KR102053596B1 - Method and apparatus of network slicing by using dynamic network traffic analysis based on software defined networking - Google Patents

Abstract

Provided are a network slicing method through SDN-based dynamic network traffic analysis and an apparatus thereof. According to an embodiment of the present invention, a network control method comprises the steps of: dynamically collecting information related to a bandwidth usage amount of a network slice included in a network from one or more SDN switches included in the network using an open-flow protocol; calculating the bandwidth usage amount of the network slice using information related to the bandwidth usage amount; calculating a bandwidth control amount of the network slice on the basis of the bandwidth usage amount; loading a static weight set on the network slice; calculating a dynamic weight using the bandwidth usage amount, a frequency of bandwidth change of the network slice, an allocation bandwidth of the network slice, and the number of times of bandwidth control reserves of the network slice; updating a bandwidth increase list by using an ID of the network slice, the bandwidth control amount, the static weight, and the dynamic weight when the bandwidth control amount is negative; aligning a bandwidth increase list using the static weight and the dynamic weight; and controlling a network using the aligned bandwidth increase list.

Description

METHOD AND APPARATUS OF NETWORK SLICING BY USING DYNAMIC NETWORK TRAFFIC ANALYSIS BASED ON SOFTWARE DEFINED NETWORKING}

The present invention relates to an SDN-based network slicing method and apparatus. More specifically, the present invention relates to a method and apparatus for performing optimal network slicing through dynamic network traffic analysis in performing network slicing in an SDN-based network.

Software Defined Networking (SDN) technology refers to a network technology that separates a control plane and a data plane from existing network equipment and forms a control plane in software to implement a centralized network architecture. In SDN, the control plane tells the network what data is going to where, and the data plane is responsible for actually sending the data to its designated destination under control of the control plane. The network devices that make up an SDN network are switches that can be programmed through an SDN controller using industry standard control protocols such as OpenFlow, and, with SDN technology, one network device for all network devices that make up an SDN network. Consistent control through the SDN controller enables network management and network efficiency. It also enables independent network management that is not dependent on the hardware design of the network equipment manufacturer.

On the other hand, efficient network resource distribution technology is needed to effectively reflect the requirements of various services. Conventionally, network resources have been distributed assuming that all nodes connected to the network generate saturated traffic. In this case, when there are a large number of nodes generating unsaturated traffic or supersaturated traffic, network resource distribution is inefficient. In particular, the delay-sensitive service providing technology based on Mobile Edge Computing (MEC) technology aims to minimize service latency by providing a service through the nearest server among a plurality of distributed servers. However, in the conventional technology, since the service was provided along a static traffic path based on access policy information, the conventional technology did not adequately cope with changes in the network environment or traffic bottlenecks, and there is no suitable method for providing a delay sensitive service. Did.

As a technology for flexible network resource distribution, network virtualization refers to a technology for configuring a plurality of logical networks by dividing a resource of a physical network infrastructure. Network virtualization can provide a flexible network that reflects the requirements of various services.

In the virtual network, the network bandwidth can be flexibly set on the logical network without being dependent on the physical network, and thus, the bandwidth can be actively adjusted for frequent traffic change and network change requests. When dividing a physical network infrastructure into multiple independent virtual networks, each separate virtual network is called a network slice, and the task of creating a network slice is called network slicing.

The present invention relates to a method and apparatus for performing optimal network slicing through dynamic network traffic analysis in a network in which SDN technology is implemented.

Korean Patent Publication No. 10-2016-0036182 (published Apr. 4, 2016)

The technical problem to be solved by the present invention relates to a method and apparatus for controlling a network using a result of analyzing network traffic between hosts in an SDN-based network.

Another technical problem to be solved by the present invention relates to a method and apparatus for slicing a network using a result of analyzing network traffic between hosts connected to the network.

Another technical problem to be solved by the present invention is to analyze a network traffic between hosts connected to a network, and then to a method and apparatus for controlling a network slice so as to increase a bandwidth of a network slice including a host with high network traffic. will be.

Another technical problem to be solved by the present invention is to analyze the network traffic between the hosts connected to the network, if a plurality of network slices need to increase the bandwidth, each network slice is given a priority to give priority to the network How to distribute network resources evenly among all hosts by increasing the bandwidth of the slice first, but giving priority to subsequent slices by assigning weights according to the number of reservations for network slices whose bandwidth increase is reserved due to lack of available bandwidth And to an apparatus.

The technical problems of the present invention are not limited to the technical problems mentioned above, and other technical problems not mentioned will be clearly understood by those skilled in the art from the following description.

Network control method according to an embodiment of the present invention for solving the technical problem, SDN (Network Defined Networking) based network control method, one or more included in the network using an OpenFlow (OpenFlow) protocol Calculating a bandwidth usage of a network slice included in the network from an SDN switch, calculating a bandwidth adjustment amount of the network slice based on the bandwidth usage, loading a static weight set in the network slice, the network slice Calculating a dynamic weight of the network slice using the number of reserved bandwidth adjustments of the network slice; if the calculated bandwidth adjustment amount is negative, the loaded static weight and the calculated dynamic weight value are used in the bandwidth increase list. network Updating the priority of the rice and controlling the network according to the priority of the updated bandwidth increase list, wherein controlling the network comprises: if the bandwidth adjustment is greater than the available bandwidth, the network slice Suspending the bandwidth adjustment and increasing the bandwidth adjustment reservation count, wherein the higher the static weight or the dynamic weight value is, the higher the priority is, and the higher the bandwidth adjustment reservation count is, the greater the dynamic weight is. Can be calculated as a value.

According to an embodiment of the present invention, the network can be efficiently controlled by using the result of analyzing network traffic between hosts in the SDN-based network.

According to another embodiment of the present invention, by slicing the network using the result of analyzing the network traffic between the hosts connected to the network, it is possible to maximize the use of network resources without lack or waste of the network.

According to another embodiment of the present invention, after analyzing the network traffic between the hosts connected to the network, by controlling the network to increase the bandwidth of the network slice including the host with a lot of network traffic to meet the lack of bandwidth in real time Can be.

According to another embodiment of the present invention, when network traffic between hosts connected to a network is analyzed and bandwidth increase is required for a plurality of network slices, priority is given to each network slice to determine a network slice having a high priority. The network slices can be distributed evenly to all hosts by increasing the bandwidth first, but giving priority to subsequent slices by assigning weights according to the number of reservations for network slices whose bandwidth increase is reserved due to the lack of available bandwidth.

According to another embodiment of the present invention,

The effect of the present invention is not limited to the above-mentioned effects, and other effects not mentioned may be exerted in the technical field of the present invention.

1 is a diagram illustrating a structure of a general SDN network system.
2 is a diagram comparing network slicing before and after.
3 is a diagram illustrating an entire network according to an embodiment of the present invention.
4 is a diagram illustrating a network slice according to an embodiment of the present invention.
5 is a diagram illustrating bandwidth adjustment of a network slice according to an embodiment of the present invention.
6 is a flowchart illustrating a network control method according to an embodiment of the present invention.
7 is a flowchart illustrating a process of changing a bandwidth using a bandwidth increase list in a network control method according to an embodiment of the present invention.
8 is a diagram illustrating a network control apparatus according to an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. Advantages and features of the present invention and methods for achieving them will be apparent with reference to the embodiments described below in detail with the accompanying drawings. However, the present invention is not limited to the embodiments disclosed below, but may be implemented in various forms, and only the embodiments are intended to complete the disclosure of the present invention, and the general knowledge in the art to which the present invention pertains. It is provided to fully inform the person having the scope of the invention, which is defined only by the scope of the claims. Like reference numerals refer to like elements throughout.

Unless otherwise defined, all terms used in the present specification (including technical and scientific terms) may be used as meanings that can be commonly understood by those skilled in the art. In addition, terms that are defined in a commonly used dictionary are not ideally or excessively interpreted unless they are specifically defined clearly. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. As used herein, the singular forms "a", "an" and "the" include the plural unless specifically stated otherwise.

As used herein, the phrase 'comprises' or 'comprising' does not exclude the presence or addition of any other configuration / step / operation other than the configuration / step / operation mentioned.

Hereinafter, embodiments of the present invention will be described with reference to the drawings.

1 is a diagram illustrating a structure of a general SDN network system.

As shown in FIG. 1, the SDN network system may include an SDN controller 110, a plurality of SDN switches 121, 122, and 123, and a plurality of network devices 131, 132, and 133.

SDN separates the control plane and data plane from existing network equipment and implements a centralized network architecture by making the control plane in software form. The SDN controller 110 may be a centralized control unit that is in charge of a control plane. The plurality of SDN switches 121, 122, 123, and the plurality of network devices 131, 132, and 133 may be network equipment that is in charge of a data plane.

The SDN controller 110 is a kind of command computer that controls the SDN system, and can perform various and complex functions, such as routing, policy declaration, and security check. The SDN controller 110 may define a flow of packets generated from a plurality of SDN switches 121, 122, and 123 in a lower layer. The SDN controller 110 may refer to a network topology and the like to allow flows to be allowed by network policy. After calculating the path (data path), the entry of the flow can be set to a switch on the path. The SDN controller 110 may communicate with the SDN switches 121, 122, 123 using a specific protocol, for example, an OpenFlow protocol. At this time, the communication channel between the SDN controller 110 and the SDN switches 121, 122, 123 may be encrypted for security.

Network devices 131, 132, 133 are controlled by SDN switches 121, 122, 123. The network devices 131, 132, and 133 may be terminal devices that transmit or receive data or information, for example, communication devices or routers that transmit and receive data. The network devices 131, 132, 133 may also be referred to as network functions that perform various functions on the network. The network function can be virtualized using NFV (Network Function Virtualiztion). NFV abstracts network functionality, enabling you to install, control, and manipulate network functionality through software running on standardized computing nodes.

2 is a diagram comparing network slicing before and after.

In the general network before performing network slicing, various services were provided through one network, and as a result, there was a problem in that it did not reflect network requirements for each service.

In the network illustrated in FIG. 2 (a), a smartphone is connected to a network to receive communication / internet services, a sensor is connected to a network to provide logistics, agriculture, and climate services, and a car is connected to a network to autonomously drive. Get service. In the network illustrated in FIG. 2 (a), it is difficult to meet network requirements for each service since all devices are provided with the service through the same network.

FIG. 2B is a diagram illustrating a network after performing network slicing on the network illustrated in FIG. 2A. As shown in FIG. 2 (b), when the network slice is divided into a network slice applied to a smartphone, a network slice applied to a sensor, and a network slice applied to a car, the network environment of each network slice is different. By setting it up, it is possible to meet the network requirements of each service.

Network slicing is described in detail. Network slicing is a method in which a physical network infrastructure is divided into several parts by using virtualization technology, like a hard disk is divided into partitions such as a C drive and a D drive. Means that. Network slicing involves applying software virtualization technology to a common physical network infrastructure to create multiple virtual networks, with multiple logical networks running on top of the physical network infrastructure. In this case, each of the separated logical networks (virtual networks) is called a network slice. The network slice is a concept similar to a virtual machine (VM) in server virtualization.

The service provided through the network differs in the required network performance depending on the type of service provided. For example, a network related to autonomous vehicles reacts immediately to risks on the road, so 1 ms (0.001 sec) vs. 'low latency' performance is important. Water companies, on the other hand, are rather slow, but the 'ultra-capacity' capability of being able to send small amounts of data simultaneously from thousands of devices is important.

Services such as Mobile Broadband, Massive IoT, and Mission-critical IoT, which are representative use-cases in the 5G standard, also have different network requirements in terms of mobility, charging, security, police control, latency, and reliability. In the case of Massive IoT service where fixed sensors that measure rainfall, rainfall, etc. are connected to the network, functions such as handover and locate update are unnecessary. Mission-critical IoT services, such as autonomous driving or remote industrial robot control, require low latency within a few ms.

As described above, when the network requirements are different for each service, a service is provided through a single network without considering the requirements of each service, or a separate network is constructed for each service to satisfy different requirements for each service. In the case of establishing a separate network for each service, when the service is terminated, the existing network has to be discarded or redesigned.

However, network slicing technology makes it easy to build multiple networks to meet service-specific requirements. The physical network infrastructure can be broken down into multiple virtual networks (network slices), and each virtual network (network slice) can be optimized for separate application, service, device and operator requirements.

Network slicing does not just divide the network into multiples, but features a customizable configuration for each network slice. For example, some network slices can be optimized for HD video streaming, some network slices are secure, and some network slices can be connected to many devices. As many services are provided through the network, the types of network traffic vary widely, and it is difficult for the existing communication networks to optimize the network according to the traffic. However, the network slicing technology enables the network to be customized as needed.

Each network slice implemented in software technology operates completely independently. If one network slice fails or a cyber attack cannot interfere with another network slice, the network of the other network slice is not affected. Therefore, in the case of services where data security is important, security can be enhanced by using a network slice specialized for security.

3 is a diagram illustrating an entire network according to an embodiment of the present invention.

As shown in FIG. 3, the entire network according to an embodiment of the present invention may include a plurality of nodes 310, 320, 330, 340, 350, and 360. Each node has one or more hosts connected to it. In this case, the host may be a server. The server may be a host that accepts connections from other devices (clients) among devices that have established a connection to the network.

An example of a network according to an embodiment of the present invention may be a KREONET (Korea Research Environment Open Network) network of the Korea Institute of Science and Technology Information (KISTI). KREONET has been in operation since 1988 and supports high-performance network infrastructure for about 200 research institutes in industries and research institutes. In the KREONET network, each research institute may be viewed as a node 310, 320, 330, 340, 350, 360, and each node may be connected with one or more servers.

As shown in FIG. 3, before performing network slicing, each node uses the same network without particular limitation. As a result, increased traffic between specific nodes affects the communication of other nodes, which may affect nodes that require more than a certain level of network performance.

4 is a diagram illustrating a network slice according to an embodiment of the present invention.

As shown in FIG. 4, a network slice according to an embodiment of the present invention may be set between a plurality of nodes 310, 320, 330, 340, 350, and 360.

In a network connecting the first node 310 and the second node 320, the network is sliced using a network virtualization technology to form a first network slice (hatched portion) between the first node 310 and the second node 320. Can be communicated through.

In a network connecting the first node 310 and the fourth node 340, the network is sliced using a network virtualization technology so that the first node 310 and the fourth node 340 are connected through a first network slice (hatched portion). To communicate.

In a network connecting the second node 320 and the third node 330, the network is sliced using a network virtualization technology so that the second node 320 and the third node 330 are connected through a second network slice (dot part). To communicate.

5 is a diagram illustrating bandwidth adjustment of a network slice according to an embodiment of the present invention.

When N network slicing is performed, all N + 1 network slices are generated including the default network slice. Each network slice behaves like a completely independent network, so different network policies can be established for each network slice.

As an example of establishing a different network policy for each network slice, a network bandwidth allocated for each network slice may be set differently.

5 illustrates the result of changing the bandwidth of each network slice when the bandwidth of the physical network is 1000 Mbps.

Since network bandwidth is a resource defined by a physical network, as the bandwidth of one network slice increases, the bandwidth of the remaining network slices decreases.

As shown in FIG. 5 (a), when the bandwidth of the physical network is 1000 Mbps and no network slicing is performed, a bandwidth of up to 1000 Mbps is guaranteed. A network that does not perform network slicing can also be viewed as a slice. For convenience, this is called a default network slice. While a network slice performs communication between defined nodes, a basic network slice performs communication between any node connected to the network.

As shown in FIG. 5 (b), when one network slicing is performed in a network having a bandwidth of 1000 Mbps, two network slices are formed including a default network slice. Bandwidth can be allocated for sliced network slices individually, and the overall bandwidth is limited by the bandwidth of the physical network. 5 (b) allocates a bandwidth of 300 Mbps to the first network slice, and as a result, the basic network slice has a bandwidth of 700 Mbps.

As shown in FIG. 5 (c), when the bandwidth of the physical network is 1000 Mbps and two network slicings are performed, three network slices are formed including a default network slice. Bandwidth can be allocated individually for each network slice, and the overall bandwidth is limited by the bandwidth of the physical network. 5 (c) allocates a bandwidth of 300 Mbps to the first network slice and a bandwidth of 400 Mbps to the second network slice. As a result, the basic network slice has a bandwidth of 300 Mbps.

The present invention relates to a network control method for measuring and analyzing available bandwidth of a network slice provided to an arbitrary user group based on SDN, and dynamically adjusting bandwidth of each network slice accordingly.

The present invention determines priorities in terms of efficient bandwidth usage, fair bandwidth allocation, minimizing the number of changes in bandwidth, and performs intelligent bandwidth management based on priorities.

The bandwidth increase / decrease setting of the network slice performs a network slice update that corrects a required bandwidth based on a related technology or protocol, for example, a meter table of OpenFlow.

Network slicing according to an embodiment of the present invention is a method of creating and providing a virtual dedicated network that guarantees a specific bandwidth for an arbitrary group of users based on SDN, and more flexible and detailed operation based on an open flow meter. E2E Slicing is performed to perform dynamic bandwidth allocation and control.

That is, the network slice according to an embodiment of the present invention may be an end to end (E2E) slice that guarantees a preset bandwidth among a plurality of hosts belonging to the network slice.

In addition, as the conventional traffic control method is limited to the SDN environment, the network control method according to an embodiment of the present invention is connected to each SDN internal gateway, in contrast to the conventional traffic control method in an environment associated with an IP network. Since the target path is appropriately selected in consideration of the external IP network information, there is an advantage that the traffic can be dynamically controlled in accordance with the network environment change by applying to the IP-based delay sensitive service providing technology.

In particular, the network control method according to an embodiment of the present invention, when the user (IP network) receives a large amount of data from the server (SDN network), the more sensitive the service is to the delay time, the more efficient the traffic control Can be. For example, when providing a delay-sensitive downstreaming service such as OTT or cloud configured based on SDN to a user from a network operator or service provider, the network control method of the present invention can significantly reduce the delay rate of the service. There is an advantage.

6 is a flowchart illustrating a network control method according to an embodiment of the present invention.

The network control method according to an embodiment of the present invention may be performed by a computer or an SDN controller connected to a network.

As shown in FIG. 6, the network control method according to an exemplary embodiment of the present invention uses information related to bandwidth usage of a network slice included in a network from one or more SDN switches included in the network using an OpenFlow protocol. Dynamically collecting (S610), calculating bandwidth usage of the network slice using information related to the bandwidth usage (S620), calculating bandwidth adjustment amount of the network slice based on the bandwidth usage (S630), and network Loading static weights set in the slice (S640), calculating dynamic weights using the bandwidth usage, the frequency of bandwidth variation of the network slice, the allocated bandwidth of the network slice, and the number of reserved bandwidth adjustments of the network slice (S650), and the bandwidth If the adjustment is negative, the network slot Updating the bandwidth increase list by using ID, bandwidth adjustment amount, static weight, and dynamic weight (S660), sorting the bandwidth increase list by using static weight and dynamic weight (S670), and sorted bandwidth increase list. It may include the step of controlling the network using (S680).

Using the OpenFlow protocol, information related to bandwidth usage of a network slice included in a network may be dynamically collected from one or more SDN switches included in the network (S610).

Dynamically collecting information related to bandwidth usage of the network slice (S610) may include receiving packet information via the corresponding SDN switch from the SDN switches belonging to each node constituting the network using the open flow meter information. It may include. There are many kinds of data that can be collected through the OpenFlow protocol, and the number is increasing with each version of OpenFlow. For example, with Openflow version 1.3, the port number of the receiving port (in_port), the receiving port's physical port number (in_phy_port) from the SDN switch, and the metadata used to pass information between tables, Ethernet destination MAC address (eth_dst), Ethernet source MAC address (eth_src), Ethernet frame type (eth_type), VLAN ID (vlan_vid), VLAN PCP (vlan_pcp), IP DSCP (ip_dscp), IP ECN (ip_ecn), IP protocol type (ip_proto), source IP address of IPv4 (ipv4_src), destination IP address of IPv4 (ipv4_dst), TCPd source port number (tcp_src), TCP destination port number (tcp_dst), UDP source port number (udp_src), UDP destination port number (udp_dst), SCTP source port number (sctp_src), SCTP destination port number (sctp_dst), ICMP Type (icmpv4_type), ICMP Code (icmpv4_code), ARP enable code (arp_op), ARP source IP address (arp_spa ), ARP's destination IP address (arp_tpa), ARP's source MAC note (arp_sha), ARP destination MAC address (arp_tha), IPv6 source IP address (ipv6_src), IPv6 destination IP address (ipv6_dst), IPv6 flow label (ipv6_flabel), ICMPv6 Type (icmpv6_type), ICMPv6 Code ( icmpv6_code), IPv6 Naver Discovery Target Address (ipv6_nd_target), IPv6 Naver Discovery Source Link Layer Address (ipv6_nd_sll), IPv6 Naver Discovery Target Link Layer Address (ipv6_nd_tll), MPLS Label (mpls_label), MPLS Traffic Class (mpls_tc), MPLS Information about a BoS bit (mpls_bos), an I-SID (pbb_isid) of an 802.1ah PBB, metadata about a logical port (tunnel_id), and a pseudo field (ipv6_exthdr) of an IPv6 extension header may be collected.

Meanwhile, when a network slice add / update / delete request is generated, the meter table information of an access switch of a host end belonging to a path of the corresponding network slice is added, updated, or deleted. If 1) the same link of network slice paths overlaps, and 2) two or more branched paths exist at both ends of the network slice paths, control is performed on all switches belonging to the network slice, not the host access switch list.

When a data request occurs between any two hosts in the network, a flow rule is generated and a meter is mapped based on the network slice information. In this case, the meter identifier may be used by generating a UUID (Universally Unique IDentifier) based on a result of performing a hash function that inputs a network identifier and a network slice identifier. By using the same meter identifier for all host access switches in the network slice, the number of meter management table entries can be reduced.

Using the OpenFlow protocol, information collected from one or more SDN switches included in a network may be used to collect information related to bandwidth usage between a plurality of hosts connected to the network in real time.

After calculating bandwidth usage among a plurality of hosts connected to the network, determining the network slice to which each host belongs, and calculating the bandwidth usage of the network slice included in each network (S620).

When the bandwidth usage is calculated, the bandwidth adjustment amount of the network slice may be calculated based on the bandwidth usage (S630).

The step S630 of calculating the bandwidth adjustment amount of the network slice based on the bandwidth usage may include loading the bandwidth allocated to the network slice and subtracting the bandwidth usage from the allocated bandwidth. That is, after checking the bandwidth allocated to the corresponding network slice, the bandwidth adjustment amount of the network slice may be calculated by subtracting the bandwidth usage of the corresponding network slice from the bandwidth allocated to the corresponding network slice.

For example, if the bandwidth allocated to a network slice is 300Mbps and the bandwidth usage of the network slice is 400Mbps, the bandwidth adjustment amount of the network slice may be calculated as -100Mbps. In this case, the bandwidth throttling amount is negative, and a negative bandwidth throttling amount indicates that the bandwidth allocated to the corresponding network slice needs to be increased.

As another example, if the bandwidth allocated to a network slice is 300Mbps and the bandwidth usage of the corresponding network slice is 100Mbps, the bandwidth adjustment amount of the network slice may be calculated as + 200Mbps. In this case, the bandwidth throttling amount is positive, and the positive bandwidth throttling amount indicates that the bandwidth allocated to the corresponding network slice can be reduced.

In order to determine a priority for adjusting bandwidth among the plurality of network slices, a static weight set in the network slice may be loaded (S640).

Loading the static weight set in the network slice (S640) includes loading a static weight value set by the network administrator according to the importance of the network slice. The static weight can be arbitrarily set by the network administrator according to the importance of the network slice. For example, the static weight may be a number between 0 and 1, and the static weight applied to the network slice to which network resources should be allocated first and foremost may be 1. A large weight value means that a network resource should be allocated to a corresponding network slice first, and when a bandwidth increase request is requested for a plurality of network slices, the bandwidth should be increased before other network slices.

Dynamic weights of the network slices may be calculated in order to determine a priority for adjusting bandwidth among the plurality of network slices (S650).

In the calculating of the dynamic weight of the network slice (S650), the dynamic weight of the network slice may be calculated using the bandwidth usage of the network slice, the frequency variation of the bandwidth of the network slice, the allocated bandwidth of the network slice, and the number of reserved bandwidth adjustments of the network slice. Can be (S650).

The formula for calculating the dynamic weight is clearly shown in Equation 1 using Equation.

In Equation 1, BWused is the bandwidth usage of the network slice, BWallocated is the bandwidth allocated to the network slice, Nskip (cur) is the number of bandwidth adjustment reservations of the network slice, and Nskip (max) is the maximum bandwidth adjustment reservation among all network slices. The number of times, Nchange (cur) is the frequency of the bandwidth change of the network slice, Nchange (max) is the maximum bandwidth change frequency of all network slices.

Nskip (max) means the maximum number of bandwidth adjustment reservations among all network slices. Specifically, Nskip (max) is the number of bandwidth adjustment reservations among the network slices of all network slices included in the current network. The number of reserved bandwidth adjustments. Nskip (cur) divided by Nskip (max) is for data normalization.

Nchange (max) is the maximum frequency of bandwidth variation among all network slices. Specifically, Nchange (max) is the frequency of bandwidth variation of the network slice with the largest bandwidth variation among all network slices included in the current network. Means. Nchange (cur) divided by Nchange (max) is for data normalization.

As shown in Equation 1, as the bandwidth usage (BWused) is larger than the allocation bandwidth (BWallocated) of a network slice, the dynamic weight of the corresponding network slice may be calculated. For example, the dynamic weighting when the allocated bandwidth is 300 Mbps and the bandwidth usage is 700 Mbps can be calculated more than when the allocated bandwidth is 300 Mbps and the bandwidth usage is 400 Mbps. This is because when the bandwidth usage (BWused) is larger than the allocation bandwidth (BWallocated), communication in the network slice can be performed smoothly only when the bandwidth allocated to the network slice is urgently changed.

As shown in Equation 1, as the number of bandwidth adjustment reservations Nskip (cur) of the corresponding network slice is larger than the maximum number of bandwidth adjustment reservations Nskip (max) among all network slices included in the network, the dynamic weight is increased. Can be. For example, 9 maximum bandwidth throttling reservations among all network slices included in the network, rather than 9 maximum bandwidth throttling reservations among all network slices in the network and 1 bandwidth throttling reservation with that network slice. And when the number of reserved bandwidth adjustments of the corresponding network slice is seven, the dynamic weight can be largely calculated. If the bandwidth of the network slice is too high, the allocation bandwidth needs to be increased.However, if the allocation bandwidth of the network slice cannot be increased due to the lack of network resources or the increase of the allocation bandwidth of another network slice with priority, the allocation bandwidth of the network slice The adjustment is reserved and the number of bandwidth adjustment reservations is increased by one.

Although network resources are preferably allocated to network slices that are currently busy, so that bandwidth can be efficiently used, fair bandwidth allocation should be considered along with efficient bandwidth usage. If the static weight and dynamic weight of one network slice are greater than the static weight and dynamic weight of another network slice, the prior assignment of network resources to only that network slice will result in the network slice monopolizing the network resources. As a result, communication in other network slices may not be smooth. Therefore, in view of fair bandwidth allocation, it is desirable to increase the priority of network resource allocation for network slices having a large number of reserved bandwidth adjustments.

As shown in Equation 1, as the bandwidth variation frequency Nchange (cur) of the corresponding network slice is larger than the maximum bandwidth variation frequency Nchange (max) among all network slices included in the network, the dynamic weight may be larger. . For example, 20 times the maximum bandwidth variation among all network slices included in the network, compared to 20 times / minute of the maximum bandwidth variation among all network slices included in the network and 10 times / minute of the bandwidth slice of the network slice. The dynamic weight can be largely calculated when / min and the frequency of the bandwidth change of the corresponding network slice is 17 times / min.

If bandwidth usage intermittently exceeds the bandwidth allocated to the network slice, it may be a result of temporary traffic increase, and thus it is not necessary to increase the allocated bandwidth of the network slice. However, if the bandwidth usage occurs more than the bandwidth allocated to the network slice, it can be regarded as continuous traffic increase, not temporary traffic increase. Therefore, it is recommended to increase the allocated bandwidth of the network slice so that smooth communication of the network slice is achieved. desirable.

If the bandwidth adjustment amount is negative, the bandwidth increase list may be updated using the bandwidth adjustment amount, the static weight, and the dynamic weight of the network slice (S660).

The bandwidth increase list is a list of network slices that require bandwidth increase. If the bandwidth adjustment amount of the network slice is negative, information about the network slice may be inserted into the bandwidth increase list. If the bandwidth adjustment amount of the network slice is positive, it is preferable to insert information about the network slice into a separate bandwidth reduction list. The bandwidth reduction list will be described later.

The bandwidth increase list may be a priority queue. The bandwidth increase list may store information about each network slice, and the information about each network slice may include ID, bandwidth adjustment amount, static weight, and dynamic weight of each network slice. Network slices that are not inserted in the bandwidth increase list may be enqueued to the bandwidth increase list by using priority of information about the network slices calculated by the static weight and the dynamic weight of the network slice. In the case of a network slice already inserted in the bandwidth increase list, the information in the bandwidth increase list may be updated, and the order in the bandwidth increase list may be adjusted by the priority calculated by the static weight and the dynamic weight.

Alternatively, the bandwidth increase list may be a queue, stack or array. The bandwidth increase list may store information about each network slice, and the information about each network slice may include ID, bandwidth adjustment amount, static weight, and dynamic weight of each network slice. For a network slice that is not inserted in the bandwidth increase list, information about the network slice may be inserted in the bandwidth increase list. In the case of a network slice already inserted in the bandwidth increase list, the information in the bandwidth increase list may be updated. After the insertion into the bandwidth increase list or the update of the information is finished, the bandwidth increase list may be sorted by the priority calculated by the static weight and the dynamic weight of each network slice included in the bandwidth increase list (S670).

If the bandwidth increase list is a priority queue, the bandwidth increase list is stored in an ordered state. If the bandwidth increase list is not a priority queue but a data structure such as another queue, stack, or array, the bandwidth increase list is stored in random order and then sorted by performing a separate sort operation.

After obtaining the sorted bandwidth increase list, the network may be controlled using the sorted bandwidth increase list (S680).

7 is a flowchart illustrating a process of changing a bandwidth using a bandwidth increase list in a network control method according to an embodiment of the present invention.

As shown in FIG. 7, the controlling of the network (S680) may include storing information about the network slice in the bandwidth reduction list when the amount of bandwidth adjustment is positive (YES in S681). S683).

The bandwidth reduction list is a list of network slices capable of bandwidth reduction. If the bandwidth adjustment amount of the network slice is positive, information about the network slice may be inserted into the bandwidth reduction list. If the bandwidth throttling amount of the network slice is positive, it is not necessary to immediately reduce the bandwidth, and it is desirable to perform the operation of reducing the bandwidth when there is another network slice requiring the bandwidth increase. The amount of bandwidth adjustment (ie, extra bandwidth) of the network slice inserted into the bandwidth reduction list can be taken into account in the available bandwidth calculation described below.

The bandwidth adjustment amount of the network slice included in the bandwidth reduction list is taken into account in the available bandwidth calculation, and the bandwidth increase of the network slice is required as the bandwidth adjustment amount of the network slice included in the bandwidth increase list is negative. The amount of bandwidth increase is calculated by adding the absolute value to the amount of bandwidth adjustment in which the bandwidth adjustment is negative. If the required bandwidth increase amount (absolute value of the bandwidth adjustment amount) is less than the available bandwidth, the bandwidth of the network slice included in the bandwidth reduction list is decreased, and then the bandwidth of the network slice included in the bandwidth increase list is increased.

Controlling the network (S680) may include determining whether the bandwidth adjustment amount is smaller than the available bandwidth if the bandwidth adjustment amount is negative (“N” in S681) (S682). .

If the bandwidth adjustment amount (absolute value of the bandwidth adjustment amount calculated as negative) is smaller than the available bandwidth (YES in S682), the bandwidth of the network slice included in the bandwidth reduction list is reduced (S684). When the bandwidth of the network slice included in the bandwidth reduction list is reduced, the bandwidth of the network slice in which the bandwidth adjustment amount is negative is increased (S686).

If the bandwidth adjustment amount is greater than the available bandwidth (No in S682), the bandwidth adjustment of the corresponding network slice is reserved. The step of suspending bandwidth adjustment of the network slice may include increasing the number of reserved bandwidth adjustments of the corresponding network slice (S685). The number of reserved bandwidth adjustments is reflected in the dynamic weight calculation, so that the larger the number of reserved bandwidth adjustments, the more priority is given to future network bandwidth increase.

8 is a block diagram illustrating a computing device according to an embodiment of the present invention.

As shown in FIG. 8, a computing device according to an embodiment of the present invention may include one or more processors 810, a memory 820 for loading a computer program executed by the processor 810, and one or more computer programs. Temporarily stored storage 830, and network interface 840.

Processor 810 controls the overall operation of each component of a computing device capable of implementing the method according to embodiments of the present invention. The processor 810 is configured to include a central processing unit (CPU), a micro processor unit (MPU), a micro controller unit (MCU), a graphics processing unit (GPU), or any type of processor well known in the art. Can be. In addition, the processor 810 may perform operations on at least one application or program for executing a method according to embodiments of the present invention. A computing device capable of implementing a network control method according to embodiments of the present invention may include one or more processors.

The memory 820 stores various data, commands, and / or information. The memory 820 may load one or more computer programs from the storage 830 to execute a network control method according to embodiments of the present invention.

Storage 830 may non-transitory store one or more computer programs. The storage 830 may be a non-volatile memory such as a read only memory (ROM), erasable programmable ROM (EPROM), electrically erasable programmable ROM (EPROM), flash memory, or the like, a hard disk, a removable disk, or a technical field to which the present invention pertains. It may comprise any known type of computer readable recording medium. The storage 830 stores a computer program for performing a network control method according to embodiments of the present invention.

The network interface 840 supports wired and wireless Internet communication of a computing device according to an embodiment of the present invention. In addition, the network interface 840 may support various communication methods other than Internet communication. To this end, the network interface 840 may be configured to include a communication module well known in the art.

The computer program for performing the network control method according to the embodiments of the present invention may be loaded into the memory 830 to include instructions for causing the processor 810 to perform the method according to the embodiments of the present invention. have.

Specifically, a computer program may dynamically obtain information regarding bandwidth usage of a network slice included in a network from at least one SDN switch included in the network using an OpenFlow protocol for a software defined network (SDN) based network. Instructions to collect, instructions to calculate bandwidth usage of network slices using information related to bandwidth usage, instructions to calculate bandwidth throttling of network slices based on bandwidth usage, instructions to load static weights set in network slices, bandwidth usage , Instructions for calculating dynamic weights using the frequency variation of the network slices, the allocated bandwidth of the network slices, and the number of times the network slices are reserved, and the amount of bandwidth adjustments is negative. Face, instructions to update the bandwidth increase list using the ID, bandwidth adjustment, static weights and dynamic weights of the network slice, instructions to sort the bandwidth increase list using static weights and dynamic weights, and an ordered list of bandwidth increase It may include instructions for controlling the network. Since the contents of each instruction are similar to those of the above-described network control method, redundant descriptions are omitted.

Those skilled in the art to which the present invention pertains may change the order of one or more steps, omit one or more steps, or one or more steps in each flowchart step of the present specification without departing from the essential characteristics of the present invention. Various steps may be variously modified and modified, such as executing steps in parallel.

The methods according to the embodiments of the present invention described so far may be performed by execution of a computer program implemented in computer readable code. The computer program may be transmitted to and installed on the second computing device from the first computing device via a network such as the Internet, and thus may be used in the second computing device. The first computing device and the second computing device both include a server device, a physical server belonging to a server pool for cloud services, and a stationary computing device such as a desktop PC.

The computer program may include a magnetic storage medium (for example, a ROM, a floppy disk, a hard disk, etc.), an optical reading medium (for example, a CD-ROM, a DVD, etc.), a flash memory (for example, a USB, an SSD), and the like. It may be stored in the same non-transitory medium.

Although the embodiments of the present invention have been described above with reference to the accompanying drawings, those skilled in the art to which the present invention pertains may be embodied in other specific forms without changing the technical spirit or essential features of the present invention. I can understand. Therefore, it is to be understood that the embodiments described above are exemplary in all respects and not restrictive.

Claims (11)

In the network defined method based on Software Defined Networking (SDN),
Calculating bandwidth usage of a network slice included in the network from at least one SDN switch included in the network using an OpenFlow protocol;
Calculating a bandwidth adjustment amount of the network slice based on the bandwidth usage;
Loading a static weight set on the network slice;
Calculating a dynamic weight of the network slice using the number of reserved bandwidth adjustments of the network slice;
If the calculated bandwidth adjustment amount is negative, updating the priority of the network slice in a bandwidth increase list using the loaded static weight and the calculated dynamic weight; And
Controlling the network according to a priority of the updated bandwidth increase list,
Controlling the network,
If the amount of bandwidth throttling is greater than the available bandwidth, suspending bandwidth throttling of the network slice and increasing the number of bandwidth throttling reservations;
The higher the static weight or the dynamic weight is, the higher the priority is,
As the number of reserved bandwidth adjustments increases, the dynamic weight is calculated to be a larger value.
Network control method. The method of claim 1,
Computing the bandwidth adjustment amount,
Loading a bandwidth set in the network slice; And
Subtracting the bandwidth usage from the set bandwidth;
Network control method. The method of claim 1,
Controlling the network,
If the amount of bandwidth adjustment is positive (+), storing information about the network slice in a bandwidth reduction list; And
If the bandwidth adjustment amount is negative and the bandwidth adjustment amount is less than the available bandwidth, reducing the bandwidth of one or more network slices included in the bandwidth reduction list and then increasing the bandwidth of the network slice according to the bandwidth adjustment amount. Comprising the steps,
Network control method. The method of claim 3,
Between the storing in the bandwidth reduction list and increasing the bandwidth, calculating the available bandwidth using a bandwidth adjustment amount for each network slice stored in the bandwidth reduction list,
Network control method. The method of claim 3,
Summing the bandwidth adjustment for each network slice stored in the bandwidth reduction list to calculate the available bandwidth, between storing in the bandwidth reduction list and increasing the bandwidth.
Network control method. The method of claim 1,
Calculating the dynamic weights,
Calculating the dynamic weight using the allocated bandwidth of the network slice further;
The greater the bandwidth usage relative to the allocated bandwidth, the greater the dynamic weight is calculated.
Network control method. The method of claim 1,
Calculating the dynamic weights,
Calculating the dynamic weight as a larger value as the number of bandwidth adjustment reservations is larger than the maximum number of bandwidth adjustment reservations for all network slices included in the network.
Network control method. The method of claim 1,
Calculating the dynamic weights,
Calculating the dynamic weight using the frequency of bandwidth variation of the network slice further;
The greater the frequency of the bandwidth variation compared to the maximum bandwidth variation frequency for all network slices included in the network, the dynamic weight is calculated to be a large value,
Network control method. The method of claim 1,
Computing the bandwidth usage,
Calculating the bandwidth usage by using open flow meter information of an SDN switch included in the network slice;
Network control method. The method of claim 1,
The network slice is an end to end (E2E) slice that guarantees a preset bandwidth among a plurality of hosts belonging to the network slice.
Network control method. In the network defined method based on Software Defined Networking (SDN),
Calculating bandwidth usage of a network slice included in the network from at least one SDN switch included in the network using an OpenFlow protocol;
Calculating a bandwidth adjustment amount of the network slice based on the bandwidth usage;
Loading a static weight set on the network slice;
Calculating a dynamic weight of the network slice using the number of reserved bandwidth adjustments of the network slice;
If it is determined that the bandwidth of the network slice needs to be increased through the calculated bandwidth adjustment amount, updating the priority of the network slice in the bandwidth increase list using the loaded static weight and the calculated dynamic weight; And
Controlling the network according to a priority of the updated bandwidth increase list,
Controlling the network,
Comparing the bandwidth adjustment amount with the available bandwidth, and if it is determined that the available bandwidth is insufficient, suspending the bandwidth adjustment of the network slice and increasing the number of the reserved bandwidth adjustments,
The higher the static weight or the dynamic weight is, the higher the priority is,
As the number of reserved bandwidth adjustments increases, the dynamic weight is calculated to be a larger value.
Network control method.

Images