A 5G-TSN industrial heterogeneous virtual network architecture and a virtual resource fine-grained scheduling method

Abstract

The present invention discloses a 5G-TSN industrial heterogeneous virtual network architecture and a fine-grained scheduling method for virtual resources, and relates to the field of industrial communications. NFV technology is used to virtualize the TSN intelligent gateway of the edge computing layer into a virtual machine with TSN functions; the SDN controller centrally controls the TSN intelligent gateway supporting SDN technology deployed at the edge computing layer, and combines with NFV technology to perform virtualized network modeling of the 5G-TSN industrial heterogeneous network from the perspective of resource scheduling, and schedules and manages network slices with different QoS belonging to different industrial applications. The present invention is oriented to the upper-layer dynamic application requirements of flexible manufacturing, takes into account the heterogeneous QoS requirements of on-site communication resources and data streams, introduces the age of information (AoI) as an indicator at the virtual level to characterize the information timeliness of 5G access network data transmission, and proposes a 5G-TSN industrial heterogeneous virtual network architecture and heterogeneous QoS mapping mechanism in combination with the characteristics of 5G data, which can realize fine-grained on-demand delivery of industrial multi-priority heterogeneous data streams.

Description

Technical Field

[0001] The present invention relates to the field of industrial communications, and in particular to a 5G-TSN industrial heterogeneous virtual network architecture and a fine-grained scheduling method for virtual resources. Background Art

[0002] Faced with a new wave of digitalization, how to realize the digitalization and intelligence of manufacturing has become one of the strategic directions for the country's future industrial competitiveness. Therefore, it is urgent to break through the barriers to the industrialization of advanced technologies and apply emerging communication technologies represented by the fifth-generation mobile communication technology (5G) and time-sensitive networking protocol (TSN) to industrial sites. Through wireless-wired integration, full-factor interconnection and end-to-end on-demand delivery of "man-machine-material-loop" and other aspects can be achieved.

[0003] In the face of future demands for small-batch, customized industrial production, traditional siloed industrial deployments are unable to meet the dynamic requirements of upper-layer applications, significantly hindering the development of industrial communication technologies. To address this, academia and industry have introduced Software-Defined Networking (SDN) and Network Function Virtualization (NFV) technologies to virtualize and allocate network resources at varying levels, regardless of the implementation details of the underlying technologies. Through centralized resource scheduling and dynamic instruction delivery, they provide a universal communication solution for applications with varying Quality of Service (QoS) requirements. Therefore, using SDN and NFV technologies to model virtualized networks in 5G-TSN industrial heterogeneous networks will help break down the technical barriers to heterogeneous protocol conversion and design scheduling mechanisms for the new 5G-TSN network architecture from a resource perspective.

[0004] While virtualized network modeling eliminates the need for technical implementation details such as protocol conversion and precise clock synchronization, integrating 5G and TSN networks from a resource perspective presents numerous challenges. First, while 5G technology offers significant improvements in latency and reliability compared to 4G-LTE and other cellular technologies, for future industrial networks centered around the integrated "sensing-transmission-computing-control" model, the ability to collect field information in real time and deliver it deterministically from end to end significantly impacts data computation efficiency and control quality. Furthermore, industrial sites face complex electromagnetic environments, numerous mobile devices, and severe multipath fading. The improvements in communication quality and efficiency offered by 5G communication mechanisms cannot fully overcome the inherent uncertainties of wireless communications. Therefore, ensuring end-to-end deterministic data delivery when integrating heterogeneous 5G and TSN networks presents a significant challenge. Secondly, for customized production requirements with varying QoS requirements, the ability to differentiate product quality is a key performance criterion. To this end, TSN and 5G each specify mechanisms such as frame preemption, Time Aware Shaper (TAS), and Per-stream Filtering and Policing (PSFP), as well as parameters such as 5QI (5G QoS Identifier) and IPV (Internal Priority Value) to provide differentiated services for multi-priority data. However, as two emerging communication technologies, achieving data QoS interoperability between them is key to 5G-TSN convergence.

