A dual-mode cpe resource mapping and scheduling method supporting network slicing

Abstract

The application discloses a kind of dual-mode CPE resource mapping and scheduling method supporting network slice, it is related to communication network resource scheduling technical field, comprising the following steps: obtaining the chain segment state data of target object and slice demand data;Dual-mode consistency encapsulation processing is carried out to the chain segment state data, generates standard chain segment data, and the standard chain segment data and the slice demand data are executed multi-section series projection processing and default probability mapping processing, and generate default portrait data.This scheme generates default portrait data by executing dual-mode consistency encapsulation to chain segment state data and combining slice demand data, realizes the probabilistic representation of multi-section series uncertainty, reduces the false triggering caused by mean value determination, generates redundancy level data to default portrait data and slice demand data and constructs boundary constraint data, realizes the collaborative determination of redundancy intensity and upper limit of overhead, inhibits the resource squeeze caused by redundancy excessive.

Description

Technical Field

[0001] This invention relates to the field of communication network resource scheduling technology, specifically to a dual-mode CPE resource mapping and scheduling method that supports network slicing. Background Technology

[0002] As communication networks continue to evolve towards higher bandwidth, lower latency, and higher reliability, network slicing technology, as an important supporting capability of the next generation of communication networks, is widely used to provide network services with mutual isolation and differentiated protection for different services on the same physical network infrastructure. In actual deployment, customer premises equipment (CPE) is gradually developing from a single access mode to multiple access and multimodal modes.

[0003] In existing technologies, CPE resource management solutions that support network slicing typically monitor the status of access links or backhaul links and map different slices to corresponding links or scheduling queues based on pre-configured slice service level requirements. When a link fails, a primary / backup switch or simple traffic replication strategy is executed. The advantage of these solutions is their relatively direct implementation, enabling basic slice isolation and service assurance in scenarios with single links or relatively stable link status changes. However, in long-distance tunnel coverage scenarios, backhaul links are typically composed of multiple communication segments connected in series. The communication status of each segment is affected by factors such as spatial environment, equipment power supply, and interference reflection, exhibiting significant uncertainty and non-stationarity. In such scenarios, the above solutions struggle to reflect the cumulative effect of overall communication risk under multi-segment series conditions. Furthermore, based on fixed thresholds or static level configurations, they lack characterization of the probability characteristics of link status fluctuations, making it easy for slice service defaults to occur even when the link average seems to meet requirements but fluctuations are frequent, resulting in insufficient accuracy of the generated scheduling scheme. Summary of the Invention

[0004] The purpose of this invention is to provide a dual-mode CPE resource mapping and scheduling method that supports network slicing, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the technical solution of the present invention is as follows: In a first aspect, this invention discloses a dual-mode CPE resource mapping and scheduling method supporting network slicing, applied to dual-mode CPE resource mapping and scheduling in tunnels, comprising the following steps: Obtain the chain segment status data and slice requirement data of the target object; The chain segment state data is subjected to dual-mode consistency encapsulation processing to generate standard chain segment data. The standard chain segment data and the slice requirement data are subjected to multi-segment serial projection processing and default probability mapping processing to generate default profile data. Risk importance coupling and level mapping processing are performed on the default profile data and the slice demand data to generate redundant level data. Resource boundary construction processing is then performed on the redundant level data and the standard chain segment data to generate boundary constraint data. The redundancy level data, the boundary constraint data, and the default profile data are subjected to adaptive selection of redundancy mode to generate redundancy mode data. Then, the redundancy mode data, the redundancy level data, the boundary constraint data, and the default profile data are subjected to dual-mode collaborative hybrid redundancy mapping to generate hybrid redundancy mapping data. The mixed redundancy mapping data, the redundancy mode data, the redundancy level data, and the boundary constraint data are processed to generate a redundancy scheduling scheme.

[0006] Secondly, this invention discloses a dual-mode CPE resource mapping and scheduling system supporting network slicing, comprising: The data acquisition module is used to acquire the chain segment status data and slice requirement data of the target object; The profiling module is used to perform dual-mode consistency encapsulation processing on the chain segment status data to generate standard chain segment data, and to perform multi-segment serial projection processing and default probability mapping processing on the standard chain segment data and the slice requirement data to generate default profile data. The constraint module is used to perform risk importance coupling processing and level mapping processing on the default profile data and the slice demand data to generate redundant level data, and to perform resource boundary construction processing on the redundant level data and the standard chain segment data to generate boundary constraint data. The mapping processing module is used to perform adaptive selection processing of redundancy mode on the redundancy level data, the boundary constraint data and the default profile data to generate redundancy mode data, and to perform dual-mode collaborative hybrid redundancy mapping processing on the redundancy mode data, the redundancy level data, the boundary constraint data and the default profile data to generate hybrid redundancy mapping data. The scheme generation module is used to perform scheduling scheme orchestration processing on the hybrid redundancy mapping data, the redundancy mode data, the redundancy level data and the boundary constraint data to generate a redundancy scheduling scheme.

[0007] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This solution performs dual-mode consistency encapsulation on chain segment state data and generates default profile data by combining it with slice requirement data. This enables probabilistic representation of uncertainties in multi-segment serial connections, reducing false triggers caused by mean judgment. Redundancy level data is generated from the default profile data and slice requirement data, and boundary constraint data is constructed to achieve coordinated determination of redundancy strength and overhead limit, suppressing resource crowding caused by excessive redundancy. Redundancy mode data is generated from the redundancy level data, boundary constraint data, and default profile data, forming hybrid redundancy mapping data. This enables adaptive switching of redundancy mode, dual-mode primary / backup, and soft concurrency, improving the stability of critical slices. Redundancy scheduling schemes are generated from the hybrid redundancy mapping data, enabling executable resource mapping and orchestration, and reducing jitter caused by frequent reconfiguration.

[0008] 2. This scheme performs scheme time window solidification processing on hybrid redundant mapping data under the constraint of periodic baseline data, generating solidified scheme data. This ensures that the parameters of the hybrid redundant mapping remain consistent within the same time window, reducing frequent reconfiguration and scheduling jitter caused by short-term fluctuations in chain segment status. It performs backoff sequence generation processing on profile shaping data, trusted label data, and sensitivity data, generating backoff sequence data. This forms a controllable backoff order when risks fall back and prioritizes the protection of highly sensitive slices, avoiding service interruptions caused by disordered decreases in concurrent parameters and redundancy levels. Under the constraints of candidate cost data and boundary constraint data, it performs amplitude limiting and protection processing on the backoff sequence data, generating amplitude limiting scheme data. It imposes upper limit constraints on concurrency and redundancy overhead and coordinates with the backoff order to suppress congestion caused by redundancy superposition. Attached Figure Description

[0009] The disclosure of this invention is illustrated with reference to the accompanying drawings. It should be understood that the drawings are for illustrative purposes only and are not intended to limit the scope of protection of this invention. In the drawings, the same reference numerals are used to refer to the same parts. Wherein: Figure 1 A flowchart illustrating the steps of a dual-mode CPE resource mapping and scheduling method supporting network slicing provided by the present invention; Figure 2 This invention provides a schematic diagram of the process for generating standard chain segment data. Figure 3 This invention provides a schematic diagram of the process for generating default profile data. Figure 4 A flowchart illustrating the process of generating redundancy level data provided by this invention; Figure 5 This is a schematic diagram of the module functions of a dual-mode CPE resource mapping and scheduling system that supports network slicing, provided by the present invention. Detailed Implementation

[0010] It is readily understood that, based on the technical solution of this invention, those skilled in the art can propose various interchangeable structural methods and implementations without altering the essential spirit of the invention. Therefore, the following detailed embodiments and accompanying drawings are merely illustrative examples of the technical solution of this invention and should not be considered as the entirety of the invention or as limitations or restrictions on the technical solution of this invention.

[0011] Application Overview: In traditional network slicing resource management technologies, for long-distance tunnel coverage scenarios, the backhaul link consists of multiple communication segments connected in series. The communication status of each segment is affected by factors such as spatial environment, equipment power supply, and interference reflection, exhibiting uncertain and non-stationary characteristics. Existing solutions are based on fixed thresholds or static level configurations, lacking characterization of the probability characteristics of link status fluctuations. This makes it difficult to accurately reflect the cumulative effect of overall communication risk under multi-segment series conditions. Consequently, when the average parameters of the links meet the requirements but fluctuations are frequent, slice service breaches are triggered, resulting in insufficient accuracy of the scheduling scheme. This, in turn, affects the isolation performance and service reliability of network slices, thus reducing their reliability.

[0012] For example, in the actual deployment of urban subway tunnel communication networks, CPE devices connect to base stations through multiple wireless links. The communication quality of each link is affected by tunnel structure reflections and train movement interference, resulting in high-frequency random fluctuations. Existing resource management mechanisms only monitor the status of a single link and execute a simple primary / backup switching strategy. When multiple links simultaneously experience temporary degradation, the cumulative risk effect is not identified, leading to interruption of data packet transmission for high-priority slice services, disruption of service continuity, and ineffective allocation of network resources.

[0013] If the above problems are not resolved, the frequency of slicing service defaults will increase, the efficiency of network resource allocation will be reduced, the overall stability of the system will decrease, and the service quality of latency-sensitive critical services will become uncontrollable, hindering the effective implementation of network slicing technology in complex communication environments.

[0014] After introducing the basic concept of the present invention, the embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Example 1:

[0015] Please see Figure 1 - Figure 4 A dual-mode CPE resource mapping and scheduling method supporting network slicing, applied to dual-mode CPE resource mapping and scheduling in tunnels, includes the following steps: Obtain the chain segment status data and slice requirement data of the target object; Perform dual-mode consistency encapsulation processing on the chain segment state data to generate standard chain segment data, and perform multi-segment concatenation projection processing and default probability mapping processing on the standard chain segment data and slice requirement data to generate default profile data. Risk importance coupling and level mapping processing are performed on default profile data and slice demand data to generate redundant level data. Resource boundary construction processing is then performed on redundant level data and standard chain segment data to generate boundary constraint data. Adaptive selection of redundancy methods is performed on redundancy level data, boundary constraint data, and default profile data to generate redundancy method data. Dual-mode collaborative hybrid redundancy mapping is then performed on redundancy method data, redundancy level data, boundary constraint data, and default profile data to generate hybrid redundancy mapping data. Perform scheduling scheme orchestration processing on mixed redundant mapping data, redundant mode data, redundant level data and boundary constraint data to generate a redundant scheduling scheme.

[0016] However, in practical applications, the chain segment status data from dual-mode CPE may have temporal inconsistencies, data volatility, and data differences between different communication modes. This poses a challenge to direct consistency encapsulation and may result in the generated standard chain segment data lacking reliability and accuracy, thereby affecting subsequent resource mapping and scheduling decisions.

[0017] In response, this application further proposes performing dual-mode consistency encapsulation processing on the chain segment state data to generate standard chain segment data, specifically including: Perform periodic consistency processing on the chain segment state data to generate periodic baseline data, and perform robust aggregation processing on the chain segment state data under the constraints of the periodic baseline data to generate robust chain segment data; Dual-mode consistency encapsulation is performed on the link segment state data under the two communication modes to generate dual encapsulation data, and trusted tag generation is performed on the robust link segment data and dual encapsulation data to generate trusted tag data. The periodic baseline data, robust chain segment data, dual encapsulated data, and trusted tag data are consistently merged and encapsulated to generate standard chain segment data.

[0018] Among them, the periodic consistency processing aims to eliminate inconsistencies in the original chain segment state data in terms of time sampling or reporting periods; Periodic baseline data is the result of periodic consistency processing. It defines a unified time reference frame or data sampling frequency and serves as the basis for subsequent data processing, ensuring that the state data of all chain segments are comparable in the time dimension. Robust aggregation processing is used to extract more representative and noise-resistant robust chain segment data from the original chain segment state data under the constraint of periodic baseline data; Robust chain segment data is chain segment state data after robust aggregation processing. It is aligned with the periodic benchmark data in time and has better stability and anti-interference ability in terms of value. Dual-mode conformance encapsulation processing is specifically designed for link state data in two communication modes (e.g., cellular network and Wi-Fi) of dual-mode CPE. Its purpose is to align and integrate data from different modes that describe the same or related link state to reveal the correlation or differences between modes. Dual encapsulation data is the result of dual-mode consistent encapsulation processing, which contains aligned and integrated link state information under two communication modes; The trusted tag generation process aims to evaluate the reliability or confidence level of robust chain segment data and dual encapsulation data, and generate corresponding trusted tag data. Trusted labeled data is a quantitative or categorical representation of the reliability of robust chain segment data and dual encapsulated data. It can be a Boolean value (trustworthy / untrustworthy), a confidence score, or a multi-level classification (high trustworthy, medium trustworthy, low trustworthy). Consistent merging and encapsulation processing is the final integration of periodic baseline data, robust chain segment data, dual encapsulation data, and trusted tag data to form unified standard chain segment data; Standard chain segment data is the final data product generated after a series of preprocessing and encapsulation processes.

