KR20190087131A - Method, program, recording medium and computing device for allocating virtualization network function resource - Google Patents

Abstract

The present invention relates to a method and a program for allocating virtual network function (VNF) resources, a computer-readable recording medium recording the same, and a computing device. According to the present invention, the computing device comprises: a virtual network function workload database to store virtual network function (VNF) workload information including identification information of virtual computers, a resource capacity of each virtual computer, and a current workload of each virtual computer; and a global resource allocator module arranged on a network function virtual orchestrator (NFVO) of a virtual network function architecture to use the VNF workload information stored in the virtual network function workload database to allocate the number of active virtual computers and a workload of the active virtual computers. According to the present invention, power of servers is turned off in an idle state, and the number of driving servers is minimized in an active state to minimize energy consumption costs and efficiently allocate a large quantity of VNF resources.

Description

TECHNICAL FIELD [0001] The present invention relates to a virtual network function resource allocation method, a program, a computer-readable recording medium on which the resource is allocated, a computer-

The present invention relates to a virtual network function resource allocation method, a program, a computer-readable recording medium storing the program, and a computing device for allocating a large amount of VNF resources.

Network Function Virtualization (NFV) is a technology that implements the necessary hardware for the network configuration at the server side. It enables the major telecom companies to virtualize the functions of the operator network equipment to increase the flexibility of the network, . This means the way network operators build and operate network and network services by integrating many types of network equipment with industry standard, high capacity servers, switches, and storage devices using evolving information technology virtualization technologies . The NFV implements network functionality using software that can be run on a set of industry-standard server hardware, thereby altering the network architecture.

In addition, network operations may be further modified because software may be moved or instantiated dynamically at different locations within the network without the need to install new devices according to the requirements.

1 is a block diagram of a conventional Network Function Virtualization (NFV) architecture.

1, the NFV reference architecture is configured to include an NFV infrastructure (NFVI infrastructure), a virtual network function (VNF), an NFV management, and an orchestrator (M & O or MANO: Management and Orchestrator) .

First, the NFV Infrastructure (NFVI) provides the virtualized resources needed to support NFV execution, including commercial off the shelf hardware, the necessary accelerator components, and the virtualization and abstraction of the underlying hardware Lt; RTI ID = 0.0 > a < / RTI >

The Virtual Network Function (NFV) is a software implementation of network function (NF) that can operate on NFVI, and it supports element management system (EMS) for understanding and managing the characteristics of single VNF and single VNF More. VNF corresponds to an entity of a network node and is expected to be provided only as hardware independent software.

NFV management and orchestrator (M & O) includes orchestrator, lifecycle management of physical and / or software resources to support infrastructure virtualization, and VNF lifecycle management. The NFV M & O interacts with the Operation Support System (OSS) / Business Support System (BSS) outside the NFV to enable the NFV to be integrated into the existing management landscape within the entire network range. Within MANO, the NFV Orchestrator (NFVO), VNF Manager (VNFM), and Virtual Infrastructure Manager (VIM) are included.

 The NFV orchestrator (NFVO) is primarily responsible for the lifecycle management of the NS to complete the network service orchestrator function and the NFVI resource orchestrator between the multiple VIMs to complete the resource coordination function.

The VNF Manager (VNFM) is responsible for the lifecycle management of the VNF instance (instane). Each VNF may have an associated VNFM, and one VNFM may be designated to manage a single VNF instance or to manage multiple VNF instances of the same type or of different types. The available capabilities of VNF Manager include correlation of VNF instance, NFVI resource configuration for VNF, VNF instance update, VNF instance scaling, NFVI performance measurement and VNF instance related events, NFVI performance measurement and event collection and VNF instance related events. Relationships, support for VNF instances or auto-healing, termination of VNF instances, integrity management for VNF instances throughout the lifecycle of VNF instances, configuration and reporting of events between NFVI and EMS, and performing overall coordination and adaptation roles .

The Virtual Infrastructure Manager (VIM) is responsible for the control and management of the computing, storage, and network resources of the NFVI and is typically within an operator's infrastructure sub-domain. A VIM can specialize in the processing of one type of NFVI resource, or it can manage several types of NFVI resources. Available functions of VIM include orchestrating allocation / upgrade / release / recycling of NFVI resources, managing the association between virtualization resources and computing, storage, and network resources, and hardware resources (computing, storage and network) And cataloging software resources (eg, hypervisors), measuring virtual resource performance and collecting and delivering events.

