KR20170079447A - Method, Apparatus, Program, and Recording Devcie for Request and Embeding Resource for Statistical Virtual Network in Intra-Datacenter Cloud Environment - Google Patents

Abstract

The present invention relates to a stochastic virtual network request method in a single data center cloud computing environment, a request receiving apparatus using the same, a resource allocation method using the same, a resource allocating apparatus, a network apparatus, a program performing the same, and a recording medium. A request reception step of receiving a virtual network request signal including a number and a bandwidth of a desired virtual machine in the tenant; And a request transmission step of transmitting a virtual network request signal to a resource allocation device which is a component of the network device, wherein the bandwidth in the virtual network request signal is in the form of a probability distribution. A stochastic virtual network requesting method in the network can be provided. According to this, a virtual network request model is defined as a probability distribution, so that a tenant can obtain a more accurate virtual network request based on their statistics.

Description

Field of the Invention < RTI ID = 0.0 > [0001] < / RTI > The present invention relates to a probabilistic virtual network request method in a single data center cloud computing environment, a request receiving apparatus using the same, and Embeding Resource for Statistical Virtual Network in Intra-Datacenter Cloud Environment}

The present invention relates to a stochastic virtual network request method in a single data center cloud computing environment, a request receiving apparatus using the same, a resource allocation method using the same, a resource allocating apparatus, a network apparatus, a program performing the same, and a recording medium. More particularly, to a method for a resource allocation device of a network device to efficiently allocate network resources according to the request when a tenant makes a virtual network request to the network device in a single data center cloud computing environment.

In the past, in order to build an IT environment, it was necessary to purchase hardware such as a server (computer) and a switch, acquire a license, and install an application. In order to expand the infrastructure, it was inevitable to purchase additional hardware. This approach has led to economic losses due to additional installation costs and processing problems after use. As a result, most companies have recently begun using cloud computing services to pay a certain amount and use their resources to companies that operate their own data centers around the world, such as Amazon, Microsoft and Google. A data center is a facility consisting of a server (computer), a repository, a network device (switch, router, cable), a power distribution system and a cooling system. Cloud computing refers to storing and sharing data and using applications over the Internet .

The data center network refers to the infrastructure that makes up the data center and can be represented by network topology, routing / switching devices, and protocols used such as Ethernet or IP. Most existing data center networks have a tree-like structure, and FIG. 1 illustrates a fat-tree topology with k = 4 as one of the tree-like structures. The pat-tree topology has k-port switches connected to k pods, each of which consists of two layers (aggregation, edge). The fat-tree topology is suitable as a topology for configuring a data center network because it can flexibly cope with network device failures when multiple switches are connected to each other.

A virtualized data center is a virtualized data center where all or a portion of the hardware (servers, routers, switches, links) is virtualized. In general, physical hardware can be divided into several independent and separate virtual objects through software called a hypervisor. For example, one server is virtualized by the hypervisor into machines with several different capacities (CPU, memory, storage space) and different operating systems and applications. Virtualization is a technology that hides the characteristics of physical computer resources from the way different systems, applications, and end users interact with resources. VMware, Parallels, and others use virtualization applications to simultaneously use multiple operating systems on a single computer. Is a representative example of virtualization.

A Virtual Data Center is a collection of virtual resources such as Virtual Machines, Virtual Switches, and Virtual Links. A virtualized data center refers to one physical data center with technology that virtualizes resources, whereas a virtual data center refers to a logical object that is made up of some of the physical data center resources. 2 is a schematic diagram illustrating virtualization of a data center. Figure 2 shows how multiple virtual data centers can be organized into a single virtualized data center.

Virtualization technology can maximize resource efficiency in that it divides available resources into smaller units and allows them to be shared among several tenants. And one that is suitable for cloud computing services in that it does not make sense to tenants that they are sharing a single resource.

A virtual network is a collection of virtual nodes (hosts, switches, routers) and virtual links. Today, cloud computing providers provide a simple interface to tenants that allows tenants to specify the amount of CPU, memory, number of servers, storage capacity, and applications they want to use. Unlike computing resources, however, it does not guarantee bandwidth within a single cloud network, and provides best-effort service for data transmission in the cloud. That is, quality of service (QoS) for transmitted data is not guaranteed. This bandwidth non-assurance causes economic damage to tenants as it makes it difficult for tenants to anticipate the performance of their network applications. In addition, current billing policy of a business reflects only the computing resources used by each user, but does not consider bandwidth usage.

Recently, a network-abstraction technique has been proposed to solve the bandwidth guarantee problem of a single cloud network. Tennants can request network resources in a single cloud through virtual data center abstraction. Virtual datacenter abstraction refers to a virtual cluster consisting of a virtual machine and virtual links. Typically, there are a pipe model (model) and a hose model (model). The pipe model is represented by the number of virtual machines and the bandwidth between each virtual machine. The hose model is expressed by the number of virtual machines and the bandwidth of the link from each machine to the switch. Because it is not easy for tenants to accurately predict the amount of traffic between each virtual machine, an easier and more flexible hose model is being studied more extensively than the pipe model. 3 shows an example of a virtual data center abstraction expressed in the form of a hose model of < N, B >, when a tenant needs a bandwidth of B between a total of N virtual machines and a virtual machine-switch.

