WO2020215504A1 - 一种需求定义网络的实现方法及网络架构 - Google Patents
一种需求定义网络的实现方法及网络架构 Download PDFInfo
- Publication number
- WO2020215504A1 WO2020215504A1 PCT/CN2019/096636 CN2019096636W WO2020215504A1 WO 2020215504 A1 WO2020215504 A1 WO 2020215504A1 CN 2019096636 W CN2019096636 W CN 2019096636W WO 2020215504 A1 WO2020215504 A1 WO 2020215504A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- network
- demand
- user
- user demand
- layer
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0803—Configuration setting
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L41/00—Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
- H04L41/08—Configuration management of networks or network elements
- H04L41/0893—Assignment of logical groups to network elements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L47/00—Traffic control in data switching networks
- H04L47/70—Admission control; Resource allocation
- H04L47/80—Actions related to the user profile or the type of traffic
- H04L47/805—QOS or priority aware
Definitions
- the present invention relates to the field of network architecture, in particular to a method for realizing a demand-defined network and a network architecture.
- the basic capabilities of the network include three dimensions of transmission, calculation and storage. Its functions are mainly manifested in the three aspects of hardware resource sharing, software resource sharing and information exchange between users. In any case, the ultimate goal of the network is to satisfy all users in the network. The various needs of the network are also the meaning of the existence and development of the network. Therefore, the demand is the fundamental factor that determines the network architecture and attributes.
- the current network meets the needs of users by the two stages of network design and network operation and maintenance.
- the network design provides a basic network construction plan, and the network operation and maintenance configures the network based on the built network and user requirements collected by the network operator.
- These two stages are not centered on dynamic and discrete user needs, but are aimed at meeting the overall needs of predicted network services. Therefore, the matching degree of network architecture design and construction with dynamic and discrete user needs is hard to say well.
- the processing of dynamic changes in the network operation and maintenance stage is also mainly achieved manually, with poor real-time performance, insufficient accuracy of demand satisfaction, poor matching with the real-time network status, and poor network resource utilization.
- user needs may also change over time. If you rely on humans to adapt to changes in demand, it is neither real-time nor accurate enough, and the implementation cost is also high.
- the purpose of the present invention is to provide a method and network architecture for implementing a demand-defined network, so that the network can dynamically adapt to dynamically changing, discrete, and differentiated user needs.
- Collect and classify user needs divide each type of user demand into a user demand set, and adjust the user demand set according to user needs in real time;
- mapping between each meta slice and the required network resources is established, and the network resources corresponding to the meta slices are configured and executed.
- the same or similar user needs are classified into one category, and the similar user needs refer to user needs that have the same content and can share the same network resources.
- adjusting the user demand set in real time according to the user demand includes: carrying its effective time period in the user demand, and deleting the user demand that exceeds the effective time period from the user demand set.
- Delete the user requirement set When the last user demand in the user demand set is deleted, Delete the user requirement set.
- priority is set for the user needs in each user demand set.
- the user needs are deleted from the user demand set in order of priority from low to high until the network resources are satisfied.
- the user demand set set the priority for each user demand set, and determine the satiability of each user demand set in order of priority from high to low.
- each meta-slice is the smallest independent logical partition in the network, established based on a strategy, and the strategy is a corresponding operation performed after satisfying the set of user requirements corresponding to the meta-slice.
- Demand layer which is used to receive user needs through user terminal equipment or network service provider systems;
- the demand brokerage layer is used to collect and classify user needs, and divide each type of user demand into a user demand set; it is also used to adjust the user demand set in real time;
- the network intelligence layer is used to determine the satisfiability of each user demand set according to all available network resources, and to establish a meta slice for each user demand set that can be satisfied; it is also used to manage the life cycle of the meta slice, and establish The mapping between meta slices and required network resources; also used to configure network resources corresponding to the meta slices;
- a control layer which is used to control network resources in the network and configure network resources corresponding to the meta slice
- the network layer which is used to provide network resources for the meta-slice and perform the configuration of the control layer
- the network management system is used to monitor the status of the demand layer, demand broker layer, network intelligence layer, network layer and control layer, and perform daily management.
- the demand brokerage layer includes an independent server or a cluster of multiple servers, and the demand brokerage layer is also used to convert user needs into a form recognized by the network intelligence layer.
- the network intelligence layer is also used to report unsatisfied user demand sets to the demand brokerage layer, and the demand broker layer is also used to adjust the user demand sets reported by the network intelligence layer.
- the demand brokerage layer sets a priority for the user demand in each user demand set, and for the user demand set that cannot be met by network resources, the user demand is deleted from the user demand set in order of priority from low to high. , Until the network intelligence layer determines that the network resources meet the user demand set.
- the network layer includes multiple network devices, and the network resources mapped by each meta-slice are implemented by one or more network devices in the network layer.
- control layer includes one or more controllers
- network intelligence layer includes a network intelligence system
- network intelligence system is implemented by one or more servers.
- the network intelligence layer sets a user demand template for establishing meta-slices
- the user demand template is a two-dimensional structure, and one of the dimensions is time, and the other dimension is user needs that change over time.
- the controller and the network intelligent system are independent of each other, the network intelligent system directly communicates with the network device, or the network intelligent system communicates with the network device through the controller.
- the meta-slice established according to the user demand set realizes the automatic adaptation to the user demand. Compared with the manual adjustment of the network demand, it has the advantages of good real-time performance, high accuracy of realizing demand, and low cost of implementation.
- Meta-slices are established according to each user's demand set that can be met. Meta-slices of different user demand sets are isolated from each other, so the security and privacy of user needs in different demand sets are enhanced.
- Meta-slicing is established on the premise that network resources can be met, and can accurately match user needs and corresponding network resources.
- Network resources include network equipment resources and network logic resources, so network resource utilization is improved.
