CN113490279B - Network slice configuration method and device - Google Patents
Network slice configuration method and device Download PDFInfo
- Publication number
- CN113490279B CN113490279B CN202110681375.9A CN202110681375A CN113490279B CN 113490279 B CN113490279 B CN 113490279B CN 202110681375 A CN202110681375 A CN 202110681375A CN 113490279 B CN113490279 B CN 113490279B
- Authority
- CN
- China
- Prior art keywords
- network
- slice
- network slice
- virtual network
- virtual
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000002955 isolation Methods 0.000 claims abstract description 54
- 230000006870 function Effects 0.000 claims description 82
- 238000004590 computer program Methods 0.000 claims description 6
- 230000002787 reinforcement Effects 0.000 claims description 6
- 230000007246 mechanism Effects 0.000 abstract description 8
- 238000005457 optimization Methods 0.000 description 12
- 230000009471 action Effects 0.000 description 7
- 238000004422 calculation algorithm Methods 0.000 description 5
- 238000005516 engineering process Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 4
- 238000013459 approach Methods 0.000 description 3
- 238000004364 calculation method Methods 0.000 description 3
- 230000006855 networking Effects 0.000 description 3
- 230000008569 process Effects 0.000 description 3
- 230000001186 cumulative effect Effects 0.000 description 2
- 238000012545 processing Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000004458 analytical method Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 230000015556 catabolic process Effects 0.000 description 1
- 230000000052 comparative effect Effects 0.000 description 1
- 238000007596 consolidation process Methods 0.000 description 1
- 238000006731 degradation reaction Methods 0.000 description 1
- 230000001419 dependent effect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 230000005055 memory storage Effects 0.000 description 1
- 239000002184 metal Substances 0.000 description 1
- 238000010606 normalization Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 238000000926 separation method Methods 0.000 description 1
- 238000004904 shortening Methods 0.000 description 1
- 230000011664 signaling Effects 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W72/00—Local resource management
- H04W72/50—Allocation or scheduling criteria for wireless resources
- H04W72/54—Allocation or scheduling criteria for wireless resources based on quality criteria
-
- 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
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W24/00—Supervisory, monitoring or testing arrangements
- H04W24/02—Arrangements for optimising operational condition
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W28/00—Network traffic management; Network resource management
- H04W28/16—Central resource management; Negotiation of resources or communication parameters, e.g. negotiating bandwidth or QoS [Quality of Service]
- H04W28/24—Negotiating SLA [Service Level Agreement]; Negotiating QoS [Quality of Service]
Abstract
The application discloses a network slice configuration method and a device, wherein the network slice configuration method is executed on a controller in a network slice management system and comprises the following steps: acquiring a network slice; wherein the network slice corresponds to a sequence of at least two virtual network function units; in the corresponding configuration thread, determining a state set of resource utilization conditions of the infrastructure network by the network slice; and deploying the virtual network function units in the sequence according to the safety requirement of the network slice bearing application and the state set, so as to realize the configuration of the network slice. The method and the device realize the customized isolation mechanism of the slice and improve the usability of the slice.
Description
Technical Field
The present disclosure relates to the field of network slicing technologies, and in particular, to a network slice configuration method and device.
Background
The 5G era has rich vertical industry application, and the requirements of each service on time delay, bandwidth, node load balance and the like are quite different. A single physical network has failed to simultaneously meet the SLA requirements of the various vertical industries described above. Personalized, differentiated business requirements drive the generation of 5G network slices.
Network slicing is an on-demand networking manner, which allows operators to separate multiple virtual end-to-end networks on a unified infrastructure, each logically isolated from the wireless network, the carrier network, and then to the core network to adapt to various types of applications. NFV (network function virtualization) which is a core of the network slicing technology, the NFV separates hardware and software parts from a traditional network, the hardware is deployed by a unified server, the software is borne by different network functions, a logical network can be generated on one physical network according to needs, and a plurality of slices can be provided for the same user to realize various bandwidth, time delay, node load balancing and slice isolation services under different needs, so that the requirements of flexible assembly service are realized.
The network slice provides a logical network of specific network functions and network characteristics. The network slice decouples the network resources from the deployment positions, so that the flexibility and the resource utilization rate of the network service are improved. How to intelligently arrange network functions according to the requirements of service scenes is an important problem for 5G network slice operation.
Disclosure of Invention
In view of this, the present application proposes a network slice configuration method and apparatus, which solves the problem how to intelligently arrange network functions according to the needs of service scenarios in 5G network slice operation.