Based on this architecture, NS (Network Service) having a specific function can be implemented using a plurality of NFs. The end-to-end NS implemented in the conventional network only includes the physical network function. In the case of end-to-end NS implemented in NFV, there are still physical network functions at both ends, but in the middle, some or all physical network functions are replaced with virtual network functions. The function implemented by each network function and the external interface of each network function are independent of whether the network function is a physical network function or a virtual network function. A topology relation linking a virtual network function and a physical network function can be described using a virtual network function transfer graph. The characteristics of each network function can be described using a corresponding network function descriptor (NFD) do.

The VNF must be implemented based on the virtual resources provided by the NFCI, and these virtual resources are acquired by performing virtualization on the corresponding physical resources. The PNF is implemented directly based on physical resources. Unlike traditional networks, where all the controls are centralized with software and hardware-integrated network devices, NFV introduces virtualization to implement the separation of software and hardware in network devices, and control of services is mainly driven by PNF and VNF levels , And the control over performance can be made mainly at the hardware resource level of NFVI, especially NFVI.

FIG. 2 is a layout diagram of a conventional virtual computer resource allocation module, and FIG. 3 is a diagram for explaining a method of allocating resources to a conventional virtual computer.

Referring to FIGS. 2 and 3, a conventional virtual computer resource allocation module may be disposed in a virtualization infrastructure manager (VIM) to distribute resources evenly to a virtual computer. At this time, since the virtual computer resource allocation module distributes resources by selecting a server having the minimum utilization rate, for example, when there are four virtual computers and seven virtual network functions (VNFs) are allocated, [0.4 0.4 0.4 0.2] to the maximum workload vector. In this case, the efficiency of each virtual computer can not be utilized to the maximum.

In particular, resource distribution to the host virtual machine (VM) runs on the NFVI, so resources are distributed to minimize the workload of the entire virtual machine. However, this resource allocation method is not suitable for mass distribution of VNF, and a resource distribution policy which is more efficient is required.

Korean Patent Laid-Open No. 10-2017-0046721 (May 27, 2017), entitled: Energy saving control method, management server and network device

The present invention minimizes the number of activated virtual machines (VMs) by distributing a Global Resource Allocator (GRA) module on the NFVO and distributes the VNF resources so that the workload of each virtual computer is maximized, A VNF (Virtual Network Function), a virtual network function resource allocation method, a program, and a computer readable recording medium and a computing device recording the same.

In order to achieve the above object, a computing apparatus according to an embodiment of the present invention includes a VNF (Vir- net Network Function (VNF), which includes identification information of a virtual computer, resource capacity of each virtual computer, A virtual network function workload database for storing workload information, and a network function virtual orchestrator (NFVO) of a virtual network function (VNF) architecture, And a Global Resource Allocator (GRA) module that allocates the number of activated virtual machines and the workload of the active virtual machine using the VNF workload information stored in the database.

Here, the GRA module uses the VNF workload information stored in the virtual network function workload database so that the number of the activated virtual machines is minimized so that the amount of work of the active virtual machine is 60% or more and less than 100% And the workload of the activation virtual machine can be allocated.

In addition, a virtual network function workload database may be located within the VNFO.

In addition, when receiving a web-based UI for uploading a multi-network service description to be allocated from a network service manager, the GRA module activates the VNF workload using the information included in the UI and the VNF workload information The number of virtual machines and the workload of the activation virtual machine can be allocated.

In addition, the GRA module may forward the UI to the NFV mass deployment module and to allocate the NFV to the NFVI through the NFV mass deployment module.

The GRA module also drives the Genetic Algorithms to determine where VNFs and VNFs are allocated and to assign VNFs to the corresponding NFVIs via VIM.

A method for allocating a virtual network function resource according to another embodiment of the present invention is a method for allocating a virtual network function resource (VNF) to a computing device, the method comprising: ) Storing workload information, and allocating the number of active virtual machines and the workload of the active virtual machine using the stored VNF workload information.