- Figure 1 is a flowchart of a method for implementing a requirement definition network according to an embodiment of the present invention
- FIG. 2 is a schematic diagram of a network architecture of a requirement definition network according to an embodiment of the present invention
- FIG. 3 is a schematic diagram of the physical architecture of a requirement definition network according to an embodiment of the present invention.
- Figure 4 is a schematic diagram of a demand brokerage function in an embodiment of the present invention.
- FIG. 5 is a schematic diagram of generating meta slices in a network intelligent system according to an embodiment of the present invention.
- Figure 6 is a schematic diagram of a network demand proxy function according to an embodiment of the invention.
- this embodiment provides a method for implementing a demand definition network, which includes the steps:
- S101 Collect and classify user needs, divide each type of user needs into a set of user needs, and adjust the set of user needs in real time according to user needs.
- User requirements can be collected from user terminal equipment or network service provider systems (such as BOSS systems). Then divide the same or similar user needs into one category to form a set of user needs, that is, a set of user needs.
- the same user demand means that the content of the demand is exactly the same; similar user demand means that the content of the demand is partly the same, but the same network resources can be used together; the user needs of different user demand sets are essentially different.
- User A needs a bandwidth of 100M at time T1, and a delay requirement of 1s; at time T2, a bandwidth of 5M and a delay requirement of 20ms.
- User B needs a bandwidth of 100M at time T1, and a delay requirement of 1s; at time T2, a bandwidth of 5M and a delay requirement of 20ms.
- User C needs a bandwidth of 50M at time T1, and a delay requirement of 1s; at time T2, a bandwidth of 10M, and a delay requirement of 20ms.
- User D needs a bandwidth of 1000M at time T1, and a delay requirement of 100ms; at time T2, a bandwidth of 1M and a delay requirement of 2ms.
- network resources include physical resources and logical resources of network devices, which can be obtained according to the status of all network devices within the jurisdiction, such as VLANs, IP addresses, and Layer 4 ports.
- a meta-slice is established based on the strategy.
- this meta-slice is also updated in real time as the user demand set changes. If the network resources cannot meet a certain set of user needs, the problem that the resources cannot be satisfied is reported to other systems, such as the system of the network service provider.
- the above-mentioned strategy for establishing meta-slices refers to: after network resources can meet a certain set of user requirements, corresponding operations can be performed.
- the terminal name is "IPhoneX Zhang”
- the MAC address is AA: BB: CC: DD: EE: FF
- the bandwidth requirement is 100M at 9 am.
- the premise is to establish the corresponding meta-slice, which is to establish the meta-slice based on the strategy.
- Meta slices can be generated using a unified meta slice template, and represented by a logical data structure describing the meta slice. For example, a certain meta slice requires port M of device A to reserve 10M bandwidth and port N of device B to reserve 10M bandwidth..., the logical data structure is: A::M:10M, B::N:10M...
- step S103 Establish a mapping between each meta-slice and the required network resources, that is, the meta-slice in step S102 is only a design and not physically implemented, so it needs to correspond to the physical resources of the network device or the network virtual resources.
- the meta slice can be matched with the required network resources. Then, according to the operation instruction and the configuration method of the network device, the network resource corresponding to the meta slice is configured and executed.
- this embodiment explains in detail how the user demand set is adjusted in real time according to the user demand.
- the new user demand When a new user demand is received, the new user demand is allocated to the user demand set to which it belongs. If the new user demand does not belong to any user demand set, a new user demand set is created for the new user demand .
- Each user demand carries the effective time period. During the effective time period, it is regarded as the effective user demand. If the user demand exceeds the effective time period, the user demand that exceeds the time period is deleted from the user demand set; when the user demand is the last user in the set When a requirement is deleted, the user requirement set is deleted.
- the meta slice corresponding to the user demand set will be updated accordingly. If a user demand set is newly created, corresponding meta slices will also be newly generated. If the user demand set is deleted, the corresponding meta slice will also be deleted. After the meta-slice is changed, the mapping corresponding to the meta-slice will also change, so as to ensure the real-time performance of network resource allocation.
- step S101 different priority levels can be set for multiple user demand sets in step S101.
- step S102 each user demand is determined one by one according to the order of priority from high to low. The satisfiability of the set ensures that the user demand set with high priority allocates network resources first.
- the priority is set for the user needs in each user demand set.
- step S102 if the network resources cannot meet a certain user demand set, it is reported to other systems that the resources are unsatisfactory. At this time, the priority can be set from low to high.
- the user needs are sequentially deleted from the user demand set in sequence until the network resources meet the user demand set. For deleted user requirements, the network service provider or network operation and maintenance unit will be automatically notified that the network resources are insufficient to meet these user sets. In this case, additional processing is required, for example, it can be processed manually.
- this embodiment provides a network architecture of a demand definition network, which includes a demand layer, a demand brokerage layer, a network intelligence layer, a network layer, a control layer, and a network management system (NMS) for controlling each layer. ).
- a demand definition network which includes a demand layer, a demand brokerage layer, a network intelligence layer, a network layer, a control layer, and a network management system (NMS) for controlling each layer.
- NMS network management system
- the demand layer is used to receive discrete, dynamically changing, and differentiated user needs, and is the source of user needs.
- the various graphs of the demand level in Figure 2 each graph represents a user demand.
- the demand layer receives user demands through user terminal equipment or network service provider systems.
- the demand broker layer is used to collect and classify user needs, and divide each type of user demand into a user demand set. It is also used to adjust the user demand set in real time according to user needs, and it is also used to convert user needs into network intelligence layer. The form of recognition.
- the same graph in the demand brokerage layer in Figure 2 represents the same or similar user needs, and the same graph is divided into one category, indicating that the same or similar user needs form a user demand set.
- the demand brokerage layer includes a demand brokerage component, and the demand brokerage component can be an independent server or a cluster composed of multiple servers.