To achieve the above object, an embodiment of the present application provides a network slice configuration method, which is executed on a controller in a network slice management system, including:
acquiring a network slice; wherein the network slice corresponds to a sequence of at least two virtual network function units;
in the corresponding configuration thread, determining a state set of resource utilization conditions of the infrastructure network by the network slice;
and deploying the virtual network function units in the sequence according to the safety requirement of the network slice bearing application and the state set, so as to realize the configuration of the network slice.
Correspondingly, in order to achieve the above object, the embodiments of the present application provide a network slice configuration device, which includes a memory and a processor; wherein the memory is configured to store computer program instructions; the processor is configured to execute the computer program instructions to implement the network slice configuration method described above.
Through the technical means, the following beneficial effects can be realized:
the application provides a network slice configuration scheme which is used for intelligent collaborative arrangement of network slices for availability and QoS optimization. Firstly, the scheme starts from a three-level safety isolation mode of the slice, and the customized isolation mechanism of the slice is realized by limiting the differentiated isolation level requirement as a constraint condition, specifically comprising inter-slice isolation and intra-slice isolation, so that the usability of the slice is improved. And secondly, by cooperating with the core data center and the edge data center, the bandwidth is effectively saved and the network delay is reduced in terms of QoS optimization. In addition, the method provides a dedicated arrangement optimization target matched with the service requirement, and differentiated requirements of different slices on time delay, bandwidth and node load balance are distinguished.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described below, it being obvious that the drawings in the following description are only some embodiments of the present application, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a diagram of an overall architecture of a 5G network slice;
fig. 2 is a flowchart of a network slice configuration method proposed in the present application;
fig. 3 is a schematic diagram of a network slice configuration device provided in the present application.
Detailed Description
The following description of the embodiments of the present application will be made clearly and fully with reference to the accompanying drawings, in which it is evident that the embodiments described are only some, but not all, of the embodiments of the present application. All other embodiments, which can be made by one of ordinary skill in the art without undue burden from the present disclosure, are within the scope of the present disclosure.
NFV, network function virtualization, network Function Virtualization. Many types of network devices (e.g., servers, switches, storage, etc.) are built as one Data Center Network, carrying many functional software processes using general-purpose hardware such as x86, and virtualization techniques. Thereby reducing the expensive equipment cost of the network. The network equipment functions are not dependent on special hardware, resources can be fully and flexibly shared, and quick development and deployment of new services are realized.
VNF, which refers to a specific virtual network function, provides a certain network service, is implemented in a software layer, and is deployed in resources on the cloud side by using an infrastructure provided by NFVI. Cloud-side resources include, but are not limited to, virtual machines, containers, or bare-metal physical machines.
Further, standard architectures for one NFV include NFV infrastructure (NFVI), MANO (Management and Orchestration) and VNFs. Wherein the VNF is a virtual network function unit in the NFV architecture. It can be understood that the process of performing function virtualization on an existing physical network element in a telecommunications service network will be deployed in the form of a software module on a virtual resource provided by NFVI, so as to implement network function virtualization. From this, it is clear that the NFV and VNF are in the same letter-to-letter order, and the meanings are quite different. NFV is a virtualization technology or concept that solves the problem of deploying network functions on general-purpose hardware.
A hardware electronic device is used as an end side device to form a network, and bit data streams are all used on the network in a galloping way. After fully researching various types of services of all terminal side devices, the method discharges priorities for the requirements of different services, preferentially guarantees the service with high network requirements, and then gives consideration to the service with low priority, which is the practical requirement for network slicing.
As shown in fig. 1, a 5G network slice overall architecture diagram is provided. The 5G end-to-end network slicing refers to flexible allocation of network resources, networking according to needs, virtual forming of a plurality of mutually isolated logic subnets with different characteristics based on the 5G network, and each end-to-end network slicing is formed by combining wireless network, transmission network and core network subslices and unified management is performed through an end-to-end slicing management system. In the uRLLC slice, scenes such as automatic driving/auxiliary driving, remote control and the like have extremely severe time delay requirements on a network. In mctc slicing, a large-scale internet of things service scenario has massive connections, and the amount of data interacted in the network is small, requiring high computing resources and low congestion. In the eMBB slice, the large-flow mobile broadband service such as 3D/ultra-high definition video has the characteristics of high user data rate and high bandwidth. Based on the method, resources are reasonably configured, limited networks are utilized, different network slices are configured through the slicing network according to different service demands on the networks, and operators can flexibly provide personalized network services for users at low cost according to third party demands and network conditions.
