Optical transport network service scheduling method and device, and electronic equipment

Abstract

The invention discloses an optical transport network service scheduling method, an optical transport network service scheduling device and electronic equipment. The method comprises the following steps: identifying a time delay requirement threshold value and a message length of a service flow; then determining the transmission priority of the service flow; the transmission priorities comprise a first priority, a second priority and a third priority, and the first priority has the highest priority; determining a scheduling decision; in the scheduling decision, when the first priority service needs to distribute an optical channel data unit time slot but has no idle time slot meeting the condition, selecting a candidate time slot from the time slots occupied by the second or third priority service flow; when and only when the net income is greater than zero, generating a preemption instruction, interrupting the transmission of the originally occupied service flow, and remapping the time slot resource to the first priority service flow; and according to the scheduling decision, mapping the data of the service flow to the allocated optical channel data unit time slot, and completing transmission through the optical transport network so as to realize the service scheduling of the optical transport network. According to the invention, low time delay guarantee and high resource utilization rate can be considered.

Description

Technical Field

[0001] This invention relates to the field of network technology, and in particular to a method, apparatus, and electronic device for scheduling services in an optical transport network. Background Technology

[0002] Optical Transport Networks (OTNs), as a key component of modern communication infrastructure, widely support services with increasingly stringent transmission performance requirements, such as 5G, industrial control, financial transactions, and cloud computing. These emerging applications generally feature low latency and high reliability, especially placing extremely high demands on the determinism of end-to-end latency—requiring not only low average latency but also low latency jitter and predictable worst-case scenarios.

[0003] However, the mainstream service scheduling mechanisms in current OTN networks have significant shortcomings in addressing such needs, mainly in the following two aspects:

[0004] First, priority-based scheduling mechanisms generally suffer from head-of-line blocking.

[0005] In traditional implementations, service flows of different priorities are mapped to different output queues, with high-priority queues enjoying scheduling priority. However, once a low-priority long packet begins to occupy the physical channel for transmission, even if a high-priority short packet arrives subsequently, it must wait for the current transmission to complete before it can receive service. This "first-in, first-out" serial transmission mode causes the actual latency of high-priority services to be highly dependent on the length and number of data packets at the front of the queue, resulting in unstable and unpredictable latency performance, failing to meet the core requirement of deterministic low latency for real-time services.

[0006] Second, resource allocation methods mostly adopt static or semi-static configuration, lacking dynamic adaptability.

[0007] In the existing OTN architecture, Optical Channel Data Unit (ODUk) time slots are typically pre-allocated by the network management system to specific services or customers. While this approach ensures bandwidth isolation, it suffers from extremely poor flexibility: when an allocated service has no data to send, its dedicated time slot resources remain idle, resulting in low spectrum efficiency; and when a burst of high-priority, low-latency services arrives, if there are no pre-configured idle time slots, transmission resources cannot be obtained in time, and the service must wait in a queue, leading to a sharp increase in latency for critical services and making it difficult to adapt to dynamic changes in service traffic.

[0008] The aforementioned shortcomings mean that existing OTN scheduling technologies cannot simultaneously guarantee low latency and high resource utilization, and in particular, they cannot effectively support the large-scale deployment of services with millisecond or even sub-millisecond deterministic latency. Summary of the Invention

[0009] The technical problem to be solved by the present invention is to address the above-mentioned shortcomings of the prior art by proposing a service scheduling method, device and electronic equipment for optical transport networks. This method can take into account both low latency guarantee and high resource utilization, thereby effectively supporting the large-scale deployment of millisecond or even sub-millisecond deterministic latency services.

[0010] In a first aspect, the present invention provides a method for scheduling services in an optical transport network, the method comprising:

[0011] In response to service flow access to the optical transport network, the latency requirement threshold and message length of the service flow are identified;

[0012] Based on the latency requirement threshold and message length, the transmission priority of the service flow is determined, and the corresponding service transmission request is initiated. The transmission priority includes first priority, second priority, and third priority, with first priority being higher than second priority, and second priority being higher than third priority.

[0013] In response to service transmission requests, scheduling decisions are made based on transmission priority and global network status information. The global network status information is generated by collecting the queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot at preset intervals.

[0014] The scheduling decision includes: when it is necessary to allocate optical channel data unit time slots for a first-priority service flow, and there are no idle optical channel data unit time slots that meet the scheduling conditions, a candidate time slot is selected from the time slots currently occupied by the second-priority or third-priority service flow; the net delay benefit of reassigning the candidate time slot from the currently occupied second-priority or third-priority service flow to the first-priority service flow is calculated; a preemption instruction for the candidate time slot is generated if and only if the calculated net delay benefit is greater than zero; the preemption instruction is used to indicate the interruption of the transmission of the originally occupied second-priority or third-priority service flow on the candidate time slot, and to remap the resources of the candidate time slot to the first-priority service flow;

[0015] Based on the scheduling decision, the data of the service flow is mapped to the allocated optical channel data unit time slot and transmitted through the optical transport network to realize the service scheduling of the optical transport network.

[0016] Furthermore, the method also includes:

[0017] If the scheduling decision includes a preemption operation, a resource compensation identifier is generated for the preempted service flow.

[0018] Create compensation record entries for preempted service flows with resource compensation identifiers, and prioritize resource allocation for preempted service flows when network load is below a preset threshold or when a dedicated compensation time slot pool is available.

[0019] Furthermore, the net delay gain ΔD is calculated as follows: ΔD = D_high_reduced – D_low_increased; where D_high_reduced is the expected reduction in end-to-end transmission delay for the first priority service flow due to obtaining a candidate time slot, and D_low_increased is the expected increase in end-to-end transmission delay for the preempted second or third priority service flow due to losing a candidate time slot.

[0020] Furthermore, based on the latency requirement threshold and message length, the transmission priority of the service flow is determined, specifically including:

[0021] If the latency requirement threshold of the service flow is not greater than the first threshold and the message length of the service flow is not greater than the preset length threshold, then the transmission priority of the service flow is determined to be the first priority.

[0022] If the latency requirement threshold of the service flow is not greater than the first threshold, and the message length of the service flow is greater than the preset length threshold, then the transmission priority of the service flow is determined to be the second priority.

[0023] If the latency requirement threshold of the service flow is greater than the first threshold, then the transmission priority of the service flow is determined to be the third priority.

[0024] Furthermore, the scheduling decision also includes setting the threshold for determining net latency benefit to be below zero when the remaining survival time of the first priority service flow is less than or equal to a preset time window.

[0025] Furthermore, the global network state information is generated in the following way:

[0026] The queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot are collected in the first preset period.