- the network intelligence layer is used to analyze the status of all available network resources, determine the satisfiability of each user demand set, establish a meta slice for each user demand set that can be satisfied, and also manage the life cycle of the meta slice.
- the network intelligence layer is also used to establish a mapping between meta-slices and required network resources, that is, to map user demand sets to the final network resource configuration based on the current status of the network.
- user demand set 1 requires 100M bandwidth to transmit data from network node A to network node B, with an end-to-end delay of 1ms; then the network intelligence layer needs to find a route from network node A to network node B, with the bandwidth of this route Not less than 100M, and the total routing delay is not more than 1ms, and then the 100M bandwidth of the route is matched with user demand set 1.
- the control layer is used to control network resources in the network and configure network resources corresponding to the meta-slice.
- the network layer is used to provide network resources for the meta-slice.
- the network resources include physical resources and logical resources of the network device; it is also used to perform the configuration of the control layer.
- the network layer includes resource agents and multi-layer networks (such as L2, L3, and L4).
- Resource agents are components that implement resource scheduling and configuration in each layer of the multi-layer network.
- the second layer (L2) resource agent corresponds to the second layer of VLAN.
- the three-tier (L3) resource agent corresponds to three-tier logical resources such as IP routing
- the four-tier (L4) resource agent corresponds to four-tier logical resources such as four-tier ports.
- the network management system is the management entity of the entire network, which is used to monitor the status of the entire network including the demand layer, demand broker layer, network intelligence layer, network layer, and control layer, and implement daily management of the network.
- the network intelligence layer is also used to report unsatisfied user demand sets to the demand brokerage layer.
- the demand brokerage layer sets priorities for user needs in each user demand set, and for user demand sets that cannot be met by network resources , Delete user needs from the user demand set in order of priority from low to high, until the network intelligence layer determines that the network resources meet the user demand set.
- the network intelligence layer may not pass through the control layer, but the network intelligence layer itself configures the network resources corresponding to the meta-slices, and sends the operation instructions corresponding to the configuration to the network layer.
- the control layer is not used to configure the function of the network resource corresponding to the meta slice.
- this embodiment provides the physical architecture of the previous embodiment, where the demand layer includes a variety of user terminals, and the demand broker layer includes an independent server.
- the network layer includes multiple network devices, such as routers, switches, etc.; the network resources mapped by each meta-slice are implemented by one or more network devices in the network layer.
- the control layer includes a controller, and the network intelligence layer includes a network intelligence system, which is implemented by one or more servers.
- the network intelligent system is implemented by a server. Since the input interface of the server has its own rules, the user demand brokerage component needs to convert the user demand set into the rules of the network intelligent system. For example, if the user demand brokerage component collects natural language descriptions, and the network intelligent system needs machine language descriptions, the user demand brokerage component converts the user demand set described in natural language into machine language descriptions.
- the controller and the network intelligent system are logically independent of each other, and are jointly set in a physical platform, such as a server or a server group.
- the network intelligent system communicates with the network equipment through the interface between the logical platform and the network equipment .
- the controller and the network intelligent system are independent of each other, and the network intelligent system directly communicates with the network equipment.
- the controller and the network intelligent system are independent of each other, and the network intelligent system communicates with the network equipment through the controller. In this case, the operation of the network intelligent system is first down to the controller, and then down to the network device from the controller.
- the network layer includes demand agent and resource agent, which are all components embedded in the network equipment.
- the basic function of the network demand agent is to execute the operation instructions issued by the network intelligent system.
- This operation instruction is required by the network intelligent system according to the meta-slicing An instruction generated by a network resource that needs to be executed by a network device.
- the operation instruction may be directly issued by the network intelligent system, or may be issued by the network intelligent system through the controller.
- the demand agent determines whether the network resource corresponding to the operation instruction is a physical resource and/or a logical resource according to the operation instruction:
- the demand agent completes the conversion of the above-mentioned operation instruction to the network resource configuration, and the network device executes the conversion after the conversion.
- the demand agent converts the operation instruction through the resource agent to configure the network resource, and the operation instruction converted by the resource agent is executed by the network device.
- the function size of the network demand agent depends on the application scenario and the network resource status of the network equipment. If the network resources are in short supply or the network equipment in the application scenario cannot be substantially changed, the network demand agent can be designed to be very simple and only implement operations Conversion of instructions to network configuration. If the network resources can support the network demand agent to implement more complex functions, and the network equipment in the application scenario can be substantially upgraded, the mapping function and even other functions in the network intelligent system can be implemented on the network equipment. For example, the lower part of Fig. 6 is the converted network configuration, which can be directly executed by the network device.
- the network intelligence layer sets up a user demand template for meta-slices, and the user demand template can be arbitrarily expanded according to application scenarios. It is a two-dimensional structure. One dimension is time. At different times, the user needs of the same user may be essentially different; the other dimension is user needs that change over time, which can be bandwidth, delay, packet loss rate, jitter , Path, environment, etc.
- this embodiment describes the function of the demand brokerage component through FIG. 4, and the demand brokerage component is responsible for the collection, conversion, and classification of discrete and dynamic user needs.
- the upper part is various user requirements
- the middle part is the demand brokerage component
- the lower part is the converted and classified user requirements.
- the conversion is the conversion of user requirements obtained from the demand layer into requirements that can be understood by the network intelligent system.
- FIG. 5 it is a schematic diagram of the network intelligence system generating meta slices.
- the basic function of the network intelligence system is to associate user demand sets with policy-based meta slices, complete the life cycle management of meta slices, and map meta slices to networks The operations the device can perform. Among them, the life cycle management of the meta slice is adjusted in real time according to the user demand set. When the last user demand in the user demand set is deleted, the user demand set is deleted. If the user demand set is deleted, the corresponding meta slice will also be deleted.
- the upper part of Figure 5 is the user demand set output by the demand brokerage layer. The user demand in each user demand set is the same or similar, but there are essential differences in user demand between different user demand sets.