[0005] Age of Information (AoI), a parameter used to measure the timeliness of information, represents the elapsed time between the generation of a data packet and its arrival at its destination. It is also related to sampling frequency and control efficiency, and is a key indicator of the timeliness of status updates in industrial systems. Therefore, using AoI to characterize the timeliness of 5G access network data transmission within the 5G-TSN industrial heterogeneous network architecture can effectively meet the integrated "sensing-transmission-computing-control" industrial communication requirements and further ensure end-to-end deterministic data delivery.

[0006] A literature search revealed the closest implementation solution, Chinese patent application number 202110702240.6, titled "A Heterogeneous Traffic Shaper for Industrial Networks Supporting 5G and TSN Interconnection." Specifically, it proposes three end-to-end (E2E) communication types for 5G and TSN interconnection. It uses a first-in, first-out (FIFO) rule for control traffic shaping, while credit-based scheduling (CBS) is used for non-control traffic. This differentiated shaping mechanism meets the varying QoS requirements of data. This method assigns control traffic to a separate high-priority queue for transmission. However, due to the FIFO rule, control traffic arriving later in the queue cannot be forwarded in real time, resulting in queuing delays. Patent application number 201980020000.9, titled "Time-Sensitive Networking Frame Preemption Across Cellular Interfaces," specifically reserves specific time and frequency resources for the transmission of high- and low-priority data. During the time period reserved for high-priority data, if high-priority data arrives, forwarding without waiting can be achieved; during the time period reserved for low-priority data, if high-priority data arrives, the low-priority data transmission can be interrupted and preempted. However, reserved resources will cause resource waste when the traffic prediction accuracy is low, and the two types of priorities cannot adapt to complex industrial situations. The patent application number is: 202110718079.1, and the name is: A 5G-TSN cross-domain QoS and resource mapping method, device and computer-readable storage medium. Its specific content is: According to the business flow, data packet arrival rate, and service rate, a 5G-TSN resource mapping conversion relationship model based on the Markov process is established, and the business flow is mapped to the 5G resources in sequence according to the priority of the TSN business flow. However, it mainly considers the mapping between physical resource blocks, and does not consider the relationship between priority mapping and latency guarantee. The patent application number is: 202110718076.8, and its name is: 5G and TSN joint scheduling method based on wireless channel information. The specific content is: adjusting the data retransmission factor in the 5G network and the transmission delay in the TSN network according to the channel quality information CQI, so as to achieve the purpose of deterministic transmission in the 5G-TSN joint network. However, it only divides the quality level of CQI according to the coarse granularity of the threshold, and does not perform fine-grained scheduling based on the TSN queue situation and 5G resource utilization. The patent application number is: 202080014340.3, and its name is: 5G system support for virtual TSN bridge management, QoS mapping and TSN Qbv scheduling. The specific content is: The specific TSN interface and its corresponding functions for QoS flow mapping in 5G and TSN are given, but no specific instructions are given for how to calculate the QoS mapping relationship.The patent application number is: 202080040110.4, and the name is: TSN and 5GS QoS mapping-user plane-based method. The specific content is: the specific 5GS interface and its corresponding functions for QoS flow mapping in 5G and TSN are given, but no specific instructions are given for how to calculate the QoS mapping relationship.

[0007] There is little existing joint research on 5G and TSN transmission technologies. Most of the research focuses on how to ensure differentiated transmission of multi-QoS data for TSN technology. Most of them use offline methods such as resource reservation and multiple mechanisms to provide differentiated services to meet the different priority requirements of TSN network data flows, which is not very flexible.

[0008] Existing research on 5G-TSN industrial heterogeneous networks mostly considers issues such as 5G and TSN clock synchronization and resource mapping in physical heterogeneous networks. There is no research on cross-domain scheduling of virtual resources in 5G-TSN industrial heterogeneous virtual network scenarios.

[0009] Existing research on TSN technology focuses on the Time Aware Gate (TAS) mechanism of IEEE 802.1Qbv technology or the Precision Time Protocol (PTP) mechanism of IEEE 802.1AS. However, no research has been conducted on mapping virtual network functions (VNFs) based on AoI at the virtual network level.

[0010] Most existing research on multi-priority data scheduling within TSN networks simply divides data into high-priority and low-priority data flows, or into control flows and non-control flows. There is little research on the problem of fine-grained multi-QoS differentiated scheduling when the number of priorities is greater than 2.

