Method and system for inter-domain communication in access network based on cross-domain QoS mapping and PSFP linkage

Abstract

This invention discloses a method and system for access network midhaul communication based on cross-domain QoS mapping and PSFP linkage. The method includes: performing system initialization and data filling processing at the CU-side TSN end to obtain encapsulated data frames; performing cross-layer semantic mapping on the encapsulated data frames and prioritizing them through flow gating normal state monitoring, flow metering normal state monitoring, and queue normal state monitoring to obtain data flow queues of different priorities; performing gating scheduling and credit-based shaping on the data flow queues of different priorities, and sending them to the DU side for data processing to realize access network midhaul communication. This invention can effectively improve the priority transmission capability of time-sensitive services and enhance the isolation between ordinary services and critical services. As an access network midhaul communication method and system based on cross-domain QoS mapping and PSFP linkage, this invention can be widely applied in the field of network communication technology.

Description

Technical Field

[0001] This invention relates to the field of network communication technology, and in particular to an access network midhaul communication method and system based on cross-domain QoS mapping and PSFP linkage. Background Technology

[0002] As 5G networks evolve towards cloudification, virtualization, and distributed architecture, the access network architecture based on CU / DU separation has become an important development trend in mobile communication systems. Under this architecture, user plane service data needs to be transmitted between the CU and DU via the midhaul network, which places higher demands on the bandwidth guarantee, latency control, jitter suppression, and service isolation capabilities of the midhaul link. Especially in scenarios such as industrial internet, remote control, spaceborne base stations, and space-ground converged networks, some data streams have significant periodicity, low latency, and high reliability transmission requirements. Traditional best-effort Ethernet forwarding methods are insufficient to simultaneously meet the differentiated carrying requirements of critical data streams and ordinary data streams.

[0003] Time-Sensitive Networking (TSN) possesses capabilities such as traffic shaping, priority scheduling, gating control, and deterministic transmission, providing new technical means for high-reliability, low-jitter transport of critical services in China Telecom networks. However, there are differences in protocol layers and control methods between QoS semantics in 5G systems and priority marking, flow policing, and queue scheduling mechanisms in TSN networks, making it difficult for the two types of networks to work directly together. Summary of the Invention

[0004] To address the aforementioned technical problems, the present invention aims to provide an access network midhaul communication method and system based on cross-domain QoS mapping and PSFP linkage, which can effectively improve the priority transmission capability of time-sensitive services and enhance the isolation between ordinary services and critical services.

[0005] The first technical solution adopted in this invention is: an access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage, comprising the following steps: Perform system initialization and CU-side TSN terminal system data filling processing to obtain the encapsulated data frame; Perform cross-layer semantic mapping on the encapsulated data frames, and mark priorities by monitoring the normal state of flow gating, the normal state of flow metering, and the normal state of queues to obtain data flow queues with different priorities. Gating and credit-based shaping are performed on data stream queues of different priorities, and the data is sent to the DU side for data processing to realize mid-transmission communication in the access network.

[0006] Furthermore, the system initialization step specifically includes: Data interaction is performed based on the centralized unit control plane, centralized unit user plane and distributed unit to confirm the TEID, UDP port number, source IP address information and destination IP address information required for tunnel encapsulation; Clear the remaining amount in the PDCP buffer queue and the remaining amount in the RLC queue; Set threshold parameters, which include normal flow metering status, normal flow gating status, and normal queue status; The scores of the 5QI parameters are normalized using an algorithm to obtain the mapping table entries.

[0007] Furthermore, the step of data filling and processing at the CU-side TSN terminal system to obtain the encapsulated data frame specifically includes: Acquire PDCP data frames, and perform statistics on the buffer status of PDCP and RLC queues to generate corresponding scheduling instructions. The buffer status includes queue balance, buffer occupancy, and amount of data to be sent. According to the scheduling instruction, the corresponding length of the data frame is read and sent to the F1-U protocol encapsulation entity. Combined with TEID and UDP port number, the data frame is encapsulated using the F1-U user plane protocol to obtain the encapsulated data frame.

[0008] Furthermore, the step of performing cross-layer semantic mapping on the encapsulated data frames and prioritizing them through flow gating normal state monitoring, flow metering normal state monitoring, and queue normal state monitoring to obtain data flow queues of different priorities specifically includes: Based on the mapping table entries, cross-layer semantic mapping is performed on the encapsulated data frames to convert the data streams corresponding to 5G-side QoS semantics into TSN-side scheduling semantics and generate corresponding priority tags. Based on threshold parameters, the transformed data stream is sequentially monitored for normal status of flow gating, normal status of flow metering, and normal status of queue. The transformed data stream is then prioritized to obtain data stream queues with different priorities.