- the middle of Figure 5 is the network intelligence system, and the lower part is the meta-slice that needs to be established to achieve the above-mentioned user demand set.
- the upper part is the various operations issued by the network intelligent system.
- the execution of these instructions is logically equivalent to the physical instantiation of a meta slice.
- the middle part of Fig. 6 is the demand agent, and the lower part of Fig. 6 is the network configuration after conversion, which can be directly executed by the network device.
- the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
- a computer-usable storage media including but not limited to disk storage, CD-ROM, optical storage, etc.
Abstract
一种需求定义网络的实现方法及网络架构，涉及网络架构领域，方法包括步骤：收集用户需求并分类，将每一类用户需求划分为一个用户需求集，实时根据用户需求调整用户需求集；根据所有可用网络资源，判断每个用户需求集的可满足性，为可满足的每个用户需求集建立一个元切片；建立每个元切片与所需网络资源的映射，将元切片对应的网络资源进行配置并执行。本发明能够使网络根据动态变化、离散、具有差异化的用户需求进行动态的适配。
Description
本发明涉及网络架构领域，具体涉及一种需求定义网络的实现方法及网络架构。
网络的基本能力包括传送、计算和存储三个维度，其功能主要表现在硬件资源共享、软件资源共享和用户间信息交换三个方面，而无论如何，网络的最终目标是为满足网络中所有用户的各种各样的需求，这也是网络存在的意义和发展的动力，因此，需求才是决定网络架构和属性的根本性的因素。
当前的网络，满足用户需求由网络设计和网络运维两个阶段共同实现。网络设计提供基本的网络建设方案，网络运维基于建设好的网络和网络运营商收集的用户需求来配置网络。这两个阶段都不是以动态的、离散的用户需求为中心，而是以满足预测的网络业务总体需要为目标。因此，网络架构设计和建设与动态的、离散的用户需求的匹配度难以言善。另一方面，网络运维阶段对动态变化的需求处理也主要依赖人工来实现，实时性不强、需求满足精准度不足、与网络实时状态匹配度不好、网络资源利用率差。同时，用户需求也可能是随着时间而变化的，如果依赖人工来实现对需求变化的适配，既不具有实时性，也不够精确，而且实现成本也高。
软件定义网络带来的网络架构上的改变对网络的发展确有较大的积极影响，但由于软件定义网络着力于网络的上层架构，对实现和 应用方面的考量不够精细，并未明确解决由谁定义比较合理、如何高效实现上述定义的问题，使得在产品上的应用面临一些障碍。
发明内容
针对现有技术中存在的缺陷，本发明的目的在于提供一种需求定义网络的实现方法及网络架构，使网络根据动态变化、离散、具有差异化的用户需求进行动态的适配。
为达到以上目的，一方面，采取一种需求定义网络的实现方法，包括步骤：
收集用户需求并分类，将每一类用户需求划分为一个用户需求集，实时根据用户需求调整用户需求集；
根据所有可用网络资源，判断每个用户需求集的可满足性，为可满足的每个用户需求集建立一个元切片；
建立每个元切片与所需网络资源的映射，将元切片对应的网络资源进行配置并执行。
优选的，将相同或相近的用户需求分为一类，所述相近的用户需求是指部分内容相同，能共用相同网络资源的用户需求。
优选的，根据用户需求实时调整用户需求集包括：所述用户需求内携带其生效时间段，将超过生效时间段的用户需求从用户需求集中删除，当用户需求集中最后一个用户需求被删除时，删除该用户需求集。
优选的，为每个用户需求集内的用户需求设置优先级，对于网络资源不能满足的用户需求集，按照优先级由低到高的顺序依次从该用户需求集中删除用户需求，直至网络资源满足该用户需求集；为每个用户需求集设置优先级，按照优先级由高到低的顺序，依次判断每个用户需求集的可满足性。
优选的，每个元切片为网络中最小的独立逻辑分区，基于策略建立，所述策略为满足元切片对应的用户需求集后执行的相应操作。
另一方面，还提供一种需求定义网络的网络架构，包括：
需求层，其用于通过用户终端设备或网络服务提供商的系统接收用户需求；
需求经纪层，其用于收集用户需求并分类，将每一类用户需求划分为一个用户需求集；还用于实时对用户需求集进行调整；
网络智能层，其用于根据所有可用网络资源，判断每个用户需求集的可满足性，为可满足的每个用户需求集建立一个元切片；还用于管理元切片的生命周期，以及建立元切片与所需网络资源的映射；还用于配置所述元切片对应的网络资源；
控制层，其用于控制网络中的网络资源和配置所述元切片对应的网络资源；
网络层，其用于为所述元切片提供网络资源并执行控制层的配置；
网络管理系统，其用于监测需求层、需求经纪层、网络智能层、网络层和控制层状态以及进行日常管理。
优选的，所述需求经纪层包括独立服务器或多台服务器集群，所述需求经纪层还用于将用户需求转换为网络智能层识别的形式。
优选的，所述网络智能层还用于将不能满足的用户需求集上报给需求经纪层，所述需求经纪层还用于调整网络智能层上报的用户需求集。
优选的，所述需求经纪层为每个用户需求集内的用户需求设置优先级，对于网络资源不能满足的用户需求集，按照优先级由低到高的顺序依次从该用户需求集中删除用户需求，直至网络智能层判断网络资源满足该用户需求集。
优选的，所述网络层包括多个网络设备，所述每个元切片映射的网络资源通过网络层中的一个或多个网络设备实现。
优选的，所述控制层包括一台或多台控制器，所述网络智能层包括网络智能系统，网络智能系统由一台或多台服务器实现。
优选的，所述网络智能层设置建立元切片的用户需求模板，用户需求模板是二维结构，且其中一个维度是时间，另一个维度是随时间发生变化的用户需求。
