A Multi-Service Broadcast Optical Reception Scheduling Method and System Based on Wavelength Allocation

Abstract

This application discloses a multi-service broadcast optical reception scheduling method and system based on wavelength allocation, relating to the field of optical network communication technology. The method includes: for broadcast WDM multi-service concurrent reception scenarios, jointly modeling the service request category, time limit, and data size constraints with wavelength resource occupancy status and interference risk; firstly, dynamically grouping available wavelengths based on service category load prediction to achieve adaptive resource boundaries according to service attributes; then constructing a comprehensive priority index including the latest reception time limit, service category, and interference risk of candidate wavelengths, performing ordered cross-priority scheduling, and arranging reception time slots on the target wavelength to generate a wavelength scheduling table and a reception scheduling table; finally, driving the receiver to tune and receive according to time slots using scheduling commands. This improves wavelength utilization while reducing reception conflicts and latency jitter and strengthening critical service assurance.

Description

Technical Field

[0001] This application relates to the field of optical network communication technology, and in particular to a multi-service broadcast optical reception scheduling method and system based on wavelength allocation. Background Technology

[0002] As optical networks evolve towards multi-wavelength, low-latency, and multi-service capabilities, WDM (Wavelength Division Multiplexing) significantly improves capacity by using multiple wavelengths as parallel optical channels. However, the services exhibit strong heterogeneity, including both latency-sensitive control services and high-bandwidth streaming media / data distribution services. Different services face significantly different constraints on latency, bandwidth, and reliability when sharing limited wavelength resources. Existing scheduling strategies are often based on assumptions of a single service or an "average service model," leading to insufficient QoS (Quality of Service) guarantees when multiple services coexist.

[0003] In broadcast WDM networks, optical signals from different directions are typically broadcast to each node via star couplers. Nodes rely on fixed or tunable optical devices to select the receiving wavelength. Limited by the number / range of wavelengths each node can simultaneously receive, a lack of fine-grained scheduling can easily lead to wavelength conflicts, latency jitter, and low channel idle rates when multiple services are broadcast concurrently. Meanwhile, the "first-come, first-served" allocation and unified scheduling rules commonly found in existing solutions often lack explicit modeling of service types and priority differences, resulting in insufficient critical service assurance capabilities and unsatisfactory resource utilization. Summary of the Invention

[0004] This application provides a multi-service broadcast optical reception scheduling method, system, storage medium, computer program product, and electronic device based on wavelength allocation, which at least solves the problems of poor service quality and low network resource utilization efficiency in multi-service concurrent scenarios in current related technologies.

[0005] In a first aspect, embodiments of this application provide a multi-service broadcast optical reception scheduling method based on wavelength allocation, applied to the scheduling control end of a broadcast WDM optical network. The method includes: acquiring a set of service requests reported by multiple service initiating nodes within the current scheduling period, wherein each service request at least includes arrival time, data size, service category, and latest reception time limit; acquiring network status data of the broadcast WDM optical network, wherein the network status data at least includes wavelength status information of available wavelength resources, and the wavelength status information at least includes the occupancy status of each wavelength within the current scheduling period; determining the load prediction result for each service category within the current scheduling period based on the service request set, and dynamically allocating the available wavelength resources according to the load prediction result to generate corresponding wavelengths for different services. The system categorizes wavelength groups; for each service request in the service request set, it calculates a comprehensive priority index for the service request based on the latest reception time limit, the data size, the service category, and the interference risk corresponding to the candidate wavelength; it sorts each service request according to the comprehensive priority index, performs ordered cross-priority scheduling based on the sorting result, allocates a target wavelength to each service request, and allocates a reception time slot for each service request on the target wavelength to generate a wavelength scheduling table and a reception scheduling table; wherein, the reception time slot is used to avoid reception conflicts on the same wavelength and meet the latest reception time limit; it generates a broadcast scheduling instruction according to the wavelength scheduling table and the reception scheduling table and sends it to the relevant nodes to control the receivers of the receiving nodes to tune to the target wavelength for data reception within the reception time slot.

[0006] Secondly, embodiments of this application provide a multi-service broadcast optical reception scheduling system based on wavelength allocation, deployed at the scheduling control end of a broadcast WDM optical network. The system includes: a service request aggregation unit, used to acquire a set of service requests reported by multiple service initiating nodes within the current scheduling period, wherein the service requests at least include arrival time, data size, service category, and latest reception time limit; a network status awareness unit, used to acquire network status data of the broadcast WDM optical network, wherein the network status data at least includes wavelength status information of available wavelength resources, and the wavelength status information at least includes the occupancy status of each wavelength within the current scheduling period; and a wavelength group partitioning unit, used to determine the load prediction results for each service category within the current scheduling period based on the service request set, and dynamically partition the available wavelength resources according to the load prediction results to generate groups corresponding to different service categories. Other wavelength groups; a priority evaluation unit, used to calculate a comprehensive priority index for each service request in the service request set based on the latest reception time limit, the data size, the service category, and the interference risk corresponding to the candidate wavelength; a resource scheduling table generation unit, used to sort each service request according to the comprehensive priority index, perform ordered cross-priority scheduling based on the sorting result, allocate target wavelengths to each service request, and allocate reception time slots to each service request on the target wavelengths to generate a wavelength scheduling table and a reception scheduling table; wherein, the reception time slots are used to avoid reception conflicts on the same wavelength and meet the latest reception time limit; a broadcast scheduling instruction distribution unit, used to generate broadcast scheduling instructions according to the wavelength scheduling table and the reception scheduling table and send them to relevant nodes to control the receivers of the receiving nodes to tune to the target wavelength for data reception within the reception time slots.

[0007] Thirdly, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform the steps of the wavelength allocation-based multi-service broadcast optical reception scheduling method of any embodiment of this application.

[0008] Fourthly, embodiments of this application provide a storage medium storing a computer program thereon, characterized in that, when the program is executed by a processor, it implements the steps of the multi-service broadcast optical reception scheduling method based on wavelength allocation according to any embodiment of this application.

[0009] Fifthly, embodiments of this application provide a computer program product, including a computer program / instructions, which, when executed by a processor, implement the steps of the wavelength allocation-based multi-service broadcast optical reception scheduling method of any embodiment of this application.

[0010] The multi-service broadcast optical reception scheduling method and system based on wavelength allocation provided in this application can achieve at least the following technical effects:

[0011] (1) Based on the load prediction of service categories, the available wavelength resources are dynamically divided into wavelength groups corresponding to different service categories. This enables the scheduling control terminal to establish an explicit mapping relationship between "service type - resource pool" within the same scheduling cycle, so that the wavelength supply and service structure can be adaptively matched on the cycle scale. When the load of a certain service increases, the size of its available wavelength group is adjusted accordingly, thereby reducing the competition density of different service categories in the same wavelength set, reducing waiting and fragmented occupation caused by resource mismatch, and thus improving the continuity and effective utilization of wavelength carrying capacity under multi-service concurrency conditions. It also provides a preliminary resource constraint basis for subsequent fine-grained scheduling based on service differences.

[0012] (2) Based on the wavelength group, a comprehensive priority index is introduced that simultaneously characterizes the latest reception deadline, data size, service category, and interference risk of candidate wavelengths. Scheduling is performed using an ordered cross-priority approach, which allows scheduling decisions to evolve from a single ranking to a joint drive of "urgency-resource adaptability." The latest reception deadline makes it easier to constrain service completion time within a controllable window; the combined effect of data size and service category helps achieve a more robust scheduling balance among services of different granularities and sensitivities; and the involvement of interference risk guides services to select more available target wavelengths to reduce uncertainty. Simultaneously, the reception time slot explicitly schedules reception behavior on the same wavelength in the time dimension, enabling reception conflicts under concurrent broadcasting to be avoided or significantly compressed at the scheduling table level. This improves the time-limit fulfillment capability of critical services, reduces receiver latency jitter, and enhances overall reception stability without increasing uncertainty negotiation at the end-side.

[0013] This technical solution uses dynamic grouping of wavelength resources driven by service load prediction as the resource layer foundation, and comprehensive priority-based wavelength / time slot joint scheduling oriented towards time constraints and interference risks as the decision-making layer means to model and implement the differences among multiple services in a consistent manner from resource organization to scheduling execution. As a result, the resulting wavelength scheduling table and receiver scheduling table can directly constrain the receiver to tune to the target wavelength within a specific time slot to complete reception in a tabular and executable manner. This enables the scheduling process to simultaneously achieve adaptive resource allocation, guaranteed key constraints, and controllable reception conflicts in multi-service concurrent scenarios. Attached Figure Description

[0014] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0015] Figure 1 A flowchart illustrating an example of a wavelength allocation-based multi-service broadcast optical reception scheduling method according to an embodiment of this application is shown.

[0016] Figure 2 This document illustrates an example of allocating target wavelengths to each service request via ordered cross-priority scheduling according to an embodiment of this application.

[0017] Figure 3 A flowchart illustrating an example of performing wavelength fragmentation defragmentation operations according to an embodiment of this application is shown.