[0027] The queue status is converted into node load data expressed as a percentage, and the occupancy status is represented as a time slot resource occupancy bitmap;

[0028] The node load data and time slot resource occupancy bitmap are integrated in a second preset period to generate global network status information. The second preset period is shorter than the first preset period.

[0029] Furthermore, the method also includes:

[0030] During service transmission, the optical power and bit error rate of the allocated optical channel are monitored in real time. When the optical power is lower than the preset power threshold or the bit error rate is higher than the preset bit error threshold, the rerouting mechanism is triggered to re-execute the scheduling decision based on global network status information.

[0031] Furthermore, after determining the transmission priority of the service flow, the method also includes:

[0032] The end-to-end transmission latency of the service flow is continuously monitored through a sliding window mechanism. If the end-to-end transmission latency exceeds the corresponding latency requirement threshold for multiple consecutive monitoring periods, the transmission priority of the service flow is increased by one level.

[0033] In a second aspect, the present invention provides an optical transport network service scheduling device, the device comprising:

[0034] The identification unit is used to identify the latency requirement threshold and message length of the service flow in response to the service flow accessing the optical transport network.

[0035] The processing unit, connected to the identification unit, is used to determine the transmission priority of the service flow based on the latency requirement threshold and the message length, and to initiate the corresponding service transmission request. The transmission priority includes a first priority, a second priority, and a third priority, with the first priority being higher than the second priority, and the second priority being higher than the third priority.

[0036] The determination unit, connected to the processing unit, is used to respond to service transmission requests and determine scheduling decisions based on transmission priority and global network status information. The global network status information is generated by collecting the queue status of each network node in the optical transport network and the occupancy status of the time slot of the optical channel data unit at a preset period.

[0037] The scheduling decision includes: when it is necessary to allocate optical channel data unit time slots for a first-priority service flow, and there are no idle optical channel data unit time slots that meet the scheduling conditions, a candidate time slot is selected from the time slots currently occupied by the second-priority or third-priority service flow; the net delay benefit of reassigning the candidate time slot from the currently occupied second-priority or third-priority service flow to the first-priority service flow is calculated; a preemption instruction for the candidate time slot is generated if and only if the calculated net delay benefit is greater than zero; the preemption instruction is used to indicate the interruption of the transmission of the originally occupied second-priority or third-priority service flow on the candidate time slot, and to remap the resources of the candidate time slot to the first-priority service flow;

[0038] The mapping transmission unit, connected to the determining unit, is used to map the data of the service flow to the allocated optical channel data unit time slots according to the scheduling decision, and complete the transmission through the optical transport network to realize the service scheduling of the optical transport network.

[0039] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and when the processor runs the computer program stored in the memory, the processor executes the optical transport network service scheduling method according to the first aspect.

[0040] This invention establishes a multi-level transmission priority system by identifying latency requirement thresholds and message lengths for service flows. It combines this with a dynamic scheduling decision-making mechanism based on global network state information (including queue status of each node and time slot occupancy status of optical channel data units). In particular, it introduces a conditional preemption strategy based on "net latency gain," effectively overcoming the latency jitter caused by queue head blocking and the lack of flexibility in static resource allocation in existing technologies. While ensuring strict latency requirements for high-priority services, it significantly improves network resource utilization efficiency, thus balancing low latency guarantees with high resource utilization. This effectively supports the large-scale deployment of millisecond-level or even sub-millisecond-level deterministic latency services. Specific beneficial effects are as follows:

[0041] 1) Achieve deterministic low-latency assurance to support large-scale deployment of critical business operations:

[0042] This invention designs an intelligent preemption mechanism based on "net latency benefit" for first-priority (highest priority) service flows. When there are no idle optical channel data unit time slots in the system, the scheduler evaluates candidate resources from the time slots already occupied by second- or third-priority services, and calculates the difference between the end-to-end latency reduction brought by reallocating the time slot to the high-priority service and the latency increase of the preempted service (i.e., "net latency benefit"). Preemption is only performed when this benefit is greater than zero. This mechanism fundamentally avoids the head-of-line blocking problem caused by low-priority long message transmission in traditional queue scheduling, ensuring that high-priority short burst services (such as industrial control instructions and high-frequency transaction messages) can obtain predictable, low-jitter transmission services, meeting their stringent requirements for millisecond or even sub-millisecond deterministic latency, and providing reliable support for the large-scale commercial deployment of critical services in OTN networks.

[0043] 2) Improve the efficiency of dynamic resource reuse while balancing overall network performance and fairness:

[0044] Unlike static or semi-static fixed time slot allocation methods, this invention allows all priority services to compete for and reuse time slot resources on demand. Low-priority services can fully utilize idle bandwidth when there is no competition from high-priority services, avoiding resource idleness; while high-priority services only trigger preemption when a quantitative assessment confirms "overall latency benefit," avoiding frequent and ineffective resource reallocation. This collaborative mechanism of "on-demand allocation + conditional preemption" ensures the SLA of critical services while minimizing interference with ordinary services, achieving an organic unity of high resource utilization and high service quality.

[0045] 3) Enhance the globality, real-time performance, and adaptive capabilities of scheduling decisions:

[0046] By continuously collecting queue depth and timeslot occupancy bitmaps of all network nodes at preset intervals, this invention constructs a highly timely global resource status view, enabling scheduling decisions to break free from the limitations of a local perspective and possessing the ability to comprehensively optimize the entire network. Simultaneously, by combining the latency thresholds and packet length characteristics of each service for priority segmentation, the scheduling strategy can accurately match the actual needs of different services. This mechanism significantly improves the adaptability of optical transport networks to bursty traffic, mixed service loads, and dynamic network environments, laying the foundation for building intelligent bearer networks for diverse low-latency applications in the future.

[0047] It should be understood that the description in this section is not intended to identify key or essential features of the embodiments of the present invention, nor is it intended to limit the scope of the invention. Other features of the invention will become readily apparent from the following description. Attached Figure Description

[0048] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used together with the embodiments of the invention to explain the invention and do not constitute a limitation thereof. The above and other features and advantages will become more apparent to those skilled in the art from the detailed description of exemplary embodiments with reference to the accompanying drawings, in which:

[0049] Figure 1 This is a schematic diagram of an optical transport network service scheduling method provided in an embodiment of the present invention;

[0050] Figure 2 This is a flowchart of optical transport network service scheduling provided in an embodiment of the present invention;

[0051] Figure 3 This is a diagram illustrating the optical transport network service scheduling framework provided in an embodiment of the present invention.

