Low latency networking method and device combined with dynamic priority preemption and time sensitive network

Abstract

This application provides a low-latency networking method and apparatus that integrates dynamic priority preemption and time-sensitive networking, relating to the field of communication network technology. It acquires the original communication service flow to be processed through the physical ingress port of the network switching node and performs service attribute parsing based on packet header metadata to construct a global attribute distribution of the service to be scheduled. A non-linear coupling relationship between service urgency and link resource occupancy is established to construct a dynamic priority preemption scheduling sequence for the service flow. If a target priority packet preempts a controlled priority packet, a preemptive heterogeneous packet slice association carrier is constructed. This carrier is input into a time-aware shaping algorithm engine to generate a time-gated list and construct a distributed time-sensitive networking scheduling gating sequence accordingly. This application improves the preemption agility of urgent service packets, enabling high stability of end-to-end transmission latency in dense node networks.

Description

Technical Field

[0001] This application relates to the field of communication network technology, and more specifically, to a low-latency networking method and apparatus that integrates dynamic priority preemption and time-sensitive networking. Background Technology

[0002] With the large-scale application of unmanned swarms in complex and dynamic environments, scenarios such as high-density node collaborative combat, intelligent munition guidance, and drone swarm search and rescue in disaster areas place extremely high demands on the real-time performance and reliability of wireless network communication. In dense unmanned node scenarios, high-priority control commands (such as emergency obstacle avoidance and task replanning) coexist with low-priority status data (such as sensor feedback and log reporting). The intense competition among multiple priority data streams and sudden traffic fluctuations pose even more stringent challenges to the deterministic control of end-to-end communication latency.

[0003] In existing low-latency transmission schemes for unmanned cooperative networks, a common architecture combines fixed-priority static scheduling with traditional carrier sense multiple access (CSM) mechanisms. This scheme first assigns static priorities to different service flows and allocates fixed time slots during network initialization. Then, during data transmission, each node competes for channel space and transmits data according to the static priority queue and a pre-defined backoff window. Finally, when encountering sudden changes in network load, a global clock is used to periodically reconstruct the network topology and resource allocation to alleviate network congestion.

[0004] However, this network control scheme, which combines static priority with contention-based access, has significant technical drawbacks. In high-frequency data interaction scenarios with dense nodes, fixed priority allocation strategies struggle to adapt to rapidly fluctuating network traffic loads. Furthermore, traditional conflict avoidance mechanisms are prone to causing high-priority emergency control commands to become queued and blocked for extended periods due to low-priority data occupying the channel during multi-priority data flow contention. Simultaneously, the lack of fine-grained packet preemption and multi-hop time-sensitive queue scheduling mechanisms results in significant uncertainty in the cumulative multi-hop transmission delay, failing to meet the deterministic low-latency network scheduling requirements of unmanned nodes under complex conditions such as sudden traffic surges. Summary of the Invention

[0005] This application provides a low-latency networking method and apparatus that integrates dynamic priority preemption and time-sensitive networking to at least alleviate the aforementioned technical problems.

[0006] A low-latency networking method integrating dynamic priority preemption and time-sensitive networking, comprising: Step 1: Obtain the original communication service flow to be networked through the physical ingress port of the network switching node and perform service attribute parsing based on the metadata of the message header to construct the global attribute distribution pattern of the scheduled service to be represented in the spatiotemporal domain of the network. Step 2: Based on the distribution of the global attributes of the services to be scheduled, establish a non-linear coupling relationship between the urgency of the services and the occupancy of the link resources, so as to generate a dynamic scheduling weight table for the service flow and construct a dynamic priority preemption scheduling sequence for the service flow to represent the order of processing of each service flow within a preset scheduling period. Step 3: Monitor whether there is a target priority message performing channel preemption processing on the controlled priority message in the dynamic priority preemption scheduling sequence of the service flow. If so, construct a preemptive heterogeneous message slice association carrier to characterize the interleaving distribution characteristics in the physical link. Step 4: Input the preemptive heterogeneous message slice association carrier into the time-aware shaping algorithm engine built based on the time-sensitive network protocol standard to generate a network transmission time slot allocation mapping relationship to generate a time gating list and construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow.

[0007] Optionally, step 1 includes: The original communication service flow is parsed based on the metadata of the message header to extract the basic service attribute feature group, which includes message arrival time, traffic burst tolerance and periodic interaction frequency. Based on the aforementioned service basic attribute feature group, a service category clustering identification based on latency sensitivity is performed to map the original communication service flow into time-sensitive service flow and non-time-sensitive service flow respectively; Based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in network bandwidth resources, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution status of different service categories in the network spatiotemporal domain.

[0008] Optionally, based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in network bandwidth resources, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution status of different service categories in the network spatiotemporal domain, including: Based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in the network bandwidth resources, the corresponding traffic packet length distribution is parameterized to obtain the traffic packet length distribution parameters; Determine the transmission time interval for network data transmission between the time-sensitive service flow and the non-time-sensitive service flow, and parameterize and characterize it to generate a transmission time interval parameter; Based on the traffic packet length distribution parameter and the transmission time interval parameter, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution pattern of different service categories in the network spatiotemporal domain.

[0009] Optionally, step 2 includes; Based on the dynamic scheduling weight table of the business flow, the feature parameters of each dimension in the global attribute distribution of the business to be scheduled are weighted to obtain the weighted feature parameters of each dimension. The weighted feature parameters of each dimension are mapped to the multidimensional network resource topology boundary space to generate a load drift evolution trajectory that characterizes the current service load deviating from the deterministic transmission range of the network. Based on the load drift evolution trajectory, a dynamic priority preemption scheduling sequence for service flows is constructed to characterize the order in which each service flow is processed within a preset scheduling period.

[0010] Optionally, based on the load drift evolution trajectory, a dynamic priority preemption scheduling sequence for service flows is constructed to characterize the order in which each service flow is processed within a preset scheduling period, including: The load drift evolution trajectory is subjected to spatiotemporal dimension projection processing to obtain spatiotemporal projection results; Based on the spatiotemporal projection results, load integration is performed to calculate the multidimensional composite load deviation degree used to characterize the intensity of flow fluctuations. Dynamic priority evaluation is performed based on the multidimensional composite working condition load deviation and the global attribute distribution of the services to be scheduled, so as to construct a dynamic priority preemption scheduling sequence for services to represent the order of processing of each service flow within a preset scheduling period.

[0011] Optionally, step 3 includes: When it is detected that a target priority message has entered the scheduling ready state and the network egress port is transmitting a controlled priority message that has a channel occupation conflict with the target priority message, the control egress queue manager issues a forced interrupt command to the controlled priority message and obtains the interrupt command execution result. When the execution result of the interruption instruction indicates that the controlled priority message is interrupted, the transmission position status of the controlled priority message is saved to obtain the status saving information of the interrupted message, so as to construct a preemptive heterogeneous message slice association carrier for characterizing the interleaving distribution characteristics in the physical link.

[0012] Optionally, a preemptive heterogeneous packet slice association carrier is constructed to characterize the interleaving distribution characteristics in the physical link, including: Based on the state storage information, the interrupted controlled priority message is subjected to data fragmentation and marking processing based on the minimum slice granularity to generate controlled priority fragmented message data. Based on the controlled priority fragmented message data and the target priority message that triggers the channel preemption process, a preemptive heterogeneous message slice association carrier for the interleaving distribution characteristics in the physical link is constructed.

[0013] Optionally, step 4 includes: The preemptive heterogeneous packet slice association carrier is input into a time-aware shaping algorithm engine built on the time-sensitive network protocol standard to perform mapping and matching calculations between packet slices and network bandwidth resources, so as to generate a network transmission time slot allocation mapping relationship. By calling the master clock in the distributed network, the time synchronization reference provided by the master clock is obtained and determined as the master clock time synchronization reference; Based on the master clock time synchronization reference, the gated scheduling time base of the time-aware shaping algorithm engine is aligned and calibrated to obtain the calibrated time synchronization reference; Based on the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, a time gating list is generated, and then a distributed time-sensitive network scheduling gating sequence for the original communication service flow is constructed.

[0014] Optionally, a time-gated list is generated based on the calibrated time synchronization reference and the network transmission time slot allocation mapping relationship, thereby constructing a distributed time-sensitive network scheduling gating sequence for the original communication service flow, including: Based on the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, the feature parameters of the time gating list are mapped to the gating switch logic in the time-aware shaping algorithm engine to obtain the feature parameter mapping result. The action timing is arranged according to the feature parameter mapping results to determine the opening and closing action status of the network egress port in each scheduling cycle. The opening and closing action states are compiled into the time gating list; Based on the time-gating list, the forwarding timing of each message slice in the preemptive heterogeneous message slice association carrier is time-series scheduled to construct a distributed time-sensitive networking scheduling gating sequence for the original communication service flow.

[0015] A low-latency networking device integrating dynamic priority preemption and time-sensitive networking, comprising: The global attribute situation construction module is used to obtain the original communication service flow to be networked through the physical inlet port of the network switching node and perform service attribute parsing based on the metadata of the message header to construct the global attribute distribution situation of the scheduled service to represent the distribution situation of different service categories in the network spatiotemporal domain. The dynamic priority scheduling sequence construction module is used to establish a non-linear coupling relationship between the urgency of the service and the occupancy of the link resources based on the distribution of the global attributes of the service to be scheduled, so as to generate a dynamic scheduling weight table for the service flow and construct a dynamic priority preemption scheduling sequence for the service flow to represent the order of processing of each service flow within a preset scheduling period. The preemptive message slice association carrier construction module is used to monitor whether there is a target priority message performing channel preemption processing on the controlled priority message in the dynamic priority preemption scheduling sequence of the service flow. If so, a preemptive heterogeneous message slice association carrier is constructed to characterize the interleaving distribution characteristics in the physical link. The gate sequence construction module is used to input the preemptive heterogeneous message slice association carrier into the time-aware shaping algorithm engine built based on the time-sensitive network protocol standard, generate network transmission time slot allocation mapping relationship to generate a time gating list, and construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow accordingly.

[0016] The technical advantages of the technical solution provided in this application are as follows: This application presents a low-latency networking method that integrates dynamic priority preemption and time-sensitive networking. Addressing the technical shortcomings of traditional static scheduling and contention-based access mechanisms with fixed priorities, such as the tendency for high-priority commands to be blocked and uncertain transmission delays, this method acquires the original communication service flow through the physical ingress ports of network switching nodes and analyzes and constructs a comprehensive attribute distribution of the services to be scheduled. This solves the problem of insufficient fine-grained perception of sudden and complex traffic in traditional solutions. Compared to traditional static scheduling methods that rely on fixed backoff windows, this application constructs a dynamic priority preemption scheduling sequence for service flows by establishing a non-linear coupling relationship between service urgency and link resource occupancy. This breaks the rigid traffic allocation strategy, allowing network resources to adaptively tilt with dynamic load evolution, significantly improving adaptability to high-density burst traffic.

[0017] Furthermore, by monitoring the scheduling evolution of priorities and constructing a preemptive heterogeneous message slice association carrier when preemption occurs, the drawback of long waiting times caused by low-priority data occupying the channel under the traditional contention mechanism is solved. In the traditional scheme, once low-priority data preempts the channel, high-priority instructions must wait for it to finish transmitting before they can access the channel; however, this application interleaves and reassembles messages of different priorities in a slice association manner, giving high-priority messages fine-grained forced interruption and instant access capabilities. Compared with the traditional uninterrupted transmission mode, it better avoids excessive queuing and channel starvation for urgent tasks, and the agility of urgent instructions accessing the channel is higher.

[0018] Finally, the preemptive heterogeneous message slice association carrier is input into a time-aware shaping algorithm engine built based on the Time-Sensitive Networking Protocol (TSN) standard to generate a time-gated list and construct a distributed time-sensitive network scheduling gating sequence, effectively solving the problem of significant uncertainty in the cumulative transmission delay of traditional multi-hop networks. Traditional collision avoidance mechanisms are prone to nonlinear amplification of queuing delays when forwarding to multiple nodes. This application combines the gating scheduling mechanism of TSN to allocate and map the preemptive slice characteristics to time slots, establishing a deterministic transmission channel with time boundaries in the physical link. Compared with traditional contention-based networking methods, this method better suppresses end-to-end delay jitter, enabling dense unmanned nodes to achieve higher determinism and stability in collaborative communication delays under sudden and complex environments. Attached Figure Description

[0019] Figure 1 This is a flowchart illustrating a low-latency networking method that integrates dynamic priority preemption and time-sensitive networking, as provided in an embodiment of this application.

[0020] Figure 2 This is a schematic diagram of a low-latency networking device that integrates dynamic priority preemption and time-sensitive networking, provided as an embodiment of this application.

[0021] Figure 3 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0022] like Figure 1 The image shows a low-latency networking method that integrates dynamic priority preemption and time-sensitive networking according to an embodiment of this application, which includes: Step 1: Obtain the original communication service flow to be networked through the physical ingress port of the network switching node and perform service attribute parsing based on the metadata of the message header to construct the global attribute distribution pattern of the scheduled service to be represented in the spatiotemporal domain of the network. Step 2: Based on the distribution of the global attributes of the services to be scheduled, establish a non-linear coupling relationship between the urgency of the services and the occupancy of the link resources, so as to generate a dynamic scheduling weight table for the service flow and construct a dynamic priority preemption scheduling sequence for the service flow to represent the order of processing of each service flow within a preset scheduling period. Step 3: Monitor whether there is a target priority message performing channel preemption processing on the controlled priority message in the dynamic priority preemption scheduling sequence of the service flow. If so, construct a preemptive heterogeneous message slice association carrier to characterize the interleaving distribution characteristics in the physical link. Step 4: Input the preemptive heterogeneous message slice association carrier into the time-aware shaping algorithm engine built based on the time-sensitive network protocol standard to generate a network transmission time slot allocation mapping relationship to generate a time gating list and construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow.

[0023] Optionally, step 1 includes: The original communication service flow is parsed based on the metadata of the message header to extract the basic service attribute feature group, which includes message arrival time, traffic burst tolerance and periodic interaction frequency. Based on the aforementioned service basic attribute feature group, a service category clustering identification based on latency sensitivity is performed to map the original communication service flow into time-sensitive service flow and non-time-sensitive service flow respectively; Based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in network bandwidth resources, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution status of different service categories in the network spatiotemporal domain.

[0024] Preferably, the specific implementation process of step 1 is as follows: In the unmanned collaborative node communication scenario involving a multi-channel adaptive array antenna data link terminal, the physical ingress port of the network switching node first registers the original communication service flow entering the physical link to form a physical ingress port reception record. The physical ingress port reception record includes at least the physical ingress port source, reception order, and local time base flag corresponding to the reception order when the original communication service flow enters the physical ingress port. The physical ingress port reception record is used to limit the reading source and reading order of subsequent packet header metadata. Subsequently, under the reading source and reading order limited by the physical ingress port reception record, the packet header metadata of the original communication service flow is read to form a parsable data object of the original communication service flow. The parsable data object of the original communication service flow simultaneously carries the physical ingress port reception record and the packet header metadata, so that the original communication service flow no longer enters the queue only according to the order of arrival, but is first converted into a parsable data object of the original communication service flow that can express the service latency requirements and network resource occupancy tendency. Thus, a sequential connection is formed between the physical ingress port receiving record, the message header metadata, and the original communication service flow parsable data object. The original communication service flow parsable data object continues to serve as input for service attribute parsing, and the service attribute parsing is further used to form subsequent service basic attribute feature groups.

[0025] Preferably, in the specific technical implementation of step 1, when parsing the service attributes of the original communication service flow based on the message header metadata, the service category indication content, priority indication content, message length indication content, source node indication content, destination node indication content, and periodic interaction indication content are first read from the message header metadata carried by the parsable data object of the original communication service flow. The service category indication content, the priority indication content, the message length indication content, the source node indication content, the destination node indication content, and the periodic interaction indication content are then summarized into the message header metadata reading result. The packet header metadata reading result is not directly used as a scheduling basis, but continues to participate in service attribute parsing. The service attribute parsing determines the initial latency requirement of the original communication service flow based on the priority indication content in the packet header metadata reading result, determines the single-time occupation tendency of the original communication service flow on network bandwidth resources based on the packet length indication content in the packet header metadata reading result, determines the network spatiotemporal domain transmission span of the original communication service flow in the network spatiotemporal domain based on the source node indication content and destination node indication content in the packet header metadata reading result, and determines the periodic interaction identification basis of the original communication service flow based on the periodic interaction indication content in the packet header metadata reading result, thus forming the service attribute parsing result. The service attribute parsing result includes the initial latency requirement, the single-time occupation tendency, the network spatiotemporal domain transmission span, and the periodic interaction identification basis. The service attribute parsing result then serves as the direct basis for extracting the basic service attribute feature group, ensuring a continuous technical source relationship between the packet header metadata reading result, the service attribute parsing result, and the basic service attribute feature group.

[0026] Preferably, when further processing the service attribute parsing result in step 1, the original communication service flows corresponding to the same source node indication content, the same destination node indication content, and the same service category indication content are first merged according to the local time base flag in the physical ingress port reception record, the reception order in the packet header metadata reading result, the source node indication content in the packet header metadata reading result, the destination node indication content in the packet header metadata reading result, and the service category indication content in the packet header metadata reading result, to form a same-source, same-direction service flow segment. The same-source, same-direction service flow segment is used to eliminate arrival time misjudgments caused by the mixing of original communication service flows corresponding to different source node indication contents and different destination node indication contents entering the physical ingress port; subsequently, the local time base flags corresponding to adjacent packets in the same-source, same-direction service flow segment are subjected to sequential differential processing to obtain the packet arrival time. The message arrival time is used to characterize the temporal density of the same-source and same-direction service flow segments at the physical ingress port. The message arrival time is further written into the service basic attribute feature group so that the service basic attribute feature group can carry the actual arrival rhythm of the original communication service flow, rather than just carrying a static service identifier.