The step of allocating the number of activated virtual machines and the workload of the activation virtual machine may be handled in a network function virtual orchestrator (NFVO) of a virtual network function (VNF) architecture.

Storing VNF (Virtyal Network Function) workload information may be handled in a Network Function Virtual Orchestrator (NFVO) of a Virtual Network Function (VNF) architecture.

The step of allocating the number of active virtual machines and the workload of the active virtual machine may drive a Genetic Algorithm to determine where VNF and VNF are allocated.

In addition, the step of storing the VNF (Virtyal Network Function) workload information and the step of assigning the number of activated virtual machines to the amount of work of the activation virtual computer require allocation from a network service manager The method of claim 1, further comprising: receiving a web-based UI for uploading a multi-network service description, wherein the step of assigning the number of active virtual machines and the workload of the active virtual computer comprises: The number of activated virtual machines and the workload of the active virtual machine can be allocated using the workload information.

In addition, the step of allocating the number of active virtual machines and the amount of work of the activation virtual computer may further include transmitting the NFVs to be allocated to the NFVIs through the NFV bulk allocation module.

A program recorded on a computer-readable recording medium for realizing the virtual network function resource allocation method according to another embodiment of the present invention can be provided.

The present invention can provide a computer readable recording medium for realizing the virtual network function resource allocation method according to another embodiment of the present invention.

According to the present invention, it is possible to minimize the energy consumption cost by minimizing the number of the drive servers in the idle state by turning off the power of the server and in the active state, and efficient resource allocation of the large VNF is possible have.

1 is a block diagram of a conventional Network Function Virtualization (NFV) architecture.
2 is a layout diagram of a conventional virtual computer resource allocation module.
3 is a diagram for explaining a method of distributing resources to a conventional virtual computer.
4 is a layout diagram of a Global Resource Allocate (GRA) module according to an embodiment of the present invention.
5 is a block diagram of an NFV architecture according to an embodiment of the present invention.
FIG. 6 is a diagram illustrating a method for determining an activated virtual machine in a GRA module according to an embodiment of the present invention. Referring to FIG.
7 to 8 are diagrams for explaining a method of distributing resources to a virtual computer according to an embodiment of the present invention.
9 is a diagram for verifying the efficiency of a method of distributing resources to a virtual machine according to an embodiment of the present invention.
10 is a diagram illustrating a genetic algorithm according to an embodiment of the present invention.

Hereinafter, a computing apparatus and a virtual network function resource allocation method according to embodiments of the present invention will be described with reference to the accompanying drawings.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings. The embodiments are described in detail to facilitate understanding of the present invention, and do not limit the scope of the present invention. Accordingly, equivalents that perform the same functions as the present invention are also within the scope of the present invention.

In the following description, the same reference numerals denote the same components, and unnecessary redundant explanations and descriptions of known technologies will be omitted. In the embodiments of the present invention, "communication "," communication network ", and "network" The three terms include wired and wireless transmission and reception networks capable of transmitting and receiving a program, a file, a packet, and the like between a user terminal, a terminal of another user, and a server.

Hereinafter, specific embodiments will be described with reference to the accompanying drawings.

4 is a layout diagram of a Global Resource Allocate (GRA) module according to an embodiment of the present invention.

4, a Global Resource Allocate (GRA) module according to an exemplary embodiment of the present invention includes a network function virtual orchestrator (NFVO) of a virtual network function (VNF) .

The service provider can place a mini-data center near the user's location and move the NFVI to the network edge. In order to be able to backup in case of disaster, and to optimize energy consumption, the GRA module should be placed in NFVO of NFV. Here, the GRA module may be formed of a hardware module or a software module performing the above functions.

5 is a block diagram of an NFV architecture according to an embodiment of the present invention.

Referring to FIG. 5, the NFV architecture according to an embodiment of the present invention has a form of adding the GRA 110 and the virtual network function workload database 120 to the NFVO 100 in the conventional NFV architecture.