With respect to the operation of the admission control module for a virtual network request, when a tenant requests as much computing and network resources as needed through virtual data center abstraction (VDC abstraction), the business entity decides whether to accept or reject the request And then allocate resources appropriately for that request. In other words, the admission control module of the provider determines whether 1) the tenant's request can be accepted (whether the tenant can provide the requested information), 2) the number of candidates that can be provided to the tenant And to select and allocate resources according to their own criteria such as maximization of the acceptance ratio, load-balancing, and the like. Therefore, the admission control module can be designed in the form of an optimization problem to find the most ideal allocation method. Figure 4 is a flowchart of cloud datacenter virtualization configured with virtual datacenter abstraction and admission control modules.

[Patent Document 1] U.S. Published Patent Application No. US20150067170 A1, Method and system to allocate bandwidth based on deadline in cloud computing networks, Telefonaktiebolaget L M Ericsson (Publ. [Patent Document 2] U.S. Published Patent Application No. US20130055091 A1, Graph-Based Virtual Data Center Requests, Cisco Technology, Inc.

However, since existing models require a certain amount of bandwidth, they can not accurately reflect the traffic statistics of tenants, and they have the disadvantage of inefficient resource management depending on the type of traffic of the tenants. Most traffic is time-varying, but since tenants are forced to request a certain amount of bandwidth, tenants usually request the maximum amount of traffic. As a result, as shown in FIG. 5 where an example of inefficient network resource management is shown, even though the amount of traffic generated by the tenants using the cloud service is less than the requested amount of bandwidth, a certain amount of bandwidth has already been allocated to them, New virtual network requests can not be accepted, which leads to the inefficiency of network resource management. Of course, the way to reallocate bandwidth to the tenants through the recalculation process has also been studied, but it is impractical to require a high complexity task on the switch or hypervisor.

Accordingly, the present invention has been made to solve the above-described problems.

It is an object of the present invention to provide a stochastic virtual network request method in a single data center cloud computing environment for probabilistic guarantee of bandwidth instead of existing hose model and pipe model in order to achieve efficient operation of network resources, , A resource allocation method using the same, a resource allocation device, a network device, a program for performing the same, and a recording medium.

Hereinafter, specific means for achieving the object of the present invention will be described.

It is an object of the present invention to provide a probabilistic virtual network request method performed by a request receiving module of a network device to provide a virtual network to a tenant in a single data center cloud computing environment, A request receiving step of receiving a virtual network request signal of the tenant including the virtual network request signal; And a request transmission step of transmitting the virtual network request signal to a resource allocation apparatus which is a component of the network apparatus, wherein the bandwidth in the virtual network request signal is in the form of a probability distribution. Can be achieved by providing a probabilistic virtual network request method in a cloud computing environment.

Also, the bandwidth in the virtual network request signal may be expressed as a normal distribution, and the virtual network request signal may include a number of desired virtual machines in the tenant, an average value of the desired bandwidth in the tenant, And a probability value at which the desired bandwidth is ensured in the tenant.

It is another object of the present invention to provide a stochastic virtual network request method performed by a request receiving module of a network device in order to provide a virtual network to a tenant in a single data center cloud computing environment, A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of a desired virtual machine in the tenant; And a request transmission step of transmitting the virtual network request signal to a resource allocation apparatus which is a component of the network apparatus, wherein the bandwidth in the virtual network request signal is in the form of a probability distribution. The present invention can be achieved by providing a recording medium on which a program for performing a stochastic virtual network request method in a cloud computing environment is recorded on a computer.

It is another object of the present invention to provide a program stored in a recording medium such that a probabilistic virtual network request method performed by a request receiving module of a network device to provide a virtual network to a tenant in a single data center cloud computing environment is performed on the computer A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of virtual machines desired in the tenant; And a request transmission step of transmitting the virtual network request signal to a resource allocation apparatus which is a component of the network apparatus, wherein the bandwidth in the virtual network request signal is in the form of a probability distribution. A method for requesting a stochastic virtual network in a cloud computing environment can be achieved by providing a program stored in a recording medium so as to be performed on the computer.

It is another object of the present invention to provide a method and system for providing a virtual network to a tenant in a single data center cloud computing environment and configured in a network device for providing a virtual network to the tenant, And a request receiving module for transmitting a request signal to a resource allocation device which is a component of the network device, wherein the virtual network request signal includes a number and a bandwidth of a desired virtual machine in the tenant, Wherein the bandwidth is in the form of a probability distribution in a single data center cloud computing environment.

Another object of the present invention is to provide a resource allocation method performed by a network device for providing a virtual network to a tenant in a single data center cloud computing environment, the method comprising the steps of: A request reception step of receiving a virtual network request signal; Determining whether the received virtual network request signal is acceptable or not; And a virtual machine allocation step of allocating a virtual machine to the tenant based on the accepted virtual network request signal, wherein the bandwidth in the virtual network request signal is in the form of a probability distribution , And a resource allocation method in a single data center cloud computing environment.

Also, the bandwidth in the virtual network request signal may be expressed as a normal distribution, and the virtual network request signal may include a number of desired virtual machines in the tenant, an average value of the desired bandwidth in the tenant, And a probability value at which the desired bandwidth is ensured in the tenant.