[0019] This solution performs periodic consistency processing on the chain segment state data to eliminate time inconsistencies caused by different data sources or sampling frequencies, thereby generating unified periodic benchmark data. Within the time frame of this periodic benchmark data, the chain segment state data is robustly aggregated, and noise and outliers are filtered out using statistical methods to obtain more representative and stable robust chain segment data. Meanwhile, taking into account the characteristics of dual-mode CPE, dual-mode consistency encapsulation processing is performed on the link segment state data under the two communication modes to align and integrate the data of different modes to form dual encapsulated data to reflect the dual-mode collaborative state. In order to evaluate the reliability of these processed data, trust tag generation processing is further performed on the robust link segment data and dual encapsulated data to assign trust tags to the data.

[0020] Finally, the periodic baseline data, robust chain segment data, dual encapsulated data, and trusted tag data are consistently merged and encapsulated to generate standard chain segment data that is structurally complete, time-aligned, numerically robust, modally coordinated, and includes a trustworthiness assessment. Through this series of meticulous processing steps, the original, heterogeneous chain segment state data is transformed into a high-quality, highly reliable unified data format, providing a solid foundation for subsequent complex analysis and decision-making, and effectively avoiding resource mapping and scheduling deviations caused by data quality issues.

[0021] In some of the above implementations, after generating standard chain segment data, it is necessary to perform multi-segment concatenation projection processing and default probability mapping processing on the standard chain segment data and the slice requirement data to generate default profile data. However, in practical applications, the dynamism and uncertainty of network chain segments, as well as the diversity of slice requirements, make it difficult to accurately capture potential default risks by directly performing projection and mapping, which may result in insufficiently accurate default profile data, thereby affecting subsequent resource scheduling decisions.

[0022] In response, this application further proposes performing multi-segment concatenated projection processing and default probability mapping processing on standard chain segment data and slice requirement data to generate default profile data, specifically including: Under the constraint of periodic benchmark data, the standard chain segment data is subjected to segment sequence serial projection processing to generate serial projection data, and the serial projection data is subjected to uncertainty weighting processing based on the trusted label data to generate conservative factor data. By using sliced ​​demand data as service constraint input, and combining concatenated projection data with conservative factor data in default probability mapping processing, default profile data is generated.

[0023] This scheme generates serial projection data by performing segment sequence projection processing on standard link segment data under the constraint of periodic reference data. This simulates the cumulative effect of network slices being transmitted serially on multiple links, and transforms the performance indicators of a single link segment into end-to-end performance predictions that reflect the entire transmission path. For example, if a slice needs to pass through three network links, the processing will comprehensively consider parameters such as the latency and bandwidth of these three links to predict the total latency and available bandwidth of the slice on the entire path.

[0024] Given the dynamic and uncertain nature of the network environment, relying solely on concatenated projection data may not be sufficient to reflect potential risks. Therefore, this application further performs uncertainty weighting processing on the concatenated projection data based on trusted marker data to generate conservative factor data. The trusted marker data indicates the reliability of the standard chain segment data. When the data reliability is low, the uncertainty weighting processing will introduce greater conservatism, for example, by increasing the margin of prediction delay or reducing the upper limit of prediction bandwidth, thereby generating a more robust conservative factor data. Subsequently, the slice demand data is used as a service constraint input, and together with the concatenated projection data and conservative factor data, it is applied to the default probability mapping process to ultimately generate default profile data. The slice demand data clarifies the specific service quality requirements of the slice and serves as the benchmark for judging whether a default has occurred. The default probability mapping process comprehensively considers the predicted end-to-end performance, the conservative estimate of uncertainty, and the service requirements of the slice to calculate the probability of a service default occurring on a specific path. For example, if the predicted end-to-end latency, after adjustment by the conservative factor, is still higher than the maximum latency required by the slice, the default probability will increase accordingly. In this way, this application can more comprehensively and accurately assess the default risk of network slices and provide a reliable risk profile for subsequent resource scheduling.

[0025] In some of the embodiments described above in this application, a dual-mode consistency encapsulation process is proposed to be performed on the chain segment state data, and default profile data is generated by combining the slice requirement data to assess potential default risks. However, in practical applications, obtaining default profile data alone cannot directly guide how to effectively allocate resource redundancy, especially when facing complex and ever-changing tunnel network environments. A mechanism is needed to quantify the importance of different risks and convert them into specific redundancy levels to ensure the rationality and efficiency of resource scheduling.

[0026] In response, this application further proposes performing risk importance coupling and level mapping processing on default profile data and slice requirement data to generate redundant level data, specifically including: Risk importance coupling processing is performed on default profile data and slice requirement data to generate risk weight data. The specific calculation formula is as follows: , In the formula, This indicates the number of slices currently participating in the scheduling. Slice The probability profile value of default. Slice Importance weights, Slice The probability profile value of default. Slice Importance weights, This represents the sensitivity coefficient to the probability of default. The sensitivity coefficient represents the importance level; all the above data have been normalized during the calculation. Then, the risk weight data and the concatenated projection data are subjected to concatenated sensitivity extraction processing to generate sensitivity data; Based on the conservative factor data, a conservative mapping adjustment process is performed on the risk weight data and sensitivity data to generate grade mapping data. Then, grade consistency processing is performed on the grade mapping data under the allowable constraints of the slice requirement data to generate redundant grade data.

[0027] Among them, risk importance coupling processing is performed on default profile data and slice requirement data to assess the impact of different default risks on the overall system or service quality, thereby providing a basis for subsequent redundancy level classification. This processing can be based on a preset business rule base, according to the service level agreement (SLA) requirements defined in the slice requirement data, combined with the risk type and severity identified in the default profile data, to assign different importance weights to various risks. For example, for slices with extremely high reliability requirements, default risks related to data integrity may be given higher weights. In addition, machine learning models can also be used to learn and dynamically calculate risk weight data by analyzing the impact of different default events on service quality in historical data. Risk weight data is a numerical representation that quantifies the impact of different default risks on a system or service. It can be a discrete level value or a continuous percentage or score. The cascade sensitivity extraction process is used to analyze the sensitivity of cascaded projection data to changes in risk weights in order to identify the most critical links in the entire chain path that have the greatest impact on risk. This process can be performed by perturbation analysis on each chain segment or node in the cascaded projection data to observe how the risk weight data changes accordingly, thereby calculating the sensitivity coefficient of each chain segment or node. For example, the critical path analysis method in graph theory can be used in combination with risk weight data to identify sensitive paths or nodes that have a greater impact on the overall risk. Sensitivity data reflects a quantitative indicator of the degree of influence of each part of the concatenated projection data on the risk weight. It can be a numerical value representing the contribution of each chain segment or path to the risk weight. The conservative mapping adjustment process adjusts the risk weight and sensitivity data based on the conservative factor data to ensure that the redundancy level setting is sufficiently conservative and can effectively cope with the uncertainty in the system. The conservative factor data comes from the trusted tag data generated during the dual-mode consistency encapsulation process of the chain segment state data, reflecting the uncertainty or reliability of the data. When the conservative factor is high, it indicates that there is a large uncertainty in the data or assessment. At this time, the risk weight or sensitivity can be increased by introducing a safety margin or penalty factor, thereby tending to a higher redundancy level. For example, fuzzy logic or interval analysis methods can be used to use the conservative factor as an adjustment parameter to dynamically adjust the risk weight and sensitivity, so that it tends to a more conservative assessment when the uncertainty is high. The level mapping data is a preliminary redundancy level representation that combines risk weights and sensitivity after conservative adjustment. Under the constraints of the slice requirement data, the level consistency processing standardizes or unifies the level mapping data to ensure that the final generated redundancy level data meets the service level agreement (SLA) requirements and resource limitations of the slice. For example, if the level mapping data suggests that a slice needs "2N" redundancy, but the slice requirement data clearly specifies that the maximum allowable redundancy of the slice is "N+1", then the level consistency processing will adjust the redundancy level to "N+1" to meet the service constraints and resource budget of the slice. Redundancy level data is the final, specific level used to guide the redundancy configuration of resources. It can be a discrete level (such as no redundancy, N+1 redundancy, 2N redundancy, etc.) or a continuous proportion or number of redundant resources.

[0028] This solution uses risk importance coupling processing on default profile data and slice demand data to identify and quantify the impact of different default risks on service quality, thereby generating risk weight data. Based on this, combined with cascaded projection data, cascaded sensitivity extraction is performed to further clarify which segments or paths are more sensitive to high risks. Subsequently, conservative factor data was introduced to conservatively map and adjust the risk weight data and sensitivity data, effectively addressing the risk assessment bias caused by data uncertainty. This made the initial level mapping data more robust. Finally, under the allowable constraints of the sliced ​​demand data, the initial redundant levels were transformed into redundant level data that conformed to the actual service level agreement and resource constraints through level consistency processing. The entire process transformed the abstract default risk into an operable redundant level with priority and constraints, providing accurate and reliable input for the subsequent dual-mode collaborative hybrid redundant mapping processing. This made the allocation of resource redundancy more refined and intelligent, thereby improving the adaptability and robustness of the entire resource mapping and scheduling method.

[0029] In some of the embodiments described above in this application, although redundancy level data, boundary constraint data, and default profile data have been generated based on the chain segment status data and slice requirement data, how to adaptively select the most suitable redundancy method based on this information for different network slice requirements and risk characteristics to ensure the effectiveness and flexibility of resource scheduling remains a problem that needs to be solved.

[0030] In response, this application further proposes to perform adaptive selection processing of redundancy methods on redundancy level data, boundary constraint data, and default profile data, generating redundancy method data specifically including: Perform permissible domain consistency processing on slice requirement data and redundancy level data to generate feasible domain data, and perform candidate cost generation processing on feasible domain data under the joint constraints of default profile data, risk weight data and boundary constraint data to generate candidate cost data. Based on conservative factor data, conservative stable selection processing is performed on the profile shaping data to generate stable selection data; among which, the profile shaping data is obtained by performing risk distribution shaping processing on the default profile data; Perform slice-level uniqueness decision processing on feasible domain data, candidate cost data, and stable selection data to generate redundant mode data.

[0031] Specifically, the process of ensuring consistency between slice requirement data and redundancy level data in the permissible domain aims to define which redundancy methods are technically feasible and meet the redundancy level requirements, based on the specific requirements of network slices for service quality, bandwidth, latency, and reliability, as well as the redundancy levels required for different slices or links determined by risk assessment and importance analysis. This process can map specific slice requirement types (e.g., high reliability, low latency) to a set of permissible redundancy technologies (e.g., N+1 backup, M:N backup, dual-active, path diversity) based on predefined rule sets or lookup tables. Alternatively, it can utilize a policy-based engine to dynamically calculate and filter a set of redundancy methods that meet the conditions based on slice requirement parameters (e.g., maximum acceptable downtime, recovery time target) and redundancy levels (e.g., high, medium, low), thereby generating permissible domain data. Candidate cost generation processing is performed on the feasible mode data under the joint constraints of default profile data, risk weight data, and boundary constraint data. The purpose is to evaluate the cost or expense of each redundancy method in the feasible mode data. These costs include not only resource consumption, but also potential risk exposure, management complexity, etc. This processing can establish a multi-dimensional cost model for each redundancy method, comprehensively considering resource consumption (bandwidth, computing, storage), deployment complexity, operation and maintenance costs, and potential losses under default profile and risk weight. Then, the comprehensive cost is calculated to generate candidate cost data. Alternatively, optimization algorithms can be used to simulate and evaluate each redundancy method in the feasible mode under boundary constraint data (such as maximum resource budget, minimum performance requirements) to obtain its expected cost under specific risk scenarios. The conservative and stable selection process is applied to the profile shaping data based on conservative factor data. This process aims to select the most stable or reliable redundancy method from the profile shaping data, taking into account the degree of uncertainty or risk aversion during the decision-making process (i.e., conservative factor data), and generate stable selection data. This process can set a conservative threshold. Only when the risk exposure of the redundancy method is below the threshold and it shows high stability in the profile shaping data is it selected as a stable choice. Alternatively, a risk aversion function or utility function can be used to apply the conservative factor data to the profile shaping data to sort and filter the redundancy methods, giving priority to those schemes that can maintain performance even in the worst case. Profile shaping data is obtained by performing risk distribution shaping on default profile data. This process can smooth, normalize, or piecewise linearize the default probability distribution to eliminate noise or simplify the model, or convert non-Gaussian default profile data into an approximate Gaussian distribution so that decision-making methods based on mean and variance can be applied. The process involves performing slice-level uniqueness decision processing on feasible domain data, candidate cost data, and stable selection data. The goal is to comprehensively consider all input data and determine a final, unique redundancy method for each network slice, generating redundancy method data. This processing can employ a multi-objective optimization algorithm, using candidate cost data as the cost function and stable selection data as the reliability constraint within the feasible domain, to make trade-offs and select the optimal redundancy method. Alternatively, it can be based on an expert system or machine learning model, combining the feasible domain, candidate cost, and stable selection data with preset business priorities and strategies to make intelligent decisions and assign a unique redundancy method to each slice.

[0032] Through the above technical solution, this application can adaptively select the most suitable redundancy method based on multiple dimensions of information such as slice requirements, redundancy level, default profile, risk weight, boundary constraints, and conservative factors. This enables more refined and efficient use of network resources, avoiding over-configuration or under-configuration of resources. Thus, while meeting the diverse service quality requirements of different network slices, it significantly improves the overall resilience and reliability of the network. This solution effectively solves the technical challenge of how to intelligently select redundancy strategies to balance cost and reliability in complex network slice environments.

[0033] This application further proposes performing dual-mode collaborative hybrid redundancy mapping processing on redundancy mode data, redundancy level data, boundary constraint data, and default profile data to generate hybrid redundancy mapping data, specifically including: Asynchronous master-slave rhythm generation processing is performed on the periodic reference data and the dual encapsulated data to generate keep-alive rhythm data. Sensitivity data and risk weight data are introduced into the default profile data to perform soft concurrency parameter calculation processing, and concurrency parameter data is generated under the constraints of candidate cost data and boundary constraint data; Perform mode-level co-mapping processing on redundancy mode data and redundancy level data to generate co-mapping data. The specific calculation formula is as follows: , In the formula, T represents the mode-level collaborative mapping matrix. The redundant data representation vector represents slice s. The redundancy level data representation vector of slice s. This indicates the Kronecker product (tensor product) operation. All the data above has been normalized before the calculation. Modal load allocation processing is performed on standard chain segment data and serial projection data to generate load allocation data. Under stable selection data constraints, boundary constraint verification processing is performed on cooperative mapping data and load allocation data to generate hybrid redundant mapping data.

[0034] The asynchronous primary / backup rhythm generation process aims to generate keep-alive rhythm data based on periodic baseline data and dual encapsulated data. The core of this process is to provide a dynamic, asynchronous activation and switching mechanism for resource redundancy in a tunnel environment for dual-mode CPEs, thereby improving resource utilization and system response speed. This process can analyze periodic baseline data (e.g., network state sampling period, service periodicity) and dual encapsulated data (e.g., real-time performance of dual-mode links, fault indications), and use state machines or event-driven models to dynamically adjust the activation, heartbeat detection, and switching frequencies of primary and backup links, generating a series of timestamps or event-triggered rules as keep-alive rhythm data. Alternatively, a machine learning-based approach can be used to train a model using historical data, predict link stability under different modes, and intelligently generate asynchronous primary / backup switching strategies and keep-alive cycles based on service priorities, optimizing keep-alive rhythm data in a data-driven manner. The soft concurrency parameter calculation process aims to analyze default profile data by introducing sensitivity data and risk weight data, and generate concurrency parameter data under the constraints of candidate cost data and boundary constraint data. Its purpose is to allow a certain degree of "soft concurrency" during resource mapping, that is, to moderately share or use resources in parallel while meeting service quality requirements, thereby improving resource utilization. Simultaneously, risk assessment is used to control the degree of concurrency and avoid service degradation caused by excessive concurrency. This process can be achieved by establishing a multi-objective optimization model, taking default profile data (representing potential default risk), sensitivity data (representing the degree of sensitivity to risk), and risk weight data (representing the importance of risk) as input, combined with candidate cost data (representing the cost of different redundancy methods) and boundary constraint data (representing resource capacity limitations), to calculate the maximum concurrency within the range of ensuring service quality and controllable risk, and generate concurrency parameter data. Alternatively, a method based on fuzzy logic or an expert system can be used to comprehensively evaluate default profiles, sensitivity, and risk weights based on preset rules and empirical knowledge, and dynamically adjust concurrency parameters in conjunction with candidate costs and boundary constraints to achieve a balance between risk and efficiency. The method-level collaborative mapping process aims to collaboratively map redundancy method data and redundancy level data to generate collaborative mapping data. Its core lies in organically combining previously determined redundancy methods (e.g., N+1, 1+1, N-way, etc.) with redundancy levels (e.g., high, medium, low) to form a comprehensive redundancy strategy. This ensures that network slices of different importance levels can obtain redundancy protection that matches their risk tolerance and business needs, avoiding excessive or insufficient redundancy. This process can be achieved by constructing a mapping table or decision matrix to cross-match different redundancy methods with redundancy levels, defining a specific collaborative mapping strategy for each combination. Alternatively, it can employ a rule engine or optimization algorithm-based approach to dynamically generate optimal collaborative mapping data based on redundancy method and redundancy level data, combined with preset business strategies and resource cost models, to achieve the best balance between resource efficiency and reliability.

[0035] Modal load balancing processing aims to perform load distribution on standard link segment data and concatenated projection data, generating load distribution data. Its purpose is to rationally allocate network traffic and resource load in a dual-mode CPE environment based on the real-time status and predicted performance of different communication modes (e.g., cellular network and Wi-Fi) and service requirements, thereby optimizing overall network performance and avoiding single-mode overload or resource idleness. This processing can generate load distribution data by real-time monitoring of performance indicators such as bandwidth, latency, and packet loss rate of the two communication modes, combining standard link segment data (representing link stability and reliability) and concatenated projection data (representing future link performance predictions), and employing dynamic load balancing algorithms (e.g., based on minimum latency, maximum bandwidth, or minimum cost) to determine the traffic distribution ratio between different modes. Alternatively, it can use a control plane based on SDN (Software-Defined Networking) or NFV (Network Functions Virtualization) to obtain modal status and service requirements through a programming interface, and dynamically adjust routing and resource allocation rules according to preset strategies or optimization goals. Boundary constraint verification processing aims to verify the collaborative mapping data and load allocation data under stable selection data constraints, generating hybrid redundant mapping data. Its purpose is to ensure that the previously generated collaborative mapping strategy and load allocation scheme do not exceed preset resource boundaries and stable operating conditions during actual deployment. This is a final verification and adjustment mechanism to ensure that the generated hybrid redundant mapping data is feasible and robust. This processing can be achieved by comparing the collaborative mapping data (which defines the redundancy strategy) and load allocation data (which defines traffic allocation) with the stable selection data (representing conservatively selected stable operating parameters). It checks whether all preset boundary constraints such as resource capacity, performance indicators, and risk thresholds are met. If any are not met, iterative adjustments are made or warnings are issued until compliant hybrid redundant mapping data is generated. Alternatively, formal verification or simulation methods can be used. The collaborative mapping data and load allocation data are taken as input and simulated in the operating environment provided by the stable selection data to verify whether their behavior under various boundary conditions meets expectations. Optimization adjustments are then made based on the simulation results to ultimately generate reliable hybrid redundant mapping data.

[0036] Through the above technical solutions, this application enables refined control and optimization of dual-mode CPE resource mapping. Asynchronous primary / backup rhythm generation effectively improves the system's response speed and resource utilization efficiency to dynamic environmental changes, avoiding the limitations of traditional synchronous mechanisms. Soft concurrency parameter calculation ensures service quality while achieving elastic resource sharing through risk awareness, significantly improving resource utilization. Mode-level collaborative mapping ensures that network slices of different importance levels receive redundancy protection precisely matched to their needs, avoiding resource waste or insufficient protection. Modal load allocation fully leverages the advantages of dual-mode CPE, achieving intelligent cross-modal traffic scheduling and optimizing overall network performance. Finally, boundary constraint verification ensures the practical feasibility and robustness of all mapping schemes, thus providing network slices with more reliable, efficient, and adaptive resource redundancy and scheduling capabilities in complex and ever-changing tunnel environments, significantly improving service continuity and user experience.

[0037] In some of the embodiments described above in this application, a dual-mode collaborative hybrid redundancy mapping process is proposed to generate hybrid redundancy mapping data. However, in its implementation, how to transform this mapping data into a specific, executable, and highly reliable redundant scheduling scheme, and ensure its stability and fault recovery capability in a dynamic network environment, remains a problem that needs to be solved.

[0038] In response, this application further proposes to perform scheduling scheme orchestration processing on hybrid redundant mapping data, redundant mode data, redundant level data, and boundary constraint data, generating a redundant scheduling scheme, specifically including: Perform scheme time window solidification processing on the hybrid redundant mapping data under the constraint of periodic reference data to generate solidified scheme data; The backtracking sequence generation process is performed on the portrait reshaping data, trusted label data, and sensitivity data to generate backtracking sequence data. Under the constraints of candidate cost data and boundary constraint data, the backtracking sequence data is subjected to amplitude limiting and protection processing to generate amplitude limiting scheme data. Perform orchestration generation processing on the collaborative mapping data and load allocation data to generate execution orchestration data; The consistency of the solidified scheme data, rollback sequence data, amplitude limiting scheme data and execution orchestration data is checked to generate a redundant scheduling scheme.

[0039] The process involves solidifying the hybrid redundant mapping data under the constraints of periodic benchmark data, generating solidified scheme data. This aims to transform the initially generated hybrid redundant mapping data into solidified scheme data with clear time attributes and stability, ensuring that the scheduling scheme is deterministic and executable within a specific time window. By constraining the periodic benchmark data, the scheme can be aligned with the periodic characteristics of system operation, improving the stability and predictability of the scheme. This process can be achieved by aligning and aggregating the hybrid redundant mapping data with the periodic benchmark data. For example, the mapping data can be segmented according to the time granularity defined by the periodic benchmark data, and the mapping results within each time granularity can be summarized or averaged to form solidified scheme data with determinism in each period. Alternatively, a time window function or filter can be introduced to smooth the hybrid redundant mapping data and limit the time window, ensuring that the mapping results are stable and consistent within each preset periodic benchmark time window, thereby generating solidified scheme data. The process involves generating backoff sequence data by performing backoff sequence generation on profile shaping data, trusted label data, and sensitivity data. Under the constraints of candidate cost data and boundary constraint data, the backoff sequence data undergoes amplitude limiting and protection processing to generate amplitude limiting scheme data. This data is used to construct a backup strategy, i.e., a backoff sequence, that can be quickly switched over when the primary scheduling scheme fails. By using profile shaping data, trusted label data, and sensitivity data, potential risks and uncertainties can be assessed, thereby generating a highly targeted backoff scheme. Subsequent amplitude limiting and protection processing ensures that the backoff scheme is feasible in terms of resources and cost and can effectively address risks. The backoff sequence generation process can identify high-risk scenarios based on profile shaping data, determine data reliability using trusted label data, and assess the risk impact using sensitivity data, thus pre-designing a series of backup strategies for each potential failure point or high-risk area. Resource allocation or path switching schemes are used to form fallback sequence data. Bandwidth limiting and protection processing can then filter and optimize the fallback sequence based on candidate cost data (such as switching costs and resource consumption) and boundary constraint data (such as maximum available bandwidth and latency limits), eliminating impractical or overly costly schemes and adding necessary protection mechanisms (such as resource reservation and fast switching instructions), ultimately generating bandwidth limiting scheme data. Alternatively, fallback sequence generation processing can use decision trees or rule engines to construct a series of fault response rules based on profile shaping data, trusted label data, and sensitivity data. Each rule corresponds to a fallback action or sequence. Bandwidth limiting and protection processing can then introduce multi-objective optimization algorithms to minimize candidate cost data while satisfying boundary constraint data, thereby selecting the optimal bandwidth-limited protection scheme from the generated fallback sequence to form bandwidth limiting scheme data. The orchestration generation process performs orchestration generation on the collaborative mapping data and load allocation data to generate execution orchestration data. This aims to transform the collaborative mapping data and load allocation data generated during the dual-mode collaborative hybrid redundancy mapping process into directly executable instructions or configurations to guide CPEs and network devices in actual resource configuration and traffic scheduling. This is a crucial step in translating logical mapping into physical operations. The orchestration generation process can generate a series of network configuration scripts or API call instructions based on the redundancy policies defined in the collaborative mapping data (such as primary / standby switchover, N+M backup) and the traffic paths and bandwidths specified in the load allocation data. For example, it can configure CPE interface priorities, routing table entries, and network-side slice resource allocation instructions, thus forming execution orchestration data. Alternatively, an orchestration engine can take the collaborative mapping data and load allocation data as input, combine them with predefined network device templates and operating specifications, and automatically generate configuration command sets conforming to specific device vendors or network protocols. These command sets can be directly distributed to CPEs and network devices, constituting execution orchestration data. The final verification step in generating a redundant scheduling scheme involves performing consistency checks on the fixed scheme data, fallback sequence data, clipping scheme data, and execution orchestration data. This ensures that there are no conflicts between all sub-schemes (fixed scheme, fallback sequence, clipping scheme, and execution orchestration) and that they can work together to form a complete, reliable, and executable redundant scheduling scheme. Consistency checks can identify and correct potential logical errors or resource conflicts. Formal verification methods can be used, such as building a model to represent the relationships and constraints between the fixed scheme data, fallback sequence data, clipping scheme data, and execution orchestration data. Model checking or theorem proofs can then be used to verify whether these data meet all preset consistency rules, such as resource limits, correct switching logic, and latency requirements. Alternatively, consistency checks can be performed through simulation or model testing. All sub-schemes are loaded into a simulation environment to simulate different network states and fault scenarios. The execution results of each scheme are observed to check for resource conflicts, logical deadlocks, or service interruptions. Adjustments are made based on the test results until all schemes are consistent, ultimately generating the redundant scheduling scheme.

[0040] Through the above technical solutions, this application can transform the initially generated hybrid redundancy mapping results into a comprehensive, executable, and highly reliable redundant scheduling scheme. The scheme's time window solidification process ensures the stability and predictability of the scheduling scheme in the time dimension, avoiding scheduling chaos caused by dynamic changes. The generation of rollback sequences and the limiting and protection processes provide pre-planned, resource-controlled backup strategies for potential failures, significantly improving service resilience and fault recovery capabilities, and effectively reducing the risk of service interruption. The orchestration generation process transforms abstract mapping and allocation into specific network operation instructions, greatly simplifying the difficulty of deploying and managing complex network slice services. The final consistency verification process ensures the coordination and conflict-free nature of each sub-scheme, ensuring the integrity and reliability of the entire redundant scheduling scheme.

[0041] This application further proposes a scheme time window solidification process for hybrid redundant mapping data under the constraint of periodic reference data, and the generated solidified scheme data specifically includes: The hybrid redundancy mapping data is aligned with the periodic boundary based on the periodic reference data to generate aligned mapping data. The aligned mapping data and the redundancy level data are then processed to achieve mapping consistency, generating consistent mapping data. Perform mode consistency processing on the consistent mapping data and the redundant mode data to generate mode consistent data. Then, under the constraints of the boundary constraint data, perform solidified parameter convergence processing on the mode consistent data to generate solidified scheme data.

[0042] Through the above technical solutions, this application can systematically and meticulously solidify complex hybrid redundancy mapping data under multiple constraints, including periodic baseline data, redundancy level data, redundancy method data, and boundary constraint data. Specifically, by aligning periodic boundaries, the problem of inconsistent time granularity is solved, ensuring the uniformity of the scheduling scheme in the time dimension. By unifying the mapping, the redundant configuration of resource allocation is precisely matched with the risk importance of the business, avoiding resource waste or insufficient redundancy. By unifying the methods, the abstract redundancy level is transformed into a specific and operable redundancy implementation mechanism, enhancing the practicality of the scheme. Finally, by solidifying parameter convergence processing, the scheme is iteratively optimized under strict resource boundary constraints, ensuring that the generated solidified scheme data is stable, feasible, and has high resource utilization. This significantly improves the accuracy, reliability, and resource utilization efficiency of the redundant scheduling scheme.

[0043] This application further proposes performing clipping and protection processing on the backoff sequence data under the constraints of candidate cost data and boundary constraint data, generating clipping scheme data specifically including: The candidate cost data and boundary constraint data are subjected to cost consistency processing to generate cost constraint data. The fixed scheme data and cost constraint data are subjected to amplitude limit rule generation processing to generate amplitude limit rule data. Perform protection sequence fusion processing on the amplitude limiting rule data and the backoff sequence data to generate protection sequence data, and perform amplitude limiting adjustment processing on the solidified scheme data based on the protection sequence data to generate amplitude limiting scheme data.

[0044] Through the above technical solution, this application can ensure that the generated rollback sequence data is controllable in terms of resource consumption and operating costs, avoiding the problem that the rollback scheme cannot be effectively executed due to exceeding system resource limits or causing excessive operating overhead. By integrating the amplitude limit rule with the rollback sequence, the rollback scheme can automatically avoid potential risks and resource overrun problems during execution, thereby significantly improving the robustness, reliability and practicality of the scheduling scheme. Example 2:

[0045] Please see Figure 5 A dual-mode CPE resource mapping and scheduling system supporting network slicing, comprising: The data acquisition module is used to acquire the chain segment status data and slice requirement data of the target object; The profiling module is used to perform dual-mode consistency encapsulation processing on the chain segment state data to generate standard chain segment data, and to perform multi-segment concatenation projection processing and default probability mapping processing on the standard chain segment data and slice requirement data to generate default profile data. The constraint module is used to perform risk importance coupling and level mapping processing on default profile data and slice requirement data to generate redundant level data, and to perform resource boundary construction processing on redundant level data and standard chain segment data to generate boundary constraint data. The mapping processing module is used to perform adaptive selection of redundancy methods on redundancy level data, boundary constraint data, and default profile data to generate redundancy method data, and to perform dual-mode collaborative hybrid redundancy mapping processing on redundancy method data, redundancy level data, boundary constraint data, and default profile data to generate hybrid redundancy mapping data. The scheme generation module is used to perform scheduling scheme orchestration processing on hybrid redundancy mapping data, redundancy mode data, redundancy level data and boundary constraint data, and generate redundant scheduling schemes.

[0046] This embodiment has the same technical effects as Embodiment 1.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. All data mentioned in this application have undergone normalization and other preprocessing to achieve dimensional uniformity during calculation.

[0048] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for mapping and scheduling dual-mode CPE resources supporting network slicing, applied to dual-mode CPE resource mapping and scheduling in tunnels, characterized in that, Includes the following steps: Obtain the chain segment status data and slice requirement data of the target object; The chain segment state data is subjected to dual-mode consistency encapsulation processing to generate standard chain segment data. The standard chain segment data and the slice requirement data are subjected to multi-segment serial projection processing and default probability mapping processing to generate default profile data. Risk importance coupling and level mapping processing are performed on the default profile data and the slice demand data to generate redundant level data. Resource boundary construction processing is then performed on the redundant level data and the standard chain segment data to generate boundary constraint data. The redundancy level data, the boundary constraint data, and the default profile data are subjected to adaptive selection of redundancy mode to generate redundancy mode data. Then, the redundancy mode data, the redundancy level data, the boundary constraint data, and the default profile data are subjected to dual-mode collaborative hybrid redundancy mapping to generate hybrid redundancy mapping data. The mixed redundancy mapping data, the redundancy mode data, the redundancy level data, and the boundary constraint data are processed to generate a redundancy scheduling scheme.