[0052] Figure 4 This is a schematic diagram of an optical transport network service scheduling device provided in an embodiment of the present invention;

[0053] Figure 5A block diagram of an electronic device provided in an embodiment of the present invention.

[0054] Reference numerals: 10, identification unit; 20, processing unit; 30, determination unit; 40, mapping and transmission unit; 100, processor; 200, memory. Detailed Implementation

[0055] It is understood that the specific embodiments and accompanying drawings described herein are merely for explaining the invention and are not intended to limit the invention.

[0056] It is understood that, without conflict, the various embodiments and features in the embodiments of the present invention can be combined with each other.

[0057] It is understood that, for ease of description, only the parts related to the present invention are shown in the accompanying drawings, while the parts unrelated to the present invention are not shown in the drawings.

[0058] It is understood that each unit or module involved in the embodiments of the present invention may correspond to only one entity structure, or may be composed of multiple entity structures, or multiple units or modules may be integrated into one entity structure.

[0059] It is understood that, without conflict, the functions and steps marked in the flowcharts and block diagrams of this invention may occur in a different order than that marked in the accompanying drawings.

[0060] It is understood that the flowcharts and block diagrams of this invention illustrate the possible architecture, functions, and operations of systems, apparatuses, devices, and methods according to various embodiments of this invention. Each block in the flowchart or block diagram may represent a unit, module, program segment, or code, containing executable instructions for implementing the specified function. Furthermore, each block or combination of blocks in the block diagram and flowchart can be implemented using a hardware-based system to achieve the specified function, or using a combination of hardware and computer instructions.

[0061] It is understood that the units and modules involved in the embodiments of the present invention can be implemented by software or by hardware. For example, the units and modules can be located in a processor.

[0062] Example 1:

[0063] Optical transport networks (OTNs), as the backbone platform of modern communication networks, have become core infrastructure supporting critical businesses such as 5G, industrial internet, financial transactions, cloud computing, and smart transportation, thanks to their high bandwidth, high reliability, and robust operation and maintenance capabilities. However, with the explosive growth of emerging applications such as remote control, high-frequency trading, immersive VR / AR, cloud gaming, autonomous driving vehicle-to-everything (V2X) communication, power differential protection, and cross-data center collaborative computing, the network's requirements for end-to-end transmission latency have upgraded from "low average latency" to "millisecond-level or even sub-millisecond-level deterministic latency"—not only requiring extremely low latency, but also emphasizing low jitter and predictable worst-case scenarios.

[0064] Against this backdrop, traditional OTN scheduling mechanisms face severe challenges: priority-based queue scheduling is susceptible to "head-of-queue blocking," causing high-priority short packets to be forced to wait for low-priority long packets to complete transmission, resulting in severe latency fluctuations; while static or semi-static ODUk time slot allocation can ensure bandwidth isolation, but resource rigidity leads to wasted idle time and cannot respond to sudden low-latency service demands, resulting in critical tasks either queuing timeouts or low resource utilization.

[0065] The optical transport network service scheduling method proposed in this embodiment is designed to address the core contradictions in the aforementioned high real-time, high dynamic, and multi-service mixed scenarios, and has broad and crucial application value:

[0066] In the field of industrial internet, it can provide sub-millisecond deterministic latency guarantee for scenarios such as collaborative control of robotic arms, synchronization of automated production lines, and precision operation of remote equipment, ensuring zero-latency and zero-jitter execution of control commands, and effectively avoiding production interruptions or safety accidents caused by latency anomalies.

[0067] In high-frequency trading systems for finance, it is possible to identify microsecond-level latency-sensitive trading instruction streams and prioritize the scheduling of key messages through an intelligent preemption mechanism based on net latency benefits, helping institutions gain a decisive speed advantage in nanosecond-level competition.

[0068] In immersive interactive applications (such as VR remote surgery, metaverse, and cloud gaming), dynamic allocation of bandwidth and time slot resources provides a stable, low-jitter transmission channel for audio and video interactive streams, significantly reducing user dizziness and operation latency, and improving the smoothness and real-time performance of the immersive experience.

[0069] In cloud data center interconnection scenarios, when faced with sudden data synchronization, disaster recovery switching or distributed AI training traffic, the optical channel data unit time slots can be quickly reorganized, which can significantly improve the spectrum utilization efficiency of the backbone link while ensuring the SLA of high-priority tasks.

[0070] In 5G uRLLC and vehicle-to-everything (V2X) applications, it efficiently carries hybrid services of enhanced mobile broadband (eMBB) and ultra-reliable low-latency communication (uRLLC), enabling instantaneous response and reliable transmission of safety-critical signals such as emergency braking commands for autonomous vehicles and collaborative warnings at intersections, thus supporting highly reliable vehicle-road cooperative services.

[0071] By integrating latency threshold and message length feature identification of integrated service flows, periodic perception of the entire network queue and time slot occupancy status, and precise preemption decision based on the condition of "net latency benefit greater than zero", this embodiment achieves efficient dynamic reuse of optical channel resources while strictly ensuring deterministic low latency for critical services. It provides a high-performance, highly flexible and intelligent OTN bearer solution for the digital transformation of various industries.

[0072] like Figure 1 As shown in the figure, this embodiment provides a service scheduling method for an optical transport network, including steps S1 to S4.

[0073] Step S1: In response to the access of the service flow to the optical transport network, identify the latency requirement threshold and message length of the service flow.

[0074] Step S2: Determine the transmission priority of the service flow based on the latency requirement threshold and the message length, and initiate the corresponding service transmission request; the transmission priority includes first priority, second priority and third priority, with the first priority being higher than the second priority, and the second priority being higher than the third priority.

[0075] As one specific implementation method, the transmission priority of the service flow is determined based on the latency requirement threshold and the message length, specifically including:

[0076] If the latency requirement threshold of the service flow is not greater than the first threshold and the message length of the service flow is not greater than the preset length threshold, then the transmission priority of the service flow is determined to be the first priority.

[0077] If the latency requirement threshold of the service flow is not greater than the first threshold, and the message length of the service flow is greater than the preset length threshold, then the transmission priority of the service flow is determined to be the second priority.

[0078] If the latency requirement threshold of the service flow is greater than the first threshold, then the transmission priority of the service flow is determined to be the third priority.

[0079] Step S3: In response to the service transmission request, determine the scheduling decision based on the transmission priority and global network status information; the global network status information is generated by collecting the queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot at a preset period.