The method may further include determining whether the received virtual network request signal is accepted when the bandwidth of the virtual link satisfies the following equation.

[Mathematical Expression]

Here, _L B is the bandwidth of the virtual _{link, μ = μ e + μ n} · min {i, N n -i}, σ 2 = σ e 2 + σ n 2 · min {i, N n -i} , 1-ε = max {1 -ε e, 1-ε n} a, F refers to the cumulative density function (cumulative density function) of the standard normal distribution _{and, <N n, μ n,} σ n, ε n> It is the imaginary mean stochastic hose model of network request signal and, i is an entire virtual machine of the particular sub-means a virtual machine assigned to the tree, and the sum of all the traffic that is already going through the virtual link is <μ _e, σ _e > and has a service quality of ε _e .

In the virtual machine allocation step, a virtual machine is allocated in the order of sub-trees having a large DELTA value in the following equation.

[Mathematical Expression]

Where DELTA denotes the statistical multiplexing gain of the sub-tree and the bandwidth of the virtual network request signal to use the link of the particular sub-tree follows a normal distribution of &_lt; [mu] _n , [sigma] _n & The sum of the traffic follows a normal distribution of <μ _e , σ _e >.

It is another object of the present invention to provide a recording medium on which a program for performing a resource allocation method performed by a network device on a computer for providing a virtual network to a tenant in a single data center cloud computing environment is recorded, A request reception step of receiving the virtual network request signal of the tenant including the number and bandwidth of the tenant; Determining whether the received virtual network request signal is acceptable or not; And a virtual machine allocation step of allocating a virtual machine to the tenant based on the accepted virtual network request signal, wherein the bandwidth in the virtual network request signal is in the form of a probability distribution , And a method of allocating resources in a single data center cloud computing environment on a computer.

Another object of the present invention is to provide a program stored in a recording medium such that a resource allocation method performed by a network device for providing a virtual network to a tenant in a single data center cloud computing environment is performed on a computer, A request reception step of receiving the virtual network request signal of the tenant including the number and bandwidth of the tenant; Determining whether the received virtual network request signal is acceptable or not; And a virtual machine allocation step of allocating a virtual machine to the tenant based on the accepted virtual network request signal, wherein the bandwidth in the virtual network request signal is in the form of a probability distribution , And a method of allocating resources in a single data center cloud computing environment is performed on a computer.

Another object of the present invention is to provide a resource allocating apparatus for providing a virtual network to a tenant in a single data center cloud computing environment, the resource allocating apparatus comprising: An acceptance control module for determining whether the received virtual network request signal is accepted; And a virtual machine allocation module for receiving the virtual network request signal accepted by the acceptance control module and for allocating a virtual machine to the tenant based on the accepted virtual network request signal, Wherein the network request signal comprises a number and bandwidth of desired virtual machines in the tenant and the bandwidth in the virtual network request signal is in the form of a probability distribution. .

Another object of the present invention is to provide a network device for providing a virtual network to a tenant in a single data center cloud computing environment, the network device comprising: a request receiving module for receiving a virtual network request signal from the tenant; An acceptance control module for receiving the virtual network request signal from the request reception module and determining whether the received virtual network request signal is accepted; And a virtual machine allocation module for receiving the virtual network request signal accepted by the acceptance control module and for allocating a virtual machine to the tenant based on the accepted virtual network request signal, Wherein the network request signal comprises the number and bandwidth of virtual machines desired by the tenant and the bandwidth in the virtual network request signal is in the form of a probability distribution. .

As described above, the present invention has the following effects.

First, according to an embodiment of the present invention, a virtual network request model is defined as a probability distribution, so that a tenant can obtain a more accurate virtual network request based on their statistics.

Second, according to an embodiment of the present invention, the service provider can increase the request acceptance ratio and provide better service quality to the tenants.

Third, according to an embodiment of the present invention, since more tenants can be accommodated in comparison with network resources, the same level of service can be provided to tenants at a lower fee.

Fourth, according to the embodiment of the present invention, the predictability of the bandwidth is improved, so that the utilization of cloud computing is expanded.

BRIEF DESCRIPTION OF THE DRAWINGS The accompanying drawings, which are incorporated in and form a part of the specification, illustrate exemplary embodiments of the invention and, together with the description, serve to provide a further understanding of the technical idea of the present invention. And should not be construed as limiting.
FIG. 1 is a schematic diagram showing a fat-tree topology with k = 4 as one of the tree-like structures. FIG.
2 is a schematic diagram illustrating virtualization of a data center,
3 is a schematic diagram showing an example of a virtual data center abstraction expressed in the form of a hose model of < N, B &
4 is a flowchart of a cloud datacenter virtualization configured with a virtual datacenter abstraction and admission control module,
5 is a schematic diagram showing an example of inefficient network resource management,
Figure 6 is a block diagram illustrating a network device in a single data center cloud computing environment in accordance with one embodiment of the present invention;
FIG. 7 is a diagram illustrating a case where a tenant according to an embodiment of the present invention has five virtual machines, an average of 100 and a standard deviation of 10 between the machine switches, A schematic diagram showing an example of a network,
FIG. 8 is a schematic diagram showing an example of a greedy algorithm used for allocation of a virtual machine in an admission control module according to an embodiment of the present invention;
FIG. 9 is a schematic diagram showing an example of a virtual network request of a probabilistic hose model of <4, 100, 10 ² ,?> According to an embodiment of the present invention;
10 is a schematic diagram showing an example of a link bandwidth condition according to an embodiment of the present invention;
11 is a flowchart of an algorithm of a resource allocation apparatus according to an embodiment of the present invention,
FIG. 12 is a flowchart illustrating a resource allocation method according to an embodiment of the present invention;
13 is a graph illustrating a simulation result of a resource allocation method according to a stochastic virtual network request according to an embodiment of the present invention.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. In the following detailed description of the operation principle of the preferred embodiment of the present invention, a detailed description of known functions and configurations incorporated herein will be omitted when it may unnecessarily obscure the subject matter of the present invention.