Specifically, the network service manager can use a web-based UI to upload multi-network service descriptions that need to be allocated. The UI can pass the request to the NFV mass deployment module and the NFV bulk deployment module to deliver the NFV from the GRA 110 to the NFVI 300. [

The GRA 110 checks the type of VNF and the number of VNFs included in each NS through the NS catalog. That is, the GRA 110 may check the VNF catalog and the NFV resource to obtain information on each VNF resource, and check the database of the VNF workload prediction catalog to obtain each VNF workload information. In addition, the GRA 110 may check the resource status catalog to obtain the resource status of each NFVI.

GRA 110 may drive Genetic Algorithms to determine where VNFs and VNFs are allocated and to allocate VNFs to corresponding NFVIs 300 via VIM 200. [ At this time, the VIM 200 transfers a request from the GRA 110 to the NFVI 300 for mapping and allocation scheduling of the VNF, and the NFVI 300 can allocate a virtual machine (VM) for driving the VNF.

FIG. 6 is a diagram illustrating a method for determining an activated virtual machine in a GRA module according to an embodiment of the present invention. Referring to FIG.

Referring to FIG. 6, a method for determining an activated virtual machine in the GRA module according to an exemplary embodiment of the present invention includes: determining whether an activated virtual machine has a minimum number of active virtual machines, You can assign the number and the workload of the activation virtual machine.

Specifically, the energy efficiency of a server depends on the load level for maximum energy efficiency (usually 60% to 100%) depending on the workload and hardware of the server. To increase the energy efficiency of the cloud server, the workload of the server can be increased by distributing the virtual machine driven service assigned to the server. The power consumption depends on the workload (CPU utilization) of the server, and the power consumption and the CPU utilization ratio are expressed by the following equation (1).

Where P is the power consumption, u is the CPU utilization, Pmax is the power consumption when the CPU utilization is 100%, k is the Pidle versus Pmax, and Pidle is the power consumption when the CPU utilization is 0%.

In the figure, it can be seen that the energy efficiency can be increased when the server is driven with a CPU utilization rate of 60% or more and 100% or less.

7 to 8 are diagrams for explaining a method of distributing resources to a virtual computer according to an embodiment of the present invention.

7 to 8, a method for distributing resources to a virtual machine according to an embodiment of the present invention includes: distributing VNF workload information stored in a virtual network function workload database in a Global Resource Allocator (GRA) Can be used to allocate the number of active virtual machines and the workload of the active virtual machine.

Specifically, in network function virtualization, when a virtual network function is requested, it is associated with a host virtual machine (VM) running in a network function virtualization infrastructure (NFVI). It is possible to minimize the number of driving servers that maintain the active state while keeping the CPU utilization rate at 60% to 100%, and reduce the energy consumption cost by turning off the power of the server in the inactive state.

For example, if there are four virtual machines and distribute seven VNF resources, only two virtual machines can be kept active and two virtual machines can be kept inactive. At this time, the workload vector of the active virtual machine is [80% 60% 0 0].

The Global Resource Allocation (GRA) module 110 requests a Virtual Network Function (VNF) distribution to the Virtual Infrastructure Manager (VIM) 200 and the Virtual Infrastructure Manager (VIM) And control and management of network resources. Each VIM 200 can distribute the delivered number of virtual network functions f1 to f7 to each virtual computer (VM). At this time, resources are allocated so that the number of virtual machines (VMs) becomes minimum and the CPU utilization rate becomes 60% to 100%. After one virtual network function is allocated at a time, After determining the status, the virtual machine to be assigned the next virtual network function is determined.

To obtain the resource status of each NFVI, you can monitor the air entities of each NFVI and monitor the server information. In addition, the VNF workload prediction database can be updated by learning VNF workload information from the monitoring server.

As described above, the GRA 110 may drive the Genetic Algorithms to determine where VNF and VNF are to be allocated and to allocate the VNF to the corresponding NFVI 300 via the VIM 200. [ The genetic algorithm is such that the sum of the VMs arranged in one server is 1 and does not exceed the capacity of each server as shown in [Equation 2] and [Equation 3].

Here, x _ij is a position indicator of the VM, x _ij = 1 indicates that VM exists in the jth server, and x _ij = 0 _otherwise .

Here, N represents the number of VFNs, d _i ^r represents resource requests, i represents 1 to N, and r represents requests for different resources (workload, expected probability, CPU, Mem, Bandwith, actve).