[0018] Figure 4 A flowchart illustrating the operational mechanism of an example of a wavelength allocation-based multi-service broadcast optical reception scheduling method according to an embodiment of this application is shown.

[0019] Figure 5 A schematic diagram showing the comparison of the average latency of different scheduling algorithms as a function of load is presented.

[0020] Figure 6 A schematic diagram showing the comparison of wavelength utilization as a function of load for different scheduling algorithms is presented.

[0021] Figure 7 A heatmap showing the average latency ratio of different baseline scheduling algorithms as a function of load is illustrated in the embodiments of this application.

[0022] Figure 8 A structural block diagram of an example of a wavelength allocation-based multi-service broadcast optical reception scheduling system according to an embodiment of this application is shown. Detailed Implementation

[0023] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0024] It should be noted that in the research and engineering practice of multi-wavelength access and distribution networks, a common approach in current technologies is to borrow the dynamic bandwidth allocation mechanism from systems such as Ethernet passive optical networks. Within a scheduling cycle, the control end, based on the queue / request information reported by each terminal, divides and allocates available transmission resources to different terminals or service flows according to certain rules. Some studies have proposed strategies such as interleaved polling, sequential filling, and water-filling allocation to reduce the computational complexity of resource configuration and the overhead caused by frequent switching. Other methods tend to concentrate the services of the same terminal onto a single wavelength to reduce switching latency, or treat multiple wavelengths as a unified capacity pool and allocate them periodically. While this type of mechanism is feasible in point-to-multipoint uplink bandwidth management, its focus is more on "how bandwidth shares are divided within a cycle." It often lacks a modeling expression coupled with physical receiving constraints to address issues in broadcast WDM architectures such as wavelength contention caused by multi-source concurrent broadcasting, limited simultaneous reception capabilities at the receiving end, and the need for explicit orchestration of receiving actions.

[0025] Regarding the dynamic allocation of multi-wavelength resources, some studies have attempted to introduce prediction and learning methods to improve resource matching capabilities under high load and fluctuating traffic. For example, some scholars have used time-series prediction networks to estimate the traffic or queue growth trend of services within a short future time window, and based on the prediction results, applied differentiated allocation rules to services of different priorities, thereby improving the latency performance of high-priority services under certain conditions. These methods share the common characteristic of using "prediction-allocation" as the core closed loop, emphasizing training and fitting capabilities based on historical samples, and achieving bias protection through priority differentiation during the resource allocation phase. However, because the adaptability of prediction models to sudden traffic and drastic short-cycle changes is affected by sample coverage and update frequency, and their resource decisions mainly focus on "how to allocate available bandwidth," their coverage of aspects such as concurrent tuning at the receiver side, wavelength conflict avoidance, and receiver timing organization in broadcast mode is relatively limited.

[0026] In broadcast scenarios with high deterministic requirements, some studies have focused on broadcast-selective WDM networks, proposing a scheduling framework that combines multi-channel round-robin and reservation. This framework maps periodic messages to a fixed round-robin sequence across multiple channels, reduces bandwidth fragmentation through parameterized round-robin periods and weight configurations, and utilizes the remaining bandwidth of periodic messages for dynamic reservation of non-periodic messages. This type of scheme improves scheduling predictability through a structured round-robin mechanism and reduces the impact of tuning delay on determinism by reducing or eliminating tunable devices. However, it typically relies on relatively known periodic message parameters and a relatively stable scheduling rhythm. When service types are more diverse and arrival characteristics are more random, maintaining predictability while considering the differentiated constraints of various services still requires more refined resource organization and receiver-side behavior orchestration mechanisms.

[0027] Furthermore, regarding the wavelength contention problem in WDM broadcast star LANs, some studies have proposed randomized scheduling and retransmission control approaches: under a single tunable receiver, strategies such as continuous retransmission or randomized delayed retransmission are used to statistically reduce the probability of collisions, and the impact of random delay on throughput and stability is analyzed. Other works discuss using a centralized controller to uniformly control wavelength selection to reduce collisions, or employing reservation protocols to avoid collisions. However, reservation mechanisms often require receivers to cooperate in meeting availability conditions, leading to a trade-off between receiver utilization efficiency and system latency. These works provide valuable insights from the perspective of media access and collision control. However, under conditions of multiple services coexisting and significantly different service constraints, there is still considerable room for engineering integration in unifying service attributes, priorities, and physical layer wavelength / receiver constraints into the scheduling decision framework.

[0028] Therefore, establishing a consistent and feasible collaborative mechanism between wavelength resource organization, service differentiation constraint expression, and receiver-side executable scheduling remains a key issue that needs further improvement.

[0029] It should be understood that the above description of the relevant technologies is intended only to help the public better understand the inventive spirit and motivation of this application, and is not intended to limit this application. Furthermore, the technical solutions described in the above-mentioned relevant technologies are not prior art, and may also be undisclosed technical solutions, such as those under research or in the laboratory stage.

[0030] The technical solutions in this application, including the collection, storage, use, processing, transmission, provision, and disclosure of users' personal information, comply with relevant laws and regulations and do not violate public order and good morals.

[0031] Figure 1 A flowchart illustrating an example of a wavelength allocation-based multi-service broadcast optical reception scheduling method according to an embodiment of this application is shown.

[0032] Regarding the execution subject of the method in the embodiments of this application, it can be any controller or processor with computing or processing capabilities. It can be the controller in the scheduling control terminal of a broadcast WDM optical network. Based on "load prediction-driven dynamic configuration of wavelength groups" as the resource layer foundation, and with "comprehensive priority constrained by cutoff constraints and interference risks + time-slotted reception orchestration" as the scheduling layer kernel, it realizes multi-service broadcast reception from coarse-grained allocation to constrained and predictable collaborative scheduling. It can more stably balance the time limit guarantee of critical services and the efficiency of wavelength resource utilization, and improve the determinism and sustainable carrying capacity of system-level scheduling.

[0033] In some examples, it may be integrated into an electronic device or terminal through software, hardware, or a combination of both, and the type of terminal or electronic device may be diverse.

[0034] like Figure 1 As shown, in step S110, a set of service requests reported by multiple service initiating nodes within the current scheduling period is obtained. The service requests include at least the arrival time, data size, service category, and latest reception time limit.

[0035] In some implementations, the scheduling control terminal collects service requests reported by multiple service initiating nodes through an independent control channel or management plane interface within a predetermined reporting window of each scheduling cycle, and aggregates them into a request set for the current cycle. To ensure subsequent scheduling can be executed, each service request carries at least the arrival time, data size, service category, and latest reception deadline. The arrival time reflects the order in which services enter the system, the data size characterizes its wavelength occupancy requirements, the service category reflects the differences in service emphasis on latency / bandwidth / reliability, and the latest reception deadline forms a hard or semi-hard completion boundary.

[0036] In addition, after receiving a request, the scheduling control terminal can perform necessary normalization processing, such as uniform clock domain calibration of timestamps, enumeration mapping of service categories, verification and error correction of abnormal fields (missing, out-of-bounds, duplicate request IDs, etc.), and write the request into the corresponding scheduling queue according to the service category, while retaining the association information between the request and the source node and the target receiving node set (if it is a broadcast / multicast service).

[0037] In step S120, network status data of the broadcast WDM optical network is obtained. The network status data includes at least wavelength status information of available wavelength resources, and the wavelength status information includes at least the occupancy status of each wavelength in the current scheduling period.

[0038] Here, the wavelength status information reflects at least the occupancy status of each wavelength in the current scheduling cycle, such as whether it is idle, whether it has been reserved by the scheduling result of the previous cycle, whether it is in protection / maintenance status, and the occupancy time range in the cycle.

[0039] In addition, in broadcast WDM scenarios, in order to match scheduling with receiver capabilities, network status data can also reflect the receiver capability constraints of receiving nodes, such as the number of receivers that a node can operate simultaneously, whether the receiver is a tunable device, the tunable range / step limit, and the settling time required for tuning switching. This information can be provided by the nodes periodically reporting or by the device capability library maintained by the management plane.

[0040] In step S130, the load prediction results of each service category in the current scheduling period are determined based on the service request set, and the available wavelength resources are dynamically divided according to the load prediction results to generate wavelength groups corresponding to different service categories.

[0041] In some implementations, the scheduling control unit, based on the aggregated set of service requests, statistically analyzes the demand intensity within the current scheduling cycle according to service category, and combines this with recent cycle arrival trends to form load forecasting results for each category. Load forecasting can be performed in various unrestricted ways and can be accomplished using methods that are currently easy to implement in engineering, such as accumulating service data volume by category, accumulating estimated duration of use, or simultaneously considering the number of requests and average size. A smoothing mechanism can also be introduced to avoid frequent fluctuations in resource allocation caused by sudden bursts in a single cycle.

[0042] Subsequently, the scheduling and control terminal dynamically divides the available wavelength resources based on the prediction results, forming wavelength groups corresponding to different service categories. On the one hand, it can configure wavelength subsets with higher availability and less fragmentation for latency-sensitive or control-related services; on the other hand, it can configure wavelength subsets with wider carrying windows for high-bandwidth or throughput-oriented services. In addition, it can set minimum guaranteed wavelength numbers or priority available bands for critical service categories to ensure that they still have schedulable resource boundaries in highly competitive scenarios.