The same reference numerals are used for portions having similar functions and functions throughout the drawings. Throughout the specification, when a part is connected to another part, it includes not only a case where it is directly connected but also a case where the other part is indirectly connected with another part in between. In addition, the inclusion of an element does not exclude other elements, but may include other elements, unless specifically stated otherwise.

Network devices in a single data center cloud computing environment

With respect to a network device in a single data center cloud computing environment, Figure 6 is a block diagram illustrating a network device in a single data center cloud computing environment in accordance with an embodiment of the present invention. As shown in FIG. 6, a network device 1 in a single data center cloud computing environment according to an embodiment of the present invention may include a request receiving module 2, a resource allocating device 3, and the like.

The request receiving module 2 is a configuration for receiving a probabilistic virtual network request in at least one tenant. According to an embodiment of the present invention, a tenant's probabilistic virtual network request can be directly transferred from the tenant to the request receiving module 2. Alternatively, a converter may be configured between the tenant and the request receiving module 2 so that the request receiving module 2 receives the converted probabilistic virtual network request. In addition, the probabilistic virtual network request according to an embodiment of the present invention can be realized by SLA (Service Level Agreement) between the tenant and the cloud computing provider.

A probabilistic virtual network request according to an embodiment of the present invention is characterized by defining a bandwidth as a probability distribution in a tenant virtual network request. A stochastic virtual network request of a tenant according to an exemplary embodiment of the present invention requires that all the virtual link bandwidths required by the tenant be represented by a probability distribution and the quality of service (QoS) There is a difference.

In connection with a concrete embodiment of the present invention relating to a tenant's virtual network request, a tenant's stochastic virtual network request according to an embodiment of the present invention is described by a probabilistic hose model and a probabilistic pipe model through a virtual data center abstraction . The first embodiment is described as a stochastic hose model, and the second embodiment can be described as a stochastic pipe model. It is also assumed that all the virtual links required by the tenant in the first and second embodiments conform to the Normal Distribution for convenience of calculation. When a probabilistic virtual network request according to an embodiment of the present invention is configured as a normal distribution, there is an advantage that a tenant can easily request based on the traffic statistics of the tenant. In addition, according to an exemplary embodiment of the present invention, when a probabilistic virtual network request is configured as a normal distribution, a regular distribution and a normal distribution are added to a normal distribution. Therefore, instead of individually servicing tenants' requests, So that the service can be made possible.

With respect to the first embodiment of the probabilistic virtual network request, the statistical hose-model is expressed as a parameter of <N, μ, σ ² , ε>. N is the number of virtual machines required by the tenant as in the existing hose model, and <μ, σ ² > is the bandwidth between the virtual machine (VM) and the virtual switch that is needed. Is defined as an independent and identically distributed normal distribution, and ε is a term for expressing quality of service (QoS), which means that bandwidth should be guaranteed with a probability of 1-ε. In order for bandwidth to be guaranteed with a probability of 1 (in any situation, bandwidth is always guaranteed), infinite bandwidth is needed, but since it is not what tenants want, ε can be expressed as a value between 0 and 1, inclusive. FIG. 7 shows a case where a tenant desires to guarantee the transmission of traffic with a probability of 95% or more when a normal distribution type traffic that follows an average of 100 and a standard deviation of 10 between five virtual machines, machine-switches flows. Therefore, in the case of the virtual network request shown in FIG. 7, it can be expressed as a probability hose model of <5, 100, 10 ² , 0.05> according to an embodiment of the present invention.

In accordance with the second embodiment of the probabilistic virtual network request, according to an embodiment of the present invention, a tenant can make a stochastic virtual network request with a pipe model other than a hose model. According to an embodiment of the present invention A stochastic pipe-model can represent the bandwidth to be guaranteed between virtual machines in the form of a matrix. For example, the average speed of the traffic sent from the virtual machine 1 (VM1) to the virtual machine 3 (VM3) in FIG. 7 is requested twice faster than the average speed from the virtual machine 1 (VM1) to the virtual machine 5 It can not be expressed by the stochastic hose model which is the first embodiment of the present invention. However, according to the stochastic pipe model of the second embodiment of the present invention, the tenant requests the bandwidth between the virtual machine 1 (VM1) and the virtual machine 3 (VM3) to be <2μ, The average speed of the traffic from the virtual machine 1 (VM1) to the virtual machine 3 (VM3) is smaller than the average speed of the traffic from the virtual machine 1 (VM1) to the virtual machine 1 (VM1) 5 &lt; / RTI &gt; (VM5).