In order to understand the existing network slice configuration scheme, the existing technical materials are searched, compared and analyzed, and the following literature scheme is found:
literature scheme 1: in Optimal virtual network function placement in multi-cloud service function chaining architecture, authors have studied the problem of virtual network function placement, aimed at achieving optimal composition SFC in a cloud distributed across geographies. Authors target minimizing inter-cloud traffic and response time in a multi-cloud scenario, and build ILP optimization problems under important constraints such as the cost of the headquarter and service level agreements. They propose affinity-based approaches to solve the problem in large network topologies.
Literature scheme 2: in Adaptive Interference-Aware VNF Placement for Service-Customized 5GNetwork policies, authors consider both edge cloud servers and core cloud servers to deploy the required VNFs on demand for each network slice, aiming to maximize the total throughput of accepted requests. And a demand model is proposed to quantify the performance degradation that may result from VNF consolidation. The main idea is to place delay sensitive slices to the edge side with a greater probability.
The above-described document scheme 1 and document scheme 2 employ a scheme using only a simple strategy when cloud-edge cooperation is considered. The former when considering deployment of core or edge clouds uses the placement restrictions imposed by SLAs, i.e. some VNFs have to be placed on the core network and some on the edge network. This approach only constrains deployment locations and does not allow flexible placement of slice network functions. The latter places delay sensitive slices to the edge side with a greater probability based on a simple greedy strategy. These methods lack the best way to flexibly achieve bandwidth savings and reduced network latency.
Literature scheme 3: in A Service-Oriented Deployment Policy of End-to-End Network Slicing Based on Complex Network Theory, authors propose a collaboration strategy for slice Service demand oriented. The overall goal is to minimize placement costs. For three typical use cases of 5G scenarios (enhanced mobile broadband (eMBB), large-scale machine type communication (emtc) and ultra-reliable and low-latency communication (ullc)), three sub-optimization objectives peculiar objectives are proposed and a corresponding greedy algorithm is designed to solve the corresponding orchestration scheme.
Literature scheme 3 designs proprietary orchestration targets for three typical end-to-end slices and proprietary orchestration algorithms that fit the slice requirements. Node deployment in the orchestration process combines infrastructure network structural features and physical node resources to define node importance indicators. But this approach does not take into account the slice isolation customization requirements of slice orchestration.
Based on this, further, current network slice studies still have some key points to consider:
first, a customized slice isolation mechanism is implemented. The 5G network slice provides differentiated on-demand service customization capability for business features of different vertical industries, and security is one of the key factors that must be considered. Unlike the privacy and closeness of traditional physical private networks, 5G network slices are virtualized private networks built on top of shared resources, which, for the security of network slices, in addition to providing traditional mobile network security mechanisms (e.g., access authentication, encryption and integrity protection of access layer and non-access layer signaling and data, etc.), also need to provide an isolation mechanism between network slices. In order to meet the isolation level requirements of different services and improve the availability of slices, network slices can comprehensively adopt different inter-slice isolation and intra-slice isolation mechanisms to realize multiple isolation protection.
And secondly, comprehensively utilizing core edge resources. With the mature business of the 5G and vertical industries, networks need to access more devices and handle massive amounts of data, however low latency and high bandwidth computational load pressures make current centralized "core" data processing modes indispensible. In addition, AR/VR, industry internet, car networking etc. put forward higher demands on communication delay. The proposal of edge computation greatly facilitates the deployment implementation of network slices. Thus, some of the functions of the network slice may be sunk from the core data center to the network edge data center, constituting edge computing capabilities. By shortening the link distance and improving the intelligent capability of the edge network, the effects of saving the backhaul bandwidth, reducing the network delay and intelligently supporting the user experience are achieved.
Again, consider distinguishing between slice traffic types. The customized network slice based on the service requirement can enable the 5G network to have good service adaptability, and can meet the differentiated requirements among different services, such as time delay, bandwidth, node load balancing and the like. Along with the remarkable demand difference among the three major types of services, in order to improve the user experience, the distinction of network slice service types should be emphasized. Network slice orchestration is required to provide orchestration targets and policies that match traffic demands, meeting flexible and diverse differentiated traffic demands.
Finally, the slice arrangement is enhanced to be intelligent. The existing slice arrangement research mostly adopts a solution optimal scheme or proposes a heuristic algorithm to solve a suboptimal solution, and the solution is not well balanced between the time complexity and the optimal solution. Furthermore, slice orchestration requires automation and intelligence to save costs, fast deployment and accommodate network changes. Based on a 5G full cloud network architecture, intelligent arrangement of a slice network can be realized by introducing a mature AI technology taking machine learning as a core. The introduction of AI can better weigh the time complexity and realize the optimal solution.