When assigning n hypervisor servers and m VNFs to the server, the above genetic algorithm is used to maximize the workload among the n hypervisor server lists and to set the hypervisor to optimize resource allocation You can choose.

9 is a diagram for verifying the efficiency of a method of distributing resources to a virtual machine according to an embodiment of the present invention.

Referring to FIG. 9, a method of distributing resources to a virtual computer according to an embodiment of the present invention may be calculated using Equation (4) below.

Where i is the number of simultaneously activated virtual machines, P [k = i] is the probability that i VMs will be activated simultaneously, N is the number of VMs that can be run on the physical server, A is the activation time . The binomial distribution E [x] is equal to the product of A and N, and the pdf curve is shown in (b). It can be confirmed that this is similar to the actual calculated real time result value.

Referring to Equation (4), the average CPU utilization of the server is 33%, which is related to the activation time probability of the VM and the number of VMs on the physical server. If the utilization rate of the server is 60% to 100% Maximum energy efficiency can be obtained.

[Table 1] shows the number (N) of VMs executed in the physical server and the activation time probability A of the VM. [Table 1] Referring to Table 1, when the activation time probability of the VM is higher than 0.6, there is no need to increase the target resources of the physical server, and it is possible to guarantee against the real server that is running with an average 60% workload. If the activation time probability of a VM is less than 0.5, then you need to use virtualization technology to increase the number of VMs running on the real server. That is, the workload can be increased or decreased according to A. Therefore, it is necessary to distribute the resources in consideration of both.

A N P < = 100% P> 100%  50% < P < = 100% Target > = 0.6 10 One 0 > 0.63 100% 0.5 12 0.9968 0.0032 0.6096 120% 0.4 15 0.9907 0.0093 0.5874 150% 0.3 20 0.9829 0.0171 0.5665 200% 0.2 30 0.9744 0.0256 0.5469 300% &Lt; = 0.1 60 > = 0.9658 <= 0.0342 &Lt; = 0.5286 600%

10 is a diagram illustrating a genetic algorithm according to an embodiment of the present invention.

Referring to FIG. 10, a Genetic Algorithm is a program for determining a VM to which a VNF is allocated.

The genetic algorithms are programmed to assign the weight to each item instead of processing the equivalence probabilities of all items that can be selected in the initialization state so that the items with the highest probability are finally automatically selected.

As described above, the virtual network function resource allocation and distribution method according to an embodiment of the present invention can be applied to an application installed in a computing device (which may include a program that is basically installed in a terminal or a operating system) And may be executed by an application (program) directly installed on the terminal by the user through an application providing server such as an application store server, an application, or a web server associated with the service.

In this regard, the virtual network function resource allocation and distribution method according to an embodiment of the present invention may be implemented as an application (program) installed in a computing device or directly installed by a user, a computer readable record Can be recorded on the medium. Such a program may be recorded on a recording medium that can be read by a computer and executed by a computer so that the above-described functions can be executed.

As described above, the program for executing the virtual network function resource allocation and distribution method according to an embodiment of the present invention can be implemented by a computer-readable medium such as a computer-readable medium, such as C, C ++, JAVA, (Code).

Such code may further include memory reference related code as to whether additional information or media needed to cause the processor of the computer to be executed should be referred to at any location (address, address) of the internal or external memory of the computer.

A functional program for implementing the present invention and a code and a code segment related thereto can be easily read by programmers in the technical field of the present invention in consideration of a system environment of a computer that reads a recording medium and executes a program, And can be inferred or altered.

A computer-readable recording medium on which the program is recorded includes a ROM, a RAM, a CD-ROM, a magnetic tape, a floppy disk, and an optical media storage device.

While the present invention has been described in connection with what is presently considered to be the most practical and preferred embodiments, it is to be understood that the invention is not limited to the disclosed embodiments. That is, within the scope of the present invention, all of the elements may be selectively coupled to at least one. In addition, although all of the components may be implemented as one independent hardware, some or all of the components may be selectively combined to perform a part or all of the functions in one or a plurality of hardware. As shown in FIG. The codes and code segments constituting the computer program may be easily deduced by those skilled in the art. Such a computer program can be stored in a computer-readable storage medium, readable and executed by a computer, thereby realizing an embodiment of the present invention. The storage medium of the computer program may include a magnetic recording medium, an optical recording medium, and the like.