The resource allocation device 3 is configured to determine whether to accept or reject a stochastic virtual network request of a received tenant and appropriately allocate VM (Virtual Machine) resources corresponding to the request, Or a controller coupled to the virtual switch. The resource allocation apparatus 3 according to an embodiment of the present invention may include a request signal converter 30, an admission control module 32, and a VM allocation module 34. [ The request signal converter 30, the acceptance control module 32, and the VM allocation module 34 according to an embodiment of the present invention may be configured to interconnect and communicate signals through at least one connection.

The request signal converter 30 is connected to the request receiving module 2 and configured to process the probable virtual network request of the received tenant so that it can be processed in the resource allocation device 3. [

The admission control module 32 is connected to the request signal converter 30 or the request receiving module 2 to determine whether to accept or reject the stochastic virtual network request of the received tenant.

The VM allocation module 34 is connected to the acceptance control module 32 and allocates or embeds a VM corresponding to the request of the tenant accepted by the acceptance control module 32.

In accordance with a concrete embodiment of the present invention relating to a resource allocation method according to a tenant's probabilistic virtual network request, a greedy algorithm can be used as a resource allocation method according to an embodiment of the present invention.

The number of virtual machines in the sub-tree that a cloud computing operator can allocate within a single data center cloud of the tree-structure is 1) the number of empty virtual machine slots in that sub-tree (condition 1) and 2) the bandwidth available for each link (condition 2). How to determine and use the above two conditions in the resource allocation method according to an embodiment of the present invention will be described in detail below.

In relation to Condition 1 above, the optimization problem that finds a way to allocate virtual machines and network resources most efficiently is known as NP-hard (NP-hard) that can not be solved within polynomial-time. In one embodiment of the present invention, a greedy algorithm is used to search from a sub-tree of a small unit as an approximate method for allocating resources. FIG. 8 illustrates an example of a greedy algorithm used for allocation of a virtual machine in an admission control module according to an embodiment of the present invention. FIG. 8 assumes a situation where a virtual network request using the hose model of <4, D> is received when one server can be divided into four virtual machines by the hypervisor. The admission control module searches from a small sub-tree consisting of leaf nodes to roots in order to allocate as many virtual machines as requested. As shown in the left figure of FIG. 8, it is impossible to assign four virtual machines to one server, so virtual machines can be allocated only after one floor is further elevated as shown in the right picture of FIG.

In addition to the condition 1 described in FIG. 8, the acceptance / control module may also consider the condition 2. With respect to Condition 2 above, the amount of bandwidth required for each link depends on the number of virtual machines to be allocated in the sub-tree connected to the link. When a probabilistic hose model request of <4, 100, 10 ² , ε> is to be allocated as shown in FIG. 9, the bandwidth required for the link L is determined by the number of virtual machines to be allocated to the sub- It is the product of the minimum number of virtual machines to be allocated to the network and the requested bandwidth. Therefore, if the normal distribution of <100, 10 ² > is expressed as D, the bandwidth required for link L is D for virtual machine allocation. This can be expressed by the following equation (1). In Equation (1), B _L denotes a bandwidth required for the link L.

For valid resource allocation, B _L must be less than or equal to the remaining bandwidth of the link.

When generalizing condition 2, when a tenant's virtual network request r enters a probabilistic hose model of r: <N _n , μ _n , σ _n , ε _n >, the resource allocation apparatus according to an embodiment of the present invention Can be considered a scenario in which i virtual machines are assigned to a particular sub-tree, and the link associated with that sub-tree is L. Since the traffic generated by each customer is independent and follows a normal distribution, the sum of all traffic over link L can be represented as one normal distribution. The sum of all traffic traversing a link L follows a normal distribution of <μ _e , σ _e > and has a quality of service (QoS) parameter of ε _e . In order to provide a stochastic bandwidth guarantee for both the new tenant and the tenant already using the service, the bandwidth B _L of the link L must satisfy the following equation (2).

In this case, μ = μ _e + μ _n min {i, N _n -i}, σ ² = σ _e ² + σ _n ² min {i, N _n -i}, 1-ε = max { _e , 1 -? _n }. In Equation (2), F denotes a cumulative density function of the standard normal distribution. 10 is an example of the link bandwidth condition described above according to an embodiment of the present invention. In order for the resource allocation device to accept a new virtual network request, the bandwidth B _L of the link _L is B _L ≥ 300 + 100 + 2500 + 100) 占 F ^-1占 (1-0.05).

According to an embodiment of the present invention, in order to increase the acceptance rate of the tenant's virtual network request and the average throughput (aspect of the service quality) of the entire service, statistical multiplexing-gain conditions of the sub- Can be added. If there are a plurality of sub-trees of the smallest size that can allocate a virtual machine, the resource allocation apparatus according to an embodiment of the present invention may be configured to select the sub-tree having the largest value of the following expression (3) have. Below equation (3), the sub-tree of the new request bandwidth to use T link L in following the normal distribution of <μ _n, σ _n>, the sum of the traffic that is already going through the link L <μ _e, σ _e &Lt; RTI ID = 0.0 &gt;&lt; / RTI &gt;

11 is a flowchart illustrating an algorithm of a resource allocation apparatus according to an embodiment of the present invention. As shown in FIG. 11, when R is received for a set of probabilistic virtual network requests for a network topology P in a particular tenant, each level of the sub-tree is allocated for each request r, and for each request r The algorithm can be configured to perform switching. In FIG. 11, N _v means the number of possible virtual machines configured in the sub-tree. More specifically, the algorithm of the resource allocation apparatus according to an embodiment of the present invention includes: determining whether the number of virtual machines available when judging from the leaf node of level 0 of the network topology is suitable for the request r; And assigning the tenant to an empty slot in the sub-tree.

The following is an example of an algorithm for calculating the number N _S of virtual machines that can be allocated to a particular server S according to an embodiment of the present invention.

The following is an example of an algorithm for calculating the number of virtual machines (N _T ) that can be assigned to a sub-tree T of each level according to an embodiment of the present invention.

According to the embodiment of the present invention, when the greedy algorithm is used in the virtual machine allocation, the following effects are generated as compared with the existing system. According to an embodiment of the present invention, when a virtual machine is allocated using a greedy algorithm that considers a sub-tree of a smaller unit preferentially, the calculation speed is fast and the link resource can be efficiently used . This is because, if all the virtual machines requested in the network are allocated to a specific server, the operation is performed only within that server. As a result, according to the embodiment of the present invention, when the virtual machine allocation is determined by the greedy algorithm, a better performance can be obtained in terms of speed as compared with the method of finding the optimal virtual machine allocation position. Also, according to the embodiment of the present invention, when the statistical multiplexing method is used as a method of determining the virtual machine allocation by the greedy algorithm, the network resource can be used more efficiently than the random method.

If a bandwidth is defined as a probability distribution among the parameters of the tenant's virtual network request as in the embodiment of the present invention, the tenants can obtain a more accurate virtual network request based on their statistics. Also, according to an embodiment of the present invention, the cloud computing provider can increase the request acceptance ratio and provide the tenant with superior service quality. In addition, according to an embodiment of the present invention, since more tenants can be accommodated in comparison with network resources, the same level of service can be provided to tenants at a smaller fee. In addition, according to an embodiment of the present invention, the predictability of bandwidth is improved, so that the utilization of cloud computing is expanded.

Hereinafter, a probabilistic virtual network request method and a resource allocation method in a network device in a single data center cloud computing environment according to an embodiment of the present invention will be described in detail.

Stochastic Virtual Network Request Method

The probabilistic virtual network requesting method according to an embodiment of the present invention can be applied to a virtual network request transmitted from a tenant to a request receiving module of a network device and transmitted from a request receiving module to a resource allocating device. The probabilistic virtual network request method according to an embodiment of the present invention includes the step of receiving the number of virtual machines desired in the tenant, the bandwidth and the service quality in the request receiving module of the network device according to an embodiment of the present invention And the bandwidth may be a value defined or expressed as a probability distribution.

Resource Allocation Method Based on Stochastic Virtual Network Request

FIG. 12 is a flowchart illustrating a resource allocation method according to an embodiment of the present invention, in connection with a resource allocation method according to a probabilistic virtual network request according to an embodiment of the present invention. As shown in FIG. 12, the resource allocation method according to an embodiment of the present invention may include a stochastic virtual network request receiving step (S10), an acceptance determining step (S20), and a VM allocating step (S30) .

The probabilistic virtual network request receiving step (S10) is a step in which the request receiving module of the network device receives a probabilistic virtual network request abstracted from the tenant. A probabilistic virtual network request according to an exemplary embodiment of the present invention may include the number and bandwidth of virtual machines requested by the tenant as described above, and the bandwidth may be defined as a probability distribution.

The acceptance decision step S20 is a step of determining, by the acceptance control module, whether to accept or reject the probabilistic virtual network request of S10 in which the bandwidth requested by the tenant is defined as a probability distribution. S20 may be configured to determine whether the number of virtual machines requested by the tenant matches the number of virtual machines presently available and whether the bandwidth requested by the tenant is appropriate in view of the remaining bandwidth.

The VM allocation step S30 is performed by the VM allocation module of the network device according to an exemplary embodiment of the present invention after the request of the tenant is accepted in S20 and corresponds to the sub-tree having the largest statistical multiplexing gain Can be performed so that tenants are assigned.

Example

13 is a graph illustrating a simulation result of a resource allocation method according to a stochastic virtual network request according to an embodiment of the present invention. In FIG. 13, Ver.A and Ver.B use probabilistic virtual network requests, and Baseline and Oktopus use existing virtual network requests. Specifically, Ver.A is a case where a sub-tree is randomly determined when there are several sub-trees of the same level capable of allocating a virtual machine, and Ver.B is a case where a sub-tree is determined considering statistical multiplexing to be. As shown in FIG. 13, when using a stochastic virtual network request (red and blue) according to an embodiment of the present invention, network resources are more efficiently used than when using an existing virtual network request (black and yellow) Can be used to accommodate more tenants.

The size, location, and arrangement of the network devices in the single data center cloud computing environment described above are not limited to the drawings, but are preferably determined according to each situation in order to achieve the object of the present invention. In addition, the bandwidth in the probabilistic virtual network request according to an embodiment of the present invention is described as a normal probability distribution for convenience of explanation, and the scope of the present invention is not limited thereto. The probabilistic virtual network request according to an exemplary embodiment of the present invention may be applied to a discrete probability distribution such as a discrete uniform distribution, a Poisson distribution, a Bernoulli distribution, a geometric distribution, a hypothesis distribution, a binomial distribution, a negative binomial distribution, May be defined as continuous probability distributions such as normal distribution, continuous uniform distribution, chi-square distribution, gamma distribution, and the like.

The apparatus described above may be implemented as a hardware component, a software component, and / or a combination of hardware components and software components. For example, the apparatus and components described in the embodiments may be implemented within a computer system, such as, for example, a processor, a controller, an arithmetic logic unit (ALU), a digital signal processor, a microcomputer, a field programmable array (FPA) A programmable logic unit (PLU), a 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. The processing device may also access, store, manipulate, process, and generate data in response to execution of the software. For ease of understanding, the processing apparatus may be described as being used singly, but those skilled in the art will recognize that the processing apparatus may have a plurality of processing elements and / As shown in FIG. For example, the processing unit may comprise a plurality of processors or one processor and one controller. Other processing configurations are also possible, such as a parallel processor.

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.

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. The program instructions to be recorded on the medium may be those specially designed and configured 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 devices described above may be configured to operate as one or more software modules to perform the operations of the embodiments, and vice versa.

The network device (for example, a router, a switch, etc.) used above may mean one of network means consisting of hardware or software for connecting another network device or an end system. The network device may also be multiple services network devices capable of including at least one of routing, bridging, VLAN switching, Layer 2 aggregation, Session border control (SBC), QoS, and subscriber management functions. The network device may also include functionality to provide multiple application services (e. G., Data, voice, video, etc.). Network devices also include subscriber end systems (e.g., servers, workstations, laptops, netbooks, tablets, mobile phones, smart phones, VoIP phones, portable media players, GPS units, gaming systems, Can access a specific content or service through the Internet, VPN, tunneling, and the like. A particular content or service may typically be provided by at least one end systems (e.g., a server end system), and such content or service may be provided by a particular provider, P2P or public webpage &Lt; / RTI &gt;

Such a network device may be specified as a media access address (MAC), an Internet Protocol (IP) address / subnet, network sockets / ports, or upper OSI layer identifiers. The network device may also include at least one line card set, a control card set, and at least one service card set (which may also be referred to as a resource card). The service card may be, for example, a firewall, an Internet Protocol Security (IPsec), an Intrusion Detection System (IDS), a Peer-to-Peer, a Voice over IP (VoIP), a Session Border Controller (SBC) Radio Service), and EPC (Evolved Packet Core).

Further, the network interface of the network device may be configured as a virtual interface or a physical interface. In addition, the physical network interface is composed of network device hardware in a network state, which is composed of, for example, a wireless network interface controller (WNIC) or a plug-in network interface controller .

As described above, those skilled in the art will appreciate that the present invention may be embodied in other specific forms without departing from the spirit or essential characteristics thereof. It is therefore to be understood that the above-described embodiments are to be considered in all respects only as illustrative and not restrictive. The scope of the present invention is defined by the appended claims rather than the detailed description, and all changes or modifications derived from the meaning and scope of the claims and their equivalents should be construed as being included in the scope of the present invention.

1: Network device
2: Request Receive Module
3: resource allocation device
30: Request signal converter
32: Accept Control Module
34: VM allocation module

Claims (13)

A method for requesting a probabilistic virtual network performed by a request receiving module of a network device to provide a virtual network to a tenant in a single data center cloud computing environment,
A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of a desired virtual machine in the tenant; And
A request transmission step of transmitting the virtual network request signal to a resource allocation device which is a constitution of the network device;
Lt; / RTI &gt;
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A Stochastic Virtual Network Request Method in a Single Data Center Cloud Computing Environment. The method according to claim 1,
In the virtual network request signal, the bandwidth may be expressed as a normal distribution,
Wherein the virtual network request signal includes a number of desired virtual machines in the tenant, an average value of the desired bandwidth in the tenant, a variance value of the desired bandwidth in the tenant, and a probability value in which the desired bandwidth in the tenant is guaranteed.
A Stochastic Virtual Network Request Method in a Single Data Center Cloud Computing Environment. There is provided a computer-readable recording medium on a computer for performing a stochastic virtual network request method performed by a request receiving module of a network device to provide a virtual network to a tenant in a single data center cloud computing environment,
A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of a desired virtual machine in the tenant; And
A request transmission step of transmitting the virtual network request signal to a resource allocation device which is a constitution of the network device;
Lt; / RTI &gt;
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A method of requesting a stochastic virtual network in a single data center cloud computing environment. A program stored in a recording medium such that a probabilistic virtual network request method performed by a request receiving module of a network device to provide a virtual network to a tenant in a single data center cloud computing environment is performed on the computer,
A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of a desired virtual machine in the tenant; And
A request transmission step of transmitting the virtual network request signal to a resource allocation device which is a constitution of the network device;
Lt; / RTI &gt;
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A program stored on a recording medium such that a probabilistic virtual network request method in a single data center cloud computing environment is performed on a computer. A network device for providing a virtual network to the tenant in a single data center cloud computing environment and receiving a virtual network request signal from the tenant and for receiving the virtual network request signal from the tenant, A request receiving module for transmitting the resource allocation request to the resource allocation apparatus;
/ RTI &gt;
Wherein the virtual network request signal includes a number and a bandwidth of a desired virtual machine in the tenant,
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A device for receiving probabilistic virtual network requests in a single data center cloud computing environment. A method of resource allocation performed by a network device to provide a virtual network to a tenant in a single data center cloud computing environment,
A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of a desired virtual machine in the tenant;
Determining whether the received virtual network request signal is acceptable or not; And
A virtual machine allocation step of allocating a virtual machine to the tenant based on the accepted virtual network request signal;
Lt; / RTI &gt;
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A method of resource allocation in a single data center cloud computing environment. The method according to claim 6,
In the virtual network request signal, the bandwidth may be expressed as a normal distribution,
Wherein the virtual network request signal includes a number of desired virtual machines in the tenant, an average value of the desired bandwidth in the tenant, a variance value of the desired bandwidth in the tenant, and a probability value in which the desired bandwidth in the tenant is guaranteed.
A method of resource allocation in a single data center cloud computing environment. 8. The method of claim 7,
In the acceptability determination step,
Determining that the received virtual network request signal is acceptable if the bandwidth of the virtual link satisfies the following equation:
A method of resource allocation in a single data center cloud computing environment.
[Mathematical Expression]

Here, _L B is the bandwidth of the virtual _{link, μ = μ e + μ n} · min {i, N n -i}, σ 2 = σ e 2 + σ n 2 · min {i, N n -i} , 1-ε = max {1 -ε e, 1-ε n} a, F refers to the cumulative density function (cumulative density function) of the standard normal distribution _{and, <N n, μ n,} σ n, ε n> It is the imaginary mean stochastic hose model of network request signal and, i is an entire virtual machine of the particular sub-means a virtual machine assigned to the tree, and the sum of all the traffic that is already going through the virtual link is <μ _e, σ _e > and has a service quality of ε _e . 8. The method of claim 7,
In the virtual machine allocation step,
Wherein a virtual machine is allocated in the order of sub-trees having a large DELTA value in the following equation: < RTI ID = 0.0 >
A method of resource allocation in a single data center cloud computing environment.
[Mathematical Expression]

Where DELTA denotes the statistical multiplexing gain of the sub-tree and the bandwidth of the virtual network request signal to use the link of the particular sub-tree follows a normal distribution of &_lt; [mu] _n , [sigma] _n & The sum of the traffic follows a normal distribution of <μ _e , σ _e >. A recording medium on which a program for performing a resource allocation method performed by a network device on a computer to provide a virtual network to a tenant in a single data center cloud computing environment,
A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of a desired virtual machine in the tenant;
Determining whether the received virtual network request signal is acceptable or not; And
A virtual machine allocation step of allocating a virtual machine to the tenant based on the accepted virtual network request signal;
Lt; / RTI &gt;
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A method of allocating resources in a single data center cloud computing environment on a computer. A program stored in a recording medium such that a resource allocation method performed by a network device is performed on a computer to provide a virtual network to a tenant in a single data center cloud computing environment,
A request reception step of receiving a virtual network request signal of the tenant including the number and the bandwidth of a desired virtual machine in the tenant;
Determining whether the received virtual network request signal is acceptable or not; And
A virtual machine allocation step of allocating a virtual machine to the tenant based on the accepted virtual network request signal;
Lt; / RTI &gt;
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A program stored on a recording medium such that a method of resource allocation in a single data center cloud computing environment is performed on a computer. 1. A resource allocation device configured in a network device for providing a virtual network to a tenant in a single data center cloud computing environment,
An acceptance control module for receiving a virtual network request signal from the request receiving module, which is a configuration of the network device, and determining whether the received virtual network request signal is accepted; And
A virtual machine allocation module for receiving the virtual network request signal accepted by the admission control module and for allocating a virtual machine to the tenant based on the accepted virtual network request signal;
/ RTI &gt;
Wherein the virtual network request signal includes a number and a bandwidth of a desired virtual machine in the tenant,
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A resource allocation device in a single data center cloud computing environment. 1. A network device for providing a virtual network to a tenant in a single data center cloud computing environment,
A request receiving module for receiving a virtual network request signal from the tenant;
An acceptance control module for receiving the virtual network request signal from the request reception module and determining whether the received virtual network request signal is accepted; And
A virtual machine allocation module for receiving the virtual network request signal accepted by the admission control module and for allocating a virtual machine to the tenant based on the accepted virtual network request signal;
/ RTI &gt;
Wherein the virtual network request signal includes a number and a bandwidth of a desired virtual machine in the tenant,
Characterized in that the bandwidth in the virtual network request signal is in the form of a probability distribution,
A network device in a single data center cloud computing environment.

Images