优选的，所述控制器和网络智能系统相互独立，网络智能系统直接与网络设备通信，或者网络智能系统通过控制器与网络设备通信。
上述技术方案中的至少一个具有如下有益效果：
1、由于用户需求是随着时间而变化的，用户需求具有动态变化、离散、差异化的特点，因此用户需求集是不断改变的，可以根据用户需求的添加、更改或删除来进行改变，网络根据用户需求集建立的元切片实现对用户需求的自动适配，与人工调整网络需求相比，具有实时性好、实现需求的精确性高、实现成本低的优点。
2、元切片是根据可满足的每个用户需求集建立的，不同用户需求集的元切片之间彼此隔离开来，是的不同需求集内用户需求的安全性和私密性得到加强。
3、元切片是在网络资源可满足的前提下建立的，可以精确匹配用户需求和对应的网络资源，网络资源包括网络设备资源和网络逻辑资源，因此网络资源利用率得到提高。
图1为本发明实施例需求定义网络的实现方法流程图；
图2为本发明实施例需求定义网络的网络架构示意图；
图3为本发明实施例需求定义网络的物理架构示意图；
图4为本发明实施例需求经纪功能示意图；
图5为本发明实施例网络智能系统生成元切片的示意图；
图6为本发明实施例网络需求代理功能示意图。
为使本发明实施例的目的、技术方案和优点更加清楚，下面将结合本发明实施例中的附图，对本发明实施例中的技术方案进行清楚、完整地描述。
如图1所示，本实施例提供一种需求定义网络的实现方法，包括步骤：
S101.收集用户需求并分类，将每一类用户需求划分为一个用户需求集，并根据用户需求实时调整用户需求集。用户需求可以从用户终端设备或者是网络服务提供商的系统(例如BOSS系统)进行收集。然后将相同或相近的用户需求分为一类，形成一个用户需求的集合，即用户需求集。其中，相同的用户需求是指需求的内容完全相同；相近的用户需求是指需求的内容有一部分相同，但是可以共同使用相同的网络资源；不同用户需求集的用户需求存在本质不同。
例如：A用户在时间T1需要带宽100M，时延要求为1s；在时间T2需要带宽5M，时延要求为20ms。B用户在时间T1需要带宽100M，时延要求为1s；在时间T2需要带宽5M，时延要求为20ms。C用户在时间T1需要带宽50M，时延要求为1s；在时间T2需要带宽10M，时延要求为20ms。D用户在时间T1需要带宽1000M，时延要求为100ms；在时间T2需要带宽1M，时延要求为2ms。
那么A和B的用户需求为相同；A和C的用户需求为相近；A和D的用户需求为不同。因此A、B和C的用户需求归入同一个用户需求集，D的用户需求归入另外一个用户需求集。
S102.根据管辖范围内所有可用网络资源，判断每个用户需求集的可满足性，为可满足的每个用户需求集建立一个元切片，并根据元切片所需网络资源生成发送给网络设备的操作指令。
其中，网络资源包括网络设备的物理资源和逻辑资源，可以根据管辖范围内所有网络设备的状态得到，逻辑资源例如VLAN，IP地址，四层端口等。对于可满足的用户需求集，则基于策略建立一个元切片。并且，这个元切片也是随着用户需求集的改变实时更新的。如果网络资源不能满足某个用户需求集，则向其他系统，例如网络服务提供商的系统报告资源不可满足的问题。
上述建立元切片的策略指的是：在网络资源可以满足某个用户需求集后，才能执行相应的操作。例如，终端名为“IPhoneX Zhang”，MAC地址为AA：BB：CC：DD：EE：FF，早上9点带宽要求为100M的，是一个用户需求集的实例，只有网络资源具备满足这个实例的前提，才能建立相应的元切片，这就是基于策略建立元切片。
上述元切片是切片的最小单位，即，最小且独立的网络逻辑分区，一个用户需求集的网络切片即为一个元切片。元切片可以采用统一的元切片模板生成，通过一个描述元切片的逻辑数据结构来表示。例如，某一元切片要求设备A的端口M保留10M带宽，要求设备B的端口N保留10M带宽…，逻辑数据结构为：A::M:10M,B::N:10M…
S103.建立每个元切片与所需网络资源的映射，也就是说，步骤S102中的元切片只是一个设计，并未物理实现，那么需要和网络设备的物理资源或网络虚拟资源对应起来。通过建立映射，就可以使元切片和所需要的网络资源对应起来。然后根据操作指令和网络设备的配置方法，将元切片对应的网络资源进行配置并执行。
基于上述实施例，本实施例详细说明用户需求集是如何根据用户 需求实时进行调整的。
当收到新的用户需求时，将新增的用户需求分配到所属的用户需求集内，若新增的用户需求不属于任何用户需求集，则为新增的用户需求建立新的用户需求集。
每个用户需求携带生效时间段，在生效时间段内，视为有效用户需求，如果用户需求超过生效时间段，那么将超过时间段的用户需求从用户需求集中删除；当用户需求集中最后一个用户需求被删除时，删除该用户需求集。
每次用户需求集改变，如增加或删除用户需求，该用户需求集对应的元切片都会进行相应的更新。如果新建立用户需求集，也会新生成对应的元切片。如果用户需求集删除，那么对应的元切片也会删除。元切片改变后，元切片对应的映射也会改变，以此来保证网络资源分配的实时性。
基于上述实施例，在另一个实施例中，步骤S101中可以对多个用户需求集设置不同的优先级，在步骤S102中，按照优先级由高到低的顺序，逐一来判断每个用户需求集的可满足性，保证优先级高的用户需求集优先分配网络资源。
另外，为每个用户需求集内的用户需求设置优先级，步骤S102中，网络资源不能满足某个用户需求集，则向其他系统报告资源不可满足，此时可以按照优先级由低到高的顺序依次从该用户需求集中删除用户需求，直至网络资源满足该用户需求集。对于被删除的用户需求，会自动通知网络服务提供商或网络运行维护单位网络资源不足以满足这些用户集，这种情况下需另行处理，比如可以通过人工进行处理。
如图2所示，本实施例提供一种需求定义网络的网络架构，其包 括需求层、需求经纪层、网络智能层、网络层、控制层、以及用于控制各层的网络管理系统(NMS)。
需求层用于接收离散、动态变化、具有差异化的用户需求，是用户需求的来源。图2中需求层的各种图形，每一个图形代表一个用户需求。通常，需求层通过用户终端设备或网络服务提供商的系统接收用户需求。
需求经纪层用于收集用户需求并分类，将每一类用户需求划分为一个用户需求集，还用于根据用户需求，实时对用户需求集进行调整，还用于将用户需求转换为网络智能层识别的形式。图2中需求经纪层中相同图形表示相同或相近的用户需求，相同的图形被划分为一类，表示相同或相近的用户需求形成一个用户需求集。所述需求经纪层包括需求经纪组件，需求经纪组件可以是一台独立服务器，也可以是由多台服务器组成的集群。