Based on the analysis, the application provides an intelligent collaborative arrangement scheme of network slices for availability and QoS optimization. Firstly, the scheme limits the differentiated network slice isolation level requirements to constraint conditions, thereby realizing customized isolation of the network slices and improving the availability of the network slices. And secondly, by cooperating with the core data center and the edge data center, the bandwidth is effectively saved and the network delay is reduced in terms of QoS optimization. In addition, a dedicated arrangement optimization target matched with the service requirement is provided, and the differentiated requirements of different network slices on time delay, bandwidth and node load balancing are distinguished.
Fig. 2 is a flowchart of a network slice configuration method proposed in the present application. The network slice configuration method is executed on a controller in the network slice management system shown in fig. 1. The method comprises the following steps:
step 201): acquiring a network slice; wherein the network slice corresponds to a sequence of at least two virtual network function units.
In this embodiment, the network slices are logically isolated networks on the same infrastructure network. Thus customizing the service corresponding to each business need into network slices s k Expressed as a collection of network functions required to contain network services Where M represents the number of functions required. In a set s of network functions k One virtual network function unit (VNF) for each network function.
Step 202): in a corresponding configuration thread, a set of states for resource utilization of the infrastructure network by the network slice is determined.
In the technical scheme, aiming at the same network slice, the network slice is split into at least two threads to execute the arrangement algorithm of the network slice in parallel, so that the aim of accelerating the configuration of the network slice is fulfilled. The number of threads may be determined based on device processor performance, with each thread executing the same network slice configuration flow.
The network slice configuration needs to consider the resource utilization of the infrastructure network, so the scheme calculates the remaining resource duty ratio of each physical node and each link of the infrastructure network. In addition, the present solution contemplates customizing network slices, and deployment decisions are also related to inter-slice isolation levels and intra-slice isolation levels. The state set of the network slice configuration can thus be represented as an m+n+2-dimensional feature:
wherein the first M elements { w 1 ，...w M The resource usage of a physical node is represented by the next N elements v 1 ，...v N The second last element represents the bandwidth usage of the linkRepresenting inter-chip separation level, last element K rank Indicating the level of on-chip isolation.
Step 203): and deploying the virtual network function units in the sequence according to the safety requirement of the network slice bearing application and the state set, so as to realize the configuration of the network slice.
The security isolation of network slices can be divided into inter-slice isolation and intra-slice isolation. According to the requirements of the application carried by the slices on the safety, such as the resource competition degree between the slices, the information protection requirement between the slices and the like, the inter-slice isolation can be further divided into physical isolation and logical isolation. Physical isolation is to allocate independent physical resources for network slices, wherein each network slice monopolizes the physical resources without affecting each other, and the physical isolation is similar to a traditional physical private network, and has application scenes such as industrial control application with strict requirements on safety. Logic isolation is realized by using an NFV-based resource isolation technology, network functions borne by different virtual machines or containers are allocated to different network slices, and isolation of the slices in a basic resource layer is realized through an isolation mechanism of the virtual machines or containers. For example, for isolation level L1, physical isolation is provided, and independent physical resources are allocated. For the isolation level L2, full logical isolation is provided, and independent network functions are allocated when deploying network slices. For the isolation level L3, shared logic isolation is provided, sharing part of the network functions. Logical isolation is a resource isolation technique using NFV based, and therefore usesRepresenting network slice s k If there is a physical isolation requirement between slices, if the network slice s k Physical isolation between slices is required ∈ ->The value is 1. Different network functions need to provide mutual isolation between the network functions according to their own security level requirements and trust relationships. In addition to flexibility and economic cost considerations, the on-chip isolation level of a network slice may be represented by a ratio of the total number of physical nodes deploying the network slice to the total number of network functions required by the network slice. Thus, K is used rank Indicating an intra-slice isolation level, the higher the value, indicating a higher intra-slice isolation requirement for the network slice. In the technical scheme, the differentiated network slice isolation level requirements are limited to constraint conditions, customized isolation of the network slices is realized, and the availability of the network slices is improved.
In another embodiment, the infrastructure network is divided into an edge data center and a core data center based on the solution shown in fig. 2. In the present solution, the infrastructure network is divided into K edge data centers and M core data centers, each of which may use weighted undirected graphsAnd the K is more than or equal to 1 and less than or equal to K+M. Wherein N is s Representing a set of physical nodes, one of which is for +.>Expressed, L s Representing a set of physical links between physical nodes, wherein physical nodes +.>And->Link between->And (3) representing. Use->Representing a set of physical nodes on an edge data center, +.>Representing a set of physical nodes on the core data center. For each physical node i, it is assumed that the available computing resources are +.>The remaining computing resources are +.>The bandwidth resource of each link is +.>The residual link bandwidth resource is->The link delay is +.>Physical nodes in core data center +.>With relatively sufficient resource capacity and high latency, while physical nodes in edge data centers are +.>Closer to the end user, the resource capacity is limited but the time delay is low.
In the network slice configuration scheme,and if so, executing the network slice arrangement algorithm. Wherein (1)>The meaning of the expression is: a sum of remaining computing resources of each physical node in the infrastructure network; />The meaning of the expression is: the network slice contains the sum of the computing resources of the corresponding physical node used by each network function at deployment.
For the technical scheme, a new problem model is provided from the perspective of slice customization isolation mechanism and core edge cooperation, and bandwidth is effectively saved and network delay is reduced in QoS optimization through cooperation of the core data center and the edge data center.
Further, since the network slice corresponds to a sequence of at least two virtual network function units. Deploying a first virtual network function unit in the sequence according to the safety requirement of the application borne by the network slice; and deploying other virtual network function units in the sequence according to the reinforcement learning strategy.
In this case, for the first virtual network function in the sequence, deploying, in the corresponding configuration thread, the first virtual network function of the network slice on the physical node of the infrastructure network if the network slice allows sharing of virtual network functions; wherein the physical nodes of the infrastructure network have deployed virtual network function units of the same type as the first virtual network function unit of the network slice, and the physical nodes of the infrastructure network also allow sharing of virtual network function units. Otherwise, deploying the first virtual network function unit on a physical node corresponding to the minimum load of the resources in the infrastructure network according to the slice type of the first virtual network function unit of the network slice.
For other virtual network function units except the first virtual network function unit in the sequence, in the corresponding configuration thread, determining the proprietary parameters of the corresponding configuration thread according to feedback values received by the deployment strategy executed on the other virtual network function units according to each state in the state set; wherein the feedback value is determined according to a feedback function; and performing deployment on other virtual network function units on physical nodes in the infrastructure network according to the reinforcement learning strategy by using the state set and the corresponding proprietary parameters.
In further detail, the scheme aims at comprehensively optimizing different requirements of three typical network slices, so that an optimized objective function consists of slice delay, load balancing of physical nodes and link bandwidth consumption. In the uRLLC slice, scenes such as automatic driving/auxiliary driving, remote control and the like have extremely severe time delay requirements on the network, and the first part of the objective function represents the minimum slice time delay. In mctc slicing, a large-scale internet of things service scenario has massive connections, and the amount of data interacted in the network is small, requiring high computing resources and low congestion. Thus, the second part of the objective function represents load balancing of the physical nodes. In the eMBB slice, the large-flow mobile broadband service such as 3D/ultra-high definition video has the characteristics of high user data rate and high bandwidth, so the third part of the objective function represents the minimum link bandwidth consumption. For customized requirements, different weights are set for the three slices to integrate the three key performance indexes, so that a specific optimization target matched with the requirements is set. The optimization objective function expression is:
wherein alpha is the time delay weight of the network slice; beta is the load balancing weight of the physical node; gamma is the bandwidth consumption weight. These three parameters are dynamically adjusted according to the requirements of each network slice. In this example, α, β, γ were set to 0.8, 0.1, respectively, for the uRLLC slice. For mctc slices, α, β, γ were set to 0.1, 0.8, 0.1, respectively. For the eMBB slices, α, β, γ were set to 0.1, 0.8, respectively. For the present technical solution, the values of α, β, γ are not limited to the above-mentioned network slices of different traffic types.
The current network slice is organized in the current configuration thread with a state set s that takes into account the resource utilization of the infrastructure network t . Where t is used to characterize the number of steps of each thread that execute the network slice configuration. For state set s t Selecting a corresponding physical node from the infrastructure network for a virtual network function to be deployed according to a feedback function, and deploying the Virtual Network Function (VNF). And physical links are deployed in all paths between the physical nodes required to deploy this VNF and the physical nodes required to deploy the last VNF. The action set can thus be expressed as: a= { n, P }, where n represents the selected physical node and P represents the selected physical link.
Executing a different set of actions a for each state receives different feedback, but the selected action may violate the constraint. Thus, when the selected action violates the constraint, the feedback value is set to-1. When the constraint condition is satisfied, defining a feedback function as using formula (b);
wherein Z is a positive number for adjusting the feedback value to be a positive number;system load balance, namely the variance of the global node resource usage, when the state set st+1 is represented; d is the propagation delay between the physical node selected when the current VNF is deployed and the physical node selected when the previous VNF is deployed; b tableShowing the physical link resource consumption. α, β, γ represent coefficients, including normalization of data.
For other virtual network function units in the corresponding sequence of network slices, except for the first virtual network function unit, according to the reinforcement learning strategy pi (a t |s t The method comprises the steps of carrying out a first treatment on the surface of the θ') selection action a t And deploying other virtual network function units. And after the action is finished, calculating a corresponding feedback value for each VNF of the network slice corresponding sequence by using the formula (b). If all the VNFs corresponding to the current network slice complete configuration or arrangement, according to s t And (3) determining the Rt value corresponding to the current configuration thread by using a formula (c). The expression of formula (c) is:
and (3) calculating a corresponding feedback value and an Rt value corresponding to the current configuration thread according to each VNF of the network slice corresponding sequence, and performing iterative calculation by using a formula (d) until the current network slice arrangement is completed. The expression of formula (d) is:
wherein t is start ＝t。
And R obtained by each iterative calculation is respectively and correspondingly accumulated according to a formula (e) and a formula (f), so that global sharing parameters corresponding to each iterative calculation are obtained. The expression of formula (e) is:
cumulative gradient wrt
The expression of formula (f) is:
cumulative gradient wrt
In the formula (e) and the formula (f)Of which, θ and θ v Are global shared parameters, theta 'and theta' v Are all proprietary parameters of the thread.
The global shared parameter corresponding to each iteration determines the proprietary parameter corresponding to the current thread. Due to the reinforcement learning strategy pi (a t |s t The method comprises the steps of carrying out a first treatment on the surface of the The special parameter theta 'of the current thread in theta') is changed, so that the technical scheme can learn on line according to the system state and feedback values given by the environment after different action set mapping is executed. Aiming at a typical slice, a dedicated arrangement optimization target matched with the service requirement corresponding to the network slice is provided, and the differentiated requirements of different slices on time delay, bandwidth and node load balancing are distinguished by adjusting weights.
Fig. 3 is a schematic diagram of a network slice configuration device according to the present application. Comprising the following steps: the system comprises a memory, a processor and a computer program stored on the memory and capable of running on the processor, wherein the processor realizes the network slice configuration method shown in fig. 2 when executing the computer program.
From the above description of embodiments, it will be apparent to those skilled in the art that the present application may be implemented in software plus a necessary general purpose hardware platform. Based on such understanding, the technical solutions of the present application may be embodied essentially or in a part contributing to the prior art in the form of a software product, which may be stored in a storage medium, such as a ROM/RAM, a magnetic disk, an optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to perform the methods described in the embodiments or some parts of the embodiments of the present application.
In this specification, each embodiment is described in a progressive manner, and identical and similar parts of each embodiment are referred to each other, and each embodiment is mainly described as different from other embodiments. In particular, for both client and server embodiments, reference may be made to the description of the embodiments of the method described above for a comparative explanation.
The application may be described in the general context of computer-executable instructions, such as program modules, being executed by a computer. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform particular tasks or implement particular abstract data types. The application may also be practiced in distributed computing environments where tasks are performed by remote processing devices that are linked through a communications network. In a distributed computing environment, program modules may be located in both local and remote computer storage media including memory storage devices.
Claims (8)
1. A network slice configuration method, performed on a controller in a network slice management system, comprising:
acquiring a network slice; wherein the network slice corresponds to a sequence of at least two virtual network function units;
in the corresponding configuration thread, determining a state set of resource utilization conditions of the infrastructure network by the network slice;
according to the safety requirement of the application borne by the network slice and the state set, deploying the virtual network function units in the sequence to realize the configuration of the network slice;
the step of deploying virtual network function units in the sequence comprises: deploying a first virtual network function unit in the sequence according to the safety requirement of the application borne by the network slice; deploying other virtual network function units in the sequence according to the reinforcement learning strategy, and determining the proprietary parameters of the corresponding configuration threads in the corresponding configuration threads according to feedback values received by the deployment strategy executed on the other virtual network function units according to each state in the state set; wherein the feedback value is determined according to a feedback function; and performing deployment on other virtual network function units on physical nodes in the infrastructure network according to the reinforcement learning strategy by using the state set and the corresponding proprietary parameters.
2. The method of claim 1, wherein the infrastructure network is divided into an edge data center and a core data center.
3. The method of claim 1, wherein the step of deploying the first virtual network function in the sequence comprises:
in a corresponding configuration thread, if the network slice allows sharing of virtual network function units, deploying a first virtual network function unit of the network slice on a physical node of the infrastructure network; wherein the physical nodes of the infrastructure network have deployed virtual network function units of the same type as the first virtual network function unit of the network slice, and the physical nodes of the infrastructure network also allow sharing of virtual network function units.
4. The method of claim 1, wherein the step of deploying the first virtual network function in the sequence comprises:
in the corresponding configuration thread, if the network slice does not allow sharing of virtual network function units already deployed by physical nodes of the infrastructure network, a first virtual network function unit of the network slice is deployed at a physical node corresponding to a minimum load of resources in the infrastructure network.
5. The method of claim 1, wherein the feedback function is determined based on an objective function; the objective function is determined according to differentiated requirements of the network slice corresponding to service requirements, which comprehensively consider minimizing slice delay, load balancing of the physical nodes and minimizing link bandwidth consumption.
6. The method of claim 1 or 2, wherein the set of states includes resource usage of physical nodes in the infrastructure network, bandwidth usage of links in the infrastructure network, a level of intra-slice isolation of the network slice, a level of inter-slice isolation of the network slice.
7. The method of claim 6, wherein the inter-slice isolation is physical isolation or complete logical isolation or shared logical isolation.
8. A network slice configuration device, wherein the network slice configuration device comprises a memory and a processor; wherein,
the memory is used for storing computer program instructions;
the processor for executing the computer program instructions to implement the network slice configuration method of any one of claims 1 to 7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110681375.9A CN113490279B (en) | 2021-06-18 | 2021-06-18 | Network slice configuration method and device |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110681375.9A CN113490279B (en) | 2021-06-18 | 2021-06-18 | Network slice configuration method and device |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113490279A CN113490279A (en) | 2021-10-08 |
| CN113490279B true CN113490279B (en) | 2024-02-06 |
Family
ID=77934088
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110681375.9A Active CN113490279B (en) | 2021-06-18 | 2021-06-18 | Network slice configuration method and device |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113490279B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115002783B (en) * | 2022-04-20 | 2024-04-05 | 燕山大学 | Industrial Internet of things resource dynamic allocation method based on network slicing |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN106792692A (en) * | 2016-12-27 | 2017-05-31 | 兴唐通信科技有限公司 | A kind of physics dicing method based on SDN technologies |
| CN107888425A (en) * | 2017-11-27 | 2018-04-06 | 北京邮电大学 | The network splitting and disposing method and apparatus of GSM |
| WO2018089634A1 (en) * | 2016-11-11 | 2018-05-17 | Intel IP Corporation | Network slice management |
| CN108495358A (en) * | 2018-03-09 | 2018-09-04 | 西安电子科技大学 | A kind of network slice selection method based on NFV |
| WO2018224151A1 (en) * | 2017-06-08 | 2018-12-13 | Huawei Technologies Co., Ltd. | Device and method for providing a network slice |
| CN109586938A (en) * | 2017-09-29 | 2019-04-05 | 华为技术有限公司 | The generation method and device of example traffic topology |
| CN110034944A (en) * | 2018-01-12 | 2019-07-19 | 华为技术有限公司 | Network splitting and disposing method and device thereof |
| CN110896355A (en) * | 2018-09-12 | 2020-03-20 | 华为技术有限公司 | Network slice selection method and device |
| WO2020258920A1 (en) * | 2019-06-26 | 2020-12-30 | 华为技术有限公司 | Network slice resource management method and apparatus |
| CN112822050A (en) * | 2021-01-05 | 2021-05-18 | 北京信息科技大学 | Method and apparatus for deploying network slices |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20160353367A1 (en) * | 2015-06-01 | 2016-12-01 | Huawei Technologies Co., Ltd. | System and Method for Virtualized Functions in Control and Data Planes |