[0027] Preferably, after determining the message arrival time, step 1 continues to perform traffic burst tolerance extraction processing on the original communication service flow based on the message length indication content in the same-source and same-direction service flow segment, the message arrival time between adjacent messages, and the network bandwidth resource occupancy record corresponding to the physical ingress port. Specifically, firstly, according to the initial latency requirement in the service attribute parsing result, a candidate transmission window for time-sensitive service flows is pre-defined. The candidate transmission window for time-sensitive service flows is used to limit the candidate transmission interval of high-latency-sensitive messages in network bandwidth resources. Then, based on the message length indication content of consecutive messages in the same-source and same-direction service flow segment, the continuous occupancy trend of the original communication service flow on network bandwidth resources within a short time range is identified. Subsequently, the continuous occupancy trend, the network bandwidth resource occupancy record, and the candidate transmission window for time-sensitive service flows are matched to determine the burst transmission amplitude that the original communication service flow can withstand without crowding out the candidate transmission window for time-sensitive service flows, thereby forming a traffic burst tolerance. After the traffic burst tolerance is written into the service basic attribute feature group, the service basic attribute feature group can simultaneously express the arrival rhythm and burst occupancy tendency of the original communication service flow, so that subsequent service category clustering identification can distinguish between periodic small packet control services and short-term centralized reporting services.

[0028] Preferably, step 1 further performs periodic interaction frequency extraction processing on the original communication service flow based on the periodic interaction indication content in the same-source and same-direction service flow segment, the periodic interaction identification basis in the service attribute parsing result, and the message arrival time. Specifically, when the periodic interaction indication content and the periodic interaction identification basis jointly indicate that the original communication service flow belongs to state synchronization, track coordination, link detection, or task control interaction between unmanned cooperative nodes, the time interval distribution between adjacent messages is first identified based on the arrival times of multiple consecutive messages, and then the stability of the time interval distribution is verified to obtain the periodic interaction frequency; when the periodic interaction indication content is missing, or the periodic interaction indication content is inconsistent with the time interval distribution, the periodic interaction frequency is corrected based on the time interval distribution. After the periodic interaction frequency is written into the service basic attribute feature group, the service basic attribute feature group can reflect the intensity of repeated interactions of the original communication service flow between unmanned cooperative nodes, and the periodic interaction frequency further participates in the time delay sensitivity discrimination processing.

[0029] Preferably, in step 1, when forming the service basic attribute feature group, the message arrival time, the traffic burst tolerance, and the periodic interaction frequency are bound according to the same source and same direction service flow segments corresponding to the same original communication service flow, and the bound message arrival time, traffic burst tolerance, and periodic interaction frequency are used as the service basic attribute feature group. The service basic attribute feature group is not a simple extraction of the message header metadata, but is derived from the message header metadata reading results, the service attribute parsing results, the physical ingress port reception records, and the network bandwidth resource occupancy records. Specifically, the message arrival time serves as the time pressure representation content in the service basic attribute feature group, the traffic burst tolerance serves as the burst occupancy pressure representation content, and the periodic interaction frequency serves as the repetitive interaction pressure representation content. The time pressure representation content, the burst occupancy pressure representation content, and the repetitive interaction pressure representation content are then used together as input for service category clustering identification, enabling the service category clustering identification to proceed around time pressure, burst occupancy pressure, and repetitive interaction pressure.

[0030] Preferably, in step 1, when performing business category clustering and identification based on the business basic attribute feature group, the time queuing tendency of the original communication service flow is first determined according to the time pressure characterization content, then the bandwidth crowding tendency of the original communication service flow is determined according to the burst occupancy pressure characterization content, and the continuous interaction tendency of the original communication service flow is determined according to the repeated interaction pressure characterization content. Then, the time queuing tendency, the bandwidth crowding tendency, and the continuous interaction tendency are jointly converted into a latency sensitivity discrimination result. The latency sensitivity discrimination result is used to characterize the comprehensive sensitivity of the original communication service flow to waiting time, bandwidth occupancy, and interaction continuity in the network spatiotemporal domain. When the latency sensitivity discrimination result indicates that the original communication service flow corresponds to emergency obstacle avoidance service, task replanning service, collaborative control service, or link detection service, the original communication service flow is mapped to a time-sensitive service flow. When the latency sensitivity discrimination result indicates that the original communication service flow corresponds to log reporting service, batch status feedback service, or non-real-time perception data feedback service, the original communication service flow is mapped to a non-time-sensitive service flow. Therefore, both the time-sensitive service flow and the non-time-sensitive service flow originate from the basic service attribute feature group and the latency sensitivity discrimination result, rather than being directly determined by a fixed priority field alone.

[0031] Preferably, after mapping the original communication service flows into time-sensitive and non-time-sensitive service flows in step 1, the distribution status of the time-sensitive and non-time-sensitive service flows in the network bandwidth resources is constructed based on network bandwidth resources. Specifically, candidate forwarding links are first determined based on the network spatiotemporal domain transmission span in the service attribute parsing results, and network egress ports are determined based on the physical ingress port source in the physical ingress port reception record, the destination node indication content in the packet header metadata reading results, and the candidate forwarding links; then, scheduling time slices are divided based on the packet arrival time and the periodic interaction frequency; subsequently, the network bandwidth resources are divided into network spatiotemporal domain units according to the physical ingress port source, the network egress port, the candidate forwarding links, and the scheduling time slices, and the time-sensitive and non-time-sensitive service flows are respectively written into the network spatiotemporal domain units corresponding to their physical ingress port source, transmission direction, candidate forwarding links, and scheduling time slices. The network spatiotemporal domain unit is used to carry the actual occupancy relationship of the time-sensitive service flow and the non-time-sensitive service flow in the network spatiotemporal domain. The actual occupancy relationship in the network spatiotemporal domain unit is further used to form the distribution status of the time-sensitive service flow and the non-time-sensitive service flow in the network bandwidth resources, so that the distribution status of the time-sensitive service flow and the non-time-sensitive service flow in the network bandwidth resources can simultaneously reflect the differences in service category, link occupancy, and time distribution.

[0032] Preferably, when constructing the overall attribute distribution status of the scheduled services based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in network bandwidth resources, step 1 first generates a status unit record for each network spatiotemporal domain unit. The status unit record includes time-sensitive service flow occupancy information, non-time-sensitive service flow occupancy information, packet arrival time aggregation information, traffic burst tolerance aggregation information, and periodic interaction frequency aggregation information; wherein, the time-sensitive service flow occupancy information originates from the time-sensitive service flows written into the network spatiotemporal domain unit, the non-time-sensitive service flow occupancy information originates from the non-time-sensitive service flows written into the network spatiotemporal domain unit, the packet arrival time aggregation information originates from the packet arrival time corresponding to each original communication service flow in the network spatiotemporal domain unit, the traffic burst tolerance aggregation information originates from the traffic burst tolerance corresponding to each original communication service flow in the network spatiotemporal domain unit, and the periodic interaction frequency aggregation information originates from the periodic interaction frequency corresponding to each original communication service flow in the network spatiotemporal domain unit. The situation unit records are then concatenated according to the order of the candidate forwarding links, so that multiple network spatiotemporal domain units traversed by the same original communication service flow during multi-hop transmission can be continuously represented; the concatenated situation unit records constitute the global attribute distribution situation of the service to be scheduled. Therefore, the global attribute distribution situation of the service to be scheduled not only represents the queue congestion situation at a certain moment, but also represents the transmission pressure transmission relationship of different service categories in multiple network spatiotemporal domain units. This transmission pressure transmission relationship continues to serve as the situation basis for subsequent dynamic priority scheduling.

[0033] Preferably, in a scenario, when step 1 is specifically implemented, if one node in the unmanned collaborative node simultaneously reports the original communication service flow corresponding to the collaborative control message, the original communication service flow corresponding to the sensing data message, and the original communication service flow corresponding to the log reporting message, then the physical ingress port first forms a physical ingress port receiving record corresponding to the original communication service flow corresponding to the collaborative control message, the original communication service flow corresponding to the sensing data message, and the original communication service flow corresponding to the log reporting message, respectively; subsequently, the physical ingress port receiving record and the message header metadata are used to form a parsable data object of the original communication service flow corresponding to the original communication service flow corresponding to the collaborative control message, the original communication service flow corresponding to the sensing data message, and the original communication service flow corresponding to the log reporting message, and the message arrival time, traffic burst tolerance, and periodic interaction frequency corresponding to each original communication service flow are extracted from the parsable data object of the original communication service flow. For original communication service flows corresponding to collaborative control messages with high periodic interaction frequency and dense message arrival times, service category clustering identification maps the original communication service flows corresponding to the collaborative control messages to time-sensitive service flows. For original communication service flows corresponding to sensing data messages that arrive in short bursts and have high traffic burst tolerance, service category clustering identification combines the message arrival time, traffic burst tolerance, and network bandwidth resource usage records of the original communication service flows corresponding to the sensing data messages to determine the degree of crowding out of the original communication service flows corresponding to the sensing data messages in the network spatiotemporal domain unit, and determines whether the original communication service flows corresponding to the sensing data messages belong to time-sensitive or non-time-sensitive service flows accordingly. For original communication service flows corresponding to log reporting messages that arrive intermittently and have low periodic interaction frequency, service category clustering identification maps the original communication service flows corresponding to the log reporting messages to non-time-sensitive service flows. The above-mentioned time-sensitive and non-time-sensitive service flows continue to be written into the network spatiotemporal domain unit and form a situation unit record. The situation unit record is finally merged into the global attribute distribution situation of the service to be scheduled.

[0034] Preferably, the global attribute distribution pattern of the service to be scheduled formed in step 1 continues to participate in the nonlinear coupling relationship establishment process between service urgency and link resource occupancy in step 2. Specifically, the packet arrival time in the global attribute distribution pattern of the service to be scheduled participates in the time-side judgment of service urgency in the form of time pressure representation; the traffic burst tolerance in the global attribute distribution pattern of the service to be scheduled participates in the bandwidth-side judgment of link resource occupancy in the form of burst occupancy pressure representation; and the periodic interaction frequency in the global attribute distribution pattern of the service to be scheduled participates in the periodic-side judgment of continuous service interaction in the form of repeated interaction pressure representation. The time pressure representation, the burst occupancy pressure representation, and the repeated interaction pressure representation are jointly entered into the subsequent dynamic scheduling weight table generation process. Therefore, step 1 does not simply complete the classification of the original communication service flow, but continuously transmits the physical ingress port reception record, the packet header metadata, the parsable data object of the original communication service flow, the packet header metadata reading result, the service attribute parsing result, the service basic attribute feature group, the time-sensitive service flow, the non-time-sensitive service flow, and the distribution status of the time-sensitive service flow and the non-time-sensitive service flow in the network bandwidth resources to the full-domain attribute distribution status of the service to be scheduled, so that the subsequent dynamic priority preemption scheduling has a technical basis directly derived from the physical ingress port reception status.

[0035] Optionally, based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in network bandwidth resources, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution status of different service categories in the network spatiotemporal domain, including: Based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in the network bandwidth resources, the corresponding traffic packet length distribution is parameterized to obtain the traffic packet length distribution parameters; Determine the transmission time interval for network data transmission between the time-sensitive service flow and the non-time-sensitive service flow, and parameterize and characterize it to generate a transmission time interval parameter; Based on the traffic packet length distribution parameter and the transmission time interval parameter, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution pattern of different service categories in the network spatiotemporal domain.

[0036] Preferably, in step 1, when constructing the global attribute distribution pattern of the service to be scheduled based on the distribution status of the time-sensitive service flow and the non-time-sensitive service flow in network bandwidth resources, the distribution status of the time-sensitive service flow and the non-time-sensitive service flow in network bandwidth resources is first read, and the time-sensitive service flow occupancy information and non-time-sensitive service flow occupancy information corresponding to each network spatiotemporal domain unit are extracted from the distribution status of the time-sensitive service flow and the non-time-sensitive service flow in network bandwidth resources. The time-sensitive service flow occupancy information is used to characterize the scheduling time slice, candidate forwarding link, and network egress port occupied by the time-sensitive service flow in the corresponding network spatiotemporal domain unit, and the non-time-sensitive service flow occupancy information is used to characterize the scheduling time slice, candidate forwarding link, and network egress port occupied by the non-time-sensitive service flow in the corresponding network spatiotemporal domain unit. Subsequently, the time-sensitive service flow occupancy information and the non-time-sensitive service flow occupancy information are merged according to network spatiotemporal domain units to form a service category occupancy merging record. The service category occupancy merging record continues to serve as the basis for generating traffic packet length distribution, so that the subsequent traffic packet length distribution can correspond to the actual occupancy position of the time-sensitive service flow and the non-time-sensitive service flow in network bandwidth resources.

[0037] Preferably, in the specific technical implementation of step 1, after forming the service category occupancy merging record, the header metadata reading results of the corresponding original communication service flow are traced back according to the service category occupancy merging record, and the message length indication content is extracted from the header metadata reading results. The message length indication content is first categorized and marked according to time-sensitive and non-time-sensitive service flows to form message length indication content with service category source; subsequently, the message length indication content with service category source is arranged in time sequence according to the scheduling time slice in the network spatiotemporal domain unit to form a category time sequence packet length record. The category time sequence packet length record is used to express the order of message length changes of different service categories within the same network spatiotemporal domain unit. The category time sequence packet length record continues to participate in the generation of traffic packet length distribution, so that the traffic packet length distribution can simultaneously reflect the relationship between service category, message length, and the order in which messages enter the scheduling time slice. Therefore, the traffic packet length distribution is not based solely on the statistics of the message length indication content, but is derived from the service category occupancy merging record, the message length indication content with the service category source, and the category time sequence packet length record.

[0038] Preferably, in step 1, when generating the traffic packet length distribution for the category-based time-series packet length records, the category-based time-series packet length records belonging to time-sensitive service flows in the same network spatiotemporal domain unit are first merged into the time-sensitive service flow packet length distribution, and the category-based time-series packet length records belonging to non-time-sensitive service flows in the same network spatiotemporal domain unit are merged into the non-time-sensitive service flow packet length distribution. The time-sensitive service flow packet length distribution is used to characterize the concentration of packet length, the range of packet length fluctuation, and the trend of continuous scheduling time slice length changes for time-sensitive service flows in the corresponding network spatiotemporal domain unit; the non-time-sensitive service flow packet length distribution is used to characterize the concentration of packet length, the range of packet length fluctuation, and the trend of continuous scheduling time slice length changes for non-time-sensitive service flows in the corresponding network spatiotemporal domain unit. Subsequently, the time-sensitive service flow packet length distribution and the non-time-sensitive service flow packet length distribution are compared and arranged within the same network spatiotemporal domain unit to obtain the traffic packet length distribution. The packet length distribution is used to express the packet length competition pattern of time-sensitive and non-time-sensitive service flows in network bandwidth resources. The concentration of packet length, the fluctuation range of packet length, and the trend of continuous occupied scheduling time slice length in the packet length distribution continue to participate in the generation of packet length distribution parameters, so that the packet length distribution parameters can take into account the packet length competition pattern in the packet length distribution.

[0039] Preferably, when parameterizing the traffic packet length distribution, step 1 does not directly perform mixed calculations on packet lengths, scheduling time slices, and candidate forwarding links of different dimensions. Instead, it first classifies the packet length distributions of time-sensitive and non-time-sensitive service flows within the traffic packet length distribution into categories with the same dimension. Specifically, it identifies concentrated packet length intervals for the time-sensitive service flow packet length distribution to obtain concentrated intervals for time-sensitive service flow packet lengths; it also identifies concentrated intervals for the non-time-sensitive service flow packet length distribution to obtain concentrated intervals for non-time-sensitive service flow packet lengths; and then, based on the overlap between the concentrated intervals for time-sensitive and non-time-sensitive service flow packet lengths, it forms packet length competition intervals. The packet length contention interval is used to characterize the range in which time-sensitive service flows and non-time-sensitive service flows compete for network bandwidth resources within a similar packet length range. The packet length contention interval is further written into the traffic packet length distribution parameter so that the traffic packet length distribution parameter can reflect the overlap relationship between the packet length concentration interval of the time-sensitive service flows and the packet length concentration interval of the non-time-sensitive service flows.

[0040] Preferably, step 1 further extracts the degree of change in the continuous packet length of time-sensitive service flows based on the continuous packet length changes in the packet length distribution of time-sensitive service flows, and extracts the degree of change in the continuous packet length of non-time-sensitive service flows based on the continuous packet length changes in the packet length distribution of non-time-sensitive service flows. The degree of change in the continuous packet length of time-sensitive service flows is used to express whether the occupation of network bandwidth resources by time-sensitive service flows tends to stabilize within a continuous scheduling time slice, and the degree of change in the continuous packet length of non-time-sensitive service flows is used to express whether non-time-sensitive service flows form bursty occupation within a continuous scheduling time slice. Subsequently, the degree of change in the continuous packet length of time-sensitive service flows, the degree of change in the continuous packet length of non-time-sensitive service flows, and the packet length contention interval are jointly parameterized to obtain the traffic packet length distribution parameters. The traffic packet length distribution parameters include the packet length contention interval, the degree of change in the continuous packet length of the time-sensitive service flow, and the degree of change in the continuous packet length of the non-time-sensitive service flow. The traffic packet length distribution parameters are then involved in the correlation processing of the transmission time interval parameters so that the transmission time interval parameters can perform time-dimensional verification of the packet length occupancy pressure in the traffic packet length distribution parameters, thereby avoiding judging the network bandwidth resource occupancy pressure solely based on packet length.