[0009] Furthermore, the flow gating normal status monitoring specifically includes a level 1 gating warning status, a level 2 gating warning status, and a level 3 serious violation status, wherein: The first-level gate control early warning status indicates that the arrival time of the data stream is within the expected gate control window. If there is a phase error greater than 10% of the gate control window length or the offset of the data stream arrival time relative to the gate control window boundary does not exceed one protection band length, a gate control violation is recorded and the cumulative count is recorded. The Level 2 gate control early warning status indicates that if any one of the following three conditions is met: the arrival time of a data stream exceeds the allowed gate control window boundary, the number of consecutive gate control violations is greater than or equal to a preset coefficient, or the number of cumulative gate control violations within a preset business cycle is greater than or equal to a preset coefficient, the data stream is marked as 1 and written into the exception cache chain. The Level 3 serious violation status indicates that if there are two consecutive preset business cycles with incorrect windows or the incorrect window offset exceeds the width of a complete gate control window, the system will switch to the backup gate control configuration file and perform temporary priority downgrading processing on the data stream according to the current business level and resource status.

[0010] Furthermore, the flow metering normal status monitoring specifically includes a first-level metering early warning status, a second-level metering early warning status, and a third-level forced degradation status, wherein: The first-level metering early warning status indicates that if either the average measurement rate meets the preset threshold or the current burst length is greater than 90% of the promised burst length, an over-limit count will be recorded for the data stream. The Level 2 metering early warning status indicates that if any one of the following four conditions is met within a consecutive preset statistical period: the preset average measurement rate is met within a consecutive preset statistical period, the preset peak measurement rate is greater than the peak information rate, the current burst length meets the preset conditions, or the consecutive over-limit count is greater than or equal to a preset number, priority downgrade processing will be performed on the data stream. The three-level forced degradation state indicates that any one of the following three conditions occurs consecutively: the data stream meets the preset average measurement rate requirement for a preset number of consecutive statistical periods, the current burst length meets the preset length requirement, or the peak rate exceeds the PIR and exceeds the preset threshold. In such cases, forced degradation processing is performed on the data stream, and it is switched to another non-primary protection queue.

[0011] Furthermore, the normal status monitoring of the queue specifically includes a first-level queue warning status. The first-level queue warning status indicates that if the current occupied queue depth meets a preset range or the average queuing waiting time reaches a preset proportion of the business latency budget, queue observation will be initiated and newly entering low-priority queues will be restricted.

[0012] Furthermore, the step of gating and scheduling data stream queues of different priorities and performing credit-based shaping, and then sending the data to the DU side for data processing to realize midhaul communication in the access network, specifically includes: The transmission time slots of data stream queues with different priorities are controlled, and priority is given to transmission within a preset time window. A credit-based shaping mechanism is introduced to constrain the transmission rate and burst behavior of the data streams before transmission to the DU-side TSN terminal system. Based on the DU-side TSN terminal system, the data stream is processed by queue reverse mapping and delivered to the corresponding RLC queue; Based on RLC queues, a cache management method based on BD linked lists is adopted to organize and cache the data streams of different bearer queues in DDR storage space to realize midhaul communication in the access network.

[0013] The second technical solution adopted in this invention is: an access network midhaul communication system based on cross-domain QoS mapping and PSFP linkage, comprising: The first module is used to perform system initialization and data filling processing on the CU-side TSN terminal system to obtain the encapsulated data frame. The second module is used to perform cross-layer semantic mapping on the encapsulated data frame and to mark the priority of data stream queues by monitoring the normal state of flow gating, the normal state of flow metering, and the normal state of queues. The third module is used to perform gating scheduling and credit-based shaping of data stream queues with different priorities, and send the data to the DU side for data processing to realize mid-pass communication in the access network.

[0014] The beneficial effects of the method and system of this invention are as follows: This invention obtains encapsulated data frames by performing system initialization and TSN terminal system data filling processing on the CU side; then, cross-layer semantic mapping is performed on the encapsulated data frames, and priority marking is performed through flow gating normal state monitoring, flow metering normal state monitoring, and queue normal state monitoring to obtain data flow queues with different priorities. By establishing a cross-layer mapping relationship between 5G service semantics and TSN scheduling semantics, service attributes such as 5QI, QFI, and DRB can be uniformly associated with PCP, VID, flow filtering, flow metering, flow gating, and queue scheduling parameters, thereby improving the priority transmission capability of key service flows in the midhaul link; finally, gating scheduling and credit-based shaping are performed on the data flow queues with different priorities, and the data is sent to the DU side for data processing to realize access network midhaul communication. This enables fine-grained identification, supervision, and scheduling of periodic time-sensitive services, and dynamically adjusts the mapping and scheduling strategy through state feedback and closed-loop backpressure mechanism, enhancing the stability and resource utilization efficiency of the system under multi-service concurrency conditions. Attached Figure Description

[0015] Figure 1 This is a flowchart of the steps of the access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage of the present invention; Figure 2 This is a structural block diagram of the access network midhaul communication system based on cross-domain QoS mapping and PSFP linkage of the present invention; Figure 3 This is a schematic diagram of a deterministic optimization architecture for a distributed spaceborne base station midhaul network provided in a specific embodiment of the present invention.

[0016] Figure reference numerals: 1. PDCP cache entity; 2. F1-U protocol encapsulation entity; 3. Cross-domain mapping and status feedback entity; 4. Gated scheduling and credit-based shaping entity; 5. E1 logical interface entity; 6. F1-C logical interface; 7. Queue reverse mapping entity; 8. DDR queue cache entity. Detailed Implementation

[0017] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments. The step numbers in the following embodiments are only for ease of explanation and do not limit the order of the steps. The execution order of each step in the embodiments can be adapted according to the understanding of those skilled in the art.

