Optimization control method and system for OTN network

Abstract

The invention relates to the technical field of optical communication networks, and discloses an optimal control method and system for an OTN (Optical Transport Network), and the method comprises the steps: dynamic network modeling, dynamic weight distribution, path evaluation and selection, adaptive routing adjustment and real-time path slicing. The system applies the method. According to the method, through dynamic network modeling, intelligent dynamic weight distribution, nonlinear path evaluation and selection, an adaptive routing strategy and a dynamic real-time path slicing technology, the intelligent level, the resource utilization efficiency and the business service quality of the OTN network are improved from multiple dimensions; the problems that a traditional OTN system is poor in adaptability, inflexible in scheduling, unreasonable in resource allocation and the like in a complex network environment are solved.

Description

Technical Field

[0001] The present application relates to the technical field of optical communication networks, and in particular to an optimization control method and system for an OTN network. Background Art

[0002] OTN, or Optical Transport Network, is a transport network based on wavelength division multiplexing technology. It organizes networks at the optical layer and provides transmission, multiplexing, routing, management, monitoring, and protection (survivability) of client signals based on optical channels. Its basic architecture includes the optical channel layer (OCH), optical multiplex section layer (OMS), optical transport section layer (OTS), optical path payload unit (OPU), optical path data unit (ODU), and optical path transport unit (OTU).

[0003] From a functional perspective, OTN utilizes all-optical transmission within subnets and optical-to-electrical-to-optical conversion at subnet boundaries. Subnets can be connected via 3R regenerators (re-amplification, re-shaping, and re-timing), forming a large optical network. At the edge of the OTN network, signals from different systems are uniformly encapsulated into the overhead-rich OTN frame structure. This enables unified operation, management, and maintenance across the entire OTN network, facilitating fault location and ensuring excellent network survivability.

[0004] While traditional OTN systems have made significant progress compared to traditional WDM networks in terms of improving wavelength / sub-wavelength service scheduling capabilities, enhancing networking flexibility, and improving protection mechanisms, they still face numerous limitations when faced with complex and changing network environments and diverse service demands. In particular, in scenarios with high loads or diverse service types, traditional OTN networks struggle to efficiently match appropriate optical channels for transmission, leading to reduced data transmission efficiency and increased risk of service interruptions. This severely restricts the application and development of OTN networks in new high-speed, high-capacity, multi-service transport scenarios.

[0005] An optimization control technology is urgently needed to improve the overall performance and service quality of OTN networks and meet the growing flexibility and reliability requirements of modern communication networks. Summary of the Invention

[0006] The purpose of this application is to provide an optimization control method and system for an OTN network to solve the technical problems raised in the above background technology.

[0007] To achieve the above objectives, this application discloses the following technical solutions:

[0008] In a first aspect, the present application discloses an optimization control method and system for an OTN network, the method comprising the following steps:

[0009] Dynamic network modeling: Based on the OTN network topology and service requirements, a multi-dimensional evaluation matrix is constructed, including nodes, links, and service flows. Parameters in the multi-dimensional evaluation matrix include link bandwidth utilization, latency, packet loss rate, link length, optical signal quality, and service priority.

[0010] Dynamic weight allocation: Based on real-time network status and service requirements, the weight coefficients of each parameter in the multi-dimensional evaluation matrix are dynamically adjusted through a machine learning model to generate a dynamic weight vector;

[0011] Path evaluation and selection: using the dynamic weight vector to perform nonlinear weighted calculation on the multidimensional evaluation matrix, determine the comprehensive evaluation value of each link, and select several links with the best and relatively good comprehensive evaluation values as candidate paths for service signal transmission;

[0012] Adaptive routing adjustment: Based on service priority and network congestion status, an adaptive routing strategy is used to dynamically adjust several candidate paths and determine a transmission path as the final path;

[0013] Real-time path slicing: During service signal transmission, the selected final path is dynamically divided into multiple sub-paths based on real-time changes in link status. Each sub-path is independently optimized to dynamically adjust end-to-end transmission efficiency.

[0014] Preferably, among the parameters in the multidimensional evaluation matrix:

[0015] The link bandwidth utilization is obtained by collecting service flow data on the link and calculating the current bandwidth utilization;

[0016] The delay is calculated based on the physical distance of the link and the signal propagation speed, combined with the node processing time, to obtain the end-to-end time delay;

[0017] The packet loss rate is obtained by monitoring the packet loss events of the link and calculating the packet loss rate per unit time;

[0018] The link length is obtained by calculating the physical length of the link based on the geographical layout of the OTN network;

[0019] The optical signal quality is obtained by evaluating the signal-to-noise ratio and bit error rate of the optical signal;

[0020] The service priority is allocated according to the service type and QoS requirements.

[0021] Preferably, the method of dynamically adjusting the weight coefficients of the parameters in the multidimensional evaluation matrix through a machine learning model to generate a dynamic weight vector specifically includes:

[0022] Initial weight setting: Set the initial weight coefficient of each parameter based on business needs and network planning;

[0023] Real-time weight adjustment: Based on the real-time data of the multi-dimensional evaluation matrix, the weight coefficient is dynamically adjusted through the machine learning model to generate a dynamic weight vector;

[0024] Weight verification: Convergence verification of dynamic weight vectors.

[0025] Preferably, the dynamically dividing the selected final path into multiple sub-paths and independently optimizing the sub-paths to achieve dynamic adjustment of the end-to-end transmission efficiency specifically includes:

[0026] Path segmentation: Divide the selected transmission path into multiple sub-paths, each sub-path corresponds to an independent link segment;

[0027] Subpath optimization: For each subpath, the comprehensive evaluation value is calculated independently, and the optimal subpath is selected based on the dynamic weight vector;

[0028] Dynamic splicing: Splice the optimized sub-paths into a complete transmission path and update the path status in real time.

[0029] Preferably, the machine learning model includes:

[0030] Input layer: receives real-time data of each parameter in the multidimensional evaluation matrix;

[0031] Hidden layer: extracts the correlation between parameters through the neural network model;

[0032] Output layer: Generates dynamic weight vectors and outputs them to the path evaluation and selection steps;

[0033] And, the loss function.

[0034] Preferably, the adaptive routing strategy includes: a load balancing strategy and a fault recovery strategy, or a priority strategy and a fault recovery strategy; wherein:

[0035] The load balancing strategy is specifically as follows: among several candidate paths, the candidate path with the smallest load is selected as the final path, and the other candidate paths are retained as backup paths;

[0036] The priority strategy is specifically as follows: based on the service priority, the candidate path that meets the high-priority service requirements is preferentially selected as the final path, and other candidate paths are retained as backup paths;

[0037] The fault recovery strategy is specifically: when a link abnormality is detected in the selected final path, automatically switching to a backup path and updating the multi-dimensional evaluation matrix.

[0038] Preferably, the fault recovery strategy specifically includes:

[0039] Anomaly detection: Detects link anomalies by monitoring the link's packet loss rate and optical signal quality;

[0040] Path switching: When a link anomaly is detected, a backup path is selected and the multi-dimensional evaluation matrix is updated;

[0041] Recovery verification: After the path is switched, the comprehensive evaluation value of the new path is verified to see if it meets the business requirements, and the dynamic weight vector is adjusted.

[0042] Preferably, the adaptive routing adjustment further includes:

[0043] Based on historical traffic data and current network status, a time series prediction model is used to predict the link load change trend in the future T time window as the prediction result, and the prediction result is input into the adaptive routing strategy.

[0044] Preferably, the load balancing strategy and the priority strategy further include:

[0045] According to the prediction results, several potential candidate paths are generated, and based on the service priority, the potential candidate paths that meet the conditions are screened out as new backup paths; the screening conditions are specifically:

[0046] For high-priority services, the comprehensive evaluation value of the potential candidate path must be greater than the preset first minimum evaluation value;

[0047] For medium / low priority services, the comprehensive evaluation value of the potential candidate path must be greater than the preset second lowest evaluation value;

[0048] The first lowest evaluation value is smaller than the second lowest evaluation value.