[0041] Preferably, in step 1, when determining the transmission time interval for network data transmission between the time-sensitive service flow and the non-time-sensitive service flow, the scheduling time slice arrangement order of each network spatiotemporal domain unit is first read from the service category occupancy merging record, and combined with the packet arrival time of adjacent packets of the same service category in the category timing packet length record, a category adjacent transmission record is generated. The category adjacent transmission record is used to describe the source of the time interval between adjacent packets of the same service category in network data transmission. Subsequently, the category adjacent transmission records belonging to the time-sensitive service flow are extracted to form the time-sensitive service flow adjacent transmission record, and the category adjacent transmission records belonging to the non-time-sensitive service flow are extracted to form the non-time-sensitive service flow adjacent transmission record. The time-sensitive service flow adjacent transmission record and the non-time-sensitive service flow adjacent transmission record are used together to determine the transmission time interval, so that the transmission time interval can simultaneously cover the time distribution differences in network data transmission between the time-sensitive service flow corresponding to the cooperative control message, the non-time-sensitive service flow corresponding to the sensing data message, and the non-time-sensitive service flow corresponding to the log reporting message.

[0042] Preferably, in step 1, when extracting the intervals of adjacent transmission records for time-sensitive service flows, the time intervals between adjacent packets of time-sensitive service flows in network data transmission are extracted according to the order of the same candidate forwarding link, the same network egress port, and adjacent scheduling time slices, to form the transmission time intervals of time-sensitive service flows. Similarly, when extracting the intervals of adjacent transmission records for non-time-sensitive service flows, the time intervals between adjacent packets of non-time-sensitive service flows in network data transmission are extracted according to the order of the same candidate forwarding link, the same network egress port, and adjacent scheduling time slices, to form the transmission time intervals of non-time-sensitive service flows. The transmission time intervals of time-sensitive service flows are used to express whether time-sensitive service flows enter network data transmission according to a relatively stable period, and the transmission time intervals of non-time-sensitive service flows are used to express whether non-time-sensitive service flows arrive in a concentrated manner or intermittently. Subsequently, the transmission time intervals of time-sensitive service flows and the transmission time intervals of non-time-sensitive service flows are merged into the transmission time interval, which continues to be used as the parameterized object of the transmission time interval parameter.

[0043] Preferably, when parameterizing the transmission time interval, step 1 first extracts the stability of the time-sensitive service flow interval from the transmission time interval of the time-sensitive service flow and the dispersion of the non-time-sensitive service flow interval from the transmission time interval of the non-time-sensitive service flow. The stability of the time-sensitive service flow interval is used to characterize the periodic maintenance state of the time-sensitive service flow in continuous scheduling time slices, and the dispersion of the non-time-sensitive service flow interval is used to characterize the burst aggregation state of the non-time-sensitive service flow in continuous scheduling time slices. Subsequently, the stability of the time-sensitive service flow interval and the dispersion of the non-time-sensitive service flow interval are compared and characterized within the same network spatiotemporal domain unit to generate the service category interval difference degree. The service category interval difference degree is used to express the degree of temporal interleaving between time-sensitive and non-time-sensitive service flows within the same network spatiotemporal domain unit. The service category interval difference degree is further written into the transmission time interval parameter, so that the transmission time interval parameter can express the competitive relationship of different service categories in the time dimension.

[0044] Preferably, in step 1, when generating the transmission time interval parameter, the stability of the interval of time-sensitive service flows, the dispersion of the interval of non-time-sensitive service flows, and the difference in the interval of service categories are parameterized and organized to form the transmission time interval parameter. The transmission time interval parameter is not an isolated record of the time difference between adjacent packets, but rather a time occupancy characterization result formed by combining the network spatiotemporal domain unit, the candidate forwarding link, the network egress port, and the scheduling time slice. Specifically, the stability of the interval of time-sensitive service flows is used to constrain the periodic transmission requirements of time-sensitive service flows; the dispersion of the interval of non-time-sensitive service flows is used to identify the burst transmission pressure of non-time-sensitive service flows; and the difference in the interval of service categories is used to describe the time interleaving state between time-sensitive and non-time-sensitive service flows. The time occupancy characterization result, through the transmission time interval parameter, participates in subsequent correlation processing with the traffic packet length distribution parameter to construct the global attribute distribution pattern of the service to be scheduled.

[0045] Preferably, in step 1, when constructing the global attribute distribution status of the service to be scheduled based on the traffic packet length distribution parameters and the transmission time interval parameters, the packet length competition interval, the degree of continuous packet length variation of time-sensitive service flows, and the degree of continuous packet length variation of non-time-sensitive service flows in the traffic packet length distribution parameters are first read within the same network spatiotemporal domain unit. Then, the interval stability of time-sensitive service flows, the interval dispersion of non-time-sensitive service flows, and the degree of difference in service category intervals in the transmission time interval parameters are read. Subsequently, the traffic packet length distribution parameters and the transmission time interval parameters within the same network spatiotemporal domain unit are correlated one-to-one to form a packet length interval correlation record. The packet length interval correlation record is used to characterize the combined effect between packet length occupancy pressure and packet time entry pressure within the same network spatiotemporal domain unit. The packet length interval correlation record is further written into the status unit record, so that the status unit record has the ability to describe both packet length occupancy pressure and packet time entry pressure. The situation unit record is formed by the time-sensitive service flow occupancy information, the non-time-sensitive service flow occupancy information, and the packet length interval association record. The situation unit record continues to serve as a unitized record for constructing the global attribute distribution situation of the service to be scheduled.

[0046] Preferably, after forming the packet length interval association record in step 1, the packet length interval association records corresponding to multiple network spatiotemporal domain units are concatenated according to the order of the candidate forwarding links to form a multi-hop packet length interval transmission record. The multi-hop packet length interval transmission record is used to characterize how the packet length occupancy pressure and packet time entry pressure of the same original communication service flow are transmitted segment by segment along the candidate forwarding link in multiple network spatiotemporal domain units. When the multi-hop packet length interval transmission record shows that time-sensitive service flows have a high degree of stability in the interval of time-sensitive service flows in the previous network spatiotemporal domain unit, and at the same time form a high degree of difference in service category interval with non-time-sensitive service flows in the next network spatiotemporal domain unit, the multi-hop packet length interval transmission record can transmit the packet time dimension crowding relationship corresponding to the packet time entry pressure to the overall attribute distribution of the service to be scheduled. When the multi-hop packet length interval transmission record shows that non-time-sensitive service flows have a high degree of continuous packet length variation and a high degree of interval dispersion of non-time-sensitive service flows, the multi-hop packet length interval transmission record can transmit the burst occupancy pressure formed by the superposition of the packet length occupancy pressure and the packet time entry pressure to the overall attribute distribution of the service to be scheduled. Therefore, the multi-hop packet long-interval transmission record enables the global attribute distribution of the scheduled service to simultaneously accommodate the packet time dimension crowding relationship and the sudden occupancy pressure.

[0047] Preferably, in a scenario, when step 1 is specifically implemented, if the same unmanned collaborative node sends time-sensitive service flows corresponding to collaborative control messages, non-time-sensitive service flows corresponding to sensing data messages, and non-time-sensitive service flows corresponding to log reporting messages respectively within adjacent scheduling time slices, then firstly, a corresponding service category occupancy merging record is generated based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in network bandwidth resources, and then a category time sequence packet length record is formed based on the service category occupancy merging record and the corresponding message length indication content. For time-sensitive service flows corresponding to collaborative control messages, the category-based time sequence packet length record shows that the message lengths are relatively close and the transmission time intervals are relatively stable, based on the packet length distribution and transmission time intervals of the time-sensitive service flows. For non-time-sensitive service flows corresponding to sensing data messages, the category-based time sequence packet length record shows that the message lengths vary greatly and the transmission time intervals arrive in a concentrated manner, based on the packet length distribution and transmission time intervals of the non-time-sensitive service flows. For non-time-sensitive service flows corresponding to log reporting messages, the category-based time sequence packet length record shows that the transmission time intervals are relatively dispersed, based on the transmission time intervals of the non-time-sensitive service flows. The aforementioned time-series packet length records participate in the generation of the traffic packet length distribution parameters and the transmission time interval parameters, respectively. The traffic packet length distribution parameters and the transmission time interval parameters further form the packet length interval association record, enabling the distribution status of the global attributes of the scheduled service to distinguish the different occupancy patterns of time-sensitive service flows corresponding to collaborative control messages, non-time-sensitive service flows corresponding to sensing data messages, and non-time-sensitive service flows corresponding to log reporting messages in the network spatiotemporal domain.

[0048] In summary, step 1 ultimately integrates the situation unit record, the packet length interval association record, and the multi-hop packet length interval transmission record into the overall attribute distribution situation of the service to be scheduled. The overall attribute distribution situation of the service to be scheduled includes time-sensitive service flow occupancy information, non-time-sensitive service flow occupancy information, traffic packet length distribution parameters, transmission time interval parameters, and packet length interval association records formed by the traffic packet length distribution parameters and the transmission time interval parameters in each network spatiotemporal domain unit. Simultaneously, the overall attribute distribution situation of the service to be scheduled also includes multi-hop packet length interval transmission records formed by concatenation along candidate forwarding links. Therefore, the overall attribute distribution situation of the service to be scheduled can characterize the distribution situation of different service categories in the network spatiotemporal domain from multiple dimensions, including service category, network bandwidth resource occupancy, traffic packet length distribution, transmission time interval, packet length occupancy pressure, packet time entry pressure, and multi-hop packet length interval transmission records. It also serves as the input basis for establishing the nonlinear coupling relationship between service urgency and link resource occupancy in step 2.

[0049] Optionally, step 2 includes; Based on the dynamic scheduling weight table of the business flow, the feature parameters of each dimension in the global attribute distribution of the business to be scheduled are weighted to obtain the weighted feature parameters of each dimension. The weighted feature parameters of each dimension are mapped to the multidimensional network resource topology boundary space to generate a load drift evolution trajectory that characterizes the current service load deviating from the deterministic transmission range of the network. Based on the load drift evolution trajectory, a dynamic priority preemption scheduling sequence for service flows is constructed to characterize the order in which each service flow is processed within a preset scheduling period.

[0050] Preferably, the specific implementation process of step 2 is as follows: After obtaining the global attribute distribution status of the service to be scheduled, first read the time-sensitive service flow occupancy information, non-time-sensitive service flow occupancy information, traffic packet length distribution parameters, transmission time interval parameters, packet length interval association records, and multi-hop packet length interval transmission records corresponding to each network spatiotemporal domain unit from the global attribute distribution status of the service to be scheduled. Then, organize the time-sensitive service flow occupancy information, the non-time-sensitive service flow occupancy information, the traffic packet length distribution parameters, the transmission time interval parameters, the packet length interval association records, and the multi-hop packet length interval transmission records into feature parameters of each dimension. The various feature parameters include service category occupancy parameters, traffic packet length occupancy parameters, transmission time interval parameters, packet length interval association parameters, and multi-hop transmission parameters. Specifically, the service category occupancy parameter originates from the occupancy information of time-sensitive and non-time-sensitive service flows; the traffic packet length occupancy parameter originates from the traffic packet length distribution parameter; the transmission time interval parameter originates from the transmission time interval parameter; the packet length interval association parameter originates from the packet length interval association record; and the multi-hop transmission parameter originates from the multi-hop packet length interval transmission record. These feature parameters are then used as weighting objects in the dynamic scheduling weight table for service flows, enabling the dynamic scheduling weight table to directly apply to the service category occupancy parameters, traffic packet length occupancy parameters, transmission time interval parameters, packet length interval association parameters, and multi-hop transmission parameters in the overall attribute distribution of the service to be scheduled, rather than only applying to fixed priority identifiers.

[0051] Preferably, in the specific technical implementation of step 2, the dynamic scheduling weight table for service flows is not a preset fixed priority table, but is dynamically generated based on the real-time occupancy relationship of different service categories in the overall attribute distribution of the services to be scheduled. Specifically, firstly, the concurrent occupancy status of time-sensitive and non-time-sensitive service flows in the same network spatiotemporal domain unit is identified based on the service category occupancy parameter; then, the packet length occupancy parameter is used to identify the packet length occupancy pressure of time-sensitive and non-time-sensitive service flows on network bandwidth resources; and finally, the time density of time-sensitive and non-time-sensitive service flows entering network data transmission is identified based on the transmission time interval parameter. Subsequently, the concurrent occupancy status, the packet length occupancy pressure, and the time density are organized into the criteria for judging the urgency of the service. The service urgency determination criterion is further correlated with the packet length interval association parameter and the multi-hop transmission parameter to form a link resource occupancy determination criterion. This criterion expresses the continuous occupancy relationship of the same original communication service flow across multiple network spatiotemporal domain units for candidate forwarding links, network egress ports, and scheduling time slices. The service urgency determination criterion and the link resource occupancy determination criterion jointly participate in the generation of the service flow dynamic scheduling weight table, enabling the table to correspond to the service urgency and link resource occupancy in the overall attribute distribution of the service to be scheduled.

[0052] Preferably, in step 2, when generating the dynamic scheduling weight table for service flows, firstly, service urgency weights are configured for time-sensitive and non-time-sensitive service flows according to the service urgency discrimination criteria, and link resource occupancy weights are configured for time-sensitive and non-time-sensitive service flows according to the link resource occupancy discrimination criteria. The service urgency weights express the scheduling requirements of the original communication service flow in terms of waiting time, interaction continuity, and control timeliness, while the link resource occupancy weights express the occupancy of the original communication service flow on candidate forwarding links, network egress ports, and scheduling time slices. Subsequently, the service urgency weights and link resource occupancy weights corresponding to the same original communication service flow are written into the dynamic scheduling weight table for service flows, so that the dynamic scheduling weight table can simultaneously reflect the service urgency corresponding to the service urgency weights and the link resource occupancy corresponding to the link resource occupancy weights. The dynamic scheduling weight table for service flows continues to be used to weight the feature parameters of each dimension, so that the feature parameters of each dimension can form a comparable scheduling expression according to the current service load.

[0053] Preferably, in step 2, when weighting the feature parameters of each dimension according to the dynamic scheduling weight table of the service flow, the service category occupancy parameter is first associated with the service urgency weight to obtain the weighted service category occupancy parameter; then, the traffic packet length occupancy parameter and the transmission time interval parameter are respectively associated with the link resource occupancy weight to obtain the weighted traffic packet length occupancy parameter and the weighted transmission time interval parameter; subsequently, the packet length interval association parameter is jointly associated with the service urgency weight and the link resource occupancy weight to obtain the weighted packet length interval association parameter, and the multi-hop transmission parameter is associated with the link resource occupancy weight to obtain the weighted multi-hop transmission parameter. The weighted service category occupancy parameter, the weighted traffic packet length occupancy parameter, the weighted transmission time interval parameter, the weighted packet length interval association parameter, and the weighted multi-hop transmission parameter together constitute the weighted feature parameters of each dimension. Therefore, the weighted feature parameters of each dimension not only retain the original distribution relationship in the overall attribute distribution of the service to be scheduled, but also carry the dynamic adjustment results of the service flow dynamic scheduling weight table on the urgency of the service and the link resource occupation within the preset scheduling period.

[0054] Preferably, after obtaining the weighted feature parameters of each dimension, step 2 maps the weighted feature parameters of each dimension to a multi-dimensional network resource topology boundary space. The multi-dimensional network resource topology boundary space is constructed based on candidate forwarding links, network egress ports, scheduling time slices, queue priorities, and bandwidth reservation boundaries in an unmanned collaborative node communication scenario, and is used to express the deterministic transmission range that the current network can carry. The multi-dimensional network resource topology boundary space includes link bandwidth boundaries, egress port queuing boundaries, scheduling time slice boundaries, queue priority boundaries, and multi-hop transmission boundaries; wherein, the link bandwidth boundary is used to limit the network bandwidth resources that can be allocated to candidate forwarding links, the egress port queuing boundary is used to limit the queuing capacity of network egress ports, the scheduling time slice boundary is used to limit the time range within which time-sensitive and non-time-sensitive service flows can enter transmission, the queue priority boundary is used to limit the order of time-sensitive and non-time-sensitive service flows in the queue, and the multi-hop transmission boundary is used to limit the cumulative transmission pressure allowed when the same original communication service flow is transmitted hop-by-hop along the candidate forwarding links. The multidimensional network resource topology boundary space continues to serve as a mapping reference for the weighted feature parameters of each dimension, enabling the weighted feature parameters of each dimension to deviate from the link bandwidth boundary, the egress port queuing boundary, the scheduling time slice boundary, the queue priority boundary, and the multi-hop transmission boundary.