[0011] Therefore, researchers in this field are committed to developing a 5G-TSN industrial heterogeneous virtual network architecture and a fine-grained scheduling method for virtual resources. Aiming at the dynamic upper-layer application requirements of flexible manufacturing, and considering the heterogeneous QoS requirements of on-site communication resources and data flows, a 5G-TSN industrial heterogeneous virtual network architecture and heterogeneous QoS mapping mechanism are provided at the virtual level to enable fine-grained, on-demand delivery of multi-priority, heterogeneous industrial data flows. Summary of the Invention

[0012] In view of the above-mentioned defects of the prior art, the technical problem to be solved by the present invention is how to provide a specific operation mode of the 5G-TSN industrial heterogeneous virtual network architecture and a cross-domain scheduling scheme of virtual resources within the 5G-TSN network, facing the production scenarios of intelligent manufacturing proposed by Industry 4.0, combining the advantages of wired TSN, wireless 5G communication and network virtualization technology; under the 5G-TSN industrial heterogeneous virtual network architecture, how to provide adaptive reliable transmission and flexible hierarchical scheduling for multi-QoS data based on network slicing technology, so as to meet the low-latency and high-reliability communication requirements of small-batch, customized flexible production; how to combine new communication indicators such as AoI to guide the differentiated on-demand mapping of network virtual resources; how to solve the communication needs of massive heterogeneous data in industrial sites and perform fine-grained elastic resource mapping at the virtual level.

[0013] To achieve the above objectives, the present invention provides a 5G-TSN industrial heterogeneous virtual network architecture, including an industrial field layer, an edge computing layer, and an SDN controller;

[0014] Edge computing layer: NFV technology is used to virtualize the TSN smart gateway at the edge computing layer into a virtual machine with TSN functions. The virtual machine has the functions of a physical TSN smart gateway: 5G micro base station and heterogeneous protocol conversion;

[0015] SDN controller: Centrally controls the SDN-enabled TSN intelligent gateway deployed at the edge computing layer, and combines it with NFV technology to schedule and manage network slices with different QoS requirements for different industrial applications.

[0016] Furthermore, the devices at the industrial field layer use 5G time-frequency resource blocks in the form of network slicing. The field devices managed by the same TSN smart gateway share spectrum resources, and there is no resource competition relationship between the field devices managed by different TSN smart gateways.

[0017] Furthermore, after the field data is generated, it can only be injected into the TSN network through the TSN intelligent gateway that manages this area. That is, the starting point for each slice data stream to be injected into the TSN network (the TSN intelligent gateway for wireless data injection into the TSN network) is fixed.

[0018] Furthermore, the 5G communication network is a virtual depiction of the physical 5G communication network. Upper-layer applications are transmitted in the 5G-TSN industrial heterogeneous network in the form of network slices composed of VNFs in sequence. The data streams corresponding to the slices use the on-site time-frequency resource blocks in a flat manner. The TSN backbone network is a virtual depiction of the physical TSN network. Each TSN intelligent gateway is virtualized into a virtual machine with TSN intelligent gateway functions, which has computing and caching capabilities. After the application slice reaches the TSN network, the internal VNF will be mapped to the TSN VM in sequence based on the performance of the TSN VM, occupying the resources of the TSN VM. Only when all VNFs belonging to a network slice are mapped can it be regarded as an application delivered to the remote end through the TSN backbone network.

[0019] Furthermore, to meet the demand for real-time sampling of data streams in industrial scenarios, the information freshness (AoI) indicator is introduced to measure the information age of sliced data streams. The AoI starts to increase from the moment the data is generated on-site and stops increasing after it is transmitted to the remote end via the 5G network and TSN network.

[0020] Furthermore, slices belonging to different applications and their corresponding data flows have different priorities, so different slices have different end-to-end age upper bounds A. E2E ,5G-TSN industrial heterogeneous virtual network schedules multi-priority slice data streams on demand according to the age upper bound of each slice.

[0021] The present invention also provides a 5G-TSN industrial heterogeneous virtual resource fine-grained scheduling method, which is characterized by comprising the following steps:

[0022] Step 1: Divide the data generated by the field device into high-priority slice data and low-priority slice data according to the slice priority;

[0023] Step 2: Divide the industrial site time-frequency resources into time-frequency resource blocks (RBs). Each RB is the smallest indivisible unit, the corresponding frequency band is the minimum bandwidth that meets the Nyquist interval, and the corresponding time period is several OFDM symbol intervals.