[0080] The scheduling decision includes: when it is necessary to allocate optical channel data unit time slots for a first-priority service flow, and there are no idle optical channel data unit time slots that meet the scheduling conditions, selecting a candidate time slot from the time slots currently occupied by a second-priority or third-priority service flow; calculating the net delay benefit of reassigning the candidate time slot from the currently occupied second-priority or third-priority service flow to the first-priority service flow; generating a preemption instruction for the candidate time slot if and only if the calculated net delay benefit is greater than zero; the preemption instruction is used to instruct the interruption of the transmission of the originally occupied second-priority or third-priority service flow on the candidate time slot, and to remap the resources of the candidate time slot to the first-priority service flow.

[0081] As a specific implementation method, the net delay gain ΔD is calculated as follows: ΔD = D_high_reduced – D_low_increased; where D_high_reduced is the expected reduction in end-to-end transmission delay of the first priority service flow due to obtaining a candidate time slot, and D_low_increased is the expected increase in end-to-end transmission delay of the preempted second or third priority service flow due to losing a candidate time slot.

[0082] In specific implementation, the optical transport network service scheduling method provided in this embodiment further includes:

[0083] If the scheduling decision includes a preemption operation, a resource compensation identifier is generated for the preempted service flow.

[0084] Create compensation record entries for preempted service flows with resource compensation identifiers, and prioritize resource allocation for preempted service flows when network load is below a preset threshold or when a dedicated compensation time slot pool is available.

[0085] As a specific implementation method, the scheduling decision also includes: when the remaining survival time of the first priority service flow is less than or equal to a preset time window, the threshold value for determining the net delay benefit is set to be lower than zero.

[0086] As a specific implementation method, global network state information is generated in the following way:

[0087] The queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot are collected in the first preset period.

[0088] The queue status is converted into node load data expressed as a percentage, and the occupancy status is represented as a time slot resource occupancy bitmap;

[0089] The node load data and time slot resource occupancy bitmap are integrated in a second preset period to generate global network status information. The second preset period is shorter than the first preset period.

[0090] Furthermore, the method also includes:

[0091] During service transmission, the optical power and bit error rate of the allocated optical channel are monitored in real time. When the optical power is lower than the preset power threshold or the bit error rate is higher than the preset bit error threshold, the rerouting mechanism is triggered to re-execute the scheduling decision based on global network status information.

[0092] As a specific implementation method, after determining the transmission priority of the service flow, the method further includes:

[0093] The end-to-end transmission latency of the service flow is continuously monitored through a sliding window mechanism. If the end-to-end transmission latency exceeds the corresponding latency requirement threshold for multiple consecutive monitoring periods, the transmission priority of the service flow is increased by one level.

[0094] For example, a sliding window mechanism is used to continuously monitor the end-to-end transmission latency of a service flow. If the end-to-end transmission latency exceeds the latency threshold specified for the corresponding sensitivity level within three consecutive monitoring periods, the transmission priority of the service flow is increased by one level. For instance, suppose a video conferencing service is marked as "high sensitivity" (latency threshold of 50 milliseconds), and the network status monitoring module samples every 1 millisecond as a monitoring period. If the actual end-to-end latency of this service is 52ms, 55ms, and 58ms in the three consecutive periods of 101ms, 102ms, and 103ms respectively, all exceeding the 50ms threshold, then the priority increase mechanism is triggered, temporarily increasing its scheduling priority from P2 (high sensitivity) to P1 (ultra-high sensitivity) to obtain better time slot resources in subsequent scheduling and ensure service quality.

[0095] Step S4: Based on the scheduling decision, the data of the service flow is mapped to the allocated optical channel data unit time slot, and the transmission is completed through the optical transport network to realize optical transport network service scheduling.

[0096] The optical transport network service scheduling flowchart provided in this embodiment is as follows: Figure 2 As shown, the process specifically includes:

[0097] Step 1: Business access and labeling.

[0098] The first step of the optical transport network (OTN) service scheduling method in this embodiment is service access and intelligent labeling. When various service flows arrive at the edge entry node of the optical transport network (OTN), the service classification and labeling module deployed at that node is immediately activated, and performs refined feature recognition and dynamic digital labeling of the service flows according to a predefined multi-level strategy.

[0099] This module first parses explicit QoS parameters in the service flow, or identifies the application type (such as voice, video conferencing, industrial control commands, financial transaction messages, etc.) based on deep packet inspection, and maps them to a four-level latency sensitivity rating system:

[0100] Level 1 (Ultra-high sensitivity business): This includes business with extremely stringent real-time requirements, such as voice calls and high-frequency financial transactions, with the end-to-end latency threshold strictly limited to within 10 milliseconds.

[0101] Level 2 (Highly Sensitive Businesses): Covers key applications such as industrial remote control, power differential protection, and high-definition video conferencing, with latency required to be stably controlled within the range of 20 to 50 milliseconds;

[0102] Level 3 (Ordinary Sensitive Business): Corresponds to regular data transmission tasks, such as web browsing and database queries, with an acceptable latency limit of 100 milliseconds;

[0103] Level 4 (Low-sensitivity services): Primarily for non-real-time background operations such as file backup and log synchronization, with no mandatory requirements for transmission timeliness and the lowest scheduling priority.

[0104] While determining the latency sensitivity level, this module further analyzes the message length characteristics of the service flow and uses 256 bytes as the threshold to classify the service flow into short message flow and long message flow. This distinction is of great significance for subsequent scheduling decisions—especially in accurately identifying "high-priority short burst" services and preventing highly sensitive long messages from excessively occupying scarce time slots due to resource contention, thus affecting overall scheduling efficiency.

[0105] To enhance the dynamic adaptability of the marking mechanism, this module employs a sliding window mechanism to continuously track the actual end-to-end transmission latency of marked service flows. If, within multiple consecutive monitoring periods, the measured latency of a service flow consistently exceeds the threshold corresponding to its current sensitivity level, an sensitivity level upgrade operation is automatically triggered. For example, a service originally classified as Level 2 may be upgraded to Level 1, thereby achieving higher assurance in subsequent scheduling.

[0106] Ultimately, the latency sensitivity level and message length category (short / long) determined through comprehensive evaluation are encoded into structured digital identifiers and embedded in the extended header of the service data packets, or transmitted to the scheduling control unit via out-of-band signaling channels. This information constitutes a standardized "service characteristic fingerprint" that runs through the entire subsequent scheduling process, providing an accurate and reliable data foundation for intelligent scheduling decisions based on multi-dimensional characteristics (including transmission priority, global network status, and net latency gains).

[0107] Step S102: Network Status Monitoring

[0108] The network status monitoring module periodically collects operational status data of key nodes in the optical transport network with a sampling period of 100 microseconds. The collected data mainly includes the following three categories: first, port queue status, including parameters reflecting the congestion level of the buffer such as queue depth and the number of traffic flows in the queue; second, service load data, covering indicators characterizing network traffic load such as real-time throughput and bandwidth utilization of each link; and third, time slot resource occupancy, which is obtained by parsing the idle or occupied status of the optical channel data unit (ODUk) time slot in the OTN frame structure.

[0109] The collected raw data is then processed and integrated to generate a highly timely global network status view. In this view, the availability of time slot resources is precisely represented in bitmap form, with each bit corresponding to a time slot and clearly indicating its idle or occupied status. The load level of each network node is normalized to a percentage value, intuitively reflecting the current resource usage intensity. Although the underlying data is collected every 100 microseconds, this global view is refreshed at a 1-millisecond update cycle, balancing state freshness with processing overhead. Finally, this view is provided to the scheduler through standardized northbound interfaces such as NETCONF, serving as the core data basis for dynamic scheduling decisions.

[0110] Step S103: Scheduling decision calculation

[0111] Upon receiving a service transmission request, the core intelligent scheduler calculates and selects the optimal Optical Channel Data Unit (ODUk) transmission time slot based on the tagging information carried by the service flow (including latency sensitivity level and packet length category) and a real-time updated global network status view, using a multi-dimensional decision-making algorithm. This scheduling decision-making process includes the following key steps:

[0112] First, the scheduler maps services to four transmission priorities according to preset rules: the highest priority (P1) corresponds to ultra-high sensitivity short message services (such as financial transaction instructions, industrial control short frames); high priority (P2) covers ultra-high sensitivity long message or high sensitivity short message services (such as high-definition voice streams or critical control signaling); normal priority (P3) includes high sensitivity long message and normal sensitivity services (such as video conferencing data or regular database interactions); and low priority (P4) corresponds to low sensitivity services (such as file backup, log synchronization, etc.).

[0113] During the time slot calculation and selection phase, the scheduler evaluates the transmission value of all candidate time slots. This evaluation comprehensively considers two core dimensions: the urgency of the time slot (i.e., the relationship between the remaining service lifetime and the current waiting latency) and transmission quality (such as link load, historical bit error rate, queue depth, and other indicators reflecting resource availability). Based on this, the optimal time slot selection algorithm prioritizes allocating the earliest available and most valuable idle time slots to high-priority services of P1 and P2 levels to minimize their end-to-end transmission latency.

[0114] When no idle time slots satisfying the scheduling conditions exist in the network, the scheduler initiates a preemption mechanism: it filters candidate resources from the time slots currently occupied by P3 or P4 services and performs a preemption benefit assessment. This assessment precisely calculates the net latency benefit of reallocating the target time slot to a high-priority service, i.e., the difference between the expected latency reduction for the high-priority service and the expected latency increase for the preempted low-priority service. Only when the net latency benefit is significantly greater than zero is a preemption command generated, authorizing the interruption of the original low-priority service's transmission, and remapping the time slot resource to a high-priority service.

[0115] In addition, the scheduler dynamically incorporates the remaining lifetime of services as a key adjustment factor. When the remaining lifetime of P1 or P2 services is lower than the preset time window (e.g., close to its latency threshold), it indicates that it is in a critical timeout state. At this time, the scheduler will appropriately relax the preemption judgment threshold (e.g., allow cases where the net latency benefit is negative but the absolute value is small) to prioritize the real-time needs of critical services and avoid SLA default due to overly conservative economic judgments.

[0116] Through the aforementioned multi-dimensional and adaptive decision-making mechanism, this scheduling strategy ensures that high-priority services receive deterministic, low-latency services while also taking into account resource utilization efficiency and overall system fairness.

[0117] Step S104: Time Slot Allocation and Service Transmission

[0118] The time slot mapping and transmission module is responsible for executing specific time slot allocation and service transmission operations based on the scheduling decisions generated by the core intelligent scheduler. In scenarios involving preemption, this module first fully saves the transmission context information of the preempted service, including its current time slot mapping status, queue position, and transmission progress, to ensure subsequent recoverability. Subsequently, it remaps the data stream of high-priority services to the preempted optical channel data unit (ODUk) time slot, completing the dynamic reallocation of resources.

[0119] The specific time slot allocation is implemented through cross-connect devices (such as electrical or optical layer cross-connect nodes) in the OTN network. To this end, the module sends standard signaling messages based on GMPLS (Generalized Multiprotocol Label Switching) to relevant network nodes to establish end-to-end low-latency optical transmission paths. In path selection, the system prioritizes allocating low-latency transmission paths to high-priority services of P1 and P2 levels. Specific strategies include selecting shorter paths with fewer physical hops, links with lower load rates, or channels with better historical transmission quality, thereby ensuring strict latency performance requirements at the routing level.

[0120] Before transmission, service data is encapsulated into an OTN frame structure conforming to the ITU-T G.709 standard and strictly mapped to the time slots specified by the scheduling decision for transmission. During service transmission, this module continuously monitors key physical layer performance parameters, including optical power and bit error rate. Once any parameter exceeds a preset threshold (e.g., optical power is lower than the receiver sensitivity or bit error rate is higher than the tolerance threshold), it is determined that the transmission quality has deteriorated, and a rerouting mechanism is automatically triggered: a new scheduling request is initiated, an alternative path is calculated based on the latest global network state information, and the reliability and quality of service of service transmission are quickly restored.

[0121] Through the aforementioned refined mapping, signaling control, path optimization, and real-time monitoring mechanisms, the time slot mapping and transmission module not only efficiently implements intelligent scheduling decisions but also provides end-to-end deterministic low-latency and high-reliability transmission guarantees for high-priority services.

[0122] Step S105: Resource Compensation and Rescheduling

[0123] To ensure long-term fairness in system scheduling, the scheduler implements a complete resource compensation mechanism for preempted low-sensitivity services (such as P3 or P4 level services). This mechanism includes the following three closely linked core components:

[0124] First, during the compensation recording phase, the scheduler accurately records the amount of resources lost by the interrupted service while performing the preemption operation, including key parameters such as the number of time slots preempted, the actual duration of occupation, and the original scheduling context. Then, a dedicated resource compensation identifier is added to the service flow as proof of its subsequent priority rescheduling rights, and it is included in the compensation management queue to ensure that the compensation behavior is traceable and quantifiable.

[0125] Secondly, in the compensation time slot allocation phase, the system proactively allocates equivalent transmission resources to services with compensation flags in subsequent scheduling cycles. Specifically, the scheduler prioritizes selecting off-peak periods with lower network loads for these services, or allocates time slot resources equivalent to the original loss from a pre-set dedicated compensation time slot pool. During this process, the system follows dynamic coordination rules: if a newly arrived ultra-high sensitivity service (such as P1 level) urgently needs resources at the time of compensation execution, the current compensation allocation is appropriately delayed to prioritize the real-time performance of the highest priority service; otherwise, the compensation operation is executed immediately to avoid unnecessary waiting.

[0126] Finally, to prevent low-priority services from falling into a "service starvation" state due to frequent preemption, the system activates a temporary priority boosting mechanism during the compensation phase. Compensated services are given a temporary priority higher than their original level during rescheduling (e.g., P4 temporarily boosted to P2), thus gaining a reasonable service opportunity in the competition. Furthermore, the scheduler periodically evaluates the compensation effect. For services that are preempted multiple times within a unit of time, their initial sensitivity level or base priority is dynamically increased to reduce the probability of preemption from the source, achieving dynamic fairness and service quality stability in the long-term operation of the system.

[0127] Through the aforementioned three-in-one compensation mechanism of recording, allocation, and rescheduling, this solution effectively safeguards the basic service rights of low-priority services while supporting the low-latency requirements of high-priority services, ensuring a sustainable balance between efficiency and fairness in the entire optical transport network scheduling system.

[0128] Figure 3The diagram shows the architecture of the low-latency Optical Transport Network (OTN) service scheduling system provided in this embodiment. The system is designed around a closed-loop architecture of "perception-decision-execution-feedback," aiming to completely solve the problems of head-of-line congestion and time slot resource rigidity in traditional OTN scheduling. The system uses a core intelligent scheduler as its central hub, coordinating five functional units: a service classification and marking module, a network status monitoring module, a time slot mapping and transmission module, a scheduling signaling module, and a performance monitoring and feedback module, all tightly integrated with the OTN network transport layer. After a service flow enters from an edge node, the service classification and marking module first performs fine-grained identification based on service type or QoS parameters: latency sensitivity is divided into four levels (ultra-high ≤10ms, high 20–50ms, normal ≤100ms, low unconstrained), and a 256-byte threshold is used to distinguish between short and long packets. Simultaneously, a sliding window mechanism is introduced to dynamically monitor the actual transmission latency; if it continuously exceeds the limit, the sensitivity level is automatically increased. Finally, the "latency sensitivity - packet length" feature is embedded in the packet header or transmitted through out-of-band signaling. Meanwhile, the network status monitoring module collects data such as port queue depth, link throughput and utilization, and ODUk time slot occupancy status of all nodes in the network at a 100-microsecond cycle. It encodes the time slot idle distribution into a bitmap, quantifies the node load into a percentage, and generates and updates a global network status view every millisecond. This view is then supplied to the scheduler in real time through standardized interfaces such as NETCONF, ensuring that decisions are based on the latest and most comprehensive network situation.

[0129] The core intelligent scheduler, acting as the system's "brain," integrates the two types of inputs mentioned above to execute multi-dimensional intelligent decisions: First, it maps services to a four-priority system of P1–P4 (P1 represents ultra-high sensitivity short messages, P2 covers ultra-high sensitivity long messages or high sensitivity short messages, P3 corresponds to high sensitivity long messages and ordinary sensitivity services, and P4 represents low sensitivity services). Then, it evaluates the transmission value of all available time slots (comprehensively considering service urgency, remaining lifetime, and link transmission quality), prioritizing the allocation of the earliest available high-value idle time slots for P1 / P2 services. When there are no idle resources, it activates an intelligent preemption mechanism based on "net latency gain > 0"—preemption is only authorized when the latency gain obtained by high-priority services is significantly greater than the latency loss of preempted low-priority services, and the preemption threshold is dynamically relaxed when high-priority services are nearing timeout to ensure critical SLA. Meanwhile, the system implements a complete resource compensation mechanism for preempted services: accurately recording the number of time slots and the duration of occupation, adding a dedicated compensation identifier, and allocating equivalent resources in subsequent scheduling cycles with priority over off-peak periods or in a dedicated compensation time slot pool; if a newly arrived ultra-sensitive service urgently needs resources, compensation is appropriately delayed; otherwise, it is executed immediately; in addition, compensated services receive a temporary priority boost during rescheduling to avoid starvation, and the initial priority of services that have been preempted multiple times is dynamically increased to maintain long-term fairness. The instructions generated by the scheduler are distributed through a dual-path system: the blue data stream is sent to the time slot mapping and transmission module for data plane operations, and the pink control signaling is transmitted to distributed nodes via the scheduling signaling module (based on protocols such as GMPLS) to achieve cross-domain collaborative control.

[0130] The time slot mapping and transmission module, acting as the execution unit, receives scheduling instructions and precisely maps service data to the designated ODUk time slots via the OTN cross-connect equipment. In preemption scenarios, it first saves the transmission context (including mapping status and progress) of the original low-priority services, then completes the resource reallocation of high-priority services, prioritizing end-to-end optical paths with fewer physical hops and lower load to enhance low-latency assurance. Simultaneously, it continuously monitors physical layer parameters such as optical power and bit error rate, automatically triggering a rerouting mechanism upon degradation. All services ultimately achieve end-to-end transmission via the OTN network transport layer, which acts as the physical channel to synchronously carry data streams and control signaling. The performance monitoring and feedback module continuously collects key indicators such as measured end-to-end latency, jitter, preemption success rate, time slot utilization, and compensation effect, forming a green feedback loop that is sent to the core intelligent scheduler and network status monitoring module, driving adaptive optimization of scheduling strategies, marking rules, and state models. The entire system, through refined business marking, dynamic preemption driven by quantitative revenue, multi-level fair compensation mechanisms, and closed-loop feedback, not only fundamentally eliminates queue head congestion and provides millisecond-level or even sub-millisecond-level deterministic low latency guarantees for scenarios such as industrial control, high-frequency financial trading, autonomous driving, 5G uRLLC, cloud gaming, and remote surgery, but also significantly improves the dynamic reuse efficiency of optical channel resources. It truly achieves synergistic optimization of low latency performance and high resource utilization, providing a highly reliable and efficient OTN bearer solution for the large-scale deployment of services with stringent latency requirements.

[0131] The technical solution of this embodiment is illustrated by a specific example below:

[0132] Suppose that in an OTN network, there are two types of services waiting to be transmitted simultaneously: Service A is a remote control command, which is of ultra-high sensitivity and short message type; Service B is video return data, which is of normal sensitivity and long message type. Under the traditional scheduling mechanism, if Service B has already occupied a certain optical channel data unit (ODUk) time slot, even if Service A arrives subsequently and is extremely sensitive to latency, it must wait for Service B to complete the transmission of the entire long message, resulting in a significant increase in the actual latency of Service A, which may not meet its millisecond-level response requirements. In the scheduling method adopted in this embodiment, when Service A enters the network, the service classification and marking module identifies it as a P1 level (highest priority) service; the core intelligent scheduler, combined with global network status information, finds that there are currently no idle high-value time slots, but the time slot occupied by Service B has not yet started actual data transmission (or is in an interruptible state). The scheduler then initiates a preemption assessment process, calculating the net latency benefit of reallocating the time slot to service A. Since service A uses short messages with a very strict latency threshold (≤10ms), while service B has a higher tolerance (≤100ms), the assessment result satisfies the condition of "net latency benefit > 0". Therefore, the scheduler immediately generates a preemption command, dynamically remapping the time slot resource to service A, enabling near-zero-wait rapid transmission and effectively ensuring the real-time performance of control commands. Simultaneously, the system automatically records the number and duration of time slots preempted for service B and marks them with a compensation tag. In subsequent scheduling cycles, the scheduler prioritizes allocating equivalent resources to service B from low-load periods or a dedicated compensation time slot pool, ensuring its transmission integrity is unaffected. Through this "precise preemption + fair compensation" mechanism, both the deterministic low-latency requirements of critical services are met, and the service fairness of ordinary services is maintained, fully demonstrating the synergistic optimization capability of this solution between performance and efficiency.

[0133] This embodiment provides a service scheduling method for low-latency Optical Transport Networks (OTN), aiming to effectively solve the problems of uncontrollable latency and low network resource utilization caused by head-of-line congestion and static resource allocation in traditional OTN scheduling mechanisms. This provides reliable support for the large-scale deployment of millisecond-level and even sub-millisecond-level deterministic latency services. The core of this method lies in constructing an end-to-end intelligent time-slot scheduling guarantee mechanism, which is collaboratively implemented by a service feature identification module, a network status acquisition module, a core intelligent scheduler (i.e., an intelligent decision-making module), and a time-slot mapping and execution module. These modules interact efficiently with each other through standardized interface protocols (such as NETCONF, GMPLS, etc.) for data exchange and command transmission. During the service access phase, service flows are finely classified using "time slot sensitivity identifiers": based on the service type, their latency sensitivity level is automatically marked (divided into four levels: ultra-high, high, normal, and low, corresponding to end-to-end latency thresholds of ≤10ms, 20–50ms, ≤100ms, and no strict requirements, respectively), and combined with message length characteristics (using 256 bytes as a threshold to distinguish between short and long messages), a structured service feature fingerprint is formed; high-sensitivity services can request to be prioritized and scheduled to the earliest available high-value time slot based on this identifier, thereby achieving near-instantaneous transmission and avoiding uncontrollable latency caused by waiting for long messages at the front of the queue to complete. This not only significantly improves the response speed of critical services but also frees up valuable bandwidth resources to process more high-value traffic. The scheduling decision-making process adopts a dual-input model of "service characteristics - network status": the core intelligent scheduler receives sensitivity level and packet length information from the service side in real time, and integrates a global status view from the network side (including the queue depth of each node port, link throughput and utilization, ODUk time slot occupancy bitmap, etc.). Through multi-dimensional algorithms, it comprehensively evaluates the urgency and transmission quality of time slots and dynamically generates the optimal time slot allocation strategy. The entire scheduling process relies on a dynamic real-time collaborative mechanism to tightly couple service scheduling decisions, time slot resource allocation and network status monitoring, ensuring that the utilization efficiency of optical channel resources is maximized while strictly guaranteeing ultra-low latency. When no idle time slots are available for high-priority services, an intelligent preemption and fair compensation mechanism is activated: The core intelligent scheduler generates a time slot preemption instruction based on the quantitative criterion of "net latency gain > 0" (i.e., the latency gain obtained by high-priority services is significantly greater than the latency loss borne by low-priority services). The time slot resources that have been occupied by low-priority services but have not yet entered the critical transmission stage are remapped to high-priority services. At the same time, a compensation mark is automatically applied to the preempted services, and equivalent resources are allocated to them in subsequent scheduling cycles from network off-peak periods or a dedicated compensation time slot pool, effectively preventing service starvation and maintaining long-term scheduling fairness.Therefore, this embodiment achieves the organic unity of deterministic low latency guarantee and high resource utilization through four major technical pillars: intelligent identification, dynamic collaboration, precise preemption and closed-loop compensation. It provides highly reliable and efficient OTN carrying capacity for application scenarios with stringent latency requirements, such as 5G URLLC, industrial remote control, high-frequency financial trading, autonomous driving, cloud gaming and remote surgery.

[0134] Example 2:

[0135] like Figure 4 As shown, this embodiment provides an optical transport network service scheduling device, which includes:

[0136] The identification unit 10 is used to identify the latency requirement threshold and message length of the service flow in response to the service flow accessing the optical transport network.

[0137] The processing unit 20, connected to the identification unit 10, is used to determine the transmission priority of the service flow based on the latency requirement threshold and the message length, and to initiate the corresponding service transmission request; the transmission priority includes a first priority, a second priority, and a third priority, with the first priority being higher than the second priority, and the second priority being higher than the third priority;

[0138] The determining unit 30, connected to the processing unit 20, is used to respond to a service transmission request and determine a scheduling decision based on transmission priority and global network status information. The global network status information is generated by collecting the queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot at a preset period.

[0139] The scheduling decision includes: when it is necessary to allocate optical channel data unit time slots for a first-priority service flow, and there are no idle optical channel data unit time slots that meet the scheduling conditions, a candidate time slot is selected from the time slots currently occupied by the second-priority or third-priority service flow; the net delay benefit of reassigning the candidate time slot from the currently occupied second-priority or third-priority service flow to the first-priority service flow is calculated; a preemption instruction for the candidate time slot is generated if and only if the calculated net delay benefit is greater than zero; the preemption instruction is used to indicate the interruption of the transmission of the originally occupied second-priority or third-priority service flow on the candidate time slot, and to remap the resources of the candidate time slot to the first-priority service flow;

[0140] The mapping transmission unit 40, connected to the determining unit 30, is used to map the data of the service flow to the allocated optical channel data unit time slot according to the scheduling decision, and complete the transmission through the optical transport network to realize the optical transport network service scheduling.

[0141] The apparatus in this embodiment is capable of performing all the methods described in the embodiment.

[0142] Example 3:

[0143] like Figure 5 As shown, this embodiment provides an electronic device, including a memory 200 and a processor 100. The memory 200 stores a computer program. When the processor 100 runs the computer program stored in the memory 200, the processor 100 executes the optical transport network service scheduling method according to Embodiment 1.

[0144] It is understood that the above embodiments are merely exemplary implementations used to illustrate the principles of the present invention, and the present invention is not limited thereto. For those skilled in the art, various modifications and improvements can be made without departing from the spirit and essence of the present invention, and these modifications and improvements are also considered to be within the scope of protection of the present invention.

Claims

1. A service scheduling method for an optical transport network, characterized in that, The method includes: In response to a service flow accessing the optical transport network, the latency requirement threshold and message length of the service flow are identified; Based on the latency requirement threshold and the message length, the transmission priority of the service flow is determined, and a corresponding service transmission request is initiated; the transmission priority includes a first priority, a second priority, and a third priority, with the first priority being higher than the second priority, and the second priority being higher than the third priority; In response to the service transmission request, a scheduling decision is determined based on the transmission priority and global network status information; the global network status information is generated by collecting the queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot at a preset period. The scheduling decision includes: when it is necessary to allocate optical channel data unit time slots for a first-priority service flow, and there are no idle optical channel data unit time slots that meet the scheduling conditions, selecting a candidate time slot from the time slots currently occupied by a second-priority or third-priority service flow; calculating the net delay benefit of reassigning the candidate time slot from the currently occupied second-priority or third-priority service flow to the first-priority service flow; generating a preemption instruction for the candidate time slot if and only if the calculated net delay benefit is greater than zero; the preemption instruction is used to instruct the interruption of the transmission of the originally occupied second-priority or third-priority service flow on the candidate time slot, and to remap the resources of the candidate time slot to the first-priority service flow; Based on the scheduling decision, the data of the service flow is mapped to the allocated optical channel data unit time slot, and the transmission is completed through the optical transport network to realize optical transport network service scheduling.

2. The optical transport network service scheduling method according to claim 1, characterized in that, The method further includes: If the scheduling decision includes a preemption operation, a resource compensation identifier is generated for the preempted service flow; A compensation record entry is created for the preempted service flow with the resource compensation identifier, and resources are allocated to the preempted service flow first when the network load is lower than a preset threshold or when a dedicated compensation time slot pool is available.

3. The optical transport network service scheduling method according to claim 1, characterized in that, The net delay gain ΔD is calculated as follows: ΔD = D_high_reduced – D_low_increased; where D_high_reduced is the expected reduction in end-to-end transmission delay for the first priority service flow due to obtaining the candidate occupied time slot, and D_low_increased is the expected increase in end-to-end transmission delay for the preempted second or third priority service flow due to losing the candidate occupied time slot.

4. The optical transport network service scheduling method according to claim 1, characterized in that, The step of determining the transmission priority of the service flow based on the latency requirement threshold and the message length specifically includes: If the latency requirement threshold of the service flow is not greater than the first threshold, and the message length of the service flow is not greater than the preset length threshold, then the transmission priority of the service flow is determined to be the first priority. If the latency requirement threshold of the service flow is not greater than the first threshold, and the message length of the service flow is greater than the preset length threshold, then the transmission priority of the service flow is determined to be the second priority. If the latency requirement threshold of the service flow is greater than the first threshold, then the transmission priority of the service flow is determined to be the third priority.

5. The optical transport network service scheduling method according to claim 1, characterized in that, The scheduling decision further includes setting the threshold value for determining the net delay benefit to be lower than zero when the remaining survival time of the first priority service flow is less than or equal to a preset time window.

6. The optical transport network service scheduling method according to claim 1, characterized in that, The global network state information is generated in the following way: The queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot are collected at a first preset period. The queue status is converted into node load data expressed as a percentage, and the occupancy status is represented as a time slot resource occupancy bitmap; The node load data and the time slot resource occupancy bitmap are integrated at a second preset period to generate global network status information. The second preset period is shorter than the first preset period.

7. The optical transport network service scheduling method according to claim 1, characterized in that, The method further includes: During service transmission, the optical power and bit error rate of the allocated optical channel are monitored in real time. When the optical power is lower than the preset power threshold or the bit error rate is higher than the preset bit error threshold, the rerouting mechanism is triggered to re-execute the scheduling decision based on global network status information.

8. The optical transport network service scheduling method according to any one of claims 1 to 7, characterized in that, After determining the transmission priority of the service flow, the method further includes: The end-to-end transmission latency of the service flow is continuously monitored using a sliding window mechanism; if the end-to-end transmission latency exceeds the corresponding latency requirement threshold in multiple consecutive monitoring periods, the transmission priority of the service flow is increased by one level.

9. A service scheduling device for an optical transport network, characterized in that, The device includes: The identification unit is used to identify the latency requirement threshold and message length of the service flow in response to the service flow accessing the optical transport network. The processing unit, connected to the identification unit, is used to determine the transmission priority of the service flow based on the latency requirement threshold and the message length, and to initiate a corresponding service transmission request; the transmission priority includes a first priority, a second priority, and a third priority, wherein the first priority is higher than the second priority, and the second priority is higher than the third priority; A determining unit, connected to the processing unit, is used to respond to the service transmission request and determine a scheduling decision based on the transmission priority and global network status information. The global network status information is generated by collecting the queue status of each network node in the optical transport network and the occupancy status of the optical channel data unit time slot at a preset period. The scheduling decision includes: when it is necessary to allocate optical channel data unit time slots for a first-priority service flow, and there are no idle optical channel data unit time slots that meet the scheduling conditions, selecting a candidate time slot from the time slots currently occupied by a second-priority or third-priority service flow; calculating the net delay benefit of reassigning the candidate time slot from the currently occupied second-priority or third-priority service flow to the first-priority service flow; generating a preemption instruction for the candidate time slot if and only if the calculated net delay benefit is greater than zero; the preemption instruction is used to instruct the interruption of the transmission of the originally occupied second-priority or third-priority service flow on the candidate time slot, and to remap the resources of the candidate time slot to the first-priority service flow; The mapping transmission unit, connected to the determining unit, is used to map the data of the service flow to the allocated optical channel data unit time slot according to the scheduling decision, and complete the transmission through the optical transport network to realize optical transport network service scheduling.

10. An electronic device comprising a memory and a processor, characterized in that, The memory stores a computer program. When the processor runs the computer program stored in the memory, the processor executes the optical transport network service scheduling method according to any one of claims 1 to 8.

Images