[0055] Preferably, in step 2, when mapping the weighted feature parameters of each dimension to the multi-dimensional network resource topology boundary space, the weighted packet length occupancy parameter is first mapped to the link bandwidth boundary to obtain the link bandwidth occupancy deviation parameter; the weighted transmission time interval parameter is mapped to the scheduling time slice boundary to obtain the scheduling time slice occupancy deviation parameter; the weighted service category occupancy parameter is mapped to the queue priority boundary to obtain the queue priority occupancy deviation parameter; the weighted packet length interval association parameter is mapped to the egress port queuing boundary to obtain the egress port queuing deviation parameter; and the weighted multi-hop transmission parameter is mapped to the multi-hop transmission boundary to obtain the multi-hop transmission deviation parameter. The link bandwidth occupancy deviation parameter, the scheduling time slice occupancy deviation parameter, the queue priority occupancy deviation parameter, the egress port queuing deviation parameter, and the multi-hop transmission deviation parameter collectively characterize the deviation state of the current service load relative to the multi-dimensional network resource topology boundary space and continue to participate in the generation of the load drift evolution trajectory. The deviation state originates from the mapping relationship between the weighted feature parameters of each dimension and the multidimensional network resource topology boundary space. The deviation state then enters the load drift evolution trajectory through the link bandwidth occupancy deviation parameter, the scheduling time slice occupancy deviation parameter, the queue priority occupancy deviation parameter, the egress port queuing deviation parameter, and the multi-hop transmission deviation parameter.

[0056] Preferably, in step 2, when generating the load shift evolution trajectory, the link bandwidth occupancy deviation parameter, the scheduling time slice occupancy deviation parameter, the queue priority occupancy deviation parameter, the egress port queuing deviation parameter, and the multi-hop transmission deviation parameter are first arranged in the order of scheduling time slices to form a scheduling cycle deviation sequence. The scheduling cycle deviation sequence is used to describe the deviation change process of the current service load on the link bandwidth boundary, the scheduling time slice boundary, the queue priority boundary, the egress port queuing boundary, and the multi-hop transmission boundary within the same preset scheduling cycle. Subsequently, the scheduling cycle deviation sequences corresponding to adjacent scheduling time slices are connected to form a load shift evolution trajectory, and the source relationship between the scheduling cycle deviation sequence and the link bandwidth occupancy deviation parameter, the scheduling time slice occupancy deviation parameter, the queue priority occupancy deviation parameter, the egress port queuing deviation parameter, and the multi-hop transmission deviation parameter is recorded simultaneously to form load shift evolution trajectory source information. The load drift evolution trajectory is used to characterize the direction, magnitude, and duration of the current service load's deviation from the network's deterministic transmission range. The deviation direction originates from the sequence of changes in deviation parameters at different network resource boundaries; the deviation magnitude originates from the degree of deviation of each deviation parameter relative to its corresponding network resource boundary; and the deviation duration originates from the consecutive occurrence of deviation parameters across multiple scheduling time slices. The load drift evolution trajectory and its source information are then used together as the basis for constructing the dynamic priority preemption scheduling sequence for the service flow.

[0057] Preferably, in step 2, when constructing the dynamic priority preemption scheduling sequence for the service flow based on the load drift evolution trajectory, the high-urgency drift segments corresponding to time-sensitive service flows and the low-urgency occupancy segments corresponding to non-time-sensitive service flows are first identified from the load drift evolution trajectory. The high-urgency drift segments originate from the continuous deviation of time-sensitive service flows in the scheduling time slice occupancy deviation parameter, the queue priority occupancy deviation parameter, and the multi-hop transmission deviation parameter, and are used to express the triggering basis for judging whether time-sensitive service flows enter channel preemption processing within a preset scheduling period. The low-urgency occupancy segments originate from the continuous occupancy of non-time-sensitive service flows in the link bandwidth occupancy deviation parameter, the egress port queuing deviation parameter, and the multi-hop transmission deviation parameter, and are used to express the triggering basis for non-time-sensitive service flows being reordered or delayed in transmission within a preset scheduling period. Subsequently, the high-urgency drift segments and the low-urgency occupancy segments are compared under the same candidate forwarding link, the same network egress port, and the same scheduling time slice to form a preemption candidate relationship record. The preemption candidate relationship record is used to record the conflict relationship between the high-urgency drift segment and the low-urgency occupied segment under the same candidate forwarding link, the same network egress port and the same scheduling time slice. The preemption candidate relationship record continues to participate in the construction of the dynamic priority preemption scheduling sequence of the service flow.

[0058] Preferably, after forming the preemption candidate relationship record in step 2, the scheduling order of the preemption candidate relationship record is arranged according to the dynamic scheduling weight table of the service flow. Specifically, when the preemption candidate relationship record represents a high-urgency drift segment corresponding to a time-sensitive service flow and a low-urgency occupancy segment corresponding to a non-time-sensitive service flow overlapping at the same network egress port, step 2 reads the service urgency weight of the corresponding time-sensitive service flow and the link resource occupancy weight of the corresponding non-time-sensitive service flow from the dynamic scheduling weight table of the service flow, and determines the service flow scheduling order mark according to the service urgency weight and the link resource occupancy weight. The service flow scheduling order mark is used to represent the order of processing of the original communication service flow within a preset scheduling period; subsequently, multiple service flow scheduling order marks are sorted according to candidate forwarding links, network egress ports, and scheduling time slices to construct the dynamic priority preemption scheduling sequence of the service flow. The dynamic priority preemption scheduling sequence for service flows is jointly generated by the load drift evolution trajectory, the source information of the load drift evolution trajectory, the preemption candidate relationship record, and the dynamic scheduling weight table for service flows, so that the dynamic priority preemption scheduling sequence for service flows can be updated as the current service load changes.

[0059] Preferably, in a scenario, when implementing step 2, if a certain unmanned collaborative node simultaneously has time-sensitive service flows corresponding to collaborative control messages, non-time-sensitive service flows corresponding to sensing data messages, and non-time-sensitive service flows corresponding to log reporting messages within the same preset scheduling period, then the overall attribute distribution status of the services to be scheduled first provides the service category occupancy parameter, traffic packet length occupancy parameter, transmission time interval parameter, packet length interval association parameter, and multi-hop transmission parameter corresponding to each of the time-sensitive service flows corresponding to collaborative control messages, non-time-sensitive service flows corresponding to sensing data messages, and non-time-sensitive service flows corresponding to log reporting messages. Step 2 then weights the above-mentioned feature parameters according to the service flow dynamic scheduling weight table to obtain the weighted feature parameters of each dimension for the time-sensitive service flows corresponding to collaborative control messages, the weighted feature parameters of each dimension for the non-time-sensitive service flows corresponding to sensing data messages, and the weighted feature parameters of each dimension for the non-time-sensitive service flows corresponding to log reporting messages. If the weighted feature parameters of the time-sensitive service flow corresponding to the coordinated control message form high-urgency drift segments at the scheduling time slice boundary and the queue priority boundary, while the weighted feature parameters of the non-time-sensitive service flow corresponding to the sensing data message form low-urgency occupancy segments at the link bandwidth boundary and the egress port queuing boundary, then step 2, based on the preemption candidate relationship record and the service flow scheduling order marker, arranges the time-sensitive service flow corresponding to the coordinated control message before the non-time-sensitive service flow corresponding to the sensing data message, and places the non-time-sensitive service flow corresponding to the log reporting message in the scheduling time slice with lower occupancy pressure. The arrangement results of the time-sensitive service flow corresponding to the coordinated control message, the non-time-sensitive service flow corresponding to the sensing data message, and the non-time-sensitive service flow corresponding to the log reporting message are further written into the service flow dynamic priority preemption scheduling sequence, so that the service flow dynamic priority preemption scheduling sequence can express the processing order of different service flows within the preset scheduling period.

[0060] In summary, the dynamic priority preemption scheduling sequence for service flows constructed in step 2 is passed to step 3 to monitor whether a target priority packet is preempting a controlled priority packet. Specifically, the service flow scheduling order flag in the dynamic priority preemption scheduling sequence is used to determine the scheduling readiness status of the target priority packet; the preemption candidate relationship record in the dynamic priority preemption scheduling sequence is used to determine whether a controlled priority packet is occupying the same network egress port; and the load drift evolution trajectory source information in the dynamic priority preemption scheduling sequence is used to record the source of network resource deviation corresponding to the channel preemption judgment. Therefore, the dynamic priority preemption scheduling sequence of the service flow formed in step 2 is not simply a result of service sorting, but carries the continuous technical association between the global attribute distribution of the service to be scheduled, the dynamic scheduling weight table of the service flow, the weighted feature parameters of each dimension, the multi-dimensional network resource topology boundary space, the load drift evolution trajectory, the source information of the load drift evolution trajectory, the high-urgency drift segment, the low-urgency occupancy segment, the preemption candidate relationship record, and the service flow scheduling order mark, and provides an executable sequential processing order for the channel preemption processing in step 3.

[0061] Optionally, based on the load drift evolution trajectory, a dynamic priority preemption scheduling sequence for service flows is constructed to characterize the order in which each service flow is processed within a preset scheduling period, including: The load drift evolution trajectory is subjected to spatiotemporal dimension projection processing to obtain spatiotemporal projection results; Based on the spatiotemporal projection results, load integration is performed to calculate the multidimensional composite load deviation degree used to characterize the intensity of flow fluctuations. Dynamic priority evaluation is performed based on the multidimensional composite working condition load deviation and the global attribute distribution of the services to be scheduled, so as to construct a dynamic priority preemption scheduling sequence for services to represent the order of processing of each service flow within a preset scheduling period.

[0062] Preferably, when constructing the dynamic priority preemption scheduling sequence of the service flow based on the load drift evolution trajectory in step 2, the load drift evolution trajectory is first subjected to spatiotemporal dimension projection processing. Specifically, the load drift evolution trajectory is formed by connecting the link bandwidth occupancy deviation parameter, scheduling time slice occupancy deviation parameter, queue priority occupancy deviation parameter, egress port queuing deviation parameter, and multi-hop transmission deviation parameter in the order of scheduling time slices. Therefore, the spatiotemporal dimension projection processing first decomposes the load drift evolution trajectory according to the candidate forwarding link, network egress port, and scheduling time slice to form a link-dimensional drift component, a port-dimensional drift component, and a time-dimensional drift component. The link-dimensional drift component originates from the link bandwidth occupancy deviation parameter and the multi-hop transmission deviation parameter, and is used to express the continuous occupancy changes of the original communication service flow on the candidate forwarding link. The port-dimensional drift component originates from the egress port queuing deviation parameter and the queue priority occupancy deviation parameter, and is used to express the queuing conflict changes of the original communication service flow at the network egress port. The time-dimensional drift component originates from the scheduling time slice occupancy deviation parameter, and is used to express the entry density of the original communication service flow in the scheduling time slice. The link-dimensional drift component, the port-dimensional drift component, and the time-dimensional drift component together constitute the spatiotemporal projection result, enabling the load drift evolution trajectory to be converted into a spatiotemporal projection result jointly expressed by the link-dimensional drift component, the port-dimensional drift component, and the time-dimensional drift component. The spatiotemporal projection result continues to serve as the input for load integration calculation.

[0063] Preferably, in the specific technical implementation of step 2, when performing spatiotemporal dimension projection processing on the load drift evolution trajectory, the link dimension drift component, the port dimension drift component, and the time dimension drift component are further labeled with service categories based on the differences in service categories between time-sensitive and non-time-sensitive service flows, so as to write service category labeling information into the spatiotemporal projection result. The service category labeling information includes service category labeling information corresponding to time-sensitive service flows and service category labeling information corresponding to non-time-sensitive service flows; after writing the service category labeling information, the spatiotemporal projection result includes the link dimension drift component corresponding to time-sensitive service flows, the port dimension drift component corresponding to time-sensitive service flows, the time dimension drift component corresponding to time-sensitive service flows, the link dimension drift component corresponding to non-time-sensitive service flows, the port dimension drift component corresponding to non-time-sensitive service flows, and the time dimension drift component corresponding to non-time-sensitive service flows. The spatiotemporal projection result continues to participate in the load integral calculation, enabling the load integral calculation to distinguish the periodic control transmission pressure of time-sensitive service flows and the burst data occupancy pressure of non-time-sensitive service flows based on the service category labeling information. The periodic control transmission pressure and the burst data occupancy pressure are then respectively included in the subsequent multi-dimensional composite load deviation calculation process, avoiding the mixing of load fluctuations of different service categories into a single congestion result within the same scheduling time slice.

[0064] Preferably, in step 2, when performing load integration calculation based on the spatiotemporal projection result, firstly, using the scheduling time slice as the integration unit, link occupancy accumulation processing is performed on the link dimension drift component in the spatiotemporal projection result to form a link load integration result; then, using the scheduling time slice as the integration unit, port queuing accumulation processing is performed on the port dimension drift component in the spatiotemporal projection result to form a port load integration result; subsequently, using the scheduling time slice as the integration unit, time density accumulation processing is performed on the time dimension drift component in the spatiotemporal projection result to form a time load integration result. The link load integration result is used to express the bandwidth occupancy continuity when the original communication service flow is transmitted along the candidate forwarding link, the port load integration result is used to express the degree of conflict accumulation when the original communication service flow queues at the network egress port, and the time load integration result is used to express the degree of time concentration when the original communication service flow enters the scheduling time slice. The link load integral result, the port load integral result, and the time load integral result are all derived from the spatiotemporal projection result, and continue to participate in the calculation of the multidimensional composite load deviation, so that the multidimensional composite load deviation can reflect the intensity of traffic fluctuations in multiple dimensions of link, port, and time.

[0065] Preferably, when calculating the multi-dimensional composite load deviation, step 2 does not directly add the link load integral result, the port load integral result, and the time load integral result indiscriminately. Instead, it first performs attribution verification under the same service flow based on the service category occupancy parameter, traffic packet length occupancy parameter, transmission time interval parameter, packet length interval association parameter, and multi-hop transmission parameter in the global attribute distribution of the service to be scheduled, to form attribution-verified link load integral result, attribution-verified port load integral result, and attribution-verified time load integral result. The attribution-verified link load integral result, the attribution-verified port load integral result, and the attribution-verified time load integral result all correspond to the same original communication service flow, the same candidate forwarding link, the same network egress port, and the same preset scheduling period. Subsequently, the attribution-verified link load integral result, the attribution-verified port load integral result, and the attribution-verified time load integral result are compositely characterized to calculate the multi-dimensional composite load deviation. Therefore, the multi-dimensional composite load deviation is not a single-dimensional congestion level, but a load deviation result formed by the same original communication service flow in terms of link occupancy, port queuing and time concentration. The multi-dimensional composite load deviation continues to be dynamically prioritized in conjunction with the global attribute distribution of the service to be scheduled.

[0066] Preferably, in step 2, after forming the multi-dimensional composite operating condition load deviation, the multi-dimensional composite operating condition load deviation is dynamically prioritized in relation to the overall attribute distribution of the services to be scheduled. Specifically, the time-sensitive service flow occupancy information, non-time-sensitive service flow occupancy information, traffic packet length distribution parameters, transmission time interval parameters, packet length interval association records, and multi-hop packet length interval transmission records corresponding to the multi-dimensional composite operating condition load deviation are first read from the overall attribute distribution of the services to be scheduled. The time-sensitive service flow occupancy information, non-time-sensitive service flow occupancy information, traffic packet length distribution parameters, transmission time interval parameters, packet length interval association records, and multi-hop packet length interval transmission records are then organized into a dynamic priority evaluation basis. The dynamic priority evaluation criteria are used to describe the source location, source service category, and source scheduling time slice of the multi-dimensional composite load deviation. The source location corresponds to the candidate forwarding link and the network egress port; the source service category corresponds to the time-sensitive service flow or the non-time-sensitive service flow; and the source scheduling time slice corresponds to the scheduling time slice to which the multi-dimensional composite load deviation belongs. Subsequently, the multi-dimensional composite load deviation is matched with the dynamic priority evaluation criteria to form a dynamic priority evaluation result for the service flow. The dynamic priority evaluation result for the service flow is used to characterize the sequential processing tendency of each original communication service flow within the same preset scheduling period and is further used to construct the dynamic priority preemption scheduling sequence for the service flow.

[0067] Preferably, in step 2, during dynamic priority evaluation, for the dynamic priority evaluation results of service flows corresponding to time-sensitive service flows, the transmission time interval parameter, time-sensitive service flow occupancy information, and multi-hop packet length interval transmission record in the dynamic priority evaluation basis are read to determine whether the time-sensitive service flows form continuous time dimension drift components in multiple scheduling time slices; for the dynamic priority evaluation results of service flows corresponding to non-time-sensitive service flows, the traffic packet length distribution parameter, non-time-sensitive service flow occupancy information, and packet length interval association record in the dynamic priority evaluation basis are read to determine whether the non-time-sensitive service flows form continuous port dimension drift components or link dimension drift components on the same network egress port. The above judgment results do not change the service category affiliation of time-sensitive and non-time-sensitive service flows, but rather change the scheduling order of time-sensitive and non-time-sensitive service flows within a preset scheduling period. Therefore, the dynamic priority evaluation results of the service flows can be distinguished from the static sorting method of fixed priority queues, allowing the same service category to have different scheduling order expressions under different network load conditions. These scheduling order expressions continue to be included in the service flow scheduling order record generation process.