| US10779186B2 (en) * | 2017-12-01 | 2020-09-15 | At&T Intellectual Property I, L.P. | Dynamic access slice pooling and software defined network controlled capabilities |
-
2021
- 2021-06-18 CN CN202110681375.9A patent/CN113490279B/en active Active
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018089634A1 (en) * | 2016-11-11 | 2018-05-17 | Intel IP Corporation | Network slice management |
| CN106792692A (en) * | 2016-12-27 | 2017-05-31 | 兴唐通信科技有限公司 | A kind of physics dicing method based on SDN technologies |
| WO2018224151A1 (en) * | 2017-06-08 | 2018-12-13 | Huawei Technologies Co., Ltd. | Device and method for providing a network slice |
| CN109586938A (en) * | 2017-09-29 | 2019-04-05 | 华为技术有限公司 | The generation method and device of example traffic topology |
| CN107888425A (en) * | 2017-11-27 | 2018-04-06 | 北京邮电大学 | The network splitting and disposing method and apparatus of GSM |
| CN110034944A (en) * | 2018-01-12 | 2019-07-19 | 华为技术有限公司 | Network splitting and disposing method and device thereof |
| CN108495358A (en) * | 2018-03-09 | 2018-09-04 | 西安电子科技大学 | A kind of network slice selection method based on NFV |
| CN110896355A (en) * | 2018-09-12 | 2020-03-20 | 华为技术有限公司 | Network slice selection method and device |
| WO2020258920A1 (en) * | 2019-06-26 | 2020-12-30 | 华为技术有限公司 | Network slice resource management method and apparatus |
| CN112822050A (en) * | 2021-01-05 | 2021-05-18 | 北京信息科技大学 | Method and apparatus for deploying network slices |
Non-Patent Citations (3)
| Title |
|---|
| Providing a Sliced，Secure，and isolated Software Infrastructure of Virtual Functions Through Blockchain Technology;Gabriel Antonio F等;2019 IEEE 20th International Conference on High Performance Switching and Routing;全文 * |
| 基于NFV和SDN的无线网络虚拟化关键技术研究;张恒升;;现代导航(03);全文 * |
| 基于网络切片的能源互联网***架构;杨德龙;万俐;李晦龄;丁慧霞;路兆铭;马宝娟;陈亚文;;电力信息与通信技术(01);全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN113490279A (en) | 2021-10-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Islam et al. | A survey on task offloading in multi-access edge computing | |
| Zhang et al. | Adaptive interference-aware VNF placement for service-customized 5G network slices | |
| US9485197B2 (en) | Task scheduling using virtual clusters | |
| CN110297699B (en) | Scheduling method, scheduler, storage medium and system | |
| Ali et al. | Real-time task scheduling in fog-cloud computing framework for iot applications: A fuzzy logic based approach | |
| Kim et al. | Multi-agent reinforcement learning-based resource management for end-to-end network slicing | |
| Chiti et al. | Virtual functions placement with time constraints in fog computing: A matching theory perspective | |
| WO2023024219A1 (en) | Joint optimization method and system for delay and spectrum occupancy in cloud-edge collaborative network | |
| Ullah et al. | Task classification and scheduling based on K-means clustering for edge computing | |
| CN116032767A (en) | Intelligent fusion identification network-oriented computing power service chain management and control system architecture | |
| WO2020134133A1 (en) | Resource allocation method, substation, and computer-readable storage medium | |
| Lin et al. | Column generation based service function chaining embedding in multi-domain networks | |
| Xiang et al. | Resource calendaring for mobile edge computing: centralized and decentralized optimization approaches | |
| CN113490279B (en) | Network slice configuration method and device | |
| Pereira et al. | Increasing the efficiency of fog nodes through of priority-based load balancing | |
| Sun et al. | Deepmigration: Flow migration for nfv with graph-based deep reinforcement learning | |
| Zhu et al. | Double-agent reinforced vNFC deployment in EONs for cloud-edge computing | |
| Elsharkawey et al. | Mlrts: multi-level real-time scheduling algorithm for load balancing in fog computing environment | |
| Ma et al. | Mobility-aware delay-sensitive service provisioning for mobile edge computing | |
| Recse et al. | Look-ahead VNF-FG embedding framework for latency-sensitive network services | |
| Yuan et al. | Virtual network function scheduling via multilayer encoding genetic algorithm with distributed bandwidth allocation | |
| Zeydan et al. | A multi-criteria decision making approach for scaling and placement of virtual network functions | |
| CN115665258B (en) | Priority perception deployment method of multi-target service function chain based on deep reinforcement learning | |
| US20230123074A1 (en) | Machine learning-based approaches for service function chain selection | |
| Narayana et al. | A research on various scheduling strategies in fog computing environment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |