A method for dynamic aggregation of optical paths in a CPO switching matrix based on service load

Abstract

The application discloses a kind of CPO exchange matrix based on optical channel dynamic aggregation method of service load, belong to optical communication and data center network technical field, this method is by real-time monitoring network delay, device response time, environmental parameter etc. Bottom factor, in turn build decision-making agility model, environmental stability model, physical response model, quantification system's control and execution capability;Further, based on the above ability dynamic generation to clock synchronization error, link detection delay and channel crosstalk Strict performance permission threshold, and the normalized link switching performance dynamic permission degree is calculated;Finally, with this permission and environmental stability coefficient as input, by extinction ratio optimization model adaptively adjusts the extinction ratio of optical signal of transmitting end.The application realizes the accurate matching of optical channel aggregation / dispersion strategy and real-time service load and system state, while ensuring nanosecond lossless switching, significantly improves network resource utilization efficiency and system energy efficiency.

Description

Technical Field

[0001] This invention belongs to the field of optical communication and data center network technology, and particularly relates to a method for dynamic aggregation of optical channels based on service load in a CPO switching matrix. Background Technology

[0002] As data centers and high-performance computing networks evolve towards ultra-high bandwidth and ultra-low latency, co-packaged optics (CPO) technology has become a key evolution direction for next-generation switching matrices due to its potential to significantly shorten electrical interconnect distances and reduce power consumption. In a CPO architecture, the core challenge for improving overall switching performance is how to efficiently and flexibly manage and schedule its numerous optical channel resources to carry dynamically changing and bursty service traffic. This leads to an urgent need for a method for dynamic optical channel aggregation based on service load in a CPO switching matrix.

[0003] Currently, channel resource management in optical networks mainly relies on static configuration or centralized SDN control strategies based on fixed thresholds. Existing technologies typically employ pre-allocated wavelengths or fixed bundled links, or, upon detecting traffic exceeding a preset threshold, the controller initiates simple channel establishment or teardown commands. At the parameter control level, fixed optical emission parameters (such as extinction ratio) are often used to ensure signal quality under worst-case conditions, lacking fine-grained perception and feedback adjustment of the real-time link status.

[0004] Therefore, there is an urgent need for a dynamic aggregation method for optical channels that can deeply integrate real-time business load with underlying multi-dimensional system states and achieve closed-loop optimization control. Summary of the Invention

[0005] To address the shortcomings of existing technologies, this invention provides a method for dynamic aggregation of optical channels based on service load in a CPO switching matrix, thus solving the aforementioned problems.

[0006] To achieve the above objectives, the present invention provides the following technical solution: a method for dynamic aggregation of optical channels based on service load in a CPO switching matrix, comprising:

[0007] A decision agility model is constructed based on the SDN controller decision delay, the southbound transmission delay of control signaling, and the local switching chip delay to obtain decision agility.

[0008] The environmental stability coefficient is obtained based on the buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation.

[0009] A fast photoelectric response coefficient is obtained based on the laser turn-on / turn-off time, the modulator drive response time, and the photodetector response time.

[0010] Based on wavelength tuning and settling time, as well as frame / packet alignment time, the function establishes response coefficients;

[0011] Based on the fast photoelectric response coefficient and functional establishment of the response coefficient, the physical response coefficient is obtained;

[0012] Based on decision agility and physical response coefficient, clock phase synchronization error, link status detection delay and inter-channel crosstalk level are used to obtain dynamic permission for link switching performance.

[0013] Based on the dynamic permission of link switching performance, environmental stability coefficient, and current signal extinction ratio, the extinction ratio of the target signal is obtained through an extinction ratio optimization model.

[0014] Based on the above technical solutions, the present invention also provides the following optional technical solutions:

[0015] A further technical solution: The extinction ratio optimization model is expressed as follows:

[0016] ;

[0017] in, Indicates the extinction ratio of the target signal. Indicates the current signal extinction ratio. Indicates the dynamic permission level for target link switching performance. Indicates the dynamic permission level for link switching performance. This represents the environmental stability coefficient.

[0018] Further technical solutions: The method for dynamic permission allocation of link switching performance is as follows:

[0019] Based on the decision agility coefficient and the physical response coefficient, the comprehensive capability coefficient is obtained through a comprehensive capability model, which is expressed as follows:

[0020] ;

[0021] in, This represents the overall capability coefficient. Indicates decision-making agility. Represents the physical response coefficient;

[0022] Based on the basic clock phase synchronization error tolerance and the overall capability coefficient, the clock phase synchronization error threshold is obtained through a clock phase synchronization error model, which is expressed as follows:

[0023] ;

[0024] in, This indicates the clock phase synchronization error threshold. This indicates the tolerance for the phase synchronization error of the base clock. Indicates the threshold tightening factor;

[0025] Based on the basic link state detection delay tolerance and the comprehensive capability coefficient, the link state detection delay threshold is obtained through a link state detection delay model, which is expressed as follows:

[0026] ;

[0027] in, Indicates the link state detection latency threshold. Indicates the latency tolerance for basic link state detection;

[0028] Based on the baseline inter-channel crosstalk level, the ideal inter-channel crosstalk level, and the overall capability coefficient, the channel crosstalk threshold is obtained through a channel crosstalk model, which is expressed as follows:

[0029] ;

[0030] in, Indicates the channel crosstalk threshold. Indicates the level of crosstalk between reference channels. This indicates the ideal level of crosstalk between channels;

[0031] The current inter-channel crosstalk level is compared with the channel crosstalk threshold to obtain the channel crosstalk level compliance.

[0032] The clock phase synchronization error threshold and the link status detection delay threshold are compared with the current clock phase synchronization error and the current link status detection delay to obtain the clock phase synchronization error compliance and the link status detection delay compliance.

[0033] Based on clock phase synchronization error compliance, link state detection delay compliance, and channel crosstalk level compliance, a dynamic link handover performance compliance model is used to obtain the dynamic compliance of link handover performance. This dynamic compliance model is expressed as follows:

[0034] ;

[0035] in, Indicates the dynamic permission level for link switching performance. Indicates the clock phase synchronization error compliance. Indicates the compliance of link state detection delay. Indicates the channel crosstalk level compliance. Represents the weight coefficient and The Furthermore, the larger the value, the higher the link state security margin.

[0036] Further technical solution: The method for obtaining the physical response coefficients is as follows:

[0037] Obtain the fast photoelectric response coefficient and the functional establishment response coefficient;

[0038] Based on the fast photoelectric response coefficient and functional establishment of the response coefficient, the physical response coefficient is obtained through a physical response model, which is expressed as follows:

[0039] ;

[0040] in, Represents the physical response coefficient. Represents the fast photoelectric response coefficient. This indicates that the function establishes response coefficients. Represents the mismatch penalty factor, the Furthermore, the larger the value, the higher the overall response agility of physical reconfiguration.

[0041] Further technical solution: The method for obtaining the environmental stability coefficient is as follows:

[0042] Acquire buffer emptying time, CPO package internal temperature change rate, and power supply voltage fluctuation;

[0043] The buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation are compared with the corresponding maximum allowable values ​​to obtain the buffer emptying time index, the temperature change rate index, and the voltage fluctuation index.

[0044] Based on the buffer emptying time exponent, temperature change rate exponent, and voltage fluctuation exponent, the environmental stability coefficient is obtained through an environmental stability model, which is expressed as follows:

[0045] ;

[0046] in, Indicates the environmental stability coefficient. This indicates the buffer emptying time index. Indicates the rate of temperature change index. Indicates the voltage fluctuation index. Represents the weight coefficient and The Furthermore, the larger the value, the more stable the environment.

[0047] Further technical solutions: The method for obtaining decision-making agility is as follows:

[0048] Acquire SDN controller decision latency, control signaling southbound transmission latency, and local switching chip latency;

[0049] The decision delay index, transmission delay index, and chip delay index of the SDN controller, the southbound transmission delay of the control signaling, and the local switching chip delay are compared with the corresponding reference values ​​to obtain the decision delay index, transmission delay index, and chip delay index.