[0068] Preferably, in step 2, when constructing the dynamic priority preemption scheduling sequence of the service flow based on the dynamic priority evaluation results, the dynamic priority evaluation results of the service flow are first merged according to the candidate forwarding links to form a link merging scheduling record; then, the link merging scheduling record is merged according to the network egress port to form a port merging scheduling record; subsequently, the port merging scheduling record is arranged sequentially according to the scheduling time slice to form a service flow scheduling order record. The link merging scheduling record is used to express the sequential processing relationship between different original communication service flows on the same candidate forwarding link, the port merging scheduling record is used to express the sequential processing relationship between different original communication service flows on the same network egress port, and the service flow scheduling order record is used to express the processing order of different original communication service flows in each scheduling time slice within the same preset scheduling period. The service flow scheduling order record continues to serve as the main content of the dynamic priority preemption scheduling sequence of the service flow, enabling the dynamic priority preemption scheduling sequence of the service flow to correspond to the candidate forwarding link, network egress port, and scheduling time slice.

[0069] Preferably, in constructing the dynamic priority preemption scheduling sequence for the service flow, step 2 further identifies preemption trigger candidate relationships based on the dynamic priority evaluation results of the service flow. Specifically, when the multi-dimensional composite load deviation of a time-sensitive service flow indicates that the time-sensitive service flow has formed continuous deviations in both the time dimension drift component and the port dimension drift component, and the multi-dimensional composite load deviation of a non-time-sensitive service flow indicates that the non-time-sensitive service flow is occupying the same network egress port, the same scheduling time slice, or the same candidate forwarding link, the correspondence between the time-sensitive service flow and the non-time-sensitive service flow is organized into preemption trigger candidate relationships. The preemption trigger candidate relationships are further written into the dynamic priority preemption scheduling sequence for the service flow, and are used in step 3 to determine whether a target priority packet performs channel preemption processing on a controlled priority packet. The preemption trigger candidate relationship does not directly execute channel preemption processing, but provides a judgment basis for step 3 from the load drift evolution trajectory, the spatiotemporal projection result and the multi-dimensional composite working condition load deviation; the judgment basis enters the service flow dynamic priority preemption scheduling sequence through the preemption trigger candidate relationship, and is used in step 3 to determine the channel occupancy conflict between the target priority message and the controlled priority message.

[0070] Preferably, in step 2, when forming the dynamic priority preemption scheduling sequence of the service flow, a consistency check is also performed on the service flow scheduling order record and the preemption trigger candidate relationship. The consistency check first verifies whether each original communication service flow in the service flow scheduling order record can find the corresponding time-sensitive service flow occupancy information or non-time-sensitive service flow occupancy information in the global attribute distribution of the service to be scheduled, and then verifies whether the time-sensitive service flows and non-time-sensitive service flows in the preemption trigger candidate relationship correspond to the same network egress port, the same candidate forwarding link, or the same scheduling time slice. The service flow scheduling order record that passes the consistency check and the preemption trigger candidate relationship that passes the consistency check together form the service flow dynamic priority preemption scheduling sequence. The service flow scheduling order record that fails the consistency check is returned to the dynamic priority evaluation basis for verification, and the preemption trigger candidate relationship that fails the consistency check is returned to the dynamic priority evaluation basis for verification, so as to prevent the scheduling order that does not correspond to the global attribute distribution of the service to be scheduled from entering the service flow dynamic priority preemption scheduling sequence, and to prevent the preemption trigger candidate relationship that does not correspond to the global attribute distribution of the service to be scheduled from entering the service flow dynamic priority preemption scheduling sequence.

[0071] Preferably, in a scenario, when step 2 is specifically implemented, if the unmanned collaborative node simultaneously has time-sensitive service flows corresponding to collaborative control messages, non-time-sensitive service flows corresponding to track status feedback messages, and non-time-sensitive service flows corresponding to log reporting messages within the same preset scheduling period, then the spatiotemporal projection results of the time-sensitive service flows corresponding to collaborative control messages, the spatiotemporal projection results of the non-time-sensitive service flows corresponding to track status feedback messages, and the spatiotemporal projection results of the non-time-sensitive service flows corresponding to log reporting messages are obtained based on the load drift evolution trajectory. For time-sensitive service flows corresponding to coordinated control messages, the spatiotemporal projection results of these flows typically show continuous deviations in both the time-dimension drift component and the port-dimension drift component. For non-time-sensitive service flows corresponding to track status feedback messages, the spatiotemporal projection results typically show high continuous occupancy in both the link-dimension drift component and the port-dimension drift component. For non-time-sensitive service flows corresponding to log reporting messages, the spatiotemporal projection results typically show intermittent entry in the time-dimension drift component. Step 2 involves performing load integration calculations on the spatiotemporal projection results of the time-sensitive service flows corresponding to the coordinated control messages, the non-time-sensitive service flows corresponding to the track status feedback messages, and the non-time-sensitive service flows corresponding to the log reporting messages. This yields the multi-dimensional composite load deviation of the time-sensitive service flows corresponding to the coordinated control messages, the multi-dimensional composite load deviation of the non-time-sensitive service flows corresponding to the track status feedback messages, and the multi-dimensional composite load deviation of the non-time-sensitive service flows corresponding to the log reporting messages. Subsequently, based on the time-sensitive service flows corresponding to the coordinated control messages... The multidimensional composite load deviation of the flow, the multidimensional composite load deviation of the non-time-sensitive service flow corresponding to the track status return message, the multidimensional composite load deviation of the non-time-sensitive service flow corresponding to the log reporting message, and the global attribute distribution of the service to be scheduled are dynamically prioritized to arrange the time-sensitive service flow corresponding to the cooperative control message before the non-time-sensitive service flow corresponding to the track status return message, and place the non-time-sensitive service flow corresponding to the log reporting message in the scheduling time slice with lower port-dimensional drift component and link-dimensional drift component, thereby forming the dynamic priority preemption scheduling sequence of the service flow.

[0072] In summary, the dynamic priority preemption scheduling sequence of the service flow generated in step 2 includes a service flow scheduling order record, preemption trigger candidate relationships, service flow dynamic priority evaluation results, multi-dimensional composite working condition load deviation, spatiotemporal projection results, and load drift evolution trajectory source information. The service flow scheduling order record characterizes the sequential processing order of each original communication service flow within a preset scheduling period. The preemption trigger candidate relationships characterize whether there is a candidate relationship for channel preemption processing between time-sensitive and non-time-sensitive service flows. The service flow dynamic priority evaluation results explain the basis for generating the service flow scheduling order record. The multi-dimensional composite working condition load deviation characterizes the intensity of traffic fluctuations. The spatiotemporal projection results explain the projection source of the multi-dimensional composite working condition load deviation. The load drift evolution trajectory source information explains that the spatiotemporal projection results originate from the link bandwidth occupancy deviation parameter, scheduling time slice occupancy deviation parameter, queue priority occupancy deviation parameter, egress port queuing deviation parameter, and multi-hop transmission deviation parameter in the load drift evolution trajectory. Therefore, the dynamic priority preemption scheduling sequence of the service flow can continuously transmit the load drift evolution trajectory, the spatiotemporal projection result, the multi-dimensional composite working condition load deviation, the global attribute distribution of the service to be scheduled, and the dynamic priority evaluation result of the service flow to step 3, so that step 3 can monitor the channel preemption processing relationship between the target priority message and the controlled priority message based on the dynamic priority preemption scheduling sequence of the service flow.

[0073] Optionally, step 3 includes: When it is detected that a target priority message has entered the scheduling ready state and the network egress port is transmitting a controlled priority message that has a channel occupation conflict with the target priority message, the control egress queue manager issues a forced interrupt command to the controlled priority message and obtains the interrupt command execution result. When the execution result of the interruption instruction indicates that the controlled priority message is interrupted, the transmission position status of the controlled priority message is saved to obtain the status saving information of the interrupted message, so as to construct a preemptive heterogeneous message slice association carrier for characterizing the interleaving distribution characteristics in the physical link.

[0074] Preferably, the specific implementation process of step 3 is as follows: After obtaining the dynamic priority preemption scheduling sequence of the service flow, the service flow scheduling order record, preemption trigger candidate relationship, service flow dynamic priority evaluation result, multi-dimensional composite working condition load deviation, spatiotemporal projection result, and load drift evolution trajectory source information are first read from the dynamic priority preemption scheduling sequence of the service flow. The service flow scheduling order record, the preemption trigger candidate relationship, the service flow dynamic priority evaluation result, the multi-dimensional composite working condition load deviation, the spatiotemporal projection result, and the load drift evolution trajectory source information are then organized into the channel preemption monitoring basis. The channel preemption monitoring criteria are used to limit the source range of target priority packets and controlled priority packets in step 3. Specifically, the service flow scheduling order record is used to determine the processing order of the original communication service flows within a preset scheduling period; the preemption trigger candidate relationship is used to determine whether there is a candidate relationship for channel preemption processing between time-sensitive and non-time-sensitive service flows; the service flow dynamic priority evaluation result is used to determine the scheduling priority corresponding to the target priority packet; and the multi-dimensional composite working condition load deviation, the spatiotemporal projection result, and the load drift evolution trajectory source information are used to determine the network resource deviation source corresponding to the channel preemption processing. Therefore, the channel preemption monitoring criteria do not solely originate from the fixed priority field in the packet header, but rather from the dynamic scheduling relationship already formed in the service flow dynamic priority preemption scheduling sequence. This dynamic scheduling relationship continues to participate in the identification of target priority packets and controlled priority packets.

[0075] Preferably, in the specific technical implementation of step 3, when identifying target priority packets based on the channel preemption monitoring criteria, firstly, time-sensitive service flows in the preceding processing position within a preset scheduling period are read according to the service flow scheduling order record. Then, the scheduling priority of the packets corresponding to the time-sensitive service flows is verified according to the service flow dynamic priority evaluation result to determine whether the packets have a higher scheduling priority than the packets being transmitted at the network egress port, thereby filtering out target priority packet candidates. Subsequently, the queue entry status, packet header reading status, and transmission permission status of the target priority packet candidates in the egress queue manager are read, and the queue entry status, packet header reading status, and transmission permission status are compiled into a scheduling ready state discrimination record. The scheduling readiness status determination record is used to determine whether the target priority packet candidate has entered a schedulable queue position, whether it has completed the packet header reading for forwarding control, and whether it has the conditions to occupy the network egress port for transmission. When the scheduling readiness status determination record indicates that the target priority packet candidate has the transmission conditions, the target priority packet candidate is determined as a target priority packet, and the target priority packet is written into the channel preemption monitoring object corresponding to the channel preemption monitoring basis. The channel preemption monitoring object is used to record target priority packets that have entered the scheduling readiness state and continue to participate in controlled priority packet identification.

[0076] Preferably, after determining the target priority message, step 3 continues to read the current transmission status record of the network egress port to determine the controlled priority message being transmitted by the network egress port. The current transmission status record of the network egress port is generated by the egress queue manager based on the current transmission process of the network egress port. The current transmission status record includes the service flow type currently occupying the network egress port, the queue position of the currently occupying network egress port, the message transmission progress of the currently occupying network egress port, and the candidate forwarding link of the currently occupying network egress port. Step 3, based on the service flow type, queue position, message transmission progress, and candidate forwarding link in the current transmission status record of the network egress port, identifies whether the message being transmitted by the network egress port belongs to a low-scheduled message in the preemption trigger candidate relationship that corresponds to the target priority message. When the current transmission status record of the network egress port indicates that the message being transmitted by the network egress port belongs to a low-scheduled message in the preemption trigger candidate relationship that corresponds to the target priority message, the message being transmitted by the network egress port is determined as a controlled priority message. Therefore, the controlled priority message is not an arbitrary low priority message, but a message being transmitted on a network egress port that is in the same network egress port or in the same candidate forwarding link conflict relationship as the target priority message.

[0077] Preferably, after determining the target priority message and the controlled priority message in step 3, channel occupancy conflict judgment is performed on the target priority message and the controlled priority message. Specifically, firstly, the target priority message is confirmed to have entered the scheduling ready state according to the scheduling ready state judgment record, and then the controlled priority message is confirmed to be occupying the network egress port according to the current transmission status record of the network egress port. Subsequently, the candidate forwarding link, network egress port, and scheduling time slice corresponding to the target priority message are compared with the candidate forwarding link, network egress port, and message transmission progress corresponding to the controlled priority message to form a channel occupancy conflict judgment result. The channel occupancy conflict determination result is used to characterize whether the target priority message needs to occupy the same network egress port before the controlled priority message has completed transmission; when the channel occupancy conflict determination result characterizes the target priority message and the controlled priority message as having a transmission overlap relationship on the same network egress port, the transmission overlap relationship is used as the conflict expression content in the channel occupancy conflict determination result, and step 3 uses the channel occupancy conflict determination result including the transmission overlap relationship as the direct basis for issuing a forced interruption command.

[0078] Preferably, step 3 controls the egress queue manager to issue a forced interruption command to the controlled priority packet based on the channel occupancy conflict determination result. The forced interruption command includes the scheduling readiness status record of the target priority packet, the current transmission status record of the network egress port of the controlled priority packet, the channel occupancy conflict determination result, and the packet transmission progress of the controlled priority packet. After receiving the forced interruption command, the egress queue manager first verifies the channel occupancy conflict determination result in the forced interruption command, then determines the interruptible transmission boundary based on the packet transmission progress of the controlled priority packet in the forced interruption command, and suspends the transmission process of the controlled priority packet occupying the network egress port at the interruptible transmission boundary to release the network egress port for the target priority packet to enter for transmission. The interruptible transmission boundary originates from the packet transmission progress of the controlled priority packet, and the interruptible transmission boundary continues to participate in the generation of the interruption command execution result, so that the interruption command execution result can reflect whether the controlled priority packet has been paused at a recoverable position.

[0079] Preferably, in step 3, after the egress queue manager executes the forced interruption instruction, the interruption instruction execution result is obtained. The interruption instruction execution result includes a forced interruption instruction reception status, a channel occupancy conflict verification status, a controlled priority message pause status, a network egress port release status, and an interruptible transmission boundary record. The forced interruption instruction reception status indicates whether the egress queue manager receives the forced interruption instruction; the channel occupancy conflict verification status indicates whether the egress queue manager confirms the channel occupancy conflict judgment result; the controlled priority message pause status indicates whether the controlled priority message stops transmitting at the interruptible transmission boundary; the network egress port release status indicates whether the network egress port switches from the controlled priority message occupancy state to a state available for the target priority message; and the interruptible transmission boundary record indicates the transmission boundary position corresponding to when the controlled priority message is paused. The forced interruption command reception status, the channel occupancy conflict verification status, the controlled priority message pause status, the network egress port release status, and the interruptible transmission boundary record are used together to determine whether the forced interruption command has been executed. When both the controlled priority message pause status and the network egress port release status in the interruption command execution result indicate that the interruption has been executed, the interruption command execution result is used to trigger the saving of the transmission location status of the controlled priority message.

[0080] Preferably, when the interrupt instruction execution result indicates that the controlled priority message is interrupted, step 3 saves the transmission position status of the controlled priority message. Specifically, firstly, based on the interruptible transmission boundary record in the interrupt instruction execution result, the boundary of the transmitted message segment and the start boundary of the untransmitted message segment of the controlled priority message are read; then, based on the current transmission status record of the network egress port of the controlled priority message, the network egress port, candidate forwarding link, queue position, and message transmission progress corresponding to the controlled priority message are read; subsequently, the boundary of the transmitted message segment, the start boundary of the untransmitted message segment, the network egress port, the candidate forwarding link, the queue position, and the message transmission progress are organized into a transmission position status. The transmission position status is used to characterize the portion of the controlled priority message that has been transmitted, the portion that has not been transmitted, and the queue position to which it needs to return when transmission is resumed when interrupted; wherein, the portion that has been transmitted corresponds to the boundary of the transmitted message segment, the portion that has not been transmitted corresponds to the start boundary of the untransmitted message segment, and the queue position to which it needs to return when transmission is resumed corresponds to the queue position. The transmission location status continues to be written into the status storage information of the interrupted message.

[0081] Preferably, when forming the state preservation information of the interrupted message in step 3, the sending position state, the interruption command execution result, the message header metadata reading result of the controlled priority message, and the source information of the dynamic priority preemption scheduling sequence of the service flow corresponding to the controlled priority message are associated and preserved. The source information of the dynamic priority preemption scheduling sequence of the service flow corresponding to the controlled priority message comes from the service flow scheduling order record, preemption trigger candidate relationship, service flow dynamic priority evaluation result, multi-dimensional composite working condition load deviation, spatiotemporal projection result, and load drift evolution trajectory source information in the channel preemption monitoring basis. The source information of the dynamic priority preemption scheduling sequence of the service flow corresponding to the controlled priority message is used to express the scheduling source of the interruption of the controlled priority message. The state preservation information of the interrupted message includes the sending location status, the interrupt command execution result, the header metadata reading result of the controlled priority message, and the source information of the dynamic priority preemption scheduling sequence of the service flow corresponding to the controlled priority message. Specifically, the sending location status expresses the sending boundary required for the controlled priority message to resume transmission; the interrupt command execution result expresses the execution process of the controlled priority message being interrupted; the header metadata reading result of the controlled priority message expresses the original communication service flow to which the controlled priority message belongs; and the source information of the dynamic priority preemption scheduling sequence of the service flow corresponding to the controlled priority message expresses the scheduling source of the interruption of the controlled priority message. The state preservation information of the interrupted message subsequently serves as the state basis for constructing the preemptive heterogeneous message slice association carrier.

[0082] Preferably, in step 3, when constructing the preemptive heterogeneous packet slice association carrier, the target priority packet, the controlled priority packet, the channel occupancy conflict judgment result, the interruption command execution result, and the state saving information of the interrupted packet are first associated to form a preemptive association basic record. The preemptive association basic record is used to express the scheduling basis for the target priority packet to enter the network egress port currently occupied by the controlled priority packet, and the state basis for the controlled priority packet to retain a recoverable transmission position after being interrupted. Subsequently, based on the preemptive association basic record, the interleaving relationship between the target priority packet and the controlled priority packet in the same physical link is organized into a carrier-based manner to construct the preemptive heterogeneous packet slice association carrier. The preemptive heterogeneous packet slice association carrier includes a scheduling readiness status record for the target priority packet, the transmission location status of the controlled priority packet, the channel occupancy conflict determination result, the interruption command execution result, and the status storage information of the interrupted packet. This allows the preemptive heterogeneous packet slice association carrier to characterize the association between the insertion and transmission of the target priority packet and the subsequent resumption of transmission of the controlled priority packet. This association is imported into the preemptive heterogeneous packet slice association carrier from the preemptive association base record and continues to serve as input for step 4, performing packet slice and network bandwidth resource mapping and matching calculations.

[0083] Preferably, in one scenario, when implementing step 3, if there are time-sensitive service flows corresponding to the collaborative control message and non-time-sensitive service flows corresponding to the trajectory status return message within the same preset scheduling period of the unmanned collaborative node, and the dynamic priority preemption scheduling sequence of the service flows indicates that the time-sensitive service flow corresponding to the collaborative control message enters the network egress port before the non-time-sensitive service flow corresponding to the trajectory status return message, then step 3 first determines the target priority message corresponding to the collaborative control message according to the dynamic priority preemption scheduling sequence of the service flows, and then determines that the controlled priority message corresponding to the trajectory status return message is occupying the network egress port through the current transmission status record of the network egress port. Subsequently, step 3 forms a channel occupancy conflict judgment result based on the scheduling readiness status judgment record of the target priority message, the current transmission status record of the network egress port of the controlled priority message, and the preemption trigger candidate relationship, and controls the egress queue manager to issue a forced interruption command based on the channel occupancy conflict judgment result. After the egress queue manager suspends the transmission of the controlled priority message according to the forced interruption command, step 3 saves the transmission position status of the controlled priority message to form the status saving information of the interrupted message. The target priority message, the controlled priority message, the channel occupancy conflict judgment result, the interruption command execution result, and the status saving information of the interrupted message are then associated to form the preemptive heterogeneous message slice association carrier. The preemptive heterogeneous message slice association carrier records the scheduling basis for inserting the target priority message corresponding to the cooperative control message into the network egress port, and also records the transmission position status required for the controlled priority message corresponding to the track status return message to resume transmission.

[0084] In summary, the preemptive heterogeneous packet slice association carrier constructed in step 3 is passed to step 4. The target priority packet in the preemptive heterogeneous packet slice association carrier is used to participate in the generation of network transmission time slot allocation mapping in step 4. The state preservation information of the controlled priority packet and the interrupted packet in the preemptive heterogeneous packet slice association carrier is used to participate in the timing scheduling for the subsequent resumption of transmission of the controlled priority packet in step 4. The channel occupancy conflict discrimination result and the interruption command execution result in the preemptive heterogeneous packet slice association carrier are used to explain the source of the interleaving distribution of the target priority packet and the controlled priority packet in the physical link. Therefore, step 3 does not complete a single interruption action, but continuously transmits the dynamic priority preemption scheduling sequence of the service flow, the channel preemption monitoring basis, the target priority message, the controlled priority message, the forced interruption instruction, the interruption instruction execution result, the sending position status, and the status saving information of the interrupted message to the preemptive heterogeneous message slice association carrier, so that the preemptive heterogeneous message slice association carrier can serve as the input basis for time-sensitive network scheduling in step 4.

[0085] Optionally, a preemptive heterogeneous packet slice association carrier is constructed to characterize the interleaving distribution characteristics in the physical link, including: Based on the state storage information, the interrupted controlled priority message is subjected to data fragmentation and marking processing based on the minimum slice granularity to generate controlled priority fragmented message data. Based on the controlled priority fragmented message data and the target priority message that triggers the channel preemption process, a preemptive heterogeneous message slice association carrier for the interleaving distribution characteristics in the physical link is constructed.

[0086] Preferably, in step 3, when performing data fragmentation marking processing on the interrupted controlled priority message based on the minimum slice granularity according to the state-saved information of the interrupted message, the sending position status, interruption command execution result, header metadata reading result of the controlled priority message, and source information of the dynamic priority preemption scheduling sequence of the service flow corresponding to the controlled priority message are first read from the state-saved information of the interrupted message. The sending position status, interruption command execution result, header metadata reading result of the controlled priority message, and source information of the dynamic priority preemption scheduling sequence of the service flow corresponding to the controlled priority message are then organized into the fragmentation marking processing basis. This fragmentation marking processing basis is used to define which message segments in the interrupted controlled priority message have been transmitted, which message segments have not been transmitted, and the corresponding candidate forwarding link and queue position required when the message segments that have not been transmitted resume transmission at the network egress port. Therefore, this fragmentation marking processing basis continues to serve as input for determining the minimum slice granularity, enabling subsequent data fragmentation marking processing to correspond to the sending position status when the controlled priority message is forcibly interrupted.

[0087] Preferably, in the specific implementation of step 3, the minimum slice granularity is not to arbitrarily divide the controlled priority message into segments, but is determined based on the boundaries of sent message segments, the starting boundaries of unsent message segments, the network egress port, the candidate forwarding links, and the message transmission progress in the fragmentation marking processing criteria. Specifically, firstly, the end position of the message segments in the controlled priority message that have occupied physical links is determined based on the boundaries of the sent message segments, and then the starting position of the message segments in the controlled priority message that have not yet occupied physical links is determined based on the starting boundaries of the unsent message segments. Subsequently, the connection position between the boundaries of the sent message segments and the starting boundaries of the unsent message segments is taken as the recoverable fragmentation boundary, and the recoverable fragmentation boundary is verified based on the network egress port, the candidate forwarding links, and the message transmission progress to form the minimum slice granularity. The minimum slice granularity is used to limit the range of message segments in the controlled priority message that can be individually marked, individually attached to the transmission position status, and participate in subsequent timing scheduling after being interrupted. The minimum slice granularity then participates in the generation of controlled priority fragmented message data.

[0088] Preferably, after obtaining the minimum slice granularity, step 3 performs data fragmentation and marking processing on the interrupted controlled priority messages. Specifically, firstly, the unsent message segments of the interrupted controlled priority messages are divided into boundaries according to the minimum slice granularity to form controlled priority message segments to be recovered; then, the controlled priority message segments to be recovered are bound to the queue position, network egress port, candidate forwarding link, and message transmission progress in the transmission position status to form controlled priority message segment marking records. The controlled priority message segment marking records are used to express the network egress port, candidate forwarding link, and queue position that each controlled priority message segment to be recovered should return to when resuming transmission; subsequently, the controlled priority message segments to be recovered and the controlled priority message segment marking records are jointly organized into controlled priority fragmented message data. The controlled priority fragmented message data therefore includes not only the controlled priority message segments to be recovered from the interrupted controlled priority messages, but also the controlled priority message segment marking records required for the controlled priority message segments to resume transmission.

[0089] Preferably, in step 3, when generating the controlled priority fragmented packet data, the interrupt instruction execution result is also written into the controlled priority packet slice marking record, so that the controlled priority fragmented packet data can retain the execution source of the controlled priority packet being forcibly interrupted. Specifically, the forced interrupt instruction reception status, channel occupancy conflict verification status, controlled priority packet pause status, network egress port release status, and interruptible transmission boundary record in the interrupt instruction execution result are respectively associated with the controlled priority pending recovery packet slice to form a controlled priority fragment execution source record. The controlled priority fragment execution source record is used to describe the interruption source of the controlled priority pending recovery packet slice being in a waiting-to-recovery transmission state, and from which interruptible transmission boundary the controlled priority pending recovery packet slice continues to participate in subsequent scheduling. The controlled priority fragment execution source record is further incorporated into the controlled priority fragmented packet data, so that the controlled priority fragmented packet data can simultaneously express the controlled priority pending recovery packet slice, the controlled priority packet slice marking record, and the controlled priority fragment execution source record.

[0090] Preferably, in step 3, after obtaining the controlled priority fragmented packet data, the target priority packet that triggered the channel preemption process is read, and the queue entry status, packet header reading status, and transmission permission status are extracted from the scheduling ready status discrimination record corresponding to the target priority packet to form a target priority packet preemption entry record. The target priority packet preemption entry record indicates that the target priority packet has met the scheduling conditions to enter the network egress port. Subsequently, the target priority packet preemption entry record is associated with the channel occupancy conflict discrimination result to form a target priority packet preemption source record. The target priority packet preemption source record expresses the source of channel occupancy conflict where the target priority packet needs to enter the same network egress port before the controlled priority packet completes transmission. The target priority packet preemption source record continues to be associated with the controlled priority fragmented packet data to construct the preemptive heterogeneous packet slice association carrier.

[0091] Preferably, in step 3, when constructing the preemptive heterogeneous packet slice association carrier based on the controlled priority fragmented packet data and the target priority packet that triggers the channel preemption process, the target priority packet, the target priority packet preemption entry record, the target priority packet preemption source record, the controlled priority fragmented packet data, and the channel occupancy conflict judgment result are first associated with the same network egress port to form an egress port preemption slice association record. The egress port preemption slice association record is used to express the scheduling conditions for the target priority packet to enter the same network egress port, and the fragmented position that the controlled priority fragmented packet data needs to recover after the target priority packet is inserted and transmitted. Subsequently, the egress port preemption slice association records are merged according to candidate forwarding links to form a link preemption slice association record. The link preemption slice association record is used to express the interleaving transmission relationship between the target priority packet and the controlled priority fragmented packet data on the same candidate forwarding link, and the link preemption slice association record continues to serve as a component of the preemptive heterogeneous packet slice association carrier.

[0092] Preferably, after forming the link preemption slice association record in step 3, the link preemption slice association record is further arranged in time sequence according to the scheduling time slice to form a time-sequence preemption slice association record. The time-sequence preemption slice association record is used to express the time position of the insertion and transmission of the target priority packet, and the time position of the resumption and transmission of the controlled priority fragment packet data; wherein, the time position of the insertion and transmission of the target priority packet originates from the target priority packet preemption entry record, and the time position of the resumption and transmission of the controlled priority fragment packet data originates from the controlled priority packet slice marking record and the controlled priority fragment execution source record. The time-sequence preemption slice association record enables the preemptive heterogeneous packet slice association carrier to express the sequential transmission relationship between the target priority packet and the controlled priority fragment packet data in the same physical link, rather than simply recording one interruption action of the target priority packet on the controlled priority packet.

[0093] Preferably, in step 3, when constructing the preemptive heterogeneous packet slice association carrier, the state storage information of the interrupted packet is also checked for consistency with the controlled priority fragmented packet data. The consistency check first verifies whether each controlled priority fragment to be recovered in the controlled priority fragmented packet data corresponds to the start boundary of the unsent packet segment in the state storage information of the interrupted packet, and then verifies whether the network egress port, candidate forwarding link, and queue position in the controlled priority packet slice marking record correspond to the sending position state. Controlled priority fragmented packet data that passes the consistency check continues to be written into the preemptive heterogeneous packet slice association carrier; controlled priority fragmented packet data that fails the consistency check returns to the fragment marking processing basis for boundary verification, to prevent controlled priority fragments to be recovered lacking a sending position state source from entering the preemptive heterogeneous packet slice association carrier.

[0094] Preferably, step 3 ultimately organizes the egress port preemption slice association record, the link preemption slice association record, the timing preemption slice association record, the controlled priority fragmented packet data, the target priority packet preemption entry record, and the target priority packet preemption source record into the preemptive heterogeneous packet slice association carrier. The preemptive heterogeneous packet slice association carrier is used to characterize the interleaving distribution characteristics of the target priority packet and the controlled priority fragmented packet data in the physical link; wherein, the egress port preemption slice association record is used to limit the network egress port where the interleaving distribution occurs, the link preemption slice association record is used to limit the candidate forwarding link corresponding to the interleaving distribution, the timing preemption slice association record is used to limit the scheduling time slice corresponding to the interleaving distribution, the controlled priority fragmented packet data is used to limit the controlled priority packet slices that need to be restored after interruption, and the target priority packet preemption entry record and the target priority packet preemption source record are used to limit the scheduling source of the target priority packet entering the physical link. The above content is collectively incorporated into the preemptive heterogeneous packet slice association carrier, enabling the preemptive heterogeneous packet slice association carrier to be used in step 4 for the mapping and matching calculation of packet slices and network bandwidth resources.

[0095] Preferably, in one scenario, when step 3 is specifically implemented, if the target priority message corresponding to the cooperative control message triggers the channel preemption process, and the controlled priority message corresponding to the track status return message is forcibly interrupted at the network egress port, then step 3 first determines the start boundary of the unsent message fragment of the controlled priority message corresponding to the track status return message based on the state preservation information of the interrupted message, and then performs data fragmentation marking processing on the unsent message fragment of the controlled priority message corresponding to the track status return message according to the minimum slice granularity to generate controlled priority fragmented message data corresponding to the track status return message. Subsequently, step 3 associates the target priority message preemption entry record, the target priority message preemption source record, and the controlled priority fragmented message data corresponding to the track status return message to form the egress port preemption slice association record, the link preemption slice association record, and the timing preemption slice association record. The resulting preemptive heterogeneous message slice association carrier can express the slice association relationship between the insertion transmission of cooperative control messages and the subsequent recovery transmission of track status return messages. The slice association relationship is jointly characterized by the exit port preemptive slice association record, the link preemptive slice association record, and the timing preemptive slice association record.

[0096] In summary, the preemptive heterogeneous packet slice association carrier constructed in step 3 is passed to step 4. The target priority packet in the preemptive heterogeneous packet slice association carrier is used to participate in the generation of network transmission time slot allocation mapping relationship in step 4. The controlled priority fragmented packet data in the preemptive heterogeneous packet slice association carrier is used to participate in the timing scheduling for the subsequent recovery transmission of the interrupted controlled priority packet in step 4. The egress port preemptive slice association record, the link preemptive slice association record, and the timing preemptive slice association record in the preemptive heterogeneous packet slice association carrier are used to explain the source and location of the interleaving distribution of the target priority packet and the controlled priority fragmented packet data in the physical link. Therefore, the preemptive heterogeneous packet slice association carrier continuously transmits the technical associations between the interrupted packet's state storage information, the minimum slice granularity, the controlled priority sliced ​​packet data, the target priority packet, the channel preemption processing, the egress port preemption slice association record, the link preemption slice association record, and the time-series preemption slice association record to step 4, enabling step 4 to perform time-sensitive network scheduling based on the preemptive heterogeneous packet slice association carrier.

[0097] Optionally, step 4 includes: The preemptive heterogeneous packet slice association carrier is input into a time-aware shaping algorithm engine built on the time-sensitive network protocol standard to perform mapping and matching calculations between packet slices and network bandwidth resources, so as to generate a network transmission time slot allocation mapping relationship. By calling the master clock in the distributed network, the time synchronization reference provided by the master clock is obtained and determined as the master clock time synchronization reference; Based on the master clock time synchronization reference, the gated scheduling time base of the time-aware shaping algorithm engine is aligned and calibrated to obtain the calibrated time synchronization reference; Based on the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, a time gating list is generated, and then a distributed time-sensitive network scheduling gating sequence for the original communication service flow is constructed.

[0098] Preferably, the specific implementation process of step 4 is as follows: The time-aware shaping algorithm engine includes a preemption carrier parsing module, a packet slice requirement compilation module, a bandwidth resource candidate matching module, a time slot mapping generation module, a master clock synchronization reading module, a gating scheduling time base calibration module, a time gating list arrangement module, and a distributed gating sequence construction module. The preemption carrier parsing module is used to perform structured parsing on the preemptive heterogeneous packet slice association carrier to extract target priority packets, controlled priority fragmented packet data, egress port preemption slice association records, link preemption slice association records, and time-series preemption slice association records; the packet slice requirement compilation module is used to generate packet slice transmission requirement records based on the target priority packets and the controlled priority fragmented packet data; the bandwidth resource candidate matching module is used to perform network bandwidth resource candidate matching based on the packet slice transmission requirement records to obtain packet slice bandwidth matching records; the time slot mapping generation module is used to generate network transmission time slot allocation mapping relationships based on the packet slice bandwidth matching records, and is used for The system invokes pre-configured gating switch constraints to perform gating executability verification on the network transmission time slot allocation mapping relationship; the master clock synchronization reading module is used to obtain the time synchronization reference provided by the master clock in the distributed network and form a master clock time synchronization reference; the gating scheduling time base calibration module is used to align and calibrate the gating scheduling time base according to the master clock time synchronization reference to obtain a calibrated time synchronization reference; the time gating list arrangement module is used to generate a time gating list according to the calibrated time synchronization reference and the network transmission time slot allocation mapping relationship; the distributed gating sequence construction module is used to construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow according to the time gating list. These modules are sequentially connected according to the data transmission order, enabling the preemptive heterogeneous packet slice association carrier to be transformed from an interleaved distribution relationship into a distributed time-sensitive network scheduling gating sequence that can be used for network egress port opening and closing control.

[0099] Preferably, in the specific technical implementation of step 4, the preemption carrier parsing module first reads the target priority packet, the controlled priority fragmented packet data, the egress port preemption slice association record, the link preemption slice association record, and the time-series preemption slice association record from the preemptive heterogeneous packet slice association carrier, and organizes the target priority packet, the controlled priority fragmented packet data, the egress port preemption slice association record, the link preemption slice association record, and the time-series preemption slice association record into a gating scheduling input record. The gating scheduling input record is used to express the network egress port, candidate forwarding link, and scheduling time slice that the target priority packet needs to be inserted for transmission, and also to express the network egress port, candidate forwarding link, and scheduling time slice that the controlled priority fragmented packet data needs to resume transmission. Subsequently, the preemptive carrier parsing module transmits the gating scheduling input record to the message slice demand compilation module, enabling the message slice demand compilation module to organize the message slice transmission requirements within the preemptive heterogeneous message slice associated carrier's defined interleaving distribution relationship, instead of regenerating a scheduling object unrelated to step 3.

[0100] Preferably, in step 4, after receiving the gating scheduling input record, the packet slice requirement compilation module performs packet slice transmission requirement parsing on the target priority packet and the controlled priority fragmented packet data to form a packet slice transmission requirement record. Specifically, the packet slice requirement compilation module first extracts the insertion transmission requirement corresponding to the target priority packet based on the target priority packet preemption entry record and the target priority packet preemption source record, and then extracts the recovery transmission requirement corresponding to the controlled priority fragmented packet data based on the controlled priority packet slice marking record and the controlled priority fragmented execution source record; subsequently, the packet slice requirement compilation module extracts the queue occupancy requirement corresponding to the network egress port based on the egress port preemption slice association record, and extracts the bandwidth occupancy requirement corresponding to the candidate forwarding link based on the link preemption slice association record, so as to organize the insertion transmission requirement, the recovery transmission requirement, the queue occupancy requirement, and the bandwidth occupancy requirement into the packet slice transmission requirement record. The packet slice transmission demand record continues to be passed to the bandwidth resource candidate matching module, enabling the bandwidth resource candidate matching module to simultaneously handle the network bandwidth resource competition relationship between target priority packet insertion transmission and controlled priority fragmented packet data recovery transmission.

[0101] Preferably, in step 4, when performing the mapping and matching calculation of packet slices and network bandwidth resources, the bandwidth resource candidate matching module first reads the queue gating status of the network egress port, the bandwidth reservation status of the candidate forwarding links, and the availability status of the scheduling time slices according to the packet slice transmission requirement record, and organizes the queue gating status, the bandwidth reservation status, and the availability status into a network bandwidth resource candidate allocation record. The network bandwidth resource candidate allocation record is used to express the priority queue of the network egress port that the target priority packet and the controlled priority fragment packet data can enter, the network bandwidth resources of the candidate forwarding links that can be occupied, and the scheduling time slices that can be included. Subsequently, the bandwidth resource candidate matching module performs resource correspondence processing on the network bandwidth resource candidate allocation record in the order of priority matching of the insertion transmission requirement corresponding to the target priority packet and continuation matching of the recovery transmission requirement corresponding to the controlled priority fragment packet data, so as to form a packet slice bandwidth matching record. The packet slice bandwidth matching record is used to express the network egress port, candidate forwarding link, and scheduling time slice corresponding to the target priority packet and the controlled priority fragmented packet data, respectively. The packet slice bandwidth matching record is then passed to the time slot mapping generation module.

[0102] Preferably, after receiving the packet slice bandwidth matching record, the time slot mapping generation module first merges the packet slice bandwidth matching record according to the network egress port to form a port time slot merging record; then, it merges the port time slot merging record according to the candidate forwarding link to form a link time slot merging record; subsequently, it arranges the link time slot merging record according to the scheduling time slice to form a network transmission time slot allocation mapping relationship. The port time slot merging record is used to express the priority queue entry relationship between the target priority packet and the controlled priority fragment packet data in the same network egress port; the link time slot merging record is used to express the bandwidth occupancy relationship between the target priority packet and the controlled priority fragment packet data in the same candidate forwarding link; and the network transmission time slot allocation mapping relationship is used to express the scheduling time slice for the insertion of the target priority packet into transmission and the scheduling time slice for the resumption of transmission of the controlled priority fragment packet data. Therefore, the network transmission time slot allocation mapping relationship is formed by the packet slice bandwidth matching record through port merging, link merging and timing arrangement, and can continue to serve as the resource basis for the time gating list arrangement module to generate the time gating list.

[0103] Preferably, when forming the network transmission time slot allocation mapping relationship, the time slot mapping generation module also performs gating executability verification on the network transmission time slot allocation mapping relationship according to the gating switch constraints in the time-aware shaping algorithm engine. The gating switch constraints are used to limit whether different priority queues in the same network egress port are allowed to open within the same scheduling time slice, the queue gating opening interval corresponding to the insertion and transmission of the target priority packet, the queue gating opening interval corresponding to the resumption and transmission of the controlled priority fragmented packet data, and the protection time boundary for queue switching. The time slot mapping generation module first verifies whether the scheduling time slice corresponding to the target priority packet matches the queue gating opening interval corresponding to the target priority packet, then verifies whether the scheduling time slice corresponding to the controlled priority fragmented packet data matches the queue gating opening interval corresponding to the controlled priority fragmented packet data, and then verifies whether there is the protection time boundary between the scheduling time slice corresponding to the target priority packet and the scheduling time slice corresponding to the controlled priority fragmented packet data. The network transmission time slot allocation mapping relationship that passes the gating executability verification continues to be passed to the time gating list orchestration module; the network transmission time slot allocation mapping relationship that fails the gating executability verification returns to the bandwidth resource candidate matching module, so that the bandwidth resource candidate matching module can re-schedule time slot matching based on the packet slice bandwidth matching record and the protection time boundary.

[0104] Preferably, in step 4, when obtaining the time synchronization reference by calling the master clock in the distributed network, the master clock synchronization reading module reads the clock count information, synchronization message sending time, and synchronization message arrival time at the network switching node from the master clock in the distributed network. It then organizes these information into a master clock synchronization reading record. This record expresses the time base source of the master clock in the distributed network and the transmission delay state when the network switching node receives the time base source. Subsequently, the master clock synchronization reading module identifies the fixed forwarding offset and port receiving offset formed during the synchronization message transmission process based on the master clock synchronization reading record. It then performs offset correction on the master clock synchronization reading record based on these offsets to form the master clock time synchronization reference. This master clock time synchronization reference is further passed to the gated scheduling time base calibration module, enabling it to use the master clock time synchronization reference as the unified time base for the time-aware shaping algorithm engine to perform gated scheduling time base alignment calibration.

[0105] Preferably, when the gated scheduling time base calibration module aligns and calibrates the gated scheduling time base of the time-aware shaping algorithm engine according to the master clock time synchronization reference, it first reads the gated scheduling time base currently used by the time-aware shaping algorithm engine and identifies the time base difference between the gated scheduling time base and the master clock time synchronization reference to form a gated scheduling time base offset record. The gated scheduling time base offset record is used to express the offset relationship between the current gate opening time, gate closing time of the time-aware shaping algorithm engine, and the master clock time synchronization reference. Subsequently, the gated scheduling time base calibration module performs synchronization correction on the gate opening time and gate closing time in the gated scheduling time base according to the gated scheduling time base offset record to form a calibrated time synchronization reference. The calibrated time synchronization reference is used to ensure that the scheduling time slices in the network transmission time slot allocation mapping relationship correspond to the gate opening time and gate closing time of the time-aware shaping algorithm engine. The calibrated time synchronization reference is then passed to the time gate list orchestration module.

[0106] Preferably, when the time gating list arrangement module generates the time gating list based on the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, it first maps the port time slot merging record, link time slot merging record, and scheduling time slice in the network transmission time slot allocation mapping relationship to the gating opening time and gating closing time in the calibrated time synchronization benchmark to form a gating time mapping record. The gating time mapping record is used to express the time position at which the priority queue of each network egress port should be opened or closed within the corresponding scheduling time slice; subsequently, the time gating list arrangement module merges the gating time mapping record according to the network egress port to form a port gating action record, and then arranges the port gating action record according to the scheduling time slice to form a time gating list. The time gating list is used to limit the queue gating opening state corresponding to the target priority packet entering the network egress port, and also to limit the queue gating opening state corresponding to the controlled priority fragment packet data recovery transmission.

[0107] Preferably, after receiving the time-gated list, the distributed gating sequence construction module first expands the time-gated list hop-by-hop according to the forwarding order of network switching nodes in the candidate forwarding links to form a hop-by-hop gating expansion record. Then, based on the calibrated time synchronization reference, it performs time slot continuation processing on the gating opening and closing times of adjacent network switching nodes in the hop-by-hop gating expansion record to form a hop-by-hop gating continuation record. The hop-by-hop gating continuation record is used to express the timing relationship where the preceding network switching node completes the reception or forwarding of a packet slice in one scheduling time slice, and the subsequent network switching node continues forwarding in an adjacent scheduling time slice. Subsequently, the distributed gating sequence construction module associates the hop-by-hop gating continuation record with the time-gated list to construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow. The distributed time-sensitive network scheduling gating sequence is used to express the gating forwarding order of the target priority packet and the controlled priority fragment packet data among multiple network switching nodes.

[0108] Preferably, when constructing the distributed time-sensitive network scheduling gating sequence, the distributed gating sequence construction module also writes the egress port preemptive slice association record, the link preemptive slice association record, and the time-series preemptive slice association record from the preemptive heterogeneous packet slice association carrier into the distributed time-sensitive network scheduling gating sequence. The egress port preemptive slice association record is used to indicate the network egress port where the target priority packet and the controlled priority fragmented packet data are interleaved in the distributed time-sensitive network scheduling gating sequence. The link preemptive slice association record is used to indicate the candidate forwarding link corresponding to the interleaved distribution in the distributed time-sensitive network scheduling gating sequence. The time-series preemptive slice association record is used to indicate the scheduling time slice relationship between the target priority packet insertion transmission and the controlled priority fragmented packet data recovery transmission in the distributed time-sensitive network scheduling gating sequence. Therefore, the distributed time-sensitive network scheduling gating sequence includes not only the port gating action record formed by the time gating list but also the packet preemption and fragmentation recovery relationship transmitted from the preemptive heterogeneous packet slice association carrier.

[0109] Preferably, in one scenario, when step 4 is specifically implemented, if the target priority message corresponding to the cooperative control message has entered the scheduling through the preemptive heterogeneous message slice association carrier, and the controlled priority fragment message data corresponding to the track status return message is waiting to resume transmission, then the preemptive carrier parsing module first reads the target priority message corresponding to the cooperative control message, the controlled priority fragment message data corresponding to the track status return message, the egress port preemptive slice association record, the link preemptive slice association record, and the timing preemptive slice association record from the preemptive heterogeneous message slice association carrier to form a gated scheduling input record; the message slice requirement compilation module then forms a message slice transmission requirement record based on the gated scheduling input record; the bandwidth resource candidate matching module then generates a message slice bandwidth matching record based on the message slice transmission requirement record; the time slot mapping generation module continues to form a network transmission time slot allocation mapping relationship based on the message slice bandwidth matching record, and performs gating executability verification on the network transmission time slot allocation mapping relationship based on the gating switch constraint. Subsequently, the master clock synchronization reading module forms a master clock time synchronization reference, the gating scheduling time base calibration module forms a calibrated time synchronization reference based on the master clock time synchronization reference, the time gating list arrangement module generates a time gating list based on the calibrated time synchronization reference and the network transmission time slot allocation mapping relationship, and the distributed gating sequence construction module finally constructs a distributed time-sensitive network scheduling gating sequence based on the time gating list. Thus, the target priority message corresponding to the cooperative control message can enter the network egress port according to the time gating list, and the controlled priority fragmented message data corresponding to the track status return message can resume transmission in subsequent scheduling time slots according to the distributed time-sensitive network scheduling gating sequence.

[0110] Preferably, the distributed time-sensitive networking scheduling gating sequence finally formed in step 4 includes the network transmission time slot allocation mapping relationship, the master clock time synchronization reference, the calibrated time synchronization reference, the time gating list, the hop-by-hop gating expansion record, the hop-by-hop gating continuation record, the egress port preemption slice association record, the link preemption slice association record, and the timing preemption slice association record. The network transmission time slot allocation mapping relationship is used to characterize the network egress port, candidate forwarding link, and scheduling time slice corresponding to the target priority packet and the controlled priority fragmented packet data; the master clock time synchronization reference is used to provide a unified time base in the distributed network; the calibrated time synchronization reference is used to make the gating scheduling time base of the time-aware shaping algorithm engine correspond to the master clock time synchronization reference; the time gating list is used to form port gating action records for network egress ports; the hop-by-hop gating unfolding record and the hop-by-hop gating continuation record are used to form gating continuation relationships in multi-hop forwarding; the egress port preemption slice association record, the link preemption slice association record, and the time-sequence preemption slice association record are used to retain the interleaving distribution relationship in the preemptive heterogeneous packet slice association carrier in the distributed time-sensitive network scheduling gating sequence, so that the distributed time-sensitive network scheduling gating sequence can serve as the gating scheduling basis for deterministic low-latency forwarding of the original communication service flow.

[0111] Optionally, a time-gated list is generated based on the calibrated time synchronization reference and the network transmission time slot allocation mapping relationship, thereby constructing a distributed time-sensitive network scheduling gating sequence for the original communication service flow, including: Based on the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, the feature parameters of the time gating list are mapped to the gating switch logic in the time-aware shaping algorithm engine to obtain the feature parameter mapping result. The action timing is arranged according to the feature parameter mapping results to determine the opening and closing action status of the network egress port in each scheduling cycle. The opening and closing action states are compiled into the time gating list; Based on the time-gating list, the forwarding timing of each message slice in the preemptive heterogeneous message slice association carrier is time-series scheduled to construct a distributed time-sensitive networking scheduling gating sequence for the original communication service flow.

[0112] Preferably, the specific implementation process of step 4 is as follows: The time gating list orchestration module in the time-aware shaping algorithm engine receives the calibrated time synchronization reference output by the gating scheduling time base calibration module, and receives the network transmission time slot allocation mapping relationship output by the time slot mapping generation module; wherein, the calibrated time synchronization reference is used to characterize the gating opening time and gating closing time that can be identified by each network switching node under the same master clock time synchronization reference, and the network transmission time slot allocation mapping relationship is used to characterize the network egress port, candidate forwarding link, and scheduling time slot corresponding to the target priority packet and the controlled priority fragment packet data, respectively. The time gating list orchestration module first reads the gating opening time and gating closing time in the calibrated time synchronization reference and the network egress port, candidate forwarding link, and scheduling time slot in the network transmission time slot allocation mapping relationship to form a gating mapping input record. The gating mapping input record continues to serve as the feature parameter mapping object of the gating switch logic, so that the subsequently generated time gating list can be simultaneously constrained by the alignment relationship between the calibrated time synchronization reference and the master clock time synchronization reference, as well as the network transmission time slot allocation mapping relationship.

[0113] Preferably, in step 4, when the time-gated list orchestration module performs feature parameter mapping on the gating switch logic within the time-aware shaping algorithm engine, it first reads the pre-configured priority queue gating status, gating start time, gating stop time, queue rotation order, and gating protection boundary in the gating switch logic. The priority queue gating status expresses the state of different priority queues in the network egress port that are allowed to open or should be closed within the scheduling time slice. The queue rotation order expresses the access sequence of different priority queues within consecutive scheduling time slices. The gating protection boundary expresses the switching interval that needs to be maintained between the insertion transmission of the target priority packet and the recovery transmission of the controlled priority fragmented packet data. Subsequently, the time-gated list orchestration module maps the priority queue gating status, the gating start time, the gating stop time, the queue rotation order, and the gating protection boundary to the gating mapping input record to obtain the feature parameter mapping result. The feature parameter mapping results are used to explain which priority queues should be opened, which priority queues should be closed, and the switching positions between different priority queues for each network egress port in each scheduling time slice.

[0114] Preferably, in the specific technical implementation of step 4, the feature parameter mapping results include a target priority packet gating parameter mapping record, a controlled priority fragmented packet data gating parameter mapping record, and a queue switching parameter mapping record. The target priority packet gating parameter mapping record originates from the target priority packet, the egress port preemption slice association record, the link preemption slice association record, and the timing preemption slice association record, and is used to express the gating start time, gating close time, and priority queue gating status of the target priority packet entering the corresponding network egress port. The controlled priority fragmented packet data gating parameter mapping record originates from the controlled priority fragmented packet data, the egress port preemption slice association record, the link preemption slice association record, and the timing preemption slice association record, and is used to express the gating start time, gating close time, and priority queue gating status of the controlled priority fragmented packet data resuming transmission after the target priority packet completes insertion transmission. The queue switching parameter mapping record originates from the gating protection boundary and is used to express the opening and closing switching relationship between the priority queue corresponding to the target priority message and the priority queue corresponding to the controlled priority fragmented message data. The target priority message gating parameter mapping record, the controlled priority fragmented message data gating parameter mapping record, and the queue switching parameter mapping record together constitute the feature parameter mapping result and continue to participate in action timing orchestration.

[0115] Preferably, in step 4, when arranging the action timing based on the feature parameter mapping results, the time-gating list arrangement module first merges the target priority packet gating parameter mapping record, the controlled priority fragmented packet data gating parameter mapping record, and the queue switching parameter mapping record according to the same network egress port to form a port gating mapping record. The port gating mapping record is used to express the sequential relationship of target priority packet insertion transmission, queue switching, and controlled priority fragmented packet data recovery transmission within the same network egress port. Subsequently, the time-gating list arrangement module arranges the port gating mapping records sequentially according to the order of scheduling time slices to determine the opening and closing action status of the network egress port within each preset scheduling period. The opening and closing action status includes the priority queue opening action status corresponding to the target priority packet, the priority queue closing action status corresponding to the controlled priority fragmented packet data, the priority queue recovery opening action status corresponding to the controlled priority fragmented packet data, and the network egress port switching action status, ensuring that the same network egress port will not simultaneously open for transmission to conflicting priority queues within the same preset scheduling period.

[0116] Preferably, in step 4, after determining the opening / closing action state, the time gating list arrangement module compiles the opening / closing action state into the time gating list. Specifically, the time gating list arrangement module first merges the opening / closing action states according to network egress ports to form port gating action records; then, it arranges the port gating action records according to scheduling time slices to form preset scheduling cycle gating action records; subsequently, it binds the preset scheduling cycle gating action records with the gating opening time and gating closing time in the calibrated time synchronization reference to form the time gating list. The time gating list includes the priority queue gating state, gating opening time, gating closing time, queue rotation order, and gating protection boundary corresponding to each network egress port in each scheduling time slice. Thus, the time gating list does not record scheduling time slices separately, but rather transforms the feature parameter mapping results and the opening / closing action states into gating scheduling content that can directly constrain the opening and closing actions of network egress ports.

[0117] Preferably, in step 4, the time-gated list orchestration module performs a gating consistency check on the time-gated list during compilation. The gating consistency check first verifies whether the priority queue opening action state corresponding to the target priority packet in the time-gated list corresponds to the target priority packet gating parameter mapping record. Then, it verifies whether the priority queue recovery opening action state corresponding to the controlled priority fragment packet data in the time-gated list corresponds to the controlled priority fragment packet data gating parameter mapping record. Finally, it verifies whether the network egress port switching action state in the time-gated list corresponds to the queue switching parameter mapping record. The time-gated list that passes the gating consistency check is then passed to the distributed gating sequence construction module; the time-gated list that fails the gating consistency check returns the feature parameter mapping result for remapping to prevent the opening / closing action states lacking the preemptive heterogeneous packet slice association carrier from entering the distributed time-sensitive network scheduling gating sequence.

[0118] Preferably, in step 4, when scheduling the forwarding timing of each packet slice in the preemptive heterogeneous packet slice association carrier according to the time-gating list, the distributed gating sequence construction module first reads the target priority packet, the controlled priority fragment packet data, the egress port preemptive slice association record, the link preemptive slice association record, and the time-series preemptive slice association record from the preemptive heterogeneous packet slice association carrier, and matches the above contents with the time-gating list to form a packet slice forwarding timing record. The packet slice forwarding timing record is used to express the insertion and transmission timing of the target priority packet at the corresponding network egress port, and the recovery and transmission timing of the controlled priority fragment packet data at the corresponding network egress port. The packet slice forwarding timing record continues to serve as the input for the distributed gating sequence construction module to perform hop-by-hop expansion, enabling the interleaving distribution relationship in the preemptive heterogeneous packet slice association carrier to enter the generation process of the hop-by-hop gating expansion record and the hop-by-hop gating continuation record.

[0119] Preferably, when the distributed gating sequence construction module performs hop-by-hop expansion based on the packet slice forwarding timing record, it first determines the candidate forwarding links where the target priority packet and the controlled priority fragment packet data are located according to the link preemption slice association record. Then, according to the forwarding order of the network switching nodes in the candidate forwarding links, it allocates the port gating action records in the time gating list to each network switching node to form a hop-by-hop gating expansion record. The forwarding order of the network switching nodes comes from the candidate forwarding links in the link preemption slice association record. The hop-by-hop gating expansion record is used to express the order in which each network switching node performs gating forwarding of the target priority packet and the controlled priority fragment packet data within the corresponding scheduling time slice. Subsequently, the distributed gating sequence construction module performs time slot continuation processing on the gating opening time and gating closing time of adjacent network switching nodes in the hop-by-hop gating expansion record according to the calibrated time synchronization reference to form a hop-by-hop gating continuation record. The hop-by-hop gated continuation record is used to express the timing relationship of the subsequent network switching node continuing to forward the target priority message or the controlled priority fragment message data according to the adjacent scheduling time slices after the previous network switching node has completed receiving or forwarding.

[0120] Preferably, in step 4, when constructing the distributed time-sensitive network scheduling gating sequence for the original communication service flow, the distributed gating sequence construction module associates the time gating list, the packet slice forwarding timing record, the hop-by-hop gating expansion record, and the hop-by-hop gating continuation record to form a distributed gating scheduling record. The distributed gating scheduling record is used to express the network egress port opening and closing order, candidate forwarding link continuation relationship, and scheduling time slice continuation relationship of the original communication service flow among multiple network switching nodes. Subsequently, the distributed gating sequence construction module writes the egress port preemption slice association record, the link preemption slice association record, and the timing preemption slice association record into the distributed gating scheduling record to construct the distributed time-sensitive network scheduling gating sequence. The distributed time-sensitive network scheduling gating sequence includes both the network egress port opening and closing action status formed by the time gating list and the interleaving distribution relationship between target priority packet insertion transmission and controlled priority fragmented packet data recovery transmission formed by the preemptive heterogeneous packet slice association carrier.

[0121] Preferably, in one scenario, when step 4 is specifically implemented, if the target priority message corresponding to the cooperative control message needs to be inserted for transmission, and the controlled priority fragment message data corresponding to the track status return message is waiting to resume transmission, then the time gating list arrangement module first performs feature parameter mapping on the priority queue gating state corresponding to the target priority message and the priority queue gating state corresponding to the controlled priority fragment message data according to the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, so as to obtain the feature parameter mapping result; then, according to the feature parameter mapping result, the action timing arrangement is performed to determine the priority queue opening action state corresponding to the target priority message corresponding to the cooperative control message, the priority queue closing action state corresponding to the controlled priority fragment message data corresponding to the track status return message, and the priority queue resuming opening action state corresponding to the controlled priority fragment message data corresponding to the track status return message; subsequently, the above opening and closing action states are compiled into the time gating list. The distributed gating sequence construction module then performs time-series scheduling on the target priority message corresponding to the cooperative control message and the controlled priority fragmented message data corresponding to the track status return message, based on the time gating list, to form the message slice forwarding timing record. Based on the message slice forwarding timing record, it forms the hop-by-hop gating deployment record and the hop-by-hop gating continuation record, thereby forming the distributed time-sensitive networking scheduling gating sequence. Thus, the target priority message corresponding to the cooperative control message can enter the network egress port in the corresponding scheduling time slice, and the controlled priority fragmented message data corresponding to the track status return message can resume transmission in subsequent scheduling time slices.

[0122] Preferably, the distributed time-sensitive networking scheduling gating sequence finally formed in step 4 includes the feature parameter mapping result, the opening and closing action status, the time gating list, the packet slice forwarding timing record, the hop-by-hop gating deployment record, the hop-by-hop gating continuation record, the egress port preemption slice association record, the link preemption slice association record, and the timing preemption slice association record. The feature parameter mapping result is used to explain how the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship are mapped to the gating switch logic; the opening and closing action status is used to explain the priority queue opening action status corresponding to the target priority packet at the network egress port in each preset scheduling period, the priority queue closing action status corresponding to the controlled priority fragment packet data, the priority queue reopening action status corresponding to the controlled priority fragment packet data, and the network egress port switching action status; the time gating list is used to convert the opening and closing action status into gating scheduling content; the packet slice forwarding timing record is used for... This document describes the forwarding timing of each packet slice in the preemptive heterogeneous packet slice association carrier; the hop-by-hop gating expansion record and the hop-by-hop gating continuation record are used to describe the gating continuation relationship in multi-hop forwarding; the egress port preemptive slice association record, the link preemptive slice association record, and the time-series preemptive slice association record are used to retain the interleaving distribution relationship between the target priority packet and the controlled priority fragmented packet data in the distributed time-sensitive networking scheduling gating sequence, so that the distributed time-sensitive networking scheduling gating sequence can serve as the gating scheduling basis for deterministic low-latency forwarding of the original communication service flow.

[0123] like Figure 2 As shown, this is a low-latency networking device that integrates dynamic priority preemption and time-sensitive networking according to an embodiment of this application, which includes: The global attribute situation construction module is used to obtain the original communication service flow to be networked through the physical inlet port of the network switching node and perform service attribute parsing based on the metadata of the message header to construct the global attribute distribution situation of the scheduled service to represent the distribution situation of different service categories in the network spatiotemporal domain. The dynamic priority scheduling sequence construction module is used to establish a non-linear coupling relationship between the urgency of the service and the occupancy of the link resources based on the distribution of the global attributes of the service to be scheduled, so as to generate a dynamic scheduling weight table for the service flow and construct a dynamic priority preemption scheduling sequence for the service flow to represent the order of processing of each service flow within a preset scheduling period. The preemptive message slice association carrier construction module is used to monitor whether there is a target priority message performing channel preemption processing on the controlled priority message in the dynamic priority preemption scheduling sequence of the service flow. If so, a preemptive heterogeneous message slice association carrier is constructed to characterize the interleaving distribution characteristics in the physical link. The gate sequence construction module is used to input the preemptive heterogeneous message slice association carrier into the time-aware shaping algorithm engine built based on the time-sensitive network protocol standard, generate network transmission time slot allocation mapping relationship to generate a time gating list, and construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow accordingly.

[0124] like Figure 3 As shown, an electronic device according to an embodiment of this application includes a processor and a memory; The memory is used to store computer programs; When the processor executes the program stored in the memory, it implements the steps of the low-latency networking method as described in this application.

[0125] Figures 2-3 For an exemplary description, please refer to the above. Figure 1 This will not be elaborated upon here.

Claims

1. A low-latency networking method integrating dynamic priority preemption and time-sensitive networking, characterized in that, include: Step 1: Obtain the original communication service flow to be networked through the physical ingress port of the network switching node and perform service attribute parsing based on the metadata of the message header to construct the global attribute distribution pattern of the scheduled service to be represented in the spatiotemporal domain of the network. Step 2: Based on the distribution of the global attributes of the services to be scheduled, establish a non-linear coupling relationship between the urgency of the services and the occupancy of the link resources, so as to generate a dynamic scheduling weight table for the service flow and construct a dynamic priority preemption scheduling sequence for the service flow to represent the order of processing of each service flow within a preset scheduling period. Step 3: Monitor whether there is a target priority message performing channel preemption processing on the controlled priority message in the dynamic priority preemption scheduling sequence of the service flow. If so, construct a preemptive heterogeneous message slice association carrier to characterize the interleaving distribution characteristics in the physical link. Step 4: Input the preemptive heterogeneous message slice association carrier into the time-aware shaping algorithm engine built based on the time-sensitive network protocol standard to generate a network transmission time slot allocation mapping relationship to generate a time gating list and construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow.

2. The low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 1, characterized in that, Step 1 includes: The original communication service flow is parsed based on the metadata of the message header to extract the basic service attribute feature group, which includes message arrival time, traffic burst tolerance and periodic interaction frequency. Based on the aforementioned service basic attribute feature group, a service category clustering identification based on latency sensitivity is performed to map the original communication service flow into time-sensitive service flow and non-time-sensitive service flow respectively; Based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in network bandwidth resources, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution status of different service categories in the network spatiotemporal domain.

3. The low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 2, characterized in that, Based on the distribution of time-sensitive and non-time-sensitive service flows in network bandwidth resources, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution of different service categories in the network's spatiotemporal domain, including: Based on the distribution status of the time-sensitive service flows and the non-time-sensitive service flows in the network bandwidth resources, the corresponding traffic packet length distribution is parameterized to obtain the traffic packet length distribution parameters; Determine the transmission time interval for network data transmission between the time-sensitive service flow and the non-time-sensitive service flow, and parameterize and characterize it to generate a transmission time interval parameter; Based on the traffic packet length distribution parameter and the transmission time interval parameter, a global attribute distribution pattern of the scheduled services is constructed to characterize the distribution pattern of different service categories in the network spatiotemporal domain.

4. The low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 1, characterized in that, Step 2 includes; Based on the dynamic scheduling weight table of the business flow, the feature parameters of each dimension in the global attribute distribution of the business to be scheduled are weighted to obtain the weighted feature parameters of each dimension. The weighted feature parameters of each dimension are mapped to the multidimensional network resource topology boundary space to generate a load drift evolution trajectory that characterizes the current service load deviating from the deterministic transmission range of the network. Based on the load drift evolution trajectory, a dynamic priority preemption scheduling sequence for service flows is constructed to characterize the order in which each service flow is processed within a preset scheduling period.

5. The low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 4, characterized in that, Based on the load drift evolution trajectory, a dynamic priority preemption scheduling sequence for service flows is constructed to characterize the sequential processing order of each service flow within a preset scheduling period, including: The load drift evolution trajectory is subjected to spatiotemporal dimension projection processing to obtain spatiotemporal projection results; Based on the spatiotemporal projection results, load integration is performed to calculate the multidimensional composite load deviation degree used to characterize the intensity of flow fluctuations. Dynamic priority evaluation is performed based on the multidimensional composite working condition load deviation and the global attribute distribution of the services to be scheduled, so as to construct a dynamic priority preemption scheduling sequence for services to represent the order of processing of each service flow within a preset scheduling period.

6. The low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 1, characterized in that, Step 3 includes: When it is detected that a target priority message has entered the scheduling ready state and the network egress port is transmitting a controlled priority message that has a channel occupation conflict with the target priority message, the control egress queue manager issues a forced interrupt command to the controlled priority message and obtains the interrupt command execution result. When the execution result of the interruption instruction indicates that the controlled priority message is interrupted, the transmission position status of the controlled priority message is saved to obtain the status saving information of the interrupted message, so as to construct a preemptive heterogeneous message slice association carrier for characterizing the interleaving distribution characteristics in the physical link.

7. The low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 6, characterized in that, Construct a preemptive heterogeneous packet slice association carrier to characterize the interleaving distribution characteristics in physical links, including: Based on the state storage information, the interrupted controlled priority message is subjected to data fragmentation and marking processing based on the minimum slice granularity to generate controlled priority fragmented message data. Based on the controlled priority fragmented message data and the target priority message that triggers the channel preemption process, a preemptive heterogeneous message slice association carrier for the interleaving distribution characteristics in the physical link is constructed.

8. The low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 1, characterized in that, Step 4 includes: The preemptive heterogeneous packet slice association carrier is input into a time-aware shaping algorithm engine built on the time-sensitive network protocol standard to perform mapping and matching calculations between packet slices and network bandwidth resources, so as to generate a network transmission time slot allocation mapping relationship. By calling the master clock in the distributed network, the time synchronization reference provided by the master clock is obtained and determined as the master clock time synchronization reference; Based on the master clock time synchronization reference, the gated scheduling time base of the time-aware shaping algorithm engine is aligned and calibrated to obtain the calibrated time synchronization reference; Based on the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, a time gating list is generated, and then a distributed time-sensitive network scheduling gating sequence for the original communication service flow is constructed.

9. A low-latency networking method integrating dynamic priority preemption and time-sensitive networking as described in claim 8, characterized in that, Based on the calibrated time synchronization reference and the network transmission time slot allocation mapping relationship, a time gating list is generated, and a distributed time-sensitive network scheduling gating sequence for the original communication service flow is constructed, including: Based on the calibrated time synchronization benchmark and the network transmission time slot allocation mapping relationship, the feature parameters of the time gating list are mapped to the gating switch logic in the time-aware shaping algorithm engine to obtain the feature parameter mapping result. The action timing is arranged according to the feature parameter mapping results to determine the opening and closing action status of the network egress port in each scheduling cycle. The opening and closing action states are compiled into the time gating list; Based on the time-gating list, the forwarding timing of each message slice in the preemptive heterogeneous message slice association carrier is time-series scheduled to construct a distributed time-sensitive networking scheduling gating sequence for the original communication service flow.

10. A low-latency networking device integrating dynamic priority preemption and time-sensitive networking, characterized in that, include: The global attribute situation construction module is used to obtain the original communication service flow to be networked through the physical inlet port of the network switching node and perform service attribute parsing based on the metadata of the message header to construct the global attribute distribution situation of the scheduled service to represent the distribution situation of different service categories in the network spatiotemporal domain. The dynamic priority scheduling sequence construction module is used to establish a non-linear coupling relationship between the urgency of the service and the occupancy of the link resources based on the distribution of the global attributes of the service to be scheduled, so as to generate a dynamic scheduling weight table for the service flow and construct a dynamic priority preemption scheduling sequence for the service flow to represent the order of processing of each service flow within a preset scheduling period. The preemptive message slice association carrier construction module is used to monitor whether there is a target priority message performing channel preemption processing on the controlled priority message in the dynamic priority preemption scheduling sequence of the service flow. If so, a preemptive heterogeneous message slice association carrier is constructed to characterize the interleaving distribution characteristics in the physical link. The gate sequence construction module is used to input the preemptive heterogeneous message slice association carrier into the time-aware shaping algorithm engine built based on the time-sensitive network protocol standard, generate network transmission time slot allocation mapping relationship to generate a time gating list, and construct a distributed time-sensitive network scheduling gating sequence for the original communication service flow accordingly.

Images