[0043] By dynamically allocating wavelength resources, the wavelength resources can be adaptively adjusted according to the business structure on a periodic scale, so that different types of businesses can form relatively clear competitive boundaries at the resource level, reducing the indiscriminate competition of multiple types of businesses within the same wavelength set.

[0044] In step S140, for each service request in the service request set, a comprehensive priority index is calculated based on the latest reception time limit, data size, service category, and interference risk corresponding to the candidate wavelength.

[0045] In some implementations, the scheduling control terminal filters a set of candidate wavelengths within the wavelength group corresponding to each service request in the service request set. The selection of candidate wavelengths must at least satisfy the hard constraint that they are "available in the current scheduling period", and can be further filtered in combination with the capability constraints of receiving nodes. For example, some receiving nodes may only support specific band ranges or have a limited number of wavelengths that can be received simultaneously, thereby avoiding the generation of an unexecutable candidate set.

[0046] Subsequently, for each service request, the scheduling control terminal calculates the comprehensive priority index of the service request based on its latest reception time limit, data size, service type and interference risk corresponding to the candidate wavelength. The interference risk can be obtained based on quality observations and historical statistics in network status data, such as crosstalk alarms, bit error trends or congestion conflict statistics, to reflect the stability risk of completing reception on the candidate wavelength.

[0047] The comprehensive priority index unifies the "urgency of the business itself (latest reception deadline)," "the impact of the business on resource consumption (data size)," "differences in business importance (business category)," and "executability risk on different candidate wavelengths (interference risk)" onto the same comparison scale, forming a ranking basis for priority.

[0048] In step S150, each service request is sorted according to the comprehensive priority index, and ordered cross-priority scheduling is performed based on the sorting result. Target wavelengths are allocated to each service request, and receive time slots are allocated to each service request on the target wavelengths to generate wavelength scheduling tables and receive scheduling tables.

[0049] Here, the scheduling control unit sorts service requests based on a comprehensive priority index, but adopts ordered cross-priority scheduling in the actual orchestration process. This ensures that the scheduling process follows priorities while maintaining consistency with the resource allocation of the aforementioned wavelength groups. Specifically, the scheduling control unit can use the "priority sorting result" as the main line, combined with the current available window and allocated progress of the service category wavelength group, to cross-use and table the requests between different service categories. When a category has occupied resources close to its allocated share in the current cycle, the scheduler can switch to processing requests with higher urgency in another category to avoid excessive resource consumption by a single category, which would prevent other categories from completing their arrangements within the available window. At the same time, for critical services that need to be prioritized, they are still allowed to obtain supplementary carrying capacity on available resources in the shared group or other categories to maintain overall availability.

[0050] In some implementations, when allocating target wavelengths and receive time slots for each request, the scheduling control terminal searches and assigns slots based on "executability," ensuring that the receive time slots are used to avoid reception conflicts on the same wavelength and meet the latest reception deadline. Specifically, a wavelength that can provide a continuous available reception window and has a lower interference risk is first selected from the candidate wavelengths as the target wavelength. Then, a specific receive time slot is allocated to the request on that wavelength. The allocation of time slots must not only avoid time overlap of scheduled services on the same wavelength, but also meet the receiving-side constraints, namely, that the number of available receivers at the relevant receiving nodes is sufficient within the time period, the stabilization time and guard interval required for receiver tuning and switching can be met, and that service reception is completed no later than its latest reception deadline. After scheduling is completed, a wavelength scheduling table (describing which services each wavelength carries in a period and the corresponding time period) and a receive scheduling table (describing when each receiving node's receiver tunes to which wavelength and which service in a period) are generated. The two tables correspond to each other to ensure end-to-end consistency.

[0051] By prioritizing, cross-scheduling, and time-slotted scheduling, the scheduler resolves potential conflicts directly during the planning phase. Receive time slots prevent concurrent reception of the same wavelength from overlapping when receivers are limited. Cross-scheduling prioritizes resources to maintain a controllable distribution among multiple services and reduces the risk of unschedulable issues caused by extreme skewness. Meanwhile, cutoff constraints (i.e., the latest reception deadline) are applied throughout the scheduling process, making it easier to keep service completion times within acceptable limits. Overall, this reduces conflicts, jitter, and improves scheduling success rate and resource utilization efficiency.

[0052] In step S160, a broadcast scheduling instruction is generated based on the wavelength scheduling table and the reception scheduling table and sent to the relevant nodes to control the receivers of the receiving nodes to tune to the target wavelength for data reception within the reception time slot.

[0053] In some implementations, broadcast scheduling instructions may be in a structured format. For example, the instruction content may include a scheduling cycle identifier and effective time boundary, a time slot sequence (including start and end times, target wavelength, associated service identifier / source node identifier) ​​for each receiving node (or each receiver), necessary tuning advance and guard interval parameters, and anomaly handling strategies (e.g., alternative wavelengths or fast backoff rules when a wavelength is temporarily unavailable).

[0054] Upon receiving an instruction, the receiving node prepares for receiver tuning in advance according to the receiving time slot arrangement specified in the instruction. At the beginning of the receiving time slot, the receiver is tuned to the target wavelength and locked until the time slot ends, thus completing the reception of the target service without introducing additional negotiation. Furthermore, if the receiving node has multiple receiver instances, different receiving tasks can be mapped to different receiver instances according to the instruction entries for parallel execution, but this is still subject to the constraints of the receiving schedule table regarding time conflicts.

[0055] By issuing scheduling results in the form of structured instructions, receiving nodes can complete tuning and reception in the specified time slots without performing disordered scanning or temporary contention. This enables the reception behavior of broadcast WDM networks to be coordinated and consistent in both time and wavelength dimensions, improving execution predictability and reducing resource waste caused by temporary conflicts and repeated tuning. As a result, the QoS guarantee and resource utilization strategies formulated by the scheduling end can be stably implemented on the network side.

[0056] Regarding the implementation details of wavelength group division in step S130, in some examples of embodiments of this application, firstly, a sliding window is used to statistically analyze the traffic arrival rate of each service category within a preset historical time period, and the current scheduling cycle is calculated accordingly. Inner Load forecast values ​​for similar services :

[0057] Equation (1)

[0058] In the formula, This represents the real-time traffic arrival rate calculated within the current sliding window. This is the load forecast value from the previous scheduling cycle. This is the exponential smoothing coefficient.

[0059] In equation (1), the load forecast value for the service category in the current scheduling period is generated using an exponential smoothing method. This involves weighting and fusing the "real-time arrival rate observed in this window" and the "forecast value formed in the previous scheduling period" according to a smoothing coefficient. The smoothing coefficient is used to control the sensitivity of the forecast to sudden changes (a larger coefficient means it follows real-time changes more closely, while a smaller coefficient means it emphasizes historical stability more), thereby avoiding severe fluctuations in the forecast caused by relying solely on single-period observations. In addition, the scheduling end can set different smoothing coefficients or different sliding window lengths for different service categories (for example, a shorter window is used for control services to respond quickly to sudden changes, and a longer window is used for high-bandwidth services to suppress short-term spikes), and writes the forecast value into the state cache after updating it in each period.

[0060] In this way, load forecasting can be updated as it changes in real time, and the exponential smoothing coefficient can suppress forecast jitter caused by short-term bursts, thus providing a stable and reliable input for subsequent resource allocation.

[0061] Then, based on the load forecast values The load balancing distribution function is used to calculate the first... Number of wavelengths to be allocated for similar services .

[0062] Equation (2)

[0063] In the formula, This represents the total number of wavelengths available as wavelength resources. This indicates the total number of business categories. The smoothing coefficient is non-zero. Indicates the current scheduling period Inner Load forecast values ​​for similar services.

[0064] Equation (2) was used to allocate wavelengths based on load ratio and to add a non-zero smoothing term. Specifically, the numerator of Equation (2) consists of the predicted load of the category and the smoothing term, and the denominator consists of the sum of the predicted load of all categories and the smoothing term. This sum is then multiplied by the total number of available wavelengths in the system, thereby achieving an adaptive allocation where "the higher the predicted load, the more wavelengths are allocated." The non-zero smoothing coefficient is used to prevent certain categories from being allocated zero wavelengths during periods of low load or short-term idle time, ensuring that they still have minimum carrying capacity and schedulable entry points. In addition, since the number of wavelengths must be an integer, the scheduling end can round down the calculation results and ensure that the sum of the number of wavelengths allocated to each category is consistent with the total number of available wavelengths through "margin backfilling / difference correction" (for example, filling the remaining wavelengths after rounding up according to the predicted load from high to low, or filling according to the decimal part). At the same time, minimum guarantee constraints can be added (such as reserving at least a certain number of wavelengths for critical business categories). Thus, wavelength resources can be proportionally adjusted according to the load changes of different business categories, reducing congestion and idleness caused by resource mismatch during multi-business concurrency.

[0065] Furthermore, based on the calculated number of wavelengths The corresponding wavelength groups are divided from the total wavelength set corresponding to the available wavelength resources, and the boundaries of each wavelength group are adjusted based on the adjacent interference rule so that the wavelength group corresponding to high priority services retains a protection interval with the adjacent wavelength group.

[0066] Equation (3)

[0067] In the formula, For the first Class of services wavelength group and adjacent first The number of wavelength intervals between wavelength groups of similar services For the first Weighting coefficients for business categories The preset high-priority judgment threshold, The preset fixed protection interval wavelength number.

[0068] In some implementations, wavelengths of the same category are preferably divided into relatively continuous segments on the spectrum to facilitate subsequent tuning and management. Subsequently, to reduce the impact of crosstalk / adjacent channel interference between adjacent wavelengths on critical services, the scheduling terminal adjusts the wavelength group boundaries according to adjacent interference rules.

[0069] In equation (3), the logic for determining whether a protection interval needs to be set between adjacent wavelength groups is given. That is, when the weight coefficient of a certain business category reaches the preset high priority determination threshold, a fixed number of protection interval wavelengths are inserted between its wavelength group and adjacent wavelength groups (which can be understood as leaving blanks or limiting the load isolation band); if the priority does not reach the threshold, the protection interval is not forced to be set.

[0070] In this way, by introducing spectrum isolation around the high-priority service wavelength group, the impact of adjacent channel interference risk on the reception quality and scheduling determinism of critical services can be significantly reduced. This makes it easier for critical services to obtain a stable transmission environment after being allocated to their wavelength group, and reduces rearrangement, retransmission or delay jitter caused by physical layer interference, thereby improving the overall availability and critical service assurance capabilities in multi-service scenarios.

[0071] Regarding the calculation details of the comprehensive priority index in step S140, in some examples of embodiments of this application, firstly, the first priority index is obtained. Latest time limit for receiving a service request Data size and the weighting coefficient of the business category Business category weighting coefficient Used to reflect the different latency and reliability requirements of various services, data size It is used to reflect the degree to which services consume wavelength occupancy and receive time slot resources.

[0072] Then, calculate the real-time urgency coefficient of the business request. .

[0073] Equation (4)

[0074] In the formula, This is the start time of the calculation for the current scheduling cycle. For the first The timestamp of each business request.

[0075] In equation (4), the calculation start time of the current scheduling cycle is used as the benchmark. The remaining time before the latest reception deadline is compared with the total allowed time window from the arrival of the service to the latest reception deadline, thus obtaining a dynamic urgency characterization that changes over time. If the current time gradually approaches the latest reception deadline, the ratio decreases accordingly, indicating that the urgency of the service is increasing. Through this processing, the scheduler can uniformly map services with different arrival times and different deadline constraints to the same comparable urgency coefficient, making the scheduling order sensitive to the "deadline constraint". This can increase the probability of timely processing of services nearing the deadline, reducing the risk of timeout and latency jitter from the source.

[0076] Furthermore, the adjacency penalty matrix is ​​determined based on the wavelength spectral adjacency relationships in the network state data, and the comprehensive priority index of service requests is calculated based on the following comprehensive evaluation model. .

[0077] Equation (5)

[0078] Equation (6)

[0079] In the formula, For normalized weighting factors, It is a tiny constant. The maximum data block size allowed by the system. For the wavelength group corresponding to this business category The average interference penalty statistic; Wavelength group The number of wavelengths included. The first in the wavelength group One wavelength, This refers to the set of wavelengths that are currently in use in the network. For set The wavelength in.

[0080] Wavelength defined in the adjacency penalty matrix With wavelength Crosstalk penalty value between:

[0081] Equation (7)

[0082] Equation (8)

[0083] In the formula, The crosstalk intensity coefficient, wavelength With wavelength The wavelength spectral interval between them and They represent wavelengths respectively. and wavelength The corresponding sequence number in the optical network wavelength grid.

[0084] It should be noted that in broadcast WDM networks, adjacent or near-adjacent wavelengths are more prone to crosstalk due to factors such as non-ideal device filtering and power spectrum leakage. Therefore, the scheduling end constructs an adjacency penalty matrix based on the wavelength spectral adjacency relationship in the network state data to quantify the "potential interference cost" of any two wavelengths in terms of spectral distance.

[0085] In some implementations, referring to Equation (8), each wavelength is first mapped to a sequence number in the wavelength grid, and the wavelength spectral interval is represented by the absolute value of the difference between the two wavelength sequence numbers; then, in Equation (7), the crosstalk intensity coefficient is further divided by the interval to define a penalty value, so that the closer the two wavelengths are, the greater the penalty, thereby characterizing the physical law that neighbors are more prone to crosstalk.

[0086] Based on this penalty definition, the scheduling terminal calculates the average interference penalty statistics for the wavelength group corresponding to a certain service category according to equation (6). Specifically, it iterates through each candidate wavelength in the wavelength group, calculates the penalty for each wavelength in the current network occupancy state, and accumulates them. Then, it normalizes the penalty by the wavelength group size to obtain the overall interference risk level of the wavelength group under the current occupancy pattern. This provides a quantitative basis for the availability quality of different wavelength groups at the current moment, enabling subsequent priority calculations to avoid directing critical services to spectrum regions with high crosstalk risk.

[0087] More specifically, the scheduling end calculates the comprehensive priority index of service requests based on equation (5). In equation (5), multiple factors are unified into a sortable scalar index: First, the urgency coefficient is given a reciprocal form and a small constant is introduced to avoid the instability of the value when the urgency approaches zero, and at the same time, the more urgent the service, the greater its contribution to this item, thereby strengthening the guiding effect of the deadline constraint on priority; Second, the data size is normalized according to the maximum data block size allowed by the system, so that factors of different dimensions can be weighted and integrated, and the impact of large-scale services on resource consumption can be explicitly perceived at the priority level, so that the scheduling end can maintain the controllability of the overall resource allocation while ensuring the deadline; Third, a service category weight term is introduced so that the differentiated guarantee of different service types can be directly reflected in the ranking basis; Fourth, the average interference penalty of the wavelength group is introduced in the form of a subtraction term in order to explicitly suppress the high interference risk in the comprehensive index, prompting the request to be more inclined to choose wavelength group resources with lower interference. Each normalized weighting factor is used to balance the contribution ratio of different factors. In engineering, it can be configured according to business objectives (such as latency priority or throughput priority) or adjusted through historical statistics.

[0088] Therefore, the comprehensive priority index simultaneously characterizes the urgency of time, the importance of the business, the scale of resource consumption, and the risk of spectrum interference. This enables the scheduling and ranking to not only provide stable priority guarantees for critical businesses, but also reduce the probability of failure and reordering triggered by crosstalk at the wavelength selection level, thereby improving the QoS determinism and overall scheduling efficiency in multi-service concurrent broadcast reception scenarios.

[0089] Figure 2 A flowchart illustrating an example of allocating target wavelengths to each service request via ordered cross-priority scheduling according to an embodiment of this application is shown.

[0090] like Figure 2 As shown, in step S210, the service requests are traversed in descending order of the comprehensive priority index.

[0091] Here, when the scheduling control terminal enters the target wavelength allocation phase, it first forms a unified list of pending service requests within the current scheduling cycle, and then iterates through them one by one according to the comprehensive priority index from high to low. To avoid result fluctuations caused by the uncertainty of requests with the same priority, stable parallel rules can be set, such as prioritizing those with the most recent reception deadline, then comparing those with earlier arrival times, or prioritizing control categories, thereby ensuring that the scheduling results are reproducible under the same input.

[0092] In step S220, for the service request to be assigned, available wavelengths are first filtered in the wavelength group of the service category to which the service request belongs, so as to determine the corresponding target wavelength.

[0093] It should be noted that for the service requests to be allocated that are traversed, the scheduling control terminal will first select available wavelengths within the wavelength group corresponding to the service category to which the request belongs, so as to achieve the allocation principle of "prioritizing the carrying within the category and isolating between categories as much as possible".

[0094] In some implementations, the available wavelengths within a wavelength group can simultaneously satisfy various scheduling requirements. For example, the wavelength is not occupied in the current scheduling period's wavelength status information, or its remaining available window is sufficient to accommodate the request; additionally, it is within the receivable / tunable range for the target receiving node and does not violate the concurrency constraints caused by the receiver number limit; furthermore, when there are adjacent channel crosstalk constraints, wavelengths that are more spectrally spaced from the currently occupied wavelengths, or located within the wavelength group rather than at the boundary, can be further prioritized to reduce the risk of adjacent interference, etc.

[0095] In step S230, when there are no available wavelengths in the wavelength group of the service category, it is determined whether to authorize the borrowing of wavelengths from wavelength groups of other service categories based on the system load status.

[0096] In some implementations, the system load status can comprehensively reflect: the occupancy level and idle capacity of each wavelength group in the current cycle, the queue backlog of each service category, the unmet demand ratio of critical services, and whether overall resources are in a state of tension or surplus. Based on this, the scheduling terminal can set authorization rules, such as allowing borrowing only when the request is highly urgent or close to the latest reception deadline, and borrowing must not exceed the minimum guarantee threshold of the borrowed category (e.g., maintaining its minimum number of wavelengths or its priority carrying capacity for critical services), and can limit the maximum amount of cross-group borrowing in a single cycle to prevent resource boundary instability. This authorization determination improves resource elasticity while maintaining the basic constraint of resource isolation, enabling the system to provide supplementary resources for urgent services even during short-term bursts or category load deviations, while avoiding a decline in overall guarantee capacity due to uncontrolled cross-group borrowing, thus balancing critical service time-limit guarantee and system stability.

[0097] In step S240, if authorized, an available wavelength is selected from the wavelength group of other service categories as the target wavelength.

[0098] In some implementations, wavelength groups with more current free space, lower predicted load, or lower service weight can be selected first to reduce the impact on subsequent scheduling of the borrowed category; and availability verification can be performed at the candidate wavelength level, including the current cycle occupancy status, the receptive range / tuning reachability of the target receiving node, and the risk of adjacent interference, etc. If necessary, a location far away from the boundary of the wavelength group can be selected first to avoid disrupting the existing protection interval.

[0099] To facilitate subsequent management and recycling, the scheduling end can mark the wavelength as "borrowed," record the borrowing source category and borrowing duration boundary, and maintain a consistent mapping when generating subsequent scheduling tables and issuing instructions to ensure that there is no ambiguity in the node's tuning and reception according to the instructions.

[0100] Therefore, by introducing available wavelengths from other wavelength groups in a controlled manner, the overall resource utilization and the schedulability of urgent services can be improved without significantly weakening the guarantee capability of the borrowed category. At the same time, by using the selection rules for source categories and candidate wavelengths and the borrowing status marking, the risk of cross-group interference and execution conflicts can be reduced, and the traceability and recoverability of scheduling behavior can be enhanced.

[0101] Regarding the implementation details of step S220, which involves filtering and determining the target wavelength within the wavelength group, in some examples of embodiments of this application, firstly, for the service request to be assigned, all available candidate wavelengths within its respective wavelength group are traversed, and the calculation of each candidate wavelength is performed. Choice Cost :

[0102] Equation (9)

[0103] In the formula, This is the time sensitivity coefficient.

[0104] In the wavelength selection process within a group, to ensure more precise and optimal utilization of available wavelengths within the same service category wavelength group, this embodiment calculates the selection cost for each available candidate wavelength and determines the target wavelength accordingly. Specifically, for a service request to be allocated, the scheduling control terminal first extracts a set of candidate wavelengths that meet the basic availability constraints from its wavelength group. For example, the wavelength has a free interval available for the service within the current scheduling period, is within the receivable / tunable range for the target receiving node, and will not significantly conflict with services already listed in the schedule.

[0105] Subsequently, each candidate wavelength is traversed and its selection cost is calculated according to Equation (9). This cost consists of two parts: one part is the wavelength switching cost, which is used to characterize the tuning burden brought about by the receiver switching from the stationary wavelength at the previous moment to the candidate wavelength; the other part is the time waiting cost, which is used to characterize the time overhead that the service needs to wait if the candidate wavelength is not yet idle. The time sensitivity coefficient is used to adjust the degree of influence of the waiting time in the total cost, so that the scheduling can weigh between "fewer switching" and "fewer waiting" according to the service's sensitivity to delay.

[0106] in, Candidate wavelength Time waiting cost:

[0107] Equation (10)

[0108] In the formula, For this candidate wavelength The moment when it transitions to an idle state within the current scheduling cycle.

[0109] In equation (10), the scheduling control terminal obtains the time when the candidate wavelength becomes idle in the current scheduling cycle from the wavelength status information / wavelength occupancy schedule, and compares it with the arrival time of the service request; if the wavelength is already idle at the arrival time, the waiting cost is 0; if the wavelength needs to be released at a later time, the waiting cost is equal to the difference between the release time and the arrival time, thus explicitly quantifying the "queueing wait caused by the wavelength being temporarily busy".

[0110] The cost of wavelength switching determined based on wavelength spectral spacing:

[0111] Equation (11)

[0112] Equation (12)

[0113] In the formula, The wavelength at which the receiving node resided in the previous moment. The switching penalty coefficient per unit wavelength interval; Candidate wavelength and The wavelength spectral interval between them and They represent wavelengths respectively. and wavelength The corresponding sequence number in the optical network wavelength grid.

[0114] In equations (11) and (12), the scheduling control terminal reads the wavelength that the receiver was stationed at in the previous moment from the state data of the receiving node as a reference, and calculates the spectral interval based on the difference between the sequence number of the candidate wavelength and the stationed wavelength in the wavelength grid. The physical law that "the larger the spectral span, the heavier the tuning action (or the more unfavorable the stabilization time)" is quantified in the cost by using the switching penalty coefficient, so that the system tends to select the candidate wavelength that is closer to the current stationed wavelength to reduce unnecessary switching.

[0115] Then, select the corresponding choice cost. The smallest candidate wavelength is selected as the target wavelength.

[0116] The scheduling control unit selects the candidate wavelength with the lowest selection cost as the target wavelength. When costs are equal, a consistent parallel selection rule can be adopted (e.g., prioritizing wavelengths available earlier or with lower interference risk) to ensure stable and reproducible scheduling results. Through this cost model, unnecessary tuning switching and waiting delays can be suppressed while maintaining availability, thereby reducing receiver tuning overhead and receiver delay jitter, and improving the stability of wavelength allocation and overall scheduling efficiency.

[0117] Regarding the implementation details of determining whether to authorize borrowing wavelengths from wavelength groups of other service categories in step S230, in some examples of embodiments of this application, firstly, the current scheduling cycle is monitored in real time. The overall system load factor .

[0118] Equation (13)

[0119] In the formula, This is the set of service requests to be scheduled within the current period. For the data size of the business request, This represents the total transmission capacity of all available wavelengths within the current period.

[0120] In step S230, to achieve controllability and recoverability of cross-service category wavelength group borrowing, the scheduling control terminal first quantifies the overall system load in the current scheduling cycle in real time and calculates the overall system load factor using equation (13). Specifically, in equation (13), the numerator is the sum of the data sizes of each request in the set of scheduled service requests for the current week, used to characterize the total service demand for this cycle; the denominator is the total transmission capacity that all available wavelengths can provide in this cycle, used to characterize the total resource supply capacity for this cycle. Thus, the obtained load factor reflects the tension of "demand / supply", which can be calculated by the scheduling terminal immediately after collecting requests and updating the status of available wavelengths and dynamically updated as scheduling progresses.

[0121] Then, based on the statistical distribution of the comprehensive priority index of all pending service requests in the current period, the dynamic borrowing threshold is calculated. .

[0122] Equation (14)

[0123] In the formula, This is the average of the overall priority index of all business requests in the current period. Standard deviation and This is the adjustment coefficient.

[0124] Subsequently, the scheduling end calculates the dynamic borrowing threshold according to formula (14) based on the statistical distribution of the comprehensive priority index of all scheduled business requests in this period. The threshold is obtained by linear combination of the average value and standard deviation of the comprehensive priority index. The average value is used to characterize the overall priority level in this period, the standard deviation is used to characterize the priority dispersion (i.e. the strength of the differentiation between high and low priorities), and the adjustment coefficient is used to control the strictness of the authorization conditions so that the threshold can be adaptively adjusted with changes in the business structure.

[0125] Furthermore, when a service request has no available wavelengths within its corresponding wavelength group, it is determined whether the overall priority index of the service request is greater than the dynamic borrowing threshold. Furthermore, the overall system load factor is less than the preset system load safety threshold. ,in This is a preset system load safety threshold.

[0126] On the one hand, if the conditions are not met, the borrowing will be rejected, and the service request will be marked as a postponement request and written into the scheduling queue for the next scheduling cycle.

[0127] Here, when a service request has no available wavelengths within its corresponding wavelength group, the scheduler checks two authorization conditions simultaneously. First, the overall priority index of the request must be greater than the dynamic borrowing threshold (indicating that it is a request that needs to be prioritized in this cycle); second, the overall system load factor must be less than the preset system load safety threshold (indicating that the system is still within the acceptable resource margin range, and cross-group borrowing will not significantly amplify the overall congestion risk). If either condition is not met, borrowing is rejected, and the request is marked as a deferred request and written into the scheduling queue for the next scheduling cycle to avoid triggering cross-group contention propagation when resources are scarce.

[0128] On the other hand, if the conditions are met, the service request is authorized to borrow idle wavelengths from wavelength groups of other service categories, and the selection cost is adjusted by applying an additional cross-group interference penalty factor to determine the target wavelength for cross-group borrowing.

[0129] Equation (15)

[0130] Equation (16)

[0131] In the formula, Candidate wavelengths borrowed across groups The cost of basic choices; Candidate wavelengths borrowed across groups The corresponding corrected choice cost; As a cross-group interference penalty factor, To borrow the base penalty coefficient across groups, Candidate wavelengths borrowed across groups The weighting coefficient of the business category to which it belongs.