[0049] In a second aspect, the present application discloses an optimization control system for an OTN network, applying the above-described optimization control method for an OTN network, the system comprising:

[0050] The dynamic network modeling module is configured to: construct a multi-dimensional evaluation matrix including nodes, links, and service flows based on the topology of the OTN network and service requirements. Parameters in the multi-dimensional evaluation matrix include link bandwidth utilization, latency, packet loss rate, link length, optical signal quality, and service priority;

[0051] A dynamic weight allocation module is configured to dynamically adjust the weight coefficients of the parameters in the multidimensional evaluation matrix through a machine learning model according to real-time network status and service requirements to generate a dynamic weight vector;

[0052] a path evaluation and selection module configured to: perform nonlinear weighted calculation on the multidimensional evaluation matrix using the dynamic weight vector to determine a comprehensive evaluation value of each link, and select several links with the best and relatively good comprehensive evaluation values as candidate paths for service signal transmission;

[0053] The adaptive routing adjustment module is configured to: combine service priority and network congestion status, use adaptive routing strategies to dynamically adjust several candidate paths and determine a transmission path as the final path to ensure the transmission quality of critical service signals;

[0054] The real-time path slicing module is configured to dynamically split the selected final path into multiple sub-paths according to the real-time changes in the link status during service signal transmission, and independently optimize the sub-paths to achieve dynamic adjustment of end-to-end transmission efficiency.

[0055] Beneficial effects: The optimization control method and system for OTN networks of the present application improve the intelligence level, resource utilization efficiency and business service quality of OTN networks from multiple dimensions through dynamic network modeling, intelligent dynamic weight allocation, nonlinear path evaluation and selection, adaptive routing strategy and dynamic real-time path slicing technology, and overcome the problems of poor adaptability, inflexible scheduling and unreasonable resource allocation of traditional OTN systems in complex network environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present application. For those skilled in the art, other drawings can be obtained based on these drawings without paying any creative work.

[0057] Figure 1 This is a flowchart of an optimization control method for an OTN network provided in an embodiment of the present application. DETAILED DESCRIPTION

[0058] The following is a clear and complete description of the technical solutions in the embodiments of this application. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0059] In this document, the term "comprising" is intended to encompass non-exclusive inclusion, such that a process, method, article, or apparatus that includes a list of elements includes not only those elements but also other elements not explicitly listed, or elements inherent to such process, method, article, or apparatus. In the absence of further limitations, an element defined by the phrase "comprising..." does not preclude the presence of additional identical elements in the process, method, article, or apparatus that includes the element.

[0060] In a first aspect, this embodiment provides a Figure 1 The optimization control method for OTN networks shown here aims to address the shortcomings of traditional OTN systems in terms of overall control of network transmission and optimal resource allocation, especially the problem of poor data transmission efficiency caused by the inability to effectively match optical channels in high-load scenarios. Specifically, this method includes the following steps:

[0061] S1-Dynamic network modeling: Based on the topology and service requirements of the OTN network, a multi-dimensional evaluation matrix including nodes, links and service flows is constructed. The parameters in the multi-dimensional evaluation matrix include link bandwidth utilization, latency, packet loss rate, link length, optical signal quality and service priority. In this step, by constructing an evaluation matrix including multi-dimensional parameters such as link bandwidth utilization, latency, packet loss rate, link length, optical signal quality and service priority, a full range of network status perception is achieved. Compared with the traditional static modeling method that relies only on a single indicator (such as bandwidth), the accuracy and adaptability of network evaluation are significantly improved, providing a more scientific decision-making basis for subsequent path selection;

[0062] S2-Dynamic Weight Allocation: Based on the real-time network status and service requirements, the weight coefficients of each parameter in the multidimensional evaluation matrix are dynamically adjusted through a machine learning model to generate a dynamic weight vector. This step uses a machine learning model to dynamically adjust the weights of each parameter in the multidimensional evaluation matrix, enabling the network to flexibly configure resource priorities based on the actual needs of different service types (such as real-time services and large data transmission services). This solves the problem of traditional OTN systems lacking differentiated processing capabilities when facing diversified services, and improves the intelligence level of network scheduling and service quality.

[0063] S3-Path evaluation and selection: Use the dynamic weight vector to perform nonlinear weighted calculations on the multidimensional evaluation matrix to determine the comprehensive evaluation value of each link, and select several links with the best and relatively good comprehensive evaluation values as candidate paths for service signal transmission. In this step, the dynamic weight vector is used to perform nonlinear weighted calculations on the multidimensional evaluation matrix, avoiding the evaluation imbalance problem that may occur in traditional linear weighting methods. By comprehensively considering multiple influencing factors, it is ensured that the selected transmission path achieves the optimal balance in terms of stability, efficiency, and priority protection, thereby improving the transmission efficiency and reliability of the entire network;

[0064] S4-Adaptive routing adjustment: Based on service priority and network congestion status, an adaptive routing strategy is used to dynamically adjust several candidate paths and determine a transmission path as the final path to ensure the transmission quality of critical service signals. In this step, based on service priority and network congestion status, an adaptive routing strategy is used to dynamically adjust the transmission path, allowing high-priority services (such as industrial control and medical remote operation) to obtain stable transmission channels even when resources are limited. This makes up for the shortcomings of traditional OTN networks in service priority assurance and significantly improves the network's service level capabilities and survivability.