网络智能层用于根据所有可用网络资源的状况进行分析，判断每个用户需求集的可满足性，为可满足的每个用户需求集建立一个元切片，还用于管理元切片的生命周期。网络智能层还用于建立元切片与所需网络资源的映射，即，基于网络的当前状况，将用户需求集映射到最终的网络资源配置。例如：用户需求集1需要100M带宽将数据从网络节点A传到网络节点B，端到端时延为1ms；那么网络智能层需要找到从网络节点A到网络节点B的一条路由，该路由带宽不小于100M，路由总时延不大于1ms，然后将该路由的100M带宽与用户需求集1对应起来。
控制层，其用于控制网络中的网络资源和配置所述元切片对应的网络资源。
网络层，其用于为所述元切片提供网络资源，网络资源包括网络 设备的物理资源和逻辑资源；还用于执行控制层的配置。网络层包括资源代理和多层网络(如L2、L3和L4)，资源代理是实现多层网络各分层中资源调度、配置的组件，如二层(L2)资源代理是对应VLAN等二层逻辑资源，三层(L3)资源代理是对应诸如IP路由的三层逻辑资源，四层(L4)资源代理是对应诸如四层端口等四层逻辑资源。
网络管理系统，是整个网络的管理实体，其用于监测包括需求层、需求经纪层、网络智能层、网络层和控制层的整个网络的状态，以及对网络实施日常管理。
除了上述功能外，网络智能层还用于将不能满足的用户需求集上报给需求经纪层，需求经纪层为每个用户需求集内的用户需求设置优先级，对于网络资源不能满足的用户需求集，按照优先级由低到高的顺序依次从该用户需求集中删除用户需求，直至网络智能层判断网络资源满足该用户需求集。
在另一种实施例示中，网络智能层可以不通过控制层，而是网络智能层自身配置所述元切片对应的网络资源，并将配置对应的操作指令下发给网络层，在此种情况下，不使用控制层配置所述元切片对应的网络资源的功能。
如图3所示，本实施例提供上一实施例的物理架构，其中，需求层包括多种用户终端，需求经纪层包括一台独立服务器。
网络层包括多个网络设备，如路由器、交换机等；每个元切片映射的网络资源通过网络层中的一个或多个网络设备实现。控制层包括一台控制器，网络智能层包括网络智能系统，通过一台或多台服务器实现。
本实施例中网络智能系统通过一台服务器实现，由于服务器的输入接口有自己的规则，用户需求经纪组件需要将用户需求集转换为网 络智能系统的规则。例如，用户需求经纪组件收集的是自然语言描述，网络智能系统需要的是机器语言描述，则用户需求经纪组件将自然语言描述的用户需求集，转换为机器语言描述。
上述网络智能系统、控制器及网络设备的关系有三种情况：
1、控制器和网络智能系统逻辑上相互独立，共同设置于一个物理平台中，例如一台服务器或一个服务器组，这种情况下网络智能系统通过逻辑平台与网络设备间的接口与网络设备通信。
2、控制器和网络智能系统相互独立，网络智能系统直接与网络设备通信。
3、控制器和网络智能系统相互独立，网络智能系统通过控制器与网络设备通信。这种情况下，网络智能系统的操作先下达到控制器，再由控制器下达到网络设备。
网络层中包括需求代理和资源代理，均是嵌入到网络设备中的一个组件，网络需求代理的基本功能是执行网络智能系统下达的操作指令，这个操作指令是网络智能系统根据元切片所需的网络资源生成的、需要让网络设备执行的指令，该操作指令可以是网络智能系统直接下达，也可以是网络智能系统通过控制器下达的。
需求代理根据操作指令判断该操指令对应的网络资源是物理资源和/或逻辑资源：
如果是物理资源，需求代理完成上述操作指令向网络资源配置的转换，转换后由网络设备执行。
如果是逻辑资源，需求代理将操作指令通过资源代理进行网络资源配置的转换，资源代理转换后的操作指令再由网络设备执行。
如果是物理资源和逻辑资源，则分别按照上述方式，各自完成转换后，由网络设备执行。
网络需求代理的功能大小取决于应用场景和网络设备的网络资源状况，如果网络资源紧缺，或者应用场景中的网络设备不能有本质性的改变，则网络需求代理可设计得很简单，仅仅实现操作指令到网络配置的转换。如果网络资源可支持网络需求代理实现较复杂的功能，并且应用场景中的网络设备能够进行实质升级，则网络智能系统中的映射功能甚至其它功能可以下沉到网络设备上实现。如附图6的下部分是转换后的网络配置，可由网络设备直接执行。
网络智能层设置建立元切片的用户需求模板，用户需求模板根据应用场景是可以任意扩充的。是一个二维结构，一个维度是时间，在不同时间，同一用户的用户需求可能有本质的不同；另一个维度是随时间发生变化的用户需求，可以是带宽、时延、丢包率、抖动、路径、环境等等。
表1
基于上述实施例，本实施例通过图4描述需求经纪组件功能，需求经纪组件负责离散、动态的用户需求的收集、转换和归类。图4中上部是各种用户需求，中间是需求经纪组件，下部是已经转换和归类好的用户需求，其中转换是把从需求层得到的用户需求转换成能被网络智能系统理解的需求。
如图5所示，是网络智能系统生成元切片的示意图，网络智能系统的基本功能是将用户需求集与基于策略的元切片对应起来，完成元切片生命周期管理，并且将元切片映射成网络设备能执行的操作。其中，元切片的生命周期管理是根据用户需求集实时调整的，当用户需求集中最后一个用户需求被删除时，删除该用户需求集，如果用户需求集删除，那么对应的元切片也会删除。图5上部是需求经纪层输出的用户需求集，每个用户需求集内的用户需求相同或相近，但是不同用户需求集之间的用户需求存在本质差异。图5中间是网络智能系统，下部是为实现上述用户需求集而需要建立的元切片。
如图6所示，其上部是网络智能系统下达的各种操作，这些指令的执行逻辑上等同于一个元切片的物理实例化，换句话说，每个元切片对应的网络操作成功实现后，该元切片就已实现。图6中间部分是需求代理，图6下部是转换后的网络配置，可以由网络设备直接执行。
本领域内的技术人员应明白，本申请的实施例可提供为方法、系统、或计算机程序产品。因此，本申请可采用完全硬件实施例、完全软件实施例、或结合软件和硬件方面的实施例的形式。而且，本申请可采用在一个或多个其中包含有计算机可用程序代码的计算机可用存储介质(包括但不限于磁盘存储器、CD-ROM、光学存储器等)上实施的计算机程序产品的形式。
本发明不局限于上述实施方式，对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也视为本发明的保护范围之内。本说明书中未作详细描述的内容属于本领域专业技术人员公知的现有技术。
Claims (13)