While the present invention has been particularly shown and described with reference to exemplary embodiments thereof, it is to be understood that the invention may be embodied otherwise without departing from the spirit and scope of the invention. Therefore, the embodiments disclosed in the present invention are intended to illustrate rather than limit the scope of the present invention, and the scope of the technical idea of the present invention is not limited by these embodiments. The scope of protection of the present invention should be construed according to the following claims, and all technical ideas within the scope of equivalents should be construed as falling within the scope of the present invention.

100: Network function Virtual orchestrator (NFVO)
110: Global Resource Allocation (GRA) module
120: Virtual network feature workload database
200: Virtualization Infrastructure Manager (VIM)
300: Network Features Virtual Infrastructure (NFVI)

Claims (14)

A virtual network function workload database that stores VNF (Virtyal Network Function) workload information including identification information of a virtual computer, resource capacity of each virtual computer, and current workload of each virtual computer; And
A network function of a Virtual Network Function (VNF) architecture. The network function is arranged in a Network Function Virtual Orchestrator (NFVO), and the VNF workload information stored in the Virtual Network Function Workload Database A Global Resource Allocator (GRA) module for allocating the workload of the activation virtual machine;
&Lt; / RTI &gt;
The method according to claim 1,
Wherein the GRA module is configured to use the VNF workload information stored in the virtual network function workload database so that the number of the activated virtual computers is minimized so that the amount of work of the activated virtual computer becomes 60% And assigning a number and a workload of the activation virtual machine.
The method according to claim 1,
Wherein the virtual network function workload database is disposed within the VNFO.
The method according to claim 1,
When receiving a web-based UI for uploading a multi-network service description to be allocated from a network service manager, the GRA module activates the VNF workload using information included in the UI and the VNF workload information And allocates the number of virtual machines and the workload of the activation virtual machine.
5. The method of claim 4,
Wherein the GRA module passes the UI to an NFV mass deployment module and to the NFV bulk deployment module to allocate an NFV to the NFV.
The method according to claim 1,
The GRA module drives Genetic Algorithms to determine where VNF and VNF are to be allocated and to assign VNF to the corresponding NFVI via VIM.
In a computing device, storing VNF (Virtyal Network Function) workload information including identification information of a virtual computer, resource capacity of each virtual computer, and a current workload of each of the virtual machines; And
Allocating a number of active virtual machines and a workload of the active virtual machine using the stored VNF workload information;
And allocating resources to the virtual network.
8. The method of claim 7,
Wherein the step of allocating the number of active virtual machines and the amount of work of the activation virtual computer comprises:
A network function of a virtual network function (VNF) architecture A virtual network function resource allocation method handled by a network function virtual orchestrator (NFVO).
8. The method of claim 7,
The step of storing the VNF (VNF: Viryal Network Function)
A network function of a virtual network function (VNF) architecture A virtual network function resource allocation method handled by a network function virtual orchestrator (NFVO).
8. The method of claim 7,
Wherein the step of allocating the number of active virtual machines and the amount of work of the activation virtual computer comprises:
A method for allocating virtual network functional resources, characterized by driving Genetic Algorithms to determine where VNF and VNF are to be allocated.
8. The method of claim 7,
Storing the VNF (Virtyal Network Function) workload information, and allocating the number of active virtual machines and the workload of the activation virtual computer,
Receiving a web-based UI for uploading a multi-network service description to be allocated from a network service manager,
The step of allocating the number of active virtual machines and the amount of work of the activation virtual computer may include allocating the number of active virtual machines and the workload of the activation virtual computer using the information included in the received UI and the VNF workload information Characterized in that the method comprises the steps of:
12. The method of claim 11,
After the step of allocating the number of active virtual machines and the workload of the activation virtual machine,
Delivering NFV to the NFVI through the NFV bulk deployment module;
Wherein the virtual network resource allocation method comprises the steps of:
A program recorded on a computer-readable recording medium for realizing a virtual network function resource allocation method according to any one of claims 7 to 13.
A computer-readable recording medium on which the program of claim 13 is recorded.

Images