[0065] S5 - Real-time Path Slicing: During service signal transmission, the selected final path is dynamically divided into multiple sub-paths based on real-time changes in link status. These sub-paths are then independently optimized to dynamically adjust end-to-end transmission efficiency. This step, using the real-time path slicing mechanism, divides the selected path into multiple sub-paths and independently optimizes them during service transmission. This allows for rapid response to link status changes (such as sudden congestion or localized failures), enabling refined management of transmission paths. This effectively alleviates the inflexibility inherent in traditional OTN networks, where once selected paths are immutable, and enhances network resilience and adaptability.

[0066] The above-mentioned optimization control method for OTN networks improves the resource scheduling efficiency and service quality of OTN networks in high-load scenarios, and enhances the network's ability to respond to emergencies.

[0067] In one embodiment, the parameters in the multidimensional evaluation matrix are obtained as follows:

[0068] The link bandwidth utilization is calculated by collecting service flow data on the link through SNMP protocol or NetFlow, and calculating the current bandwidth utilization to obtain ρ;

[0069] The delay is based on the physical distance of the link and the signal propagation speed, combined with the node processing time, to calculate the end-to-end time delay to obtain Δt;

[0070] The packet loss rate is obtained by monitoring the packet loss events of the link and calculating the packet loss rate per unit time to obtain γ;

[0071] The link length is based on the geographical layout of the OTN network, and the physical length of the link is calculated to obtain s;

[0072] The optical signal quality is evaluated by the signal-to-noise ratio and bit error rate of the optical signal to obtain the optical signal quality Q;

[0073] The service priority is allocated to obtain P according to the service type (such as IP, Ethernet, SDH, etc.) and QoS requirements.

[0074] After collecting all parameters, the data is standardized and dimensionalized, and all parameters are integrated into a six-dimensional vector, which serves as the basis for subsequent evaluation. This enables the construction of a comprehensive and accurate network status assessment system, providing a reliable basis for intelligent path selection.

[0075] Furthermore, the method of dynamically adjusting the weight coefficients of the parameters in the multidimensional evaluation matrix through the machine learning model to generate a dynamic weight vector specifically includes:

[0076] Initial weight setting: Set the initial weight coefficients of each parameter based on business needs and network planning (such as historical data and expert experience);

[0077] Real-time weight adjustment: Based on the real-time data of the multi-dimensional evaluation matrix, the weight coefficient is dynamically adjusted through the machine learning model to generate a dynamic weight vector;

[0078] Weight verification: Use the loss function of the machine learning model to verify the convergence of the dynamic weight vector. If the dynamic weight vector does not converge, continue to iterate to ensure the stability and effectiveness of the weight adjustment.

[0079] It is feasible that the dynamic weight vector adjustment formula is:

[0080]

[0081] Among them, w i (t) is the weight of the i-th parameter at time t; f i (t) is the real-time evaluation value of the i-th parameter (such as bandwidth utilization ρ, latency Δt, etc.); α and β are dynamic adjustment coefficients, satisfying α + β = 1, with α used to retain historical weights and β used to introduce the influence of the current evaluation value. This embodiment dynamically adapts to changes in network status through a nonlinear combination of historical weights and real-time evaluation values.

[0082] By using the above-mentioned dynamic weight vector acquisition method, differentiated resource scheduling for different service types can be achieved, thereby improving network flexibility and intelligence.

[0083] In one embodiment, the dynamic division of the selected final path into multiple sub-paths and independent optimization of the sub-paths to achieve dynamic adjustment of the end-to-end transmission efficiency specifically includes:

[0084] Path segmentation: Divide the selected transmission path into multiple sub-paths. The division can be based on any existing technology, such as determining it based on the link length ratio. Each sub-path corresponds to an independent link segment.

[0085] Subpath optimization: For each subpath, the comprehensive evaluation value is calculated independently, and the optimal subpath is selected based on the dynamic weight vector;

[0086] Dynamic splicing: Splice the optimized sub-paths into a complete transmission path and update the path status in real time.

[0087] It should be noted that during signal transmission, if changes in link status are monitored in real time (such as an increase in packet loss rate), the path slicing mechanism can be triggered again to perform path optimization.

[0088] It is feasible that the calculation formula of the comprehensive evaluation value of the subpath is:

[0089]

[0090] Among them, C sub is the comprehensive evaluation value of the subpath; w k (t) is the dynamic weight vector of the kth parameter; f k is the real-time value of the kth parameter; f max,k is the maximum threshold of the kth parameter; ε k is the nonlinear correction coefficient of the kth parameter (dynamically adjusted by the machine learning model). In this embodiment, the nonlinear correction coefficient ε k Normalization parameters Perform index correction and combine it with the dynamic weight vector w k (t) are multiplied, and finally the comprehensive evaluation value C of the sub-path is obtained by summing sub , to flexibly adapt to the sensitivity of different parameters.

[0091] The above-mentioned sub-path slicing and independent optimization design improves end-to-end transmission efficiency, enhances the network's ability to respond to emergencies, and avoids service interruptions caused by failure of the entire path.

[0092] In one embodiment, the machine learning model includes:

[0093] Input layer: receives real-time data of each parameter in the multidimensional evaluation matrix;

[0094] Hidden layer: extracts the correlation between parameters through the neural network model;

[0095] Output layer: Generates dynamic weight vectors and outputs them to the path evaluation and selection steps;

[0096] And, the loss function.

[0097] Based on this, the specific implementation process of the machine learning model in this embodiment is:

[0098] Build a neural network model with the network state vector as input;

[0099] Use the loss function for backpropagation training;

[0100] Output dynamic weight vector for use by path evaluation module;

[0101] Model parameters are updated regularly to adapt to network changes.

[0102] It is feasible that the loss function of the neural network model is:

[0103]

[0104] in, The weights predicted by the model; The target weight is dynamically generated by the network state; σ is a smoothing coefficient (optimized through experiments). In this embodiment, the loss function avoids excessive sensitivity to outliers by designing an exponential decay term, thereby improving the robustness of the model.

[0105] Based on the design of the above machine learning model, the model's anti-interference ability is improved, ensuring that the weight distribution is closer to actual network requirements.

[0106] In one embodiment, the adaptive routing strategy includes: a load balancing strategy and a fault recovery strategy, or a priority strategy and a fault recovery strategy; wherein:

[0107] The load balancing strategy is specifically as follows: among several candidate paths, the candidate path with the smallest load (the smallest link utilization) is selected as the final path, and the other candidate paths are retained as backup paths;

[0108] The priority strategy is specifically as follows: based on the service priority, the candidate path that meets the high-priority service requirements (i.e., the path that meets QoS) is preferentially selected as the final path, and other candidate paths are retained as backup paths;

[0109] The fault recovery strategy is specifically: when a link abnormality is detected in the selected final path, automatically switching to a backup path and updating the multi-dimensional evaluation matrix.

[0110] Based on this, the implementation process of the adaptive routing strategy is as follows:

[0111] Regularly collect link load and service priority across the entire network;

[0112] If a link is congested, the load balancing policy is enabled;

[0113] If high-priority services need to be protected, enable the priority policy;

[0114] During the execution of the load balancing policy and the priority policy, if a link abnormality is detected, the backup path will be switched.

[0115] Through the design of the above adaptive routing strategy, the service quality of key businesses is guaranteed and the survivability and robustness of the network are improved.

[0116] Furthermore, the fault recovery strategy specifically includes:

[0117] Anomaly detection: Detects link anomalies by monitoring the link's packet loss rate and optical signal quality;

[0118] Path switching: When a link anomaly is detected, a backup path is selected and the multi-dimensional evaluation matrix is updated;

[0119] Recovery verification: After the path is switched, the comprehensive evaluation value of the new path (i.e., the selected backup path) is verified to see if it meets the service requirements, and the dynamic weight vector is adjusted.

[0120] It is feasible that the formula for recovery verification is:

[0121]

[0122] Among them, R is the recovery verification index, ranging from (0,1]; C new is the comprehensive evaluation value of the new path; C old is the comprehensive evaluation value of the original path; t recovery is the path switching time; t max is the maximum allowed recovery time. Among them, the larger the calculated recovery verification index is, the better the comprehensive performance and recovery efficiency of the new path are. Rmin is set according to business needs (for example, Rmin = 0.8). When R≥Rmin is calculated, it means that the new path meets the business needs. Otherwise, it is necessary to reselect a new path or adjust the fault recovery strategy. The feasibility of path switching is quantified by a nonlinear combination of the comprehensive evaluation value ratio and the recovery time. Secondly, in this embodiment, the link anomaly is determined by obtaining the packet loss rate γ and the optical signal quality Q through hardware sensors, protocol packets (such as BFD, ICMPPing), log analysis (such as Prometheus+Grafana), etc., and calculating the link anomaly detection index. If the link anomaly detection index is greater than the dynamic threshold T dynamic The link is judged to be abnormal, where T dynamicDynamically adjust based on the mean and standard deviation of historical link anomaly detection indicators. The link anomaly detection formula is:

[0123]

[0124] Where D is the link anomaly detection indicator; γ is the current packet loss rate; γ th is the packet loss rate threshold (such as γ th =1%); Q is the current optical signal quality; Q th is the optical signal quality threshold (such as Q th =-30dBm); η is a dynamic correction factor (dynamically adjusted based on historical anomaly data). In this embodiment, the packet loss rate and optical signal quality are respectively calculated by ratioing them to their thresholds, i.e., normalizing the comprehensive packet loss rate and optical signal quality to quantify the degree of deviation between the current state and the threshold. The dynamic correction factor η is dynamically adjusted based on historical anomaly data to achieve nonlinear detection of link anomalies, thereby improving the flexibility and accuracy of detection.

[0125] By designing the above fault recovery strategy, the rationality of path switching can be effectively evaluated to avoid service interruption caused by switching failure.

[0126] As a preferred implementation of this embodiment, the adaptive routing adjustment further includes:

[0127] Based on historical traffic data and current network status, a time series prediction model (such as LSTM or Prophet) is used to predict the link load change trend in the future T time window as the prediction result, and the prediction result is input into the adaptive routing strategy.

[0128] Based on the acquisition of prediction results, the adaptive routing strategy can increase the consideration of future change trends during execution, thereby improving the adaptability to dynamic network changes.

[0129] Furthermore, the load balancing strategy and the priority strategy further include:

[0130] According to the prediction results, several potential candidate paths are generated, and based on the service priority, the potential candidate paths that meet the conditions are screened out as new backup paths; the screening conditions are specifically:

[0131] For high-priority services, the comprehensive evaluation value of the potential candidate path must be greater than the preset first minimum evaluation value;

[0132] For medium / low priority services, the comprehensive evaluation value of the potential candidate path must be greater than the preset second lowest evaluation value;

[0133] The first lowest evaluation value is smaller than the second lowest evaluation value.

[0134] By adding new backup paths and utilizing the aforementioned prediction results, we can provide a reference for future network changes in the execution of fault recovery strategies, thereby selecting backup paths that are more in line with network change trends, ensuring the service quality of key businesses and improving network survivability.

[0135] In summary, the optimization and control method for OTN networks in this embodiment improves the intelligence level, resource utilization efficiency, and service quality of OTN networks from multiple dimensions through technologies such as dynamic network modeling, intelligent dynamic weight allocation, nonlinear path evaluation and selection, adaptive routing strategies, and dynamic real-time path slicing. It overcomes the problems of traditional OTN systems in complex network environments, such as poor adaptability, inflexible scheduling, and unreasonable resource allocation. It has good application prospects and promotion value.

[0136] In a second aspect, this embodiment provides an optimization control system for an OTN network. The optimization control method for an OTN network described above is applied to the system. The system includes:

[0137] The dynamic network modeling module is configured to: construct a multi-dimensional evaluation matrix including nodes, links, and service flows based on the topology of the OTN network and service requirements. Parameters in the multi-dimensional evaluation matrix include link bandwidth utilization, latency, packet loss rate, link length, optical signal quality, and service priority;

[0138] A dynamic weight allocation module is configured to dynamically adjust the weight coefficients of the parameters in the multidimensional evaluation matrix through a machine learning model according to real-time network status and service requirements to generate a dynamic weight vector;

[0139] a path evaluation and selection module configured to: perform nonlinear weighted calculation on the multidimensional evaluation matrix using the dynamic weight vector to determine a comprehensive evaluation value of each link, and select several links with the best and relatively good comprehensive evaluation values as candidate paths for service signal transmission;

[0140] The adaptive routing adjustment module is configured to: combine service priority and network congestion status, use adaptive routing strategies to dynamically adjust several candidate paths and determine a transmission path as the final path to ensure the transmission quality of critical service signals;

[0141] The real-time path slicing module is configured to dynamically split the selected final path into multiple sub-paths according to the real-time changes in the link status during service signal transmission, and independently optimize the sub-paths to achieve dynamic adjustment of end-to-end transmission efficiency.

[0142] It should be noted that the optimization control system for the OTN network of this embodiment corresponds to the aforementioned optimization control method for the OTN network. Therefore, the parts that are specifically described in the optimization control system for the OTN network of this embodiment (including but not limited to specific implementation technical means and technical effects) can refer to the relevant records in the aforementioned optimization control method for the OTN network, and this document will not elaborate on them here.

[0143] In the embodiments provided herein, it should be understood that the embodiments described herein can be implemented in hardware, software, firmware, middleware, code, or any appropriate combination thereof. For hardware implementation, the processor can be implemented in one or more of the following units: an application-specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a processor, a controller, a microcontroller, a microprocessor, other electronic units designed to implement the functions described herein, or a combination thereof. For software implementation, part or all of the processes of the embodiments can be completed by instructing the relevant hardware through a computer program. When implemented, the above program can be stored in a computer-readable storage medium or transmitted as one or more instructions or codes on a computer-readable storage medium. Computer-readable storage media include computer storage media and communication media, wherein the communication media include any medium that facilitates the transmission of a computer program from one place to another. The storage medium can be any available medium that a computer can access. The computer-readable storage medium can include, but is not limited to, RAM, ROM, EEPROM, CD-ROM or other optical disk storage, magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store the desired program code in the form of instructions or data structures and can be accessed by a computer.

[0144] Finally, it should be noted that the above is only a preferred embodiment of the present application and is not intended to limit the present application. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or make equivalent replacements for some of the technical features therein. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application.

Claims

1. An optimization control method for an OTN network, characterized in that: The method comprises the following steps: Dynamic network modeling: Based on the OTN network topology and service requirements, a multi-dimensional evaluation matrix is constructed, including nodes, links, and service flows. Parameters in the multi-dimensional evaluation matrix include link bandwidth utilization, latency, packet loss rate, link length, optical signal quality, and service priority. Dynamic weight allocation: Based on real-time network status and service requirements, the weight coefficients of each parameter in the multi-dimensional evaluation matrix are dynamically adjusted through a machine learning model to generate a dynamic weight vector; Path evaluation and selection: using the dynamic weight vector to perform nonlinear weighted calculation on the multidimensional evaluation matrix, determine the comprehensive evaluation value of each link, and select several links with the best and relatively good comprehensive evaluation values as candidate paths for service signal transmission; Adaptive routing adjustment: Based on service priority and network congestion status, an adaptive routing strategy is used to dynamically adjust several candidate paths and determine a transmission path as the final path; Real-time path slicing: During service signal transmission, the selected final path is dynamically divided into multiple sub-paths based on real-time changes in link status. Each sub-path is independently optimized to dynamically adjust end-to-end transmission efficiency.

2. The optimization control method for an OTN network according to claim 1, characterized in that: Among the parameters in the multidimensional evaluation matrix: The link bandwidth utilization is obtained by collecting service flow data on the link and calculating the current bandwidth utilization; The delay is calculated based on the physical distance of the link and the signal propagation speed, combined with the node processing time, to obtain the end-to-end time delay; The packet loss rate is obtained by monitoring the packet loss events of the link and calculating the packet loss rate per unit time; The link length is obtained by calculating the physical length of the link based on the geographical layout of the OTN network; The optical signal quality is obtained by evaluating the signal-to-noise ratio and bit error rate of the optical signal; The service priority is allocated according to the service type and QoS requirements.

3. The optimization control method for an OTN network according to claim 2, characterized in that: The method of dynamically adjusting the weight coefficients of the parameters in the multidimensional evaluation matrix through the machine learning model to generate a dynamic weight vector specifically includes: Initial weight setting: Set the initial weight coefficient of each parameter based on business needs and network planning; Real-time weight adjustment: Based on the real-time data of the multi-dimensional evaluation matrix, the weight coefficient is dynamically adjusted through the machine learning model to generate a dynamic weight vector; Weight verification: Convergence verification of dynamic weight vectors.

4. The optimization control method for an OTN network according to any one of claims 1 to 3, characterized in that: The dynamic division of the selected final path into multiple sub-paths and independent optimization of the sub-paths to achieve dynamic adjustment of the end-to-end transmission efficiency specifically includes: Path segmentation: Divide the selected transmission path into multiple sub-paths, each sub-path corresponds to an independent link segment; Subpath optimization: For each subpath, the comprehensive evaluation value is calculated independently, and the optimal subpath is selected based on the dynamic weight vector; Dynamic splicing: Splice the optimized sub-paths into a complete transmission path and update the path status in real time.

5. The optimization control method for an OTN network according to claim 1 or 4, characterized in that: The machine learning model includes: Input layer: receives real-time data of each parameter in the multidimensional evaluation matrix; Hidden layer: extracts the correlation between parameters through the neural network model; Output layer: Generates dynamic weight vectors and outputs them to the path evaluation and selection steps; And, the loss function.

6. The optimization control method for an OTN network according to claim 1, characterized in that: The adaptive routing strategy includes: a load balancing strategy and a fault recovery strategy, or a priority strategy and a fault recovery strategy; wherein: The load balancing strategy is specifically as follows: among several candidate paths, the candidate path with the smallest load is selected as the final path, and the other candidate paths are retained as backup paths; The priority strategy is specifically as follows: based on the service priority, the candidate path that meets the high-priority service requirements is preferentially selected as the final path, and other candidate paths are retained as backup paths; The fault recovery strategy is specifically: when a link abnormality is detected in the selected final path, automatically switching to a backup path and updating the multi-dimensional evaluation matrix.

7. The optimization control method for an OTN network according to claim 6, characterized in that: The fault recovery strategy specifically includes: Anomaly detection: Detects link anomalies by monitoring the link's packet loss rate and optical signal quality; Path switching: When a link anomaly is detected, a backup path is selected and the multi-dimensional evaluation matrix is updated; Recovery verification: After the path is switched, the comprehensive evaluation value of the new path is verified to see if it meets the business requirements, and the dynamic weight vector is adjusted.

8. The optimization control method for an OTN network according to claim 6, characterized in that: The adaptive routing adjustment further includes: Based on historical traffic data and current network status, a time series prediction model is used to predict the link load change trend in the future T time window as the prediction result, and the prediction result is input into the adaptive routing strategy.

9. The optimization control method for an OTN network according to claim 8, characterized in that: The load balancing strategy and the priority strategy further include: According to the prediction results, several potential candidate paths are generated, and based on the service priority, the potential candidate paths that meet the conditions are screened out as new backup paths; the screening conditions are specifically: For high-priority services, the comprehensive evaluation value of the potential candidate path must be greater than the preset first minimum evaluation value; For medium / low priority services, the comprehensive evaluation value of the potential candidate path must be greater than the preset second lowest evaluation value; The first lowest evaluation value is smaller than the second lowest evaluation value.

10. An optimization control system for an OTN network, applying the optimization control method for an OTN network according to any one of claims 1 to 9, characterized in that: The system includes: The dynamic network modeling module is configured to: construct a multi-dimensional evaluation matrix including nodes, links, and service flows based on the topology of the OTN network and service requirements. Parameters in the multi-dimensional evaluation matrix include link bandwidth utilization, latency, packet loss rate, link length, optical signal quality, and service priority; A dynamic weight allocation module is configured to dynamically adjust the weight coefficients of the parameters in the multidimensional evaluation matrix through a machine learning model according to real-time network status and service requirements to generate a dynamic weight vector; a path evaluation and selection module configured to: perform nonlinear weighted calculation on the multidimensional evaluation matrix using the dynamic weight vector to determine a comprehensive evaluation value of each link, and select several links with the best and relatively good comprehensive evaluation values as candidate paths for service signal transmission; The adaptive routing adjustment module is configured to: combine service priority and network congestion status, use adaptive routing strategies to dynamically adjust several candidate paths and determine a transmission path as the final path to ensure the transmission quality of critical service signals; The real-time path slicing module is configured to dynamically split the selected final path into multiple sub-paths according to the real-time changes in the link status during service signal transmission, and independently optimize the sub-paths to achieve dynamic adjustment of end-to-end transmission efficiency.

Images