[0018] As 5G mobile communication networks evolve towards cloudification, virtualization, and split access networks, a bearer architecture based on the separation of centralized unit (CU) and distributed unit (DU) has become an important implementation method in existing systems. In this type of architecture, user plane services are typically transmitted between the CU and DU via the F1-U interface, and the bearer network mostly uses packet switching based on Ethernet and IP protocols to forward service data. Existing technologies typically encapsulate service data using protocols such as GTP-U, UDP, and IP, and utilize mechanisms such as priority queues and DSCP to achieve basic traffic classification and differentiated bearer capabilities. For example: 1) Differentiated forwarding strategy based on priority queues.

[0019] In the bearer network, service messages are divided into different transmission queues according to their priority markings, and high-priority services are forwarded first through strict priority scheduling or weighted scheduling to reduce the queuing delay of critical services.

[0020] 2) Priority classification strategy based on DSCP.

[0021] By setting a DSCP field at the IP layer, different types of data streams are marked, and network nodes or border nodes perform classification, queuing, and priority forwarding based on the marked data.

[0022] 3) Traffic shaping strategies based on bandwidth reservation or rate limiting.

[0023] The impact of sudden traffic surges on critical services can be mitigated by allocating minimum bandwidth to high-priority services, limiting the rate of low-priority services, or using methods such as token buckets and leaky buckets to shape the data flow.

[0024] However, existing mid-haul transport solutions under the CU / DU separation architecture primarily rely on Ethernet and IP protocols to complete F1-U user plane data transmission. While they can achieve basic service classification and differentiated forwarding through GTP-U, UDP, IP encapsulation, priority queues, and DSCP, they are essentially still best-effort-oriented transport methods, lacking deterministic priority guarantee mechanisms for periodic, time-sensitive services. Especially in scenarios such as the Industrial Internet, remote control, spaceborne base stations, and space-ground converged networks, some data streams have higher requirements for transmission latency, jitter, and reliability. Existing mid-haul solutions typically only achieve coarse-grained priority differentiation, making it difficult to finely identify, monitor, and control the timing of critical data streams. This results in critical data streams being susceptible to competition from ordinary data streams during concurrent multi-service transmission, making it difficult to obtain stable priority transmission guarantees.

[0025] Furthermore, existing technologies lack a unified cross-domain association mechanism between service-bearing semantics and Ethernet-side scheduling semantics. Current solutions typically classify services solely based on DSCP or simple priority queues, failing to further map service semantics such as 5QI, QFI, and DRB_ID in 5G systems to deterministic scheduling parameters on the TSN side, such as PCP, VID, flow filtering, flow metering, and flow gating. Consequently, they cannot fully utilize the TSN's capabilities in priority guarantee, bandwidth shaping, and gating scheduling. Especially in midhaul scenarios characterized by high latency, limited resources, and mixed service transmission, existing solutions struggle to simultaneously meet the priority transmission requirements of critical data flows and the effective carrying requirements of ordinary data flows, exhibiting insufficient service isolation capabilities and limited deterministic guarantee capabilities.

[0026] Based on this, this embodiment specifically relates to a method and apparatus for cross-domain mapping, demapping, and deterministic transmission of data streams in CU / DU midhaul scenarios. More specifically, it relates to associating QoS attributes in 5G systems with priority marking, flow policing mechanisms, and queue scheduling mechanisms in time-sensitive networks to achieve collaborative bearing and deterministic transmission of data streams between heterogeneous networks. This is particularly suitable for midhaul scenarios with high latency and limited resources in spaceborne base stations or space-ground converged networks.

[0027] First, it should be noted that, in order to improve the deterministic carrying capacity of distributed spaceborne base station midhaul networks for time-sensitive services, this embodiment of the invention designs a deterministic optimization architecture for midhaul networks that incorporates a TSN mechanism and cross-layer semantic mapping, such as... Figure 3As shown, PDCP cache entity 1 is used to classify and cache PDCP PDUs according to 5G service semantics and maintain the cache status information of each service queue. The F1-U protocol encapsulation entity 2 is used to encapsulate the service data to be transmitted using the F1-U user plane protocol based on the bearer parameters and tunnel parameters issued by the control plane. The cross-domain mapping and status feedback entity 3 is used to generate cross-domain flow object (CDFO) entries based on 5G-side service semantic information and establish a mapping relationship between 5G service semantics and TSN scheduling semantics. This entity is also used to monitor the operating status during service processing, including cache level status, gating violation status, and Meter over-limit status, and to report feedback information to the policy generation end when the relevant status reaches a preset threshold. The gating scheduling and credit-based shaping entity 4 is used at the transmission exit of the TSN end system to perform gating scheduling on different priority queues according to the pre-configured gating control list (GCL), and to perform rate constraints and traffic shaping on specific service flows through a credit-based shaping mechanism. The E1 logical interface entity 5 is used to implement logical control interaction between CU-CP and CU-UP. The F1-C logical interface entity 6 is used to implement control signaling interaction between CU-CP and DU to support midhaul bearer configuration, control parameter distribution, and related status information interaction. The queue demapping entity 7 is used to recover the corresponding 5G-side service semantics on the DU side based on the received TSN-side semantic information. The DDR queue cache entity 8 is used to organize and cache RLC queue data using a buffer descriptor BD linked list.

[0028] This embodiment presents a deterministic optimization architecture for midhaul networks using distributed spaceborne base stations. Its core lies in introducing a Time-Sensitive Networking (TSN) mechanism to address the high latency, resource constraints, and concurrent multi-service characteristics of spaceborne midhaul links. This provides low-latency, low-jitter, and highly reliable transmission guarantees for critical service flows. Based on CU / DU separation, the architecture deploys a TSN end system at the midhaul boundary. Through a cross-layer semantic mapping mechanism, service attributes such as 5QI, QFI, and DRB in the 5G system are mapped to PCP, VID, and scheduling parameters such as flow filtering, metering, and gating on the TSN side, thereby achieving the conversion from 5G bearer semantics to TSN deterministic scheduling semantics. Simultaneously, by combining priority queuing, flow policing, and gating shaping mechanisms, the priority transmission capability of time-sensitive services can be effectively improved, and the isolation between ordinary services and critical services can be enhanced.

[0029] Reference Figure 1 This invention provides an access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage, the method comprising the following steps: S100: Perform system initialization and CU-side TSN terminal system data filling processing to obtain the encapsulated data frame; S110, System initialization; Specifically, data interaction is performed based on the centralized unit control plane, centralized unit user plane, and distributed unit to confirm the TEID, UDP port number, source IP address information, and destination IP address information required for tunnel encapsulation; the remaining queue space in the PDCP buffer area and the RLC is cleared; threshold parameters are set, including normal state of flow metering, normal state of flow gating, and normal state of queues; and the 5QI parameters are normalized and scored to obtain mapping entries.

[0030] In this embodiment, it specifically includes: 1) Queue initialization: The control plane first provides the configuration parameters of each sub-layer. The CU-CP (centralized unit control plane) interacts with the CU-UP (centralized unit user plane) and DU (distributed unit) to confirm the TEID, UDP port number, source IP / destination IP address and other information required for tunnel encapsulation.

[0031] 2) Cache initialization: Clear the remaining amount of the PDCP cache queue to zero, initialize the DDR module in DU and clear the remaining amount of the RLC queue to zero.

[0032] 3) Set threshold parameters: Set the normal state of flow metering to ,in The average rate over a statistical period. To guarantee the information rate; set the normal state of flow gating to... ,in For phase error; set the normal queue state to Q_occ_ratio<60%, where Q_occ_ratio is the current occupancy / queue depth.

[0033] 4) Default mapping algorithm: The default mapping entries are obtained by normalizing the scores of the 5QI parameters.

[0034] The S120 and CU-side TSN terminal system data filling and processing are used to obtain the encapsulated data frame.

[0035] Specifically, PDCP data frames are acquired, and the buffer status of PDCP and RLC queues is statistically analyzed to generate corresponding scheduling instructions. The buffer status includes queue balance, buffer occupancy, and amount of data to be sent. Data frames of the corresponding length are read according to the scheduling instructions, sent to the F1-U protocol encapsulation entity, and F1-U user plane protocol encapsulation is performed on the data frames in combination with TEID and UDP port number to obtain encapsulated data frames.

[0036] In this embodiment, it specifically includes: 1) The PDCP cache entity receives PDCP data frames and performs statistics on the cache status of each service queue. The cache status includes at least the queue balance, cache occupancy, and amount of data to be sent. After the relevant status information is reported to the control plane, the control plane generates corresponding scheduling instructions or grant size information to control the reading process of cached data.

[0037] 2) The PDCP buffer entity reads the data frame of the corresponding length according to the scheduling instruction and sends it to the F1-U protocol encapsulation entity; the F1-U protocol encapsulation entity encapsulates the data frame with F1-U user plane protocol according to the tunnel parameters and header parameters obtained in the initialization phase.

[0038] 3) The encapsulated data frame is further sent to the cross-domain mapping and status feedback entity ③ to perform subsequent cross-domain semantic mapping, priority tag generation and status feedback processing.

[0039] S200: Perform cross-layer semantic mapping on the encapsulated data frame, and mark the priority of the data stream queues by monitoring the normal state of the flow gating, the normal state of the flow metering, and the normal state of the queue. S210. According to the mapping table entries, perform cross-layer semantic mapping on the encapsulated data frame, convert the data stream corresponding to the 5G side QoS semantics into the TSN side scheduling semantics, and generate the corresponding priority flag. In this embodiment, the system first performs cross-layer semantic mapping on the service flow based on the pre-configured default CDFO entries, converting the data flow corresponding to the 5G-side QoS semantics into the TSN-side scheduling semantics, and generating corresponding priority tags and regulatory parameters. After semantic conversion, the service flow sequentially passes through processing modules such as flow filtering, flow gating, and flow metering to achieve traffic regulation and priority scheduling in the initial state.

[0040] S220. Based on the threshold parameter, the converted data stream is monitored for normal status of flow gating, normal status of flow metering, and normal status of queue in sequence. The converted data stream is marked with priority to obtain data stream queues with different priorities.

[0041] Specifically, the normal state monitoring of flow gating includes Level 1 gating warning state, Level 2 gating warning state, and Level 3 serious violation state, among which: 1) The first-level gate control early warning status indicates that the arrival time of the data stream is within the expected gate control window. If there is a phase error greater than 10% of the gate control window length or the data stream arrival time offset from the gate control window boundary does not exceed one protection band length, a gate control violation is recorded and the cumulative count is recorded. 2) The second-level gate control early warning status indicates that if any one of the following three conditions is met: the arrival time of a data stream exceeds the allowed gate control window boundary, the number of consecutive gate control violations is greater than or equal to a preset coefficient, or the number of cumulative gate control violations within a preset business cycle is greater than or equal to a preset coefficient, the data stream is marked as 1 and written into the exception cache chain. 3) The Level 3 serious violation status means that if there is either a wrong window occurring for a consecutive preset number of business cycles or the wrong window offset exceeds the width of a complete gate control window, the system will switch to the backup gate control configuration file and perform temporary priority downgrade processing on the data stream according to the current business level and resource status.

[0042] In this embodiment, the normal state of the gating feedback mechanism has been configured and described during the system initialization process. The following focuses on the gating warning and violation handling states.

[0043] For Level 1 gated alert status: When the arrival time of a service flow is still within the expected gated window, but meets any of the following conditions, the system determines that it has entered Level 1 gated alert status: 1) Phase error Greater than 10% of the gated window length (WindowLength); 2) The arrival time offset relative to the gate window boundary does not exceed one guard band length.

[0044] In this state, the system records a cumulative gate violation count (GateViolationCnt), but allows the service flow to continue passing without triggering remapping. The phase error is also included. The calculation formula is: ; Expected window center The calculation formula is: ; For the Level 2 gating warning state, the system determines that a service flow has entered the Level 2 gating warning state when any of the following conditions are met: 1) Arrival time exceeds the allowed gated window boundary; 2) The consecutive gate violation count (ConsecutiveGateViolationCnt) is greater than or equal to 2; 3) The cumulative number of access control violations is greater than or equal to 3 in the most recent 5 business cycles.

[0045] In this state, the system marks the service flow as gate_violation=1 and writes it into the exception cache chain.

[0046] For Level 3 Critical Violation status, the system determines that a business flow has entered the Level 3 Critical Violation status when any of the following conditions are met: 1) Misalignment occurs for 3-5 consecutive business cycles; 2) The offset of the misaligned window exceeds the width of a complete gated window.

[0047] In this state, the system switches to the backup gate configuration file GateProfile_compensation and can perform temporary priority downgrade processing on the service flow if necessary, based on the current service level and resource status.

[0048] Furthermore, the normal status monitoring of flow metering specifically includes Level 1 metering early warning status, Level 2 metering early warning status, and Level 3 mandatory degradation status, among which: 1) The first-level metering early warning status means that if either the average measurement rate meets the preset threshold or the current burst length is greater than 90% of the promised burst length, an over-limit count will be recorded for the data stream. 2) The second-level metering early warning status means that if any one of the following four conditions is met within a consecutive preset statistical period: the preset average measurement rate is met within a consecutive preset statistical period, the preset peak measurement rate is greater than the peak information rate, the current burst length meets the preset conditions, or the consecutive over-limit count is greater than or equal to the preset number of times, priority downgrade processing will be performed on the data stream. 3) The three-level forced degradation state means that if any one of the following three conditions occurs consecutively for a preset number of statistical periods, the preset average measurement rate requirement is met, the current burst length meets the preset length requirement, or the peak rate exceeds the PIR, a forced degradation process is performed on the data stream, and it is switched to other non-primary protection queues.

[0049] The metering feedback mechanism has already been configured during the system initialization phase under normal conditions. The following focuses on the metering early warning and degradation processing status.

[0050] For Level 1 metering early warning status, the system determines that a business flow has entered Level 1 metering early warning status when any of the following conditions are met: 1) Average measurement rate satisfy: ; The current burst length (BurstLen) is greater than 90% of the committed burst length (CBS). Here, BurstLen represents the current burst length, and CBS represents the committed burst length.

[0051] In this state, the system records an over-limit count (MeterExceedCnt), but does not immediately perform remapping; only mild rate limiting or shaping measures can be initiated.

[0052] For the Level 2 metering early warning status, the system determines that the target queue has entered the Level 2 queue early warning status when any of the following conditions are met: 1) Meets the following criteria for 3 to 5 consecutive statistical periods: ; 2) Meets the following criteria for 2-3 consecutive statistical periods: ; 3) The current burst length satisfies: ; 4) The consecutive over-limit count ConsecutiveMeterExceedCnt is greater than or equal to 3.

[0053] In this state, the system triggers a service flow priority downgrade and can switch to a standby queue or standby metering / gating configuration.

[0054] For Level 3 mandatory degradation, the system determines that a business flow has entered a Level 3 serious violation state when any of the following conditions are met: 1) Meets the following criteria for 5 to 8 consecutive statistical periods: ; 2) The current burst length satisfies: ; 3) The peak rate exceeds the PIR more than 3 times consecutively.

[0055] In this state, the system performs a forced degradation process on the service flow, switching it to another non-primary guarantee queue to prevent it from continuously occupying critical priority resources.

[0056] Furthermore, the normal status monitoring of the queue specifically includes a first-level queue warning status. The first-level queue warning status indicates that if the current occupied queue depth meets a preset range or the average queuing waiting time reaches a preset proportion of the business latency budget, queue observation will be initiated and new low-priority queues will be restricted.

[0057] In this embodiment, a level 1 queue warning state is activated. When 60% ≤ Q_occ_ratio < 80% or Q_wait_avg reaches 30%~50% of the service latency budget, queue observation is initiated and new low-priority queues are restricted.

[0058] S300 performs gating scheduling and credit-based shaping on data stream queues of different priorities, and sends the data to the DU side for data processing to realize mid-transmission communication in the access network.

[0059] S310. Control the transmission time slots of data stream queues with different priorities, prioritize transmission within a preset time window, and introduce a credit-based shaping mechanism to constrain the transmission rate and burst behavior of the data streams before sending them to the DU-side TSN terminal system. In this embodiment, service data, after queue mapping and traffic policing, enters the corresponding priority queue in the TSN terminal system. At the transmission exit, Time-Aware Shaping (TAS) gating scheduling and credit-based shaping are further performed. Specifically, the system controls the transmission time slots of different priority queues according to a pre-configured Gated Control List (GCL), ensuring that critical service flows are transmitted preferentially within a specified time window. Simultaneously, a credit-based shaping mechanism is introduced for specific service flows to constrain their transmission rate and burst behavior, mitigating the impact of low-priority or bursty services on critical service transmission.

[0060] S320, based on the DU-side TSN terminal system, performs queue reverse mapping processing on the data stream and delivers it to the corresponding RLC queue; The S330, based on RLC queues, adopts a cache management method based on BD linked lists to organize and cache the data streams of different bearer queues in the DDR storage space, thereby realizing midhaul communication in the access network.

[0061] In this embodiment, service data enters the DU-side TSN terminal system after transmission via the F1-U link. The DU side first performs queue demapping processing on the received service flow, restoring the priority semantics, flow identifier, and related scheduling semantics of the TSN side to 5G bearer semantics recognizable by the DU side, and then delivers the data to the corresponding RLC queue. Subsequently, the RLC module adopts a BD-based linked list-based cache management method to organize and cache data from different bearer queues in the DDR storage space to support the unified scheduling and transmission of different service data by the subsequent MAC layer. This processing method helps improve the data caching efficiency and scheduling efficiency of the DU side under multi-service concurrency conditions.

[0062] In summary, this embodiment of the invention provides a priority guarantee scheme for CU / DU midhaul networks to improve the transmission determinism of time-sensitive services. In this embodiment, a CDFO (Cross-Domain FlowObject) generator, a policy orchestrator, and a state feedback collector are set up on the CU-CP side to generate and dynamically maintain cross-layer semantic mapping policies. On the CU-UP side, a cross-domain flow identification entry point, a PSFP fusion execution module, and a Qbv / Qav scheduling module are set up to realize service flow priority identification, traffic policing, and scheduling control. On the DU side, a queue de-mapping module and a BD linked list cache manager are set up to complete the recovery of TSN-side service semantics to 5G bearer queues and cache management. Furthermore, this embodiment also introduces a closed-loop backpressure and state feedback mechanism. By collecting PSFP execution status, queue occupancy status, cache level, and receiver-side queue congestion information, this information is transmitted back to the CU-CP side for policy rearrangement and dynamic adjustment of mapping parameters, thereby achieving closed-loop collaborative control between the transmitting-side mapping policy, policing parameters, and receiver-side bearer status. Through the collaboration of the above modules, high-priority business data can be prioritized for transport in the data transmission link, and the stability and resource utilization efficiency of the system can be improved in multi-service concurrent scenarios.

[0063] The purpose of this invention is to address the problem that the midhaul link in the existing CU / DU separation architecture lacks effective priority guarantee and deterministic transmission capability for periodic time-sensitive services. It proposes a method and device for cross-domain mapping and deterministic transmission of data streams in the scenario of 5G bearer network and time-sensitive network integration, so as to meet the bearing requirements of key data streams for low latency, low jitter and high reliability transmission in scenarios such as industrial Internet, remote control, spaceborne base stations and space-ground integrated networks.

[0064] This invention proposes a deterministic transmission scheme that deeply integrates cross-layer semantic mapping and PSFP. By constructing a unified cross-domain flow object on the control side, it binds 5QI, QFI, DRB_ID, TEID, and service timing attributes from the 5G system with PCP, VID, flow filtering, flow metering, flow gating, queue identifiers, and anti-mapping identifiers on the TSN side. This cross-domain flow object is then distributed to the TSN end systems on the CU and DU sides. During the PSFP flow identification phase, the TSN end system directly loads the cross-domain flow object, completing priority marker generation, filtering, metering, and gating instance binding, sending and receiving queue selection, and anti-mapping identifier configuration. This integrates the mapping process, flow supervision process, and queue scheduling process into a unified execution chain. Furthermore, the system dynamically adjusts the priority, queue affiliation, and supervision parameters of the data flow based on the PSFP execution status, queue occupancy status, and receiver-side anti-mapping status, achieving a closed-loop deterministic transport of cross-domain data flows across heterogeneous networks through "mapping-supervision-feedback-remapping."

[0065] Meanwhile, this invention also aims to achieve reverse mapping processing on the receiving side, enabling the service tags and flow states on the TSN side to be restored to the corresponding 5G bearer queue or RLC queue, and to manage the service data on the DU side through the BD linked list, thereby forming a closed-loop mechanism for bidirectional collaboration between transmission and reception, so as to improve the deterministic guarantee capability and overall bearer efficiency of time-sensitive services between heterogeneous networks.

[0066] Therefore, this invention introduces a time-sensitive networking (TSN) mechanism into the CU / DU midhaul network and establishes a cross-layer mapping relationship between 5G service semantics and TSN scheduling semantics. This enables service attributes such as 5QI, QFI, and DRB to be uniformly associated with PCP, VID, flow filtering, flow metering, flow gating, and queue scheduling parameters, thereby improving the priority transmission capability of critical data flows in the midhaul link. Compared to existing schemes that rely solely on DSCP or priority queues for coarse-grained differentiation, this invention can achieve fine-grained identification, monitoring, and scheduling of periodic time-sensitive services. It also dynamically adjusts mapping and scheduling strategies through state feedback and closed-loop backpressure mechanisms, enhancing system stability and resource utilization efficiency under multi-service concurrency conditions. Furthermore, this invention achieves efficient recovery and cache optimization of TSN-side service semantics to 5G bearer queues through DU-side reverse mapping and DDR cache management based on BD linked lists. Therefore, it is particularly suitable for scenarios with high requirements for latency, jitter, and reliability, such as industrial internet, remote control, satellite base stations, and space-ground converged networks.

[0067] Reference Figure 2 The access network midhaul communication system based on cross-domain QoS mapping and PSFP linkage includes: The first module 201 is used to perform system initialization and CU-side TSN terminal system data filling processing to obtain the encapsulated data frame. The second module 202 is used to perform cross-layer semantic mapping on the encapsulated data frame and to perform priority marking through flow gating normal state monitoring, flow metering normal state monitoring and queue normal state monitoring to obtain data flow queues with different priorities. The third module 203 is used to perform gating scheduling and credit-based shaping on data stream queues of different priorities, and send them to the DU side for data processing to realize mid-transmission communication in the access network.

[0068] Furthermore, the overall architecture of this invention can be abstracted into the following four major modules: 1) Adaptation Processing Module for CU-Side Midhaul Boundary: This module is deployed at the CU-side midhaul boundary to perform cache management, service classification, and priority bearer adaptation of service data from the 5G user plane. It combines priority queues, time-aware gating scheduling, and credit-based traffic shaping mechanisms in time-sensitive networks to achieve differentiated transmission and deterministic delivery of different types of service flows. Specifically, this module can collaboratively process cross-layer mapping results with mechanisms such as flow filtering, flow metering, and flow gating, thereby enhancing the priority forwarding capability of critical service flows in the midhaul link.

[0069] 2) Adaptation processing module for DU-side mid-haul boundary: This module is deployed at the DU-side mid-haul boundary and is used to receive, process, identify priorities, and reverse-map queues of TSN service flows from the mid-haul link, restoring the priority semantics and flow processing semantics of the TSN side to bearer semantics that can be recognized by the DU side.

[0070] 3) Cross-layer queue mapping strategy generation module based on software control side: This module is deployed on the CU-CP or control plane side and is used to generate mapping strategies related to time-sensitive network side priority marking, flow filtering, flow metering, flow gating and queue scheduling based on service semantics such as 5QI, QFI and DRB identifiers in 5G system.

[0071] 4) RLC queue cache management module based on BD linked list: This module is designed for the queue management needs of the DU side. It organizes and caches the data of different DRB queues in external storage resources such as DDR through a buffer descriptor linked list, so as to achieve unified caching, fast indexing and efficient scheduling of multiple service queues.

[0072] The content of the above method embodiments is applicable to this system embodiment. The specific functions implemented in this system embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0073] The above is a detailed description of the preferred embodiments of the present invention. However, the present invention is not limited to the embodiments described. Those skilled in the art can make various equivalent modifications or substitutions without departing from the spirit of the present invention. All such equivalent modifications or substitutions are included within the scope defined by the claims of this application.

Claims

1. A midhaul communication method for access networks based on cross-domain QoS mapping and PSFP linkage, characterized in that, Includes the following steps: Perform system initialization and CU-side TSN terminal system data filling processing to obtain the encapsulated data frame; Perform cross-layer semantic mapping on the encapsulated data frames, and mark priorities by monitoring the normal state of flow gating, the normal state of flow metering, and the normal state of queues to obtain data flow queues with different priorities. Gating and credit-based shaping are performed on data stream queues of different priorities, and the data is sent to the DU side for data processing to realize mid-transmission communication in the access network.

2. The access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage according to claim 1, characterized in that, The system initialization step specifically includes: Data interaction is performed based on the centralized unit control plane, centralized unit user plane and distributed unit to confirm the TEID, UDP port number, source IP address information and destination IP address information required for tunnel encapsulation; Clear the remaining amount in the PDCP buffer queue and the remaining amount in the RLC queue; Set threshold parameters, which include normal flow metering status, normal flow gating status, and normal queue status; The scores of the 5QI parameters are normalized using an algorithm to obtain the mapping table entries.

3. The access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage according to claim 2, characterized in that, The step of data filling and processing at the CU-side TSN terminal system to obtain the encapsulated data frame specifically includes: Acquire PDCP data frames, and perform statistics on the buffer status of PDCP and RLC queues to generate corresponding scheduling instructions. The buffer status includes queue balance, buffer occupancy, and amount of data to be sent. According to the scheduling instruction, the corresponding length of the data frame is read and sent to the F1-U protocol encapsulation entity. Combined with TEID and UDP port number, the data frame is encapsulated using the F1-U user plane protocol to obtain the encapsulated data frame.

4. The access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage according to claim 3, characterized in that, The step of performing cross-layer semantic mapping on the encapsulated data frames and prioritizing them through flow gating normal state monitoring, flow metering normal state monitoring, and queue normal state monitoring to obtain data flow queues of different priorities specifically includes: Based on the mapping table entries, cross-layer semantic mapping is performed on the encapsulated data frames to convert the data streams corresponding to 5G-side QoS semantics into TSN-side scheduling semantics and generate corresponding priority tags. Based on threshold parameters, the transformed data stream is sequentially monitored for normal status of flow gating, normal status of flow metering, and normal status of queue. The transformed data stream is then prioritized to obtain data stream queues with different priorities.

5. The access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage according to claim 4, characterized in that, The normal state monitoring of the flow gating specifically includes a level 1 gating warning state, a level 2 gating warning state, and a level 3 serious violation state, wherein: The first-level gate control early warning status indicates that the arrival time of the data stream is within the expected gate control window. If there is a phase error greater than 10% of the gate control window length or the offset of the data stream arrival time relative to the gate control window boundary does not exceed one protection band length, a gate control violation is recorded and the cumulative count is recorded. The Level 2 gate control early warning status indicates that if any one of the following three conditions is met: the arrival time of a data stream exceeds the allowed gate control window boundary, the number of consecutive gate control violations is greater than or equal to a preset coefficient, or the number of cumulative gate control violations within a preset business cycle is greater than or equal to a preset coefficient, the data stream is marked as 1 and written into the exception cache chain. The Level 3 serious violation status indicates that if there are two consecutive preset business cycles with incorrect windows or the incorrect window offset exceeds the width of a complete gate control window, the system will switch to the backup gate control configuration file and perform temporary priority downgrading processing on the data stream according to the current business level and resource status.

6. The access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage according to claim 5, characterized in that, The flow metering normal status monitoring specifically includes a first-level metering early warning status, a second-level metering early warning status, and a third-level forced degradation status, wherein: The first-level metering early warning status indicates that if either the average measurement rate meets the preset threshold or the current burst length is greater than 90% of the promised burst length, an over-limit count will be recorded for the data stream. The Level 2 metering early warning status indicates that if any one of the following four conditions is met within a consecutive preset statistical period: the preset average measurement rate is met within a consecutive preset statistical period, the preset peak measurement rate is greater than the peak information rate, the current burst length meets the preset conditions, or the consecutive over-limit count is greater than or equal to a preset number, priority downgrade processing will be performed on the data stream. The three-level forced degradation state indicates that any one of the following three conditions occurs consecutively: the data stream meets the preset average measurement rate requirement for a preset number of consecutive statistical periods, the current burst length meets the preset length requirement, or the peak rate exceeds the PIR and exceeds the preset threshold. In such cases, forced degradation processing is performed on the data stream, and it is switched to another non-primary protection queue.

7. The access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage according to claim 6, characterized in that, The normal status monitoring of the queue specifically includes a first-level queue warning status. The first-level queue warning status indicates that if the current occupied queue depth meets a preset range or the average queuing waiting time reaches a preset proportion of the business latency budget, queue observation will be initiated and new low-priority queues will be restricted.

8. The access network midhaul communication method based on cross-domain QoS mapping and PSFP linkage according to claim 7, characterized in that, The step of gating and scheduling data stream queues of different priorities and performing credit-based shaping, and then sending the data to the DU side for processing to realize midhaul communication in the access network, specifically includes: The transmission time slots of data stream queues with different priorities are controlled, and priority is given to transmission within a preset time window. A credit-based shaping mechanism is introduced to constrain the transmission rate and burst behavior of the data streams before transmission to the DU-side TSN terminal system. Based on the DU-side TSN terminal system, the data stream is processed by queue reverse mapping and delivered to the corresponding RLC queue; Based on RLC queues, a cache management method based on BD linked lists is adopted to organize and cache the data streams of different bearer queues in DDR storage space to realize midhaul communication in the access network.

9. An access network midhaul communication system based on cross-domain QoS mapping and PSFP linkage, characterized in that, Includes the following modules: The first module is used to perform system initialization and data filling processing on the CU-side TSN terminal system to obtain the encapsulated data frame. The second module is used to perform cross-layer semantic mapping on the encapsulated data frame and to mark the priority of data stream queues by monitoring the normal state of flow gating, the normal state of flow metering, and the normal state of queues. The third module is used to perform gating scheduling and credit-based shaping of data stream queues with different priorities, and send the data to the DU side for data processing to realize mid-pass communication in the access network.

Images