- 一种需求定义网络的实现方法，其特征在于，包括步骤：收集用户需求并分类，将每一类用户需求划分为一个用户需求集，实时根据用户需求调整用户需求集；根据所有可用网络资源，判断每个用户需求集的可满足性，为可满足的每个用户需求集建立一个元切片；建立每个元切片与所需网络资源的映射，将元切片对应的网络资源进行配置并执行。
- 如权利要求1所述的需求定义网络的实现方法，其特征在于：将相同或相近的用户需求分为一类，所述相近的用户需求是指部分内容相同，能共用相同网络资源的用户需求。
- 如权利要求1所述的需求定义网络的实现方法，其特征在于，根据用户需求实时调整用户需求集包括：所述用户需求内携带其生效时间段，将超过生效时间段的用户需求从用户需求集中删除，当用户需求集中最后一个用户需求被删除时，删除该用户需求集。
- 如权利要求1所述的需求定义网络的实现方法，其特征在于：为每个用户需求集内的用户需求设置优先级，对于网络资源不能满足的用户需求集，按照优先级由低到高的顺序依次从该用户需求集中删除用户需求，直至网络资源满足该用户需求集；为每个用户需求集设置优先级，按照优先级由高到低的顺序，依次判断每个用户需求集的可满足性。
- 如权利要求1所述的需求定义网络的实现方法，其特征在于：每个元切片为网络中最小的独立逻辑分区，基于策略建立，所述策略为满足元切片对应的用户需求集后执行的相应操作。
- 一种需求定义网络的网络架构，其特征在于，包括：需求层，其用于通过用户终端设备或网络服务提供商的系统接收用户需求；需求经纪层，其用于收集用户需求并分类，将每一类用户需求划分为一个用户需求集；还用于实时对用户需求集进行调整；网络智能层，其用于根据所有可用网络资源，判断每个用户需求集的可满足性，为可满足的每个用户需求集建立一个元切片；还用于管理元切片的生命周期，以及建立元切片与所需网络资源的映射；还用于配置所述元切片对应的网络资源；控制层，其用于控制网络中的网络资源和配置所述元切片对应的网络资源；网络层，其用于为所述元切片提供网络资源并执行控制层的配置；网络管理系统，其用于监测需求层、需求经纪层、网络智能层、网络层和控制层状态以及进行日常管理。
- 如权利要求6所述的需求定义网络的网络架构，其特征在于：所述需求经纪层包括独立服务器或多台服务器集群，所述需求经纪层还用于将用户需求转换为网络智能层识别的形式。
- 如权利要求6所述的需求定义网络的网络架构，其特征在于：所述网络智能层还用于将不能满足的用户需求集上报给需求经纪层，所述需求经纪层还用于调整网络智能层上报的用户需求集。
- 如权利要求8所述的需求定义网络的网络架构，其特征在于：所述需求经纪层为每个用户需求集内的用户需求设置优先级，对于网络资源不能满足的用户需求集，按照优先级由低到高的顺序依次从该用户需求集中删除用户需求，直至网络智能层判断网络资源满足该用户需求集。
- 如权利要求6所述的需求定义网络的网络架构，其特征在于：所述网络层包括多个网络设备，所述每个元切片映射的网络资源通过网络层中的一个或多个网络设备实现。
- 如权利要求6所述的需求定义网络的网络架构，其特征在于：所述控制层包括一台或多台控制器，所述网络智能层包括网络智能系统，网络智能系统由一台或多台服务器实现。
- 如权利要求6所述的需求定义网络的网络架构，其特征在于：所述网络智能层设置建立元切片的用户需求模板，用户需求模板是二维结构，且其中一个维度是时间，另一个维度是随时间发生变化的用户需求。
- 如权利要求6所述的需求定义网络的网络架构，其特征在于：所述控制器和网络智能系统相互独立，网络智能系统直接与网络设备通信，或者网络智能系统通过控制器与网络设备通信。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910324639.8 | 2019-04-22 | ||
| CN201910324639.8A CN110086662B (zh) | 2019-04-22 | 2019-04-22 | 一种需求定义网络的实现方法及网络架构 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2020215504A1 true WO2020215504A1 (zh) | 2020-10-29 |
Family
ID=67415933
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2019/096636 WO2020215504A1 (zh) | 2019-04-22 | 2019-07-19 | 一种需求定义网络的实现方法及网络架构 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN110086662B (zh) |
| WO (1) | WO2020215504A1 (zh) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115632925B (zh) * | 2022-12-26 | 2023-03-14 | 北京交通大学 | 一种时间内生确定性融合网络架构及数据通信方法 |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106549806A (zh) * | 2016-10-26 | 2017-03-29 | 清华大学 | 一种网络切片管理器及其管理方法 |
| CN106657194A (zh) * | 2015-11-02 | 2017-05-10 | 中兴通讯股份有限公司 | 一种网络切片能力开放的方法、装置及系统 |
| CN109218046A (zh) * | 2017-06-30 | 2019-01-15 | 中国移动通信有限公司研究院 | 网络切片的管理方法及系统和存储介质 |
| WO2019062836A1 (zh) * | 2017-09-30 | 2019-04-04 | 华为技术有限公司 | 网络切片管理方法及其装置 |
Family Cites Families (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN100536565C (zh) * | 2007-08-24 | 2009-09-02 | 北京市朝阳区信息网络中心 | 图像信息管理系统用户分类分级管理系统 |
| US10223654B2 (en) * | 2015-03-23 | 2019-03-05 | Accenture Global Services Limited | Automated, accelerated prototype generation system |
| CN107566145B (zh) * | 2016-06-30 | 2020-11-10 | 华为技术有限公司 | 用于管理网络切片的方法和装置 |
| CN108933686B (zh) * | 2017-05-25 | 2022-12-20 | 中兴通讯股份有限公司 | 一种网络切片方法、计算机设备及存储介质 |
| CN109218047B (zh) * | 2017-06-30 | 2021-09-10 | 中国移动通信有限公司研究院 | 网络切片处理方法及装置、通信系统及存储介质 |
| CN109219020A (zh) * | 2018-09-14 | 2019-01-15 | 云迅智能科技南京有限公司 | 一种网络切片方法及系统 |
-
2019
- 2019-04-22 CN CN201910324639.8A patent/CN110086662B/zh active Active
- 2019-07-19 WO PCT/CN2019/096636 patent/WO2020215504A1/zh active Application Filing
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106657194A (zh) * | 2015-11-02 | 2017-05-10 | 中兴通讯股份有限公司 | 一种网络切片能力开放的方法、装置及系统 |
| CN106549806A (zh) * | 2016-10-26 | 2017-03-29 | 清华大学 | 一种网络切片管理器及其管理方法 |
| CN109218046A (zh) * | 2017-06-30 | 2019-01-15 | 中国移动通信有限公司研究院 | 网络切片的管理方法及系统和存储介质 |
| WO2019062836A1 (zh) * | 2017-09-30 | 2019-04-04 | 华为技术有限公司 | 网络切片管理方法及其装置 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110086662A (zh) | 2019-08-02 |
| CN110086662B (zh) | 2020-09-01 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108259367B (zh) | 一种基于软件定义网络的服务感知的流策略定制方法 | |
| US11281486B2 (en) | Bandwidth control in multi-tenant virtual networks | |
| CN108809857B (zh) | 一种基于sdn的流量监控与业务服务质量保障策略的方法 | |
| WO2020258920A1 (zh) | 一种网络切片资源管理方法及设备 | |
| CN107395501B (zh) | 一种网络服务功能链的跨域部署方法 | |
| WO2018086569A1 (zh) | 一种基于虚拟网络的应用感知的动态sdn配置方法 | |
| CN107659419A (zh) | 网络切片方法和系统 | |
| CN114756340A (zh) | 算力调度系统、方法、装置和存储介质 | |
| CN112491619B (zh) | 一种基于sdn的服务定制网络资源自适应分配方法 | |
| CN104092756B (zh) | 一种基于dht机制的云存储系统的资源动态分配方法 | |
| CN108777633B (zh) | 支持数据调度的意图型工业sdn北向接口系统及交互方法 | |
| CN106533806B (zh) | 多租户SDN网络中基于应用感知提供跨层QoS的方法 | |
| WO2021169308A1 (zh) | 一种数据流类型识别模型更新方法及相关设备 | |
| CN109150604B (zh) | 一种基于sdn的电力通信网络系统及跨域切片方法 | |
| CN110213175B (zh) | 一种面向知识定义网络的智能管控系统及管控方法 | |
| CN105871745A (zh) | 一种sdn网络中基于混合业务的调度方法 | |
| CN113259355A (zh) | 一种基于sdn的工业互联网标识切片管理系统 | |
| WO2020215504A1 (zh) | 一种需求定义网络的实现方法及网络架构 | |
| KR20210016802A (ko) | 소프트웨어 정의 네트워킹 환경에서 서버-클라이언트 기반의 네트워크 서비스를 위한 플로우 테이블을 최적화하는 방법 및 이를 위한 sdn 스위치 | |
| CN107360483B (zh) | 一种用于软件定义光网络的控制器负载均衡算法 | |
| CN110351204A (zh) | 云数据中心网络资源动态分配的方法及系统 | |
| WO2022016969A1 (zh) | 一种数据处理方法及装置 | |
| Savchenko et al. | Computer networks monitoring and management methods | |
| Pang et al. | Research on dynamic load balancing of data center network based on SDN architecture | |
| CN117097681B (zh) | 网络资源的调度方法、装置和存储介质及电子设备 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 19926606 Country of ref document: EP Kind code of ref document: A1 |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| 122 | Ep: pct application non-entry in european phase |
Ref document number: 19926606 Country of ref document: EP Kind code of ref document: A1 |