[0050] Decision agility is obtained through a decision agility model based on the decision delay index, transmission delay index, and chip delay index. The decision agility model is expressed as follows:

[0051] ;

[0052] in, Indicates decision-making agility. Indicates the decision delay index, Indicates the transmission delay index. Indicates the chip delay index, the Furthermore, the larger the value, the more agile the decision-making.

[0053] Further technical solution: The method for obtaining the fast photoelectric response coefficient is as follows:

[0054] Acquire the laser turn-on / turn-off time, modulator drive response time, and photodetector response time;

[0055] After performing maximum-minimum normalization on the laser turn-on / off time, modulator drive response time, and photodetector response time, and taking their complements, the laser stabilization time index, modulator drive response time index, and photodetector response time index are obtained.

[0056] Based on the laser stabilization time index, the modulator drive response time index, and the photodetector response time index, the fast photoelectric response coefficient is obtained through a fast photoelectric response model, which is expressed as follows:

[0057] ;

[0058] in, Represents the fast photoelectric response coefficient. Indicates the laser settling time index. Indicates the modulator drive response time exponent. This represents the response time index of a photodetector. Represents the weight coefficient and The The larger the value, the faster the device response.

[0059] Further technical solution: The method for obtaining the response coefficient is as follows:

[0060] Obtain wavelength tuning and settling time, as well as frame / packet alignment time;

[0061] The wavelength tuning and stabilization time and the frame / packet alignment time are compared with the corresponding expected values ​​after the difference is processed and then compared with the allowable deviation from the expected value to obtain the wavelength tuning and stabilization time index and the alignment time index. The wavelength tuning and stabilization time and the frame / packet alignment time are both greater than or equal to the corresponding expected values.

[0062] Based on the wavelength tuning and settling time exponents and the alignment time exponent, the function establishment response coefficients are obtained through a function establishment response model, which is expressed as follows:

[0063] ;

[0064] in, This indicates that the function establishes response coefficients. Indicates the wavelength tuning and settling time index. The alignment time index is represented by the following. The larger the value, the faster the function is established.

[0065] This invention provides a method for dynamic aggregation of optical channels based on service load in a CPO switching matrix, which has the following advantages compared with the prior art:

[0066] 1. This invention quantifies the "decision-execution-environment" status of the system in real time through a multi-layer model and dynamically generates switching performance permission thresholds, so that the channel aggregation / deaggregation actions are highly adapted to the actual load capacity and link quality of the network, transforming passive response into intelligent prediction, which greatly improves resource utilization and agility in dealing with sudden traffic.

[0067] 2. This invention evaluates key physical layer and data link layer factors that affect handover timeliness and success by incorporating them into a unified model, and uses the obtained "dynamic permission level of link handover performance" as an accurate criterion for secure handover, thereby ensuring to the greatest extent possible at the system level that data packets are not lost or out of order during the nanosecond-level channel reconfiguration process;

[0068] 3. The extinction ratio optimization model in this invention dynamically adjusts the emitted optical power based on real-time switching permission and environmental stability, breaking the limitations of the traditional fixed parameter mode. It automatically saves energy when the link conditions are good and enhances signal quality when the conditions are poor or the environment is harsh, thus realizing Pareto optimization of the overall system performance under complex working conditions. Attached Figure Description

[0069] Figure 1 This is a schematic diagram of the process of the present invention. Detailed Implementation

[0070] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0071] The specific implementation of the present invention will be described in detail below with reference to specific embodiments.

[0072] Please see Figure 1 The present invention provides a method for dynamic aggregation of optical channels based on service load in a CPO switching matrix, comprising:

[0073] A decision agility model is constructed based on the SDN controller decision delay, the southbound transmission delay of control signaling, and the local switching chip delay to obtain decision agility.

[0074] The environmental stability coefficient is obtained based on the buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation (absolute value of the power supply voltage fluctuation amplitude).

[0075] The fast photoelectric response coefficient is obtained based on the laser turn-on / turn-off time (the time it takes for the laser to stabilize from the change of bias current to the stability of the output optical power), the modulator drive response time, and the photodetector response time.

[0076] Based on wavelength tuning and settling time (the time it takes for a tunable laser to reach the target wavelength and stabilize within the tolerance range after changing its wavelength) and frame / packet alignment time (the time it takes to adjust the data streams of different channels to the same reference frame start point during aggregation), the response coefficient is established using the acquisition function.

[0077] Based on the fast photoelectric response coefficient and functional establishment of the response coefficient, the physical response coefficient is obtained;

[0078] Based on decision agility and physical response coefficient, clock phase synchronization error, link status detection delay and inter-channel crosstalk level (after aggregation, the maximum crosstalk value of adjacent channels due to insufficient isolation, which is generally negative and in dB), obtain the dynamic permission of link switching performance.

[0079] Based on the dynamic permission of link switching performance, environmental stability coefficient, and current signal extinction ratio, the extinction ratio of the target signal is obtained through an extinction ratio optimization model.

[0080] Through the aforementioned series of dynamic evaluation and optimization processes, the CPO switching matrix can adaptively adjust optical channel aggregation strategies and optical transmission parameters based on real-time service load and multi-dimensional system status. This enables dynamic, efficient, and flexible management of optical channel resources, effectively responding to sudden surges in service traffic and optimizing overall switching performance. The various technical features are interconnected, forming a closed-loop optimization control system that ensures the reliability and performance of optical channel aggregation.

[0081] Existing technologies for optical channel management in CPO switching matrices typically employ static configuration or SDN control strategies based on fixed thresholds, with optical emission parameters (such as extinction ratio) often set to fixed values ​​to address worst-case scenarios. This approach lacks the ability to perceive and fine-tune real-time dynamic changes in the network. In contrast, this method introduces a multi-dimensional dynamic evaluation mechanism to achieve closed-loop optimization control of dynamic aggregation of optical channels in the CPO switching matrix. Specifically:

[0082] First, this method dynamically evaluates the SDN controller's responsiveness at the decision-making level by constructing a decision agility model. This differs from existing technologies that simply rely on fixed SDN control strategies, and it more accurately reflects the real-time performance of the control system, providing a basis for subsequent aggregation decisions.

[0083] Secondly, this method introduces an environmental stability coefficient, dynamically evaluating the operating environment of the CPO switching matrix by monitoring physical environmental parameters such as buffer emptying time, CPO package internal temperature change rate, and power supply voltage fluctuations. This goes beyond the rough consideration of environmental factors in existing technologies, enabling the system to sense and adapt to changes in the actual operating environment, thereby improving the reliability of aggregation operations.

[0084] Furthermore, this method provides a refined evaluation of the physical reconfiguration capability of the optical channel by obtaining the fast photoelectric response coefficient and the functional establishment response coefficient, and further obtaining the physical response coefficient. This differs from existing technologies that may only focus on the simple opening / closing of the optical channel. This embodiment delves into the response speed of devices such as lasers, modulators, and photodetectors, as well as the functional establishment process such as wavelength tuning and frame / packet alignment, providing a more comprehensive view of the physical layer response capability.

[0085] Furthermore, this method dynamically obtains the dynamic permission level for link switching performance based on key performance indicators such as decision agility, physical response coefficient, clock phase synchronization error, link status detection delay, and inter-channel crosstalk level. This enables the system to evaluate the safety and efficiency of optical channel aggregation operations in real time, avoiding misjudgments or performance bottlenecks that may occur due to fixed threshold judgments in existing technologies.

[0086] Most importantly, this method introduces an extinction ratio optimization model, which adaptively calculates and obtains the target signal extinction ratio based on dynamically acquired link switching performance dynamic permission, environmental stability coefficient, and the current signal extinction ratio. This contrasts sharply with the fixed extinction ratio strategy used in existing technologies. By dynamically optimizing the extinction ratio, this embodiment can optimize energy efficiency while ensuring optical signal quality and link reliability, avoiding unnecessary power consumption waste or performance deficiencies that may result from a fixed extinction ratio.

[0087] In summary, the technical solution of this method overcomes the limitations of existing technologies in optical channel management, which lack fine perception and adaptive adjustment capabilities, by conducting multi-dimensional, dynamic, and real-time evaluation of decision-making agility, environmental stability, physical response capability, and link switching performance, and on this basis, performing closed-loop optimization of key parameters of optical channel aggregation (such as extinction ratio). This achieves efficient, flexible, and reliable dynamic aggregation of optical channel resources in the CPO switching matrix, thereby improving the overall switching efficiency.

[0088] Preferably, the method for obtaining decision agility is as follows:

[0089] Acquire SDN controller decision latency, control signaling southbound transmission latency, and local switching chip latency;

[0090] The decision delay index, transmission delay index, and chip delay index of the SDN controller, the southbound transmission delay of the control signaling, and the local switching chip delay are compared with the corresponding reference values ​​to obtain the decision delay index, transmission delay index, and chip delay index.

[0091] Decision agility is obtained through a decision agility model based on the decision delay index, transmission delay index, and chip delay index. The decision agility model is expressed as follows:

[0092] ;

[0093] in, Indicates decision-making agility. Indicates the decision delay index, Indicates the transmission delay index. Indicates the chip delay index, the Furthermore, the larger the value, the more agile the decision-making.

[0094] SDN controller decision latency refers to the time required for a Software-Defined Networking (SDN) controller to complete a decision and generate corresponding control commands from receiving a service request or network status change information. This latency can be accurately measured by analyzing the SDN controller's internal logs or by setting timestamps at key processing nodes. Control signaling southbound transmission latency refers to the time required for control commands issued by the SDN controller to be transmitted through the southbound interface to the execution devices (such as optical switching units) in the CPO switching matrix. This latency can be obtained by embedding timestamps in the control commands and recording the timestamps when the commands arrive at the execution devices, calculating the difference between the two; or by real-time measurement using network performance monitoring tools. Local switching chip latency refers to the time required for the switching chips within the CPO switching matrix to complete internal processing, configuration, and forwarding operations after receiving control commands. This latency can usually be obtained by consulting the chip manufacturer's technical specifications or datasheets, or by benchmarking the chips in a laboratory environment. Ratioing these latency values ​​to corresponding reference values ​​aims to normalize latency data from different sources and with different dimensions, making them dimensionless exponents. For example, a ratio can be obtained by dividing the actual measured latency value by a preset maximum allowable latency value or expected latency value. These reference values ​​can be set based on system design specifications, industry standards, or historical operating data. The decision latency index, transmission latency index, and chip latency index obtained through ratio processing can intuitively reflect the deviation of each latency from its reference level, providing a unified input for subsequent comprehensive evaluation. The decision agility model is used to comprehensively evaluate the overall response speed of the CPO switching matrix from decision to execution under changes in business load. This model uses an exponential function to weight and sum the various latency indices, taking the negative exponent, so that an increase in any latency index will lead to a decrease in decision agility. The decrease in latency, and the accelerated rate of decrease with increasing latency exponential, aligns with the sensitivity of system agility to latency. Decision agility The value ranges from 0 to 1. A larger value indicates a faster and more agile overall response in system decision-making and execution.

[0095] The following is a concrete example. In a CPO switching matrix, to obtain decision agility, the system first acquires the SDN controller decision latency, the southbound transmission latency of control signaling, and the local switching chip latency. For example, the SDN controller decision latency can be obtained by recording the time difference between receiving a service request (such as a new data stream arriving) and generating an optical channel aggregation command in the controller software; let's assume it's measured to be 80 milliseconds. The southbound transmission latency of control signaling can be calculated by timestamping the SDN controller when it issues the command and then timestamping it again when the optical engine of the CPO switching matrix receives the command; let's assume it's measured to be 20 milliseconds. The local switching chip latency can be determined based on the chip specifications or actual test results; let's assume it's 5 milliseconds. Next, these latency values ​​are compared with corresponding reference values. For example, the reference value for the SDN controller decision latency can be set to 100 milliseconds, the reference value for the southbound transmission latency of control signaling to 30 milliseconds, and the reference value for the local switching chip latency to 10 milliseconds. Then, the decision latency index... The transmission delay index is 80 / 100 = 0.8. The chip delay index is 20 / 30 = 0.67. The result is 5 / 10 = 0.5. Finally, these indices are substituted into the decision agility model, i.e. Calculated The value represents the current decision agility, which will serve as an input parameter for subsequent calculations of the dynamic permission level for link switching performance.

[0096] Through the above technical solution, this application can comprehensively and quantitatively evaluate the decision-related delays affecting the dynamic aggregation efficiency of optical channels in the CPO switching matrix. By converting the SDN controller decision delay, the southbound transmission delay of control signaling, and the local switching chip delay into a unified delay index, and using a decision agility model for comprehensive calculation, the system's decision agility can be dynamically obtained. This accurate agility assessment allows for a more accurate reflection of the system's responsiveness to changes in service load when calculating the dynamic permissibility of link switching performance, thereby optimizing the aggregation strategy of optical channels. This helps avoid resource waste or service quality degradation caused by decision lag or slow execution, significantly improving the adaptability and robustness of the CPO switching matrix in complex service environments, and ensuring the timeliness and effectiveness of dynamic aggregation of optical channels.

[0097] Preferably, the method for obtaining the environmental stability coefficient is as follows:

[0098] Obtain the buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation (absolute value of the power supply voltage fluctuation amplitude).

[0099] The buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation are compared with the corresponding maximum allowable values ​​to obtain the buffer emptying time index, the temperature change rate index, and the voltage fluctuation index.

[0100] Based on the buffer emptying time exponent, temperature change rate exponent, and voltage fluctuation exponent, the environmental stability coefficient is obtained through an environmental stability model, which is expressed as follows:

[0101] ;

[0102] in, Indicates the environmental stability coefficient. This indicates the buffer emptying time index. Indicates the rate of temperature change index. Indicates the voltage fluctuation index. Represents the weight coefficient and The Furthermore, the larger the value, the more stable the environment.

[0103] The buffer emptying time refers to the time required for all data in the data buffer to be processed and sent out. It reflects the system's efficiency in processing data streams and the current load. A longer emptying time usually indicates system congestion or insufficient processing capacity. The buffer emptying time can be calculated by real-time monitoring of the buffer status and data throughput, for example, by recording the timestamps of data entering and leaving the buffer and estimating it in conjunction with the current amount of data in the buffer; or by directly reading it from the system's internal performance counters. The CPO package internal temperature change rate refers to the rate at which the internal temperature of the CPO (Co-packaged Optics) package changes over time. CPO modules are sensitive to temperature, and drastic temperature changes can affect their optical performance and reliability. The CPO package internal temperature change rate can be obtained by real-time monitoring using a temperature sensor integrated inside the CPO package and differential calculation of continuous temperature readings; or by estimating it using a thermal model combined with power consumption data. Supply voltage fluctuation (absolute value of supply voltage fluctuation amplitude) refers to the instantaneous change amplitude of the supply voltage of the CPO module. Stable supply is the basic guarantee for the normal operation of the CPO module. Voltage fluctuation may lead to device performance degradation or even failure. Supply voltage fluctuation can be obtained by real-time sampling of the voltage sensor integrated in the power management unit (PMU) and calculation of its maximum instantaneous fluctuation amplitude; or by periodic measurement through an external high-precision voltmeter.

[0104] The process of comparing the buffer emptying time, the rate of temperature change inside the CPO package, and the supply voltage fluctuation with their respective maximum allowable values ​​aims to unify and dimensionless these environmental parameters with different physical dimensions and numerical ranges, enabling comprehensive evaluation within a unified model. For example, if the buffer emptying time exceeds the maximum allowable value, its exponent will be greater than 1, indicating a high degree of environmental instability. These maximum allowable values ​​are preset thresholds based on the performance requirements, reliability standards, and practical operating experience of the CPO switching matrix. For instance, the maximum allowable buffer emptying time can be set as the maximum delay time the system can tolerate; the maximum allowable rate of temperature change inside the CPO package can be set as the rate of temperature change that does not significantly degrade optical performance; and the maximum allowable supply voltage fluctuation can be set as the voltage fluctuation range that does not affect the normal operation of the device.

[0105] An environmental stability coefficient is obtained using an environmental stability model based on the buffer emptying time exponent, temperature change rate exponent, and voltage fluctuation exponent. This model aims to integrate multiple normalized environmental parameters to quantify the overall stability of the current working environment of the CPO exchange matrix. The model employs a reciprocal form to ensure that the environmental stability coefficient increases as each environmental indices increase (i.e., the more unstable the environment). A decrease in the weighting coefficients indicates an increase in the stability of the system, while a greater increase indicates a greater increase. This nonlinear relationship better reflects the comprehensive impact of environmental factors on system stability. Among these factors, the weighting coefficients... This is used to adjust the relative importance of different environmental factors when assessing overall environmental stability. For example, if the rate of temperature change inside the CPO package has the greatest impact on system performance, it can be assigned a larger weight. These weights can be determined and optimized through expert experience, historical data analysis, machine learning algorithms, or system simulation to ensure that the model accurately reflects the stability of the actual environment.

[0106] As a specific implementation, the control unit of the CPO switching matrix can periodically perform the process of acquiring environmental stability coefficients. For example, every 100 milliseconds, the control unit acquires the current buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation by reading internal sensors and system logs. The buffer emptying time can be obtained by monitoring the average waiting time of data packets in the output queue; the internal temperature change rate of the CPO package can be obtained by continuously sampling with a digital temperature sensor (such as a PT1000 or a thermistor) integrated in the CPO module and calculating the ratio of the temperature difference between adjacent sampling points to the time interval; the power supply voltage fluctuation can be determined by high-frequency sampling of the power rail voltage using the ADC (analog-to-digital converter) built into the power management chip (such as a PMIC) and calculating the absolute value of the difference between the maximum and minimum voltages within the sampling period. After acquiring this raw data, the control unit compares it with a preset maximum allowable value. For example, if the maximum allowable buffer emptying time is 50 microseconds and the current measurement is 25 microseconds, then the buffer emptying time index... The value is 0.5. Similarly, assuming that the current temperature change rate is 2.5℃ / s as monitored in real time by a temperature sensor integrated inside the CPO package, while the maximum allowable temperature change rate preset according to the system thermal design specifications is 5℃ / s, then the temperature change rate exponent is 0.5. Assuming the power supply voltage fluctuation is sampled by the power management unit (PMU) voltage sensor and found to be 0.03V (30mV), while the maximum allowable power supply voltage fluctuation is preset to 0.05V (50mV) according to the device power supply specifications, then the voltage fluctuation index... Finally, the control unit substitutes these indices into the environmental stability model, where the weighting coefficients... It can be pre-configured using expert experience; for example, it can be set... The environmental stability coefficient at the current moment can be obtained through calculation. This coefficient can then be used in subsequent extinction ratio optimization models to guide the dynamic aggregation of optical channels.

[0107] Through the above technical solution, this application can comprehensively and dynamically evaluate the operational environment stability of the CPO switching matrix. By comprehensively considering three key environmental factors—buffer emptying time, internal temperature change rate of the CPO package, and power supply voltage fluctuation—and quantifying them into environmental stability coefficients, it effectively solves the problem of single or inaccurate environmental assessment dimensions in traditional methods. This refined environmental stability assessment provides a more accurate and reliable input for subsequent calculation of the target signal extinction ratio, enabling the decision-making of dynamic optical channel aggregation to better adapt to the actual operating conditions of the CPO switching matrix under different loads and environmental conditions. Therefore, when combined with the above-mentioned service load-based dynamic optical channel aggregation method, it can significantly improve the robustness and adaptability of the aggregation strategy, ensuring that efficient and stable optical channel aggregation performance can be maintained even in variable environments, thereby optimizing the overall operational efficiency and reliability of the CPO switching matrix.

[0108] Preferably, the method for obtaining the fast photoelectric response coefficient is as follows:

[0109] Acquire the laser turn-on / turn-off time (the time it takes for the laser to stabilize from a change in bias current to a stable output optical power), the modulator drive response time, and the photodetector response time.

[0110] After performing maximum-minimum normalization on the laser turn-on / off time, modulator drive response time, and photodetector response time, and taking their complements, the laser stabilization time index, modulator drive response time index, and photodetector response time index are obtained.

[0111] Based on the laser stabilization time index, the modulator drive response time index, and the photodetector response time index, the fast photoelectric response coefficient is obtained through a fast photoelectric response model, which is expressed as follows:

[0112] ;

[0113] in, Represents the fast photoelectric response coefficient. Indicates the laser settling time index. Indicates the modulator drive response time exponent. This represents the response time index of a photodetector. Represents the weight coefficient and The The larger the value, the faster the device response.

[0114] The laser turn-on / turn-off time refers to the time required for the laser's output optical power to reach a steady state after a change in bias current. This time is a key parameter for measuring the laser's response speed, and its length directly affects the establishment speed of the optical signal. This time can be obtained by monitoring the laser's drive current and optical power output using a high-precision oscilloscope and recording the time required for the optical power to stabilize after the current change; alternatively, it can be determined by performing multiple measurements under different drive conditions and taking the average value using dedicated laser testing equipment. The modulator drive response time refers to the time required for the modulator's optical output characteristics (such as light intensity and phase) to reach a stable change after receiving an electrical drive signal. This time reflects the modulator's ability to follow electrical signals and is an important factor affecting the optical signal modulation rate. This time can be obtained by inputting a step electrical signal to the modulator and measuring the rise / fall time of the modulator's output optical signal using a high-speed photodetector and oscilloscope; alternatively, it can be obtained by measuring the modulator's bandwidth using a network analyzer and then calculating its response time. The response time of a photodetector refers to the time required for a photodetector to convert a received optical signal into an electrical signal and achieve a stable output. This time determines the photodetector's ability to capture changes in the optical signal and is a key parameter affecting the rate of the optical receiver. This time can be obtained by inputting an ultrashort optical pulse into the photodetector and measuring the rise / fall time of its electrical output signal using a high-bandwidth oscilloscope; alternatively, it can be calculated by measuring the 3dB bandwidth of the photodetector and considering the relationship between bandwidth and response time.

[0115] The laser turn-on / turn-off time, modulator drive response time, and photodetector response time are subjected to max-min normalization and their complements are taken. This aims to unify response times of different dimensions and ranges within the [0,1] interval, ensuring that their values ​​are positively correlated with response speed. The fast photoelectric response model employs a weighted linear combination approach, combining the response exponents of each device to form a single coefficient that comprehensively characterizes the response speed of the entire photoelectric link. Among these, the weighting coefficients... This is used to adjust the relative importance of the response times of different devices in the overall fast photoelectric responsivity. For example, if the laser's response speed has a greater impact on the overall link performance, it can be assigned a larger weight. In this way, the fast photoelectric responsivity... It can accurately reflect the overall response agility of the optoelectronic link; the larger the value, the faster the device response.

[0116] The following example illustrates this point. Suppose we need to evaluate the fast response capability of the optoelectronic links in a CPO switching matrix. Based on preset weighting coefficients... (For example, it can be set according to the criticality of the device in the link or actual test data, and meets the following requirements.) The fast photoelectric response coefficient was calculated using a fast photoelectric response model. For example, if And the calculated ,but .

[0117] The above technical solution yields a comprehensive and quantified fast photoelectric response coefficient, which more accurately reflects the actual response capability of photoelectric links in the CPO switching matrix. This allows for full consideration of the comprehensive performance of key photoelectric components such as lasers, modulators, and photodetectors when calculating the physical response coefficient, thereby improving its accuracy. Furthermore, this facilitates a more precise assessment of the dynamic permissibility of link switching performance, providing a more reliable decision-making basis for the CPO switching matrix to dynamically aggregate optical channels when service load changes. It effectively avoids aggregation failures or performance degradation caused by inaccurate photoelectric response speed assessments, thus improving the stability and efficiency of the entire system.

[0118] Preferably, the method for obtaining the function to establish the response coefficient is as follows:

[0119] Acquire wavelength tuning and settling time (the time it takes for a tunable laser to reach the target wavelength and stabilize within the tolerance range after changing its wavelength) and frame / packet alignment time (the time it takes to adjust the data streams of different channels to the same reference frame start point during aggregation).

[0120] The wavelength tuning and stabilization time and the frame / packet alignment time are compared with the corresponding expected values ​​after the difference is processed and then compared with the allowable deviation from the expected value to obtain the wavelength tuning and stabilization time index and the alignