumulated in the processing queue of the path of the edge server, and the traffic accumulated in the processing queue of the path of the edge server is processed by the edge server's process scheduling decision, and the fronthaul connected from the edge server to the user accumulating in the networking queue of links and the radio access network
characterized in that, a dynamic resource allocation device.

Images