2. The dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 1, characterized in that: Performing dual-mode consistency encapsulation processing on the chain segment state data to generate standard chain segment data specifically includes: Perform periodic consistency processing on the chain segment state data to generate periodic benchmark data, and perform robust aggregation processing on the chain segment state data under the constraints of the periodic benchmark data to generate robust chain segment data; The link segment state data under the two communication modes are subjected to dual-mode consistency encapsulation processing to generate dual encapsulation data, and the robust link segment data and the dual encapsulation data are subjected to trust tag generation processing to generate trust tag data. The periodic baseline data, the robust chain segment data, the dual encapsulation data, and the trusted tag data are subjected to a consistency merging and encapsulation process to generate standard chain segment data.

3. The dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 2, characterized in that: Performing multi-segment concatenated projection processing and default probability mapping processing on the standard chain segment data and the slice demand data to generate default profile data specifically includes: Under the constraints of the periodic reference data, the standard chain segment data is subjected to segment sequence serial projection processing to generate serial projection data, and the serial projection data is subjected to uncertainty weighting processing based on the trusted label data to generate conservative factor data. The sliced ​​demand data is used as the service constraint input, and the concatenated projection data and the conservative factor data are used together in the default probability mapping process to generate default profile data.

4. The dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 3, characterized in that: Performing risk importance coupling and level mapping processing on the default profile data and the slice demand data to generate redundant level data specifically includes: The default profile data and the slice demand data are subjected to risk importance coupling processing to generate risk weight data, and the risk weight data and the serial projection data are subjected to serial sensitivity extraction processing to generate sensitivity data. Based on the conservative factor data, the risk weight data and sensitivity data are subjected to conservative mapping adjustment processing to generate grade mapping data. Then, the grade mapping data is subjected to grade consistency processing under the allowable constraints of the slice requirement data to generate redundant grade data.

5. The dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 4, characterized in that: The redundant level data, the boundary constraint data, and the default profile data are subjected to adaptive selection of redundancy methods to generate redundancy method data, specifically including: The slice requirement data and the redundancy level data are subjected to allowable domain consistency processing to generate feasible domain data. Then, under the joint constraints of the default profile data, the risk weight data and the boundary constraint data, the feasible domain data are subjected to candidate cost generation processing to generate candidate cost data. Based on the conservative factor data, conservative and stable selection processing is performed on the profile shaping data to generate stable selection data; wherein, the profile shaping data is obtained by performing risk distribution shaping processing on the default profile data; Slice-level uniqueness decision processing is performed on the feasible domain data, the candidate cost data, and the stable selection data to generate redundant mode data.

6. The dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 5, characterized in that: Performing dual-mode collaborative hybrid redundancy mapping processing on the redundancy mode data, the redundancy level data, the boundary constraint data, and the default profile data to generate hybrid redundancy mapping data specifically includes: Asynchronous master-slave rhythm generation processing is performed on the periodic reference data and the dual encapsulated data to generate keep-alive rhythm data; The sensitivity data and risk weight data are introduced into the default profile data to perform soft concurrency parameter calculation processing, and concurrency parameter data is generated under the constraints of the candidate cost data and the boundary constraint data; Perform mode-level co-mapping processing on the redundancy mode data and the redundancy level data to generate co-mapping data; Modal load allocation processing is performed on the standard chain segment data and the serial projection data to generate load allocation data. Under the constraints of the stable selection data, boundary constraint verification processing is performed on the cooperative mapping data and the load allocation data to generate hybrid redundant mapping data.

7. The dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 6, characterized in that: Performing scheduling scheme orchestration processing on the hybrid redundancy mapping data, the redundancy mode data, the redundancy level data, and the boundary constraint data to generate a redundancy scheduling scheme specifically includes: The hybrid redundant mapping data is subjected to scheme time window solidification processing under the constraints of the periodic reference data to generate solidified scheme data; The image reshaping data, the trusted label data, and the sensitivity data are subjected to backtracking sequence generation processing to generate backtracking sequence data. Under the constraints of the candidate cost data and the boundary constraint data, the backtracking sequence data is subjected to amplitude limiting and protection processing to generate amplitude limiting scheme data. The coordinated mapping data and the load allocation data are subjected to orchestration generation processing to generate execution orchestration data; The consistency of the solidification scheme data, the rollback sequence data, the amplitude limiting scheme data, and the execution orchestration data is checked to generate a redundant scheduling scheme.

8. The dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 7, characterized in that: The hybrid redundant mapping data is subjected to scheme time window solidification processing under the constraints of the periodic reference data to generate solidified scheme data, specifically including: The periodic boundary is aligned with the hybrid redundancy mapping data based on the periodic reference data to generate aligned mapping data, and the alignment mapping data and the redundancy level data are subjected to mapping consistency processing to generate consistent mapping data. The consistent mapping data and the redundant mode data are processed to achieve mode consistency, generating mode consistent data. Under the constraints of the boundary constraint data, the mode consistent data is processed to achieve solidified parameter convergence, generating solidified scheme data.

9. A dual-mode CPE resource mapping and scheduling method supporting network slicing according to claim 7, characterized in that: Under the constraints of the candidate cost data and the boundary constraint data, the backoff sequence data is subjected to clipping and protection processing to generate clipping scheme data, specifically including: The candidate cost data and the boundary constraint data are subjected to cost consistency processing to generate cost constraint data, and the fixed scheme data and the cost constraint data are subjected to amplitude limiting rule generation processing to generate amplitude limiting rule data. The limiting rule data and the rollback sequence data are subjected to protection sequence fusion processing to generate protection sequence data. The limiting scheme data is then subjected to limiting adjustment processing based on the protection sequence data to generate limiting scheme data.

10. A dual-mode CPE resource mapping and scheduling system supporting network slicing, characterized in that, include: The data acquisition module is used to acquire the chain segment status data and slice requirement data of the target object; The profiling module is used to perform dual-mode consistency encapsulation processing on the chain segment status data to generate standard chain segment data, and to perform multi-segment serial projection processing and default probability mapping processing on the standard chain segment data and the slice requirement data to generate default profile data. The constraint module is used to perform risk importance coupling processing and level mapping processing on the default profile data and the slice demand data to generate redundant level data, and to perform resource boundary construction processing on the redundant level data and the standard chain segment data to generate boundary constraint data. The mapping processing module is used to perform adaptive selection processing of redundancy mode on the redundancy level data, the boundary constraint data and the default profile data to generate redundancy mode data, and to perform dual-mode collaborative hybrid redundancy mapping processing on the redundancy mode data, the redundancy level data, the boundary constraint data and the default profile data to generate hybrid redundancy mapping data. The scheme generation module is used to perform scheduling scheme orchestration processing on the hybrid redundancy mapping data, the redundancy mode data, the redundancy level data and the boundary constraint data to generate a redundancy scheduling scheme.

Images