[0132] In some implementations, if both conditions are met, the idle wavelength is authorized to be borrowed from other service category wavelength groups. Specifically, the scheduling terminal first uses the basic selection cost (e.g., the basic cost obtained by waiting cost and switching cost according to equation (9)) for cross-group candidate wavelengths, and then applies cross-group interference penalty correction to the basic cost according to equation (15). The correction term is given by equation (16), which is composed of the cross-group borrowing basic penalty coefficient and the service category weight coefficient of the candidate wavelength, so that the cost of borrowing from the higher weight category wavelength group is higher. Thus, when borrowing across groups, the idle resources of the lower weight category are consumed first and the disturbance to the boundary of the high guarantee category resources is reduced. Finally, the candidate wavelength with the smallest correction cost is selected as the target wavelength for cross-group borrowing.

[0133] Through the above design, cross-group elastic resources are provided for high-priority and urgent requests under the premise of controllable system load. At the same time, dynamic thresholds and cross-group penalties are used to achieve dual constraints on the borrowing trigger frequency and borrowing source, thereby improving scheduling stability, the probability of timely completion of key business and the balance of overall resource utilization.

[0134] Figure 3 A flowchart illustrating an example of performing wavelength fragmentation sorting operations according to an embodiment of this application is shown.

[0135] like Figure 3 As shown, in step S310, the time slot allocation of each wavelength in the current scheduling period is detected, and non-contiguous idle time slot segments between occupied time slots are identified.

[0136] In some implementations, the scheduling control terminal performs a consistency scan of the time slot occupancy sequence for each wavelength based on the wavelength scheduling table / receive scheduling table generated in the current scheduling cycle. For each wavelength, the occupied time slots corresponding to the allocated service data blocks (including necessary guard intervals and tuning reservations) are read in chronological order, and the interval between adjacent occupied time slots is calculated to identify non-contiguous idle time slot segments located between two occupied time slots that are insufficient in length to directly carry large services or are scattered. Thus, the scheduling terminal obtains a quantitative description of "idle resources on the same wavelength being divided into multiple segments".

[0137] In step S320, for service data blocks that have been allocated target wavelengths but have not yet started transmission, a sliding rearrangement operation is performed to merge scattered idle time slot segments. The sliding rearrangement operation includes shifting the service data blocks along the time axis to an earlier time while maintaining the order of each service data block and without violating the latest reception time limit, so as to update the start time of the service data blocks. The updated start time is not earlier than the data preparation time of the service data block at the transmitting node, so as to meet the physical transmission constraints.

[0138] In some implementations, during sliding rearrangement, the scheduler shifts data blocks forward along the time axis in units of wavelength (or in units of the task set associated with the same receiving node). While maintaining the original order of the service data blocks, priority is given to "attaching" later data blocks to earlier idle segments, shifting their start times as far forward as possible to fill fragments. Simultaneously, each shift is constrained to ensure no overlap with already occupied time slots of the same wavelength, no concurrent conflicts on the receiving side, no violation of the latest reception time limit, and that the updated start time is not earlier than the data readiness time of the transmitting node (and, if necessary, physical constraints such as tuning stabilization time and guard intervals must also be met).

[0139] Once a data block is successfully moved forward, the scheduler synchronously updates the occupied range of that wavelength and the boundary of adjacent free segments, and iteratively processes subsequent data blocks until the free segments are merged into fewer, longer consecutive free windows or there is no more space to move.

[0140] Therefore, without changing the order of business and the guarantee of time limits, the scattered idle time slots are compressed and integrated into a continuous available window, reducing the fragmentation on the wavelength time axis, improving the success rate of carrying large data blocks or urgent business, and reducing waiting and scheduling jitter caused by fragmentation, thereby improving the overall resource utilization efficiency and the executability of the scheduling table.

[0141] In some examples of embodiments of this application, feedback-based closed-loop optimization steps are proposed.

[0142] More specifically, firstly, the actual received bit error rate reported by each receiving node after completing data reception is received. and receive delayed data.

[0143] Specifically, after each scheduling cycle is completed, the scheduling control terminal returns feedback information from each receiving node regarding the services it has actually received. The feedback information includes at least the actual received bit error rate. In addition to reception delay statistics (e.g., the actual delay from the start of a service to the completion of reception or its percentile value), it preferably also carries the reception wavelength, reception time slot, and adjacent occupied wavelengths that may cause adjacent channel interference within the same period, so as to accurately attribute “quality degradation” to specific wavelength combinations or spectral adjacency relationships.

[0144] Then, the adjacency penalty matrix corresponding to the interference risk is corrected online using the actual received bit error rate to update the crosstalk penalty value for the next scheduling cycle. The correction rule is as follows:

[0145] Equation (17)

[0146] In the formula, The crosstalk penalty value before correction. This is the corrected crosstalk penalty value. This is the forgetting factor, used to adjust the retention weight of historical data; Here is the penalty mapping function, used to map the detected bit error rate to the corresponding penalty increment, where when When the preset warning bit error rate threshold is exceeded, Output a penalty increment greater than zero to increase the penalty cost of the corresponding wavelength combination, thereby reducing the probability of that wavelength combination being selected in subsequent scheduling.

[0147] In equation (17), the first term reflects the retention of historical statistics. As a forgetting factor, it is used to control the "response speed" of updates ( The larger the value, the faster it follows the latest observations; the smaller the value, the more emphasis is placed on long-term stability. This is a penalty mapping function used to convert the measured bit error rate into a penalty increment or an equivalent penalty level, and when... When the preset warning threshold is exceeded, a penalty increment greater than zero is output, which increases the interference risk / selection cost of the wavelength combination in subsequent scheduling, thereby reducing the probability of it being selected again.

[0148] Through the aforementioned closed-loop online correction, the scheduling end can dynamically inject the actual physical layer transmission quality (affected by filter characteristics, power spectrum leakage, aging drift, transient interference, etc.) into the interference risk model. This allows the adjacency penalty matrix to no longer rely on static empirical parameters but can be adaptively calibrated according to the network operating status. Consequently, in subsequent wavelength group selection, comprehensive priority calculation, and target wavelength screening, it tends to avoid spectrum combinations with high bit error risk, reducing the probability of bit errors and retransmission triggering, improving reception stability, and indirectly improving latency jitter and on-time completion rate.

[0149] Figure 4 A flowchart illustrating the operational mechanism of an example of a wavelength allocation-based multi-service broadcast optical reception scheduling method according to an embodiment of this application is shown.

[0150] like Figure 4 As shown, the scheduling control terminal first completes traffic demand collection, service classification, and network status monitoring, generating service request information, service category / weight information, and network status information such as wavelength occupancy, and inputs this information into the scheduling decision link. Specifically, service requests and network status jointly drive the wavelength allocation module to dynamically allocate available wavelength resources, generating wavelength groups for different service categories. Based on this, priority index calculation is performed using the service classification results to obtain the comprehensive priority basis for each service request, which is further input into the ordered cross-priority scheduling module to complete the ordered orchestration of multiple service requests and resource contention coordination.

[0151] Simultaneously, the wavelength allocation results are used to generate broadcast scheduling schemes, and together with the priority calculation / scheduling results, they act on receiver control and wavelength allocation to form control commands for receiver tuning and receiving time slots at receiving nodes, which are then issued and executed. After the scheduling table is formed, fragmentation can be triggered for allocated but unexecuted services to merge scattered idle time slots and improve subsequent schedulability. Through the collaboration of the above modules, performance improvements are ultimately achieved, including reduced receiving conflicts, increased resource utilization, and enhanced service latency assurance capabilities.

[0152] To verify the effectiveness of the algorithm proposed in this application, a 64-wavelength broadcast star optical network simulation experiment was constructed and compared with baselines such as First-Fit scheduling, water-filling algorithm (WF-DBA), and Max-Min Ant System (MMAS).

[0153] Network load Take 7 test points ( For each load point, 10,000 service requests are randomly generated (∈{0.3,0.4,…,0.9}). Service categories are generated in a ratio of 1:2:3 (control:streaming:background). Service sizes follow an exponential distribution. Relative deadlines are Gaussian distributed with a lower bound truncation to ensure the validity of the deadline constraint. First-Fit allocates resources sequentially on available wavelengths according to arrival order. WF-DBA uses a water-filling approach to evenly allocate resources across multiple wavelengths. MMAS schedules resources using a round-robin and dynamic reservation strategy. The algorithm OCPS (Ordered Cross-Priority Scheduling) of this invention uses dynamic wavelength grouping and a comprehensive priority index for scheduling. Evaluation metrics include: average service reception completion delay, wavelength utilization (the proportion of "wavelength-time" resources occupied within the scheduling period), and relative delay ratio (OCPS average delay / corresponding baseline average delay, calculated separately for each baseline).

[0154] Figure 5 A schematic diagram showing the comparison of average latency as load changes for different scheduling algorithms is presented.

[0155] like Figure 5 As shown in the curves, with the normalized load increasing from 0.3 to 0.9, the average latency of each algorithm generally increases with the load, but the magnitude of the increase varies significantly. The OCPS algorithm of this invention maintains the lowest average latency across the entire load range, and its latency growth is relatively gradual in the high-load region (e.g., 0.7~0.9). In contrast, the average latency of First-Fit and WF-DBA increases faster with increasing load. Although MMAS is superior to First-Fit and WF-DBA, its average latency is still significantly higher than that of OCPS.

[0156] Furthermore, the change in the curve spacing shows that the higher the load, the more obvious the gap between the OCPS algorithm of this invention and the baseline algorithms. This indicates that OCPS can still more effectively suppress the latency accumulation caused by queuing and scheduling conflicts when resources are tight, thus demonstrating better high-load robustness and latency performance advantages.

[0157] Figure 6A schematic diagram showing the comparison of wavelength utilization as load changes for different scheduling algorithms is presented.

[0158] like Figure 6 As shown, the wavelength utilization of each algorithm generally increases with increasing network load. Throughout the entire load range, the curve of the OCPS algorithm of this invention remains at the top, indicating that it achieves a higher wavelength occupancy rate at each load point. In contrast, First-Fit has the lowest utilization rate at each load point, while WF-DBA and MMAS have utilization rates in between and are generally lower than OCPS. In the high-load region, the difference in utilization rates among the algorithms narrows, but OCPS still maintains its lead.

[0159] Considering the resource configuration of broadcast WDM receiver scheduling, the above differences can be understood as follows: First-Fit sequential placement is more likely to create scattered idle segments on the wavelength time axis, making it difficult to fully and continuously utilize available resources; while WF-DBA improves the overall utilization level through balanced allocation, it is limited by the degree of matching between allocation granularity and available windows; MMAS can maintain high utilization under medium to high loads, but cross-cycle reservations may lead to some idleness in some scenarios. In contrast, OCPS guides resource allocation across service categories more appropriately through dynamic wavelength grouping and improves the continuity of idle windows through fragmentation, thus maintaining high wavelength utilization across the entire load range.

[0160] Figure 7 A heatmap comparing the average latency ratios of different baseline scheduling algorithms using the method of this application is shown, varying with load. The horizontal axis represents the normalized load (0.3~0.9), and the vertical axis represents the ratios between the average latency of OCPS and the average latency of First-Fit (FF), WF-DBA, and MMAS, respectively. Each square in the graph shows the ratio value for the corresponding load point. The color bars indicate the magnitude of the ratios; darker colors indicate higher latency ratios, meaning a smaller latency difference between OCPS and the corresponding baseline; conversely, lighter colors indicate lower ratios, meaning OCPS has a greater relative advantage.

[0161] like Figure 7 As shown, the three ratios are all less than 1 (approximately 0.62~0.83) at each load point, indicating that the average latency of OCPS is lower than that of each baseline algorithm under all test loads. Meanwhile, OCPS has the smallest ratio relative to First-Fit (its advantage is most significant), and the largest ratio relative to MMAS (MMAS is closer to OCPS). Furthermore, as the load increases, the ratios generally show a slight upward trend, indicating that the latency gap between the baseline algorithm and OCPS converges somewhat in the high-load area, but OCPS still maintains a stable latency advantage.

[0162] This paper addresses the resource contention and QoS differentiation requirements caused by multi-service concurrency in broadcast WDM optical networks. It proposes a wavelength allocation-based multi-service broadcast optical receiver scheduling method, forming a closed-loop framework of "prediction-driven resource partitioning - priority modeling - ordered scheduling table placement - execution optimization - feedback adaptation." Specifically, based on service load prediction and combined with adjacent interference penalties, wavelength groups for different service categories are dynamically partitioned. When resources are insufficient, controlled cross-group borrowing is introduced to balance service isolation and resource utilization. A comprehensive priority index is constructed and coupled with an adaptive threshold, allowing the scheduling intensity to be dynamically adjusted according to system load and service urgency, thereby achieving differentiated protection for different services.

[0163] At the scheduling execution level, wavelength selection and temporal arrangement are optimized collaboratively through ordered cross-priority scheduling. Furthermore, fragmentation is combined with sliding rearrangement of unexecuted tasks to increase the proportion of continuously available time slots, further enhancing schedulability and stability under high load. Simultaneously, an online correction mechanism based on receiver feedback is introduced to iteratively update the interference risk model and related statistics, achieving adaptive optimization for complex operating environments. Theoretical analysis and simulation results show that this method outperforms comparable algorithms such as First-Fit, WF-DBA, and MMAS in terms of average delay and wavelength utilization, with more significant advantages in high-load areas. It can be applied to broadcast WDM network scenarios with multiple concurrent services, such as avionics, financial data distribution, and cloud data centers, providing a feasible scheduling approach for all-optical network resource management.

[0164] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of combined actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Secondly, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application. In the above embodiments, the descriptions of each embodiment have their own emphasis; for parts not described in detail in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0165] Figure 8 A structural block diagram of an example of a wavelength allocation-based multi-service broadcast optical reception scheduling system according to an embodiment of this application is shown, the system being deployed at a remote wireless node.

[0166] like Figure 8As shown, the multi-service broadcast optical reception scheduling system 800 based on wavelength allocation is deployed at the scheduling control end of the broadcast WDM optical network. It includes a service request aggregation unit 810, a network status awareness unit 820, a wavelength group division unit 830, a priority evaluation unit 840, a resource scheduling table generation unit 850, and a broadcast scheduling instruction distribution unit 860.

[0167] The service request aggregation unit 810 is used to obtain a set of service requests reported by multiple service initiating nodes within the current scheduling period. The service requests include at least the arrival time, data size, service category, and latest reception time limit.

[0168] The network status awareness unit 820 is used to acquire network status data of the broadcast WDM optical network. The network status data includes at least wavelength status information of available wavelength resources, and the wavelength status information includes at least the occupancy status of each wavelength in the current scheduling period.

[0169] The wavelength group partitioning unit 830 is used to determine the load prediction results of each service category in the current scheduling period based on the service request set, and to dynamically partition the available wavelength resources according to the load prediction results to generate wavelength groups corresponding to different service categories.

[0170] The priority evaluation unit 840 is used to calculate the comprehensive priority index of each service request in the service request set based on the latest reception time limit, the data size, the service category, and the interference risk corresponding to the candidate wavelength.

[0171] The resource scheduling table generation unit 850 is used to sort each service request according to the comprehensive priority index, perform ordered cross-priority scheduling based on the sorting result, allocate target wavelengths to each service request, and allocate reception time slots to each service request on the target wavelengths to generate a wavelength scheduling table and a reception scheduling table; wherein, the reception time slots are used to avoid reception conflicts on the same wavelength and meet the latest reception time limit.

[0172] The broadcast scheduling instruction distribution unit 860 is used to generate broadcast scheduling instructions according to the wavelength scheduling table and the reception scheduling table and send them to relevant nodes to control the receiver of the receiving node to tune to the target wavelength for data reception in the reception time slot.

[0173] In some embodiments, this application provides a non-volatile computer-readable storage medium storing one or more programs including execution instructions. The execution instructions can be read and executed by electronic devices (including but not limited to computers, servers, or network devices) to perform the steps of any of the wavelength allocation-based multi-service broadcast optical reception scheduling methods described above.

[0174] In some embodiments, this application also provides a computer program product, the computer program product including a computer program stored on a non-volatile computer-readable storage medium, the computer program including program instructions, which, when executed by a computer, cause the computer to perform the steps of any of the above-described multi-service broadcast optical reception scheduling methods based on wavelength allocation.

[0175] In some embodiments, this application also provides an electronic device comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to perform steps of a wavelength allocation-based multi-service broadcast optical reception scheduling method.

[0176] The above-described product can perform the methods provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects for performing the methods. Technical details not described in detail in this embodiment can be found in the methods provided in the embodiments of this application.

[0177] The electronic devices in this application can exist in various forms, including but not limited to: mobile communication devices, ultra-mobile personal computer devices, portable entertainment devices, or other airborne electronic devices with data interaction functions.

[0178] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0179] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented using software plus a general-purpose hardware platform, or of course, using hardware. Based on this understanding, the above technical solutions, in essence or the parts that contribute to the related technology, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0180] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A multi-service broadcast optical reception scheduling method based on wavelength allocation, applied to the scheduling and control end of a broadcast WDM optical network, characterized in that, The method includes: Obtain a set of service requests reported by multiple service initiating nodes within the current scheduling period. The service requests include at least the arrival time, data size, service category, and latest reception time limit. Obtain network status data of the broadcast WDM optical network, wherein the network status data includes at least wavelength status information of available wavelength resources, and the wavelength status information includes at least the occupancy status of each wavelength in the current scheduling period; Based on the set of service requests, the load prediction results for each service category in the current scheduling period are determined, and the available wavelength resources are dynamically divided according to the load prediction results to generate wavelength groups corresponding to different service categories. For each service request in the service request set, a comprehensive priority index is calculated based on the latest reception time limit, the data size, the service category, and the interference risk corresponding to the candidate wavelength. The service requests are sorted according to the comprehensive priority index, and ordered cross-priority scheduling is performed based on the sorting results. Target wavelengths are allocated to each service request, and reception time slots are allocated to each service request on the target wavelengths to generate wavelength scheduling tables and reception scheduling tables. The reception time slots are used to avoid reception conflicts on the same wavelength and to meet the latest reception time limit. Broadcast scheduling instructions are generated based on the wavelength scheduling table and the reception scheduling table and sent to relevant nodes to control the receivers of the receiving nodes to tune to the target wavelength for data reception within the reception time slot; The step of performing ordered cross-priority scheduling based on the sorting results to allocate target wavelengths to each service request includes: The service requests are traversed in descending order of the comprehensive priority index. For service requests to be assigned, available wavelengths are first filtered from the wavelength group of the service category to which the service request belongs in order to determine the corresponding target wavelength; When there are no available wavelengths in the wavelength group of the business category to which the wavelength belongs, the system load status will determine whether to authorize the borrowing of wavelengths from wavelength groups of other business categories. If authorized, an available wavelength is selected from the wavelength group of other business categories as the target wavelength.

2. The method according to claim 1, characterized in that, The step of determining the load prediction results for each service category within the current scheduling period based on the service request set, and dynamically dividing the available wavelength resources according to the load prediction results to generate wavelength groups corresponding to different service categories, includes: The sliding window is used to calculate the traffic arrival rate of each service category within a preset historical time period, and this rate is then used to calculate the current scheduling cycle. Inner Load forecast values ​​for similar services : ， In the formula, This represents the real-time traffic arrival rate calculated within the current sliding window. This is the load forecast value from the previous scheduling cycle. It is the exponential smoothing coefficient; Based on the load forecast value The load balancing distribution function is used to calculate the first... Number of wavelengths to be allocated for similar services : ， In the formula, This represents the total number of wavelengths available for wavelength resource. This indicates the total number of business categories. The smoothing coefficient is non-zero. Indicates the current scheduling period Inner Load forecast values ​​for similar services; Based on the calculated number of wavelengths From the total wavelength set corresponding to the available wavelength resources, corresponding wavelength groups are divided, and the boundaries of each wavelength group are adjusted based on the adjacent interference rule so that a protection interval is maintained between the wavelength group corresponding to high-priority services and adjacent wavelength groups. ， In the formula, For the first Class of services wavelength group and adjacent first The number of wavelength intervals between wavelength groups of similar services For the first Weighting coefficients for business categories The preset high-priority judgment threshold, The preset fixed protection interval wavelength number.

3. The method according to claim 1, characterized in that, The calculation of the comprehensive priority index for the service request based on the latest reception time limit, the data size, the service category, and the interference risk corresponding to the candidate wavelength includes: Get the Latest time limit for receiving a service request Data size and the weighting coefficient of the business category ; Calculate the real-time urgency coefficient of the service request. : ， In the formula, This is the start time of the calculation for the current scheduling cycle. For the first The timestamp of each business request; The adjacency penalty matrix is ​​determined based on the wavelength spectral adjacency relationships in the network state data, and the comprehensive priority index of the service request is calculated based on the following comprehensive evaluation model. : ， ， In the formula, For normalized weighting factors, It is a tiny constant. The maximum data block size allowed by the system. For the wavelength group corresponding to this business category The average interference penalty statistic; Wavelength group The number of wavelengths included. The first in the wavelength group One wavelength, This refers to the set of wavelengths that are currently in use in the network. For set wavelength in; The wavelength defined in the adjacency penalty matrix With wavelength Crosstalk penalty value between: ， ， In the formula, The crosstalk intensity coefficient, wavelength With wavelength The wavelength spectral interval between them and They represent wavelengths respectively. and wavelength The corresponding sequence number in the optical network wavelength grid.

4. The method according to claim 3, characterized in that, For the service request to be allocated, the available wavelengths are first filtered from the wavelength group of the service category to which the service request belongs in order to determine the corresponding target wavelength, including: For each service request to be assigned, iterate through all available candidate wavelengths within its wavelength group and calculate the value of each candidate wavelength. Choice Cost : ， In the formula, This is the time sensitivity coefficient; Candidate wavelength Time waiting cost: ， In the formula, For this candidate wavelength The moment when it transitions to an idle state within the current scheduling cycle; The cost of wavelength switching determined based on wavelength spectral spacing: ， ， In the formula, The wavelength at which the receiving node resided in the previous moment. The switching penalty coefficient per unit wavelength interval; Candidate wavelength and The wavelength spectral interval between them and They represent wavelengths respectively. and wavelength The corresponding sequence number in the optical network wavelength grid; Select the corresponding selection cost The smallest candidate wavelength is taken as the target wavelength.

5. The method according to claim 4, characterized in that, The determination of whether to authorize the borrowing of wavelengths from other service categories based on system load status includes: Real-time monitoring of the current scheduling cycle The overall system load factor : ， In the formula, This is the set of service requests to be scheduled within the current period. For the data size of the business request, This represents the total transmission capacity of all available wavelengths within the current period. Based on the statistical distribution of the comprehensive priority index of all pending service requests in the current period, the dynamic borrowing threshold is calculated. : ， In the formula, This is the average of the overall priority index of all business requests in the current period. Standard deviation and This is the adjustment coefficient; When a service request has no available wavelengths within its wavelength group, it is determined whether the overall priority index of the service request is greater than the dynamic borrowing threshold. Furthermore, the overall system load factor is less than the preset system load safety threshold. ,in The preset system load safety threshold; If the conditions are not met, the borrowing will be rejected, and the service request will be marked as a deferred request and written into the scheduling queue for the next scheduling cycle. If satisfied, the service request is authorized to borrow an idle wavelength from wavelength groups of other service categories, and the selection cost is adjusted by applying an additional cross-group interference penalty factor to determine the target wavelength for cross-group borrowing: ， ， In the formula, Candidate wavelengths borrowed across groups The cost of basic choices; Candidate wavelengths borrowed across groups The corresponding corrected choice cost; As a cross-group interference penalty factor, To borrow the base penalty coefficient across groups, Candidate wavelengths borrowed across groups The weighting coefficient of the business category to which it belongs.

6. The method according to claim 1, characterized in that, After generating the wavelength scheduling table and the receive scheduling table, the method further includes performing a wavelength fragmentation defragmentation operation, specifically including: Detect the time slot allocation of each wavelength in the current scheduling cycle and identify non-contiguous idle time slot segments between occupied time slots; For service data blocks that have been allocated target wavelengths but have not yet started transmission, a sliding rearrangement operation is performed to merge scattered idle time slot segments. The sliding rearrangement operation includes shifting the service data blocks along the time axis to an earlier time while maintaining the order of each service data block and without violating the latest reception time limit, so as to update the start time of the service data blocks. The updated start time is not earlier than the data preparation time of the service data block at the transmitting node, so as to meet the physical transmission constraints.

7. The method according to claim 3, characterized in that, The method further includes a feedback-based closed-loop optimization step, specifically including: Receive the actual received bit error rate reported by each receiving node after completing data reception. and receive delayed data; The adjacency penalty matrix corresponding to the interference risk is corrected online using the actual received bit error rate to update the crosstalk penalty value for the next scheduling period. The correction rule is as follows: ， In the formula, The crosstalk penalty value before correction. This is the corrected crosstalk penalty value. This is the forgetting factor, used to adjust the retention weight of historical data; Here is the penalty mapping function, used to map the detected bit error rate to the corresponding penalty increment, where when When the preset warning bit error rate threshold is exceeded, Output a penalty increment greater than zero to increase the penalty cost of the corresponding wavelength combination, thereby reducing the probability of that wavelength combination being selected in subsequent scheduling.

8. A multi-service broadcast optical receiver scheduling system based on wavelength allocation, deployed at the scheduling control end of a broadcast WDM optical network, for implementing the method as described in any one of claims 1-7; characterized in that, The system includes: The service request aggregation unit is used to obtain a set of service requests reported by multiple service initiating nodes within the current scheduling period. The service requests include at least the arrival time, data size, service category, and latest reception time limit. A network status awareness unit is used to acquire network status data of the broadcast WDM optical network. The network status data includes at least wavelength status information of available wavelength resources, and the wavelength status information includes at least the occupancy status of each wavelength in the current scheduling period. The wavelength group partitioning unit is used to determine the load prediction results of each service category in the current scheduling period based on the service request set, and to dynamically partition the available wavelength resources according to the load prediction results to generate wavelength groups corresponding to different service categories. The priority evaluation unit is used to calculate the comprehensive priority index of each service request in the service request set based on the latest reception time limit, the data size, the service category, and the interference risk corresponding to the candidate wavelength. The resource scheduling table generation unit is used to sort each service request according to the comprehensive priority index, perform ordered cross-priority scheduling based on the sorting result, allocate target wavelengths to each service request, and allocate reception time slots to each service request on the target wavelengths to generate a wavelength scheduling table and a reception scheduling table; wherein, the reception time slots are used to avoid reception conflicts on the same wavelength and meet the latest reception time limit. The broadcast scheduling instruction distribution unit is used to generate broadcast scheduling instructions based on the wavelength scheduling table and the reception scheduling table and send them to relevant nodes to control the receivers of the receiving nodes to tune to the target wavelength for data reception within the reception time slot.

Images