[0024] Step 3: When the upper application slice is issued, the corresponding field equipment is triggered and the data belonging to the slice is generated. The data priority is consistent with the application slice priority. E2E Tiling uses on-site time-frequency resource blocks. Slices with lower latency requirements can be tiled in the time domain, while those with higher latency requirements can be tiled in the frequency domain. Since high-priority slice data has very high latency requirements, tiling is only allowed in the frequency domain and occupies only one RB in the time domain. When high-priority slice data arrives, it can preempt the already tiled low-priority slice data resources.

[0025] Step 4: The age of data increases linearly with time after it is generated on-site. After the transmission is completed, the age A of each slice data in the 5G network can be obtained. 5G , according to the upper limit A of the age of each slice E2E and its age in 5G networks 5G , perform fine-grained priority adjustments as follows:

[0026]

[0027] Step 5: Set the priority of each VNF belonging to the slice to p TSN That is, the TSN queue priority that the VNF needs to inject its data into when mapping the TSN VM. After all VNFs are mapped to the virtual TSN network in sequence, the 5G-TSN full-process age of the slice can be obtained, completing the end-to-end on-demand delivery of on-site data to the remote end.

[0028] Furthermore, in step 1, the real-time requirement of the low-priority slice data is lower than that of the high-priority slice data, but the real-time requirements of different low-priority slice data vary, and the low-priority slice data uses time-frequency resource blocks in a tiled manner.

[0029] Furthermore, in step 4, the high-priority slice data preempts the low-priority data, which can achieve fast forwarding without waiting in the 5G network; the preempted low-priority slice data is transmitted after the high-priority slice data transmission is completed, and the transmission delay in the 5G network is extended accordingly.

[0030] Furthermore, the fine-grained QoS mapping in step 5 fine-grainedly divides the original two priorities into eight priorities through queue injection, providing elastic transmission services based on the characteristics of the TSN network.

[0031] In a preferred embodiment of the present invention, the purpose of the present invention is to meet the upper-level dynamic application needs of flexible manufacturing, consider the on-site communication resources and data flow heterogeneous QoS requirements, and provide a 5G-TSN industrial heterogeneous virtual network architecture and heterogeneous QoS mapping mechanism from a virtual level, which can realize fine-grained on-demand delivery of industrial multi-priority heterogeneous data flows.

[0032] A 5G-TSN industrial heterogeneous virtual network architecture based on network function virtualization includes the following parts:

[0033] Industrial field layer: It is mainly composed of field equipment such as communication nodes, sensors, controllers, and actuators. The equipment communicates with each other using 5G. It is responsible for monitoring and uploading on-site working conditions, and executing corresponding decisions based on instructions from the upper layer.

[0034] The edge computing layer is primarily composed of multiple TSN smart gateways, connected via the TSN backbone. Each TSN smart gateway possesses significant computing power and can perform basic computing activities. Based on its physical location, it manages nearby field nodes, collects and processes field data. In this architecture, the TSN smart gateways function as edge computing nodes.

[0035] Data communication is carried out between the industrial field layer and the edge computing layer through 5G technology. The field devices managed by the same TSN smart gateway share spectrum resources, resulting in resource competition; there is no resource competition relationship between the field devices managed by different TSN smart gateways.

[0036] After field data is generated, it can only be injected into the TSN network through the TSN intelligent gateway that manages the area. That is, the starting point for each slice data stream to be injected into the TSN network is fixed.

[0037] The 5G communication network is a virtual portrayal of the physical 5G communication network. Specifically, upper-layer applications are transmitted in the 5G-TSN industrial heterogeneous network in the form of network slices composed of VNFs in sequence, and the data streams corresponding to the slices use on-site time-frequency resource blocks in a flat manner.

[0038] The TSN backbone network is a virtual depiction of the physical TSN network. Specifically, each TSN intelligent gateway is virtualized into a virtual machine (VM) with TSN intelligent gateway functions, which has computing and caching capabilities. After the application slice reaches the TSN network, its internal VNF will be mapped to the TSN VM in sequence based on the performance of the TSN VM, occupying the memory of the TSN VM. Only when all VNFs belonging to a network slice are mapped can the application slice be considered to be delivered to the remote end through the TSN backbone network.

[0039] Considering the need for real-time sampling of data streams in industrial scenarios, we introduce the AoI (Advanced Information) metric to measure the real-time performance of sliced data streams. Data begins aging from the moment it is generated on-site and stops aging after being transmitted to the remote endpoint via 5G and TSN networks.

[0040] Due to the different real-time requirements of upper-layer applications, the location, function, data volume, etc. of industrial field equipment are also different. Slices belonging to different applications and their corresponding data streams have different priorities. Therefore, different slices have different end-to-end age upper bounds A. E2E The above-mentioned 5G-TSN industrial heterogeneous virtual network needs to perform on-demand scheduling of multi-priority slice data streams based on the age upper bound of each slice.

[0041] A multi-priority virtual resource fine-grained scheduling method includes the following steps:

[0042] Step 1: Divide the data generated by the field device into high-priority slice data and low-priority slice data according to the slice priority.

[0043] Step 2: Divide the industrial site time-frequency resources into time-frequency resource blocks (RBs). Each RB is the smallest indivisible unit, its corresponding frequency band is the minimum bandwidth that meets the Nyquist interval, and its corresponding time period is several OFDM symbol intervals.

[0044] Step 3: When the upper application slice is issued, the corresponding field device is triggered and generates data belonging to the slice. The data priority is consistent with the application slice priority. E2E Tiling uses live time-frequency resource blocks. Slices with lower latency requirements can be tiled in the time domain, while those with higher latency requirements can be tiled in the frequency domain. Since high-priority slice data has very high latency requirements, tiling is only allowed in the frequency domain, occupying only one RB in the time domain. Therefore, when high-priority slice data arrives, it can preempt already tiled low-priority slice data resources.

[0045] Step 4: After the data is generated on-site, its age increases linearly with time. After the transmission is completed, the age A of each slice data in the 5G network can be obtained. 5G , according to the upper limit A of the age of each slice E2E and its age in 5G networks 5G , and fine-grained priority adjustments are as follows:

[0046]

[0047] Step 5: Set the priority of each VNF belonging to the slice to p TSN That is, the TSN queue priority that the VNF needs to inject its data into when mapping the TSN VM. All VNFs are mapped in sequence to obtain the 5G-TSN full-process age of the slice. At this point, the end-to-end on-demand delivery of field data to the remote end has been completed.

[0048] The real-time requirement of the low-priority slice data in step 1 is lower than that of the high-priority slice data, but the real-time requirements of different low-priority slice data are different, which is why the low-priority slice data uses time-frequency resource blocks in a tiled manner.

[0049] The time period corresponding to the RB in step 2 may be selected to include several OFDM symbols.

[0050] For the high-priority slice data preempting the low-priority data described in step 3, the preempted low-priority slice data is transmitted after the high-priority slice data is completed, and its transmission delay in the 5G network is extended accordingly.

[0051] For the fine-grained QoS mapping described in step 4, the original two priorities are finely divided into eight priorities through queue injection, which can better provide elastic transmission services based on the characteristics of the TSN network. Therefore, the priority of each slice data flow in the TSN network is different from its priority in the 5G network.

[0052] In another preferred embodiment of the present invention, the slice data flow priority is adjusted for each VNF when it is mapped to the TSN VM, rather than only when it is injected into the TSN network.

[0053] Compared with the prior art, the present invention has the following obvious substantial features and significant advantages:

[0054] 1. Combining the ubiquitous access characteristics of 5G communications and the deterministic transmission characteristics of TSN networks, and considering the flexibility of network virtualization in physical resource scheduling, the innovative 5G-TSN industrial heterogeneous virtual network architecture is proposed. From a virtual perspective, it describes how to provide elastic transmission services for different application slices.

[0055] 2. Considering the transmission uncertainty brought about by the open environment of 5G communication and the severe electromagnetic interference in industrial sites, the 5G-TSN fine-grained priority mapping mechanism flexibly adjusts transmission services to provide deterministic on-demand delivery of multi-QoS data from the site to the remote end.

[0056] 3. An innovative virtual resource scheduling process under the 5G-TSN industrial heterogeneous virtual network architecture is presented, including network slice tiling within the 5G network and mapping VNFs within the TSN network to TSN virtual machines, and a specific virtual network transmission path is designed.

[0057] 4. The use of network function virtualization technology provides a more flexible industrial transmission solution, eliminating the need to consider the underlying technical implementation details of the 5G-TSN combination (clock synchronization, protocol conversion, etc.), thereby improving the feasibility of deploying industrial heterogeneous network transmission solutions.

[0058] The concept, specific structure and technical effects of the present invention will be further described below in conjunction with the accompanying drawings to fully understand the purpose, characteristics and effects of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 This is a diagram of the 5G-TSN industrial heterogeneous virtual network architecture of a preferred embodiment of the present invention;

[0060] Figure 2 This is a 5G network slice tiling diagram of a preferred embodiment of the present invention;

[0061] Figure 3 This is a 5G-TSN fine-grained QoS mapping diagram of a preferred embodiment of the present invention;

[0062] Figure 4 This is a flowchart of multi-priority data slice 5G scheduling in a preferred embodiment of the present invention;

[0063] Figure 5 This is a multi-priority data slice TSN scheduling flowchart of a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0064] The following describes several preferred embodiments of the present invention with reference to the accompanying drawings to make its technical content clearer and easier to understand. The present invention can be embodied in many different forms of embodiments, and the scope of protection of the present invention is not limited to the embodiments mentioned herein.

[0065] In the drawings, components with identical structures are denoted by the same reference numerals, and components with similar structures or functions are denoted by similar reference numerals. The size and thickness of each component shown in the drawings are arbitrary and are not limited by the present invention. For clarity, the thickness of components in some places in the drawings is appropriately exaggerated.

[0066] The present invention provides a 5G-TSN industrial heterogeneous virtual network architecture based on network function virtualization, including the following parts:

[0067] SDN controller: Centrally controls the SDN-enabled TSN intelligent gateway deployed at the edge computing layer, and combines it with NFV technology to schedule and manage network slices with different QoS requirements for different industrial applications.

[0068] Edge computing layer: NFV technology is used to virtualize the TSN smart gateway at the edge computing layer into a virtual machine with TSN functions, considering only its own computing and storage resources without considering the underlying operating details. At the same time, the virtual machine must have the functions of a physical TSN smart gateway: 5G micro base station and heterogeneous protocol conversion, etc.

[0069] Devices at the industrial field layer use 5G time-frequency resource blocks in the form of network slicing. Field devices managed by the same TSN smart gateway share spectrum resources, and there is no resource competition between field devices managed by different TSN smart gateways.

[0070] After the field data is generated, it can only be injected into the TSN network through the TSN intelligent gateway that manages the area. That is, the starting point for each slice data stream to be injected into the TSN network (the TSN intelligent gateway for wireless data injection into the TSN network) is fixed.

[0071] The 5G communication network is a virtual representation of the physical 5G communication network. Specifically, upper-layer applications are transmitted across the 5G-TSN industrial heterogeneous network in the form of network slices composed of sequentially ordered VNFs. The data streams corresponding to the slices use on-site time-frequency resource blocks in a flat manner. The TSN backbone network is a virtual representation of the physical TSN network. Specifically, each TSN intelligent gateway is virtualized into a virtual machine with TSN intelligent gateway functions, equipped with computing and caching capabilities. When an application slice arrives on the TSN network, its internal VNFs are sequentially mapped to the TSN VM based on the performance of the TSN VM, occupying the TSN VM's memory. Only when all VNFs belonging to a network slice are fully mapped can the application be considered delivered to the remote end via the TSN backbone network.

[0072] Considering the demand for real-time sampling of data streams in industrial scenarios, the information freshness (AoI) indicator is introduced to measure the information age of sliced data streams. The AoI starts to increase from the moment the data is generated on-site and stops increasing after it is transmitted to the remote end via the 5G network and TSN network.

[0073] Due to the different real-time requirements of upper-layer applications, the location, function, data volume, etc. of industrial field equipment are also different. Slices belonging to different applications and their corresponding data streams have different priorities. Therefore, different slices have different end-to-end age upper bounds A. E2E ,The above-mentioned 5G-TSN industrial heterogeneous virtual network needs to perform on-demand scheduling of multi-priority slice data streams based on the ,age upper bound of each slice.

[0074] The present invention provides a multi-priority virtual resource fine-grained scheduling method, comprising the following steps:

[0075] Step 1: Divide the data generated by the field device into high-priority slice data and low-priority slice data according to the slice priority.

[0076] Step 2: Divide the industrial site time-frequency resources into time-frequency resource blocks (RBs). Each RB is the smallest indivisible unit, its corresponding frequency band is the minimum bandwidth that meets the Nyquist interval, and its corresponding time period is several OFDM symbol intervals.

[0077] Step 3: When the upper application slice is issued, the corresponding field device is triggered and generates data belonging to the slice. The data priority is consistent with the application slice priority. E2ETiling uses live time-frequency resource blocks. Slices with lower latency requirements can be tiled in the time domain, while those with higher latency requirements can be tiled in the frequency domain. Since high-priority slice data has very high latency requirements, tiling is only allowed in the frequency domain, occupying only one RB in the time domain. Therefore, when high-priority slice data arrives, it can preempt already tiled low-priority slice data resources.

[0078] Step 4: After the data is generated on-site, its age increases linearly with time. After the transmission is completed, the age A of each slice data in the 5G network can be obtained. 5G , according to the upper limit A of the age of each slice E2E and its age in 5G networks 5G , and fine-grained priority adjustments are as follows:

[0079]

[0080] Step 5: Set the priority of each VNF belonging to the slice to p TSN That is, the TSN queue priority that the VNF needs to inject its data into when mapping the TSN VM. All VNFs are mapped in sequence to obtain the 5G-TSN full-process age of the slice. At this point, the end-to-end on-demand delivery of field data to the remote end has been completed.

[0081] The real-time requirement of the low-priority slice data in step 1 is lower than that of the high-priority slice data, but the real-time requirements of different low-priority slice data vary, which is why the low-priority slice data uses time-frequency resource blocks in a tiled manner.

[0082] For the high-priority slice data described in step 4 that preempts the low-priority data, it can achieve fast forwarding without waiting within the 5G network; the preempted low-priority slice data is transmitted after the high-priority slice data is transmitted, and its transmission delay within the 5G network is extended accordingly.

[0083] For the fine-grained QoS mapping described in step 5, the original two priorities are finely divided into eight priorities through queue injection, which can better provide elastic transmission services based on the characteristics of the TSN network. Therefore, the priority of each slice data flow in the TSN network is different from its priority in the 5G network.

[0084] like Figure 1 As shown in the figure, taking a steel mill whose main industry is hot rolling as an example, a 5G-TSN industrial heterogeneous virtual network architecture mainly consists of the following two parts:

[0085] Industrial field layer: It is mainly composed of communication nodes, temperature and humidity sensors, vibration sensors, PLCs, cameras, rollers and other field equipment. The equipment communicates with each other using 5G. It is responsible for monitoring and uploading the on-site working conditions, roller conditions, thermal imaging, etc., and executing corresponding decisions according to instructions from the upper layer.

[0086] Edge computing layer: Primarily composed of TSN intelligent gateways, these gateways are connected via the TSN backbone network. Each TSN intelligent gateway possesses significant computing power and can perform basic computing activities. Based on their physical location, they are responsible for managing nearby field nodes, collecting and processing data. In this architecture, TSN intelligent gateways function as edge computing nodes. Data communication between the industrial field layer and the edge computing layer occurs via 5G technology. Field devices managed by the same TSN intelligent gateway share spectrum resources, while field devices managed by different TSN intelligent gateways do not compete for resources. Once generated, field data can only be injected into the TSN network via the TSN intelligent gateway managing that area. This means that each slice of data stream has a fixed starting point for injection into the TSN network.

[0087] A fine-grained scheduling method for virtual resources Figure 4 As shown, Figure 5 As shown, the following steps are included:

[0088] Step 1: Divide the data generated by the field device into high-priority slice data and low-priority slice data according to the slice priority.

[0089] Step 2: Divide the industrial site time-frequency resources into time-frequency resource blocks (RBs). Each RB is the smallest indivisible unit, its corresponding frequency band is the minimum bandwidth that meets the Nyquist interval, and its corresponding time period is 2 OFDM symbol intervals.

[0090] Step 3: If Figure 2 As shown, when the upper application slice is issued, the corresponding field device is triggered and generates data belonging to the slice. The data priority is consistent with the application slice priority. According to the upper limit A of the age of each slice itself E2E Tiling uses on-site time-frequency resource blocks. Those with lower latency requirements can be tiled in the time domain, while those with higher latency requirements can be tiled in the frequency domain.

[0091] Step 4: If Figure 2As shown in the figure, since high-priority slice data has very high latency requirements, it is only allowed to be tiled in the frequency domain, and it only occupies one RB time period in the time domain. Therefore, when high-priority slice data arrives, it can preempt the already tiled low-priority slice data resources. The preempted low-priority slice data will be transmitted after the high-priority slice data is transmitted, and its transmission latency in the 5G network is correspondingly extended.

[0092] Step 5: Figure 3 As shown in the figure, the age of data increases linearly with time after it is generated on site. After the transmission is completed, the age A of each slice data in the 5G network can be obtained. 5G , according to the upper limit A of the age of each slice E2E and its age in 5G networks 5G , use the following formula to adjust its fine-grained priority:

[0093]

[0094] Step 6: Set the priority of each VNF belonging to the slice to p TSN That is, the TSN queue priority that the VNF needs to inject its data into when mapping the TSN VM. All VNFs are mapped in sequence to obtain the 5G-TSN full-process age of the slice. At this point, the end-to-end on-demand delivery of field data to the remote end has been completed.

[0095] The preferred embodiments of the present invention have been described in detail above. It should be understood that numerous modifications and variations based on the concepts of the present invention are possible without inventive effort by those skilled in the art. Therefore, any technical solution that can be derived by one skilled in the art through logical analysis, reasoning, or limited experimentation based on the concepts of the present invention and the prior art should be within the scope of protection defined by the claims.

Claims

1. A 5G-TSN industrial heterogeneous virtual resource fine-grained scheduling method, characterized by: The following steps are involved: Step 1: Divide the data generated by the field device into high-priority slice data and low-priority slice data according to the slice priority; Step 2: Divide the industrial site time-frequency resources into time-frequency resource blocks (RBs). Each RB is the smallest indivisible unit, the corresponding frequency band is the minimum bandwidth that meets the Nyquist interval, and the corresponding time period is several OFDM symbol intervals. Step 3: When the upper application slice is issued, the corresponding field equipment is triggered and the data belonging to the slice is generated. The data priority is consistent with the application slice priority. E2E Tiling uses on-site time-frequency resource blocks. Slices with lower latency requirements can be tiled in the time domain, while those with higher latency requirements can be tiled in the frequency domain. Since high-priority slice data has very high latency requirements, tiling is only allowed in the frequency domain and occupies only one RB in the time domain. When high-priority slice data arrives, it can preempt the already tiled low-priority slice data resources. Step 4: The age of data increases linearly with time after it is generated on-site. After the transmission is completed, the age A of each slice data in the 5G network can be obtained. 5G , according to the upper limit A of the age of each slice E2E and its age in 5G networks 5G , perform fine-grained priority adjustments as follows: Step 5: Set the priority of each VNF belonging to the slice to p TSN , that is, the TSN queue priority that the VNF needs to inject its data into when mapping the TSN VM. After all VNFs are mapped to the virtual TSN network in sequence, the 5G-TSN full process age of the slice can be obtained, completing the end-to-end on-demand delivery of on-site data to the remote end; In step 4, the high-priority slice data preempts the low-priority data, which can achieve fast forwarding without waiting in the 5G network; the preempted low-priority slice data is transmitted after the high-priority slice data is transmitted, and the transmission delay in the 5G network is correspondingly extended; The fine-grained QoS mapping in step 5 divides the original two priorities into eight priorities through queue injection, providing elastic transmission services based on the characteristics of the TSN network.

2. The 5G-TSN industrial heterogeneous virtual resource fine-grained scheduling method according to claim 1 is characterized in that: In step 1, the real-time requirement of the low-priority slice data is lower than that of the high-priority slice data, but the real-time requirements of different low-priority slice data vary. The low-priority slice data uses time-frequency resource blocks in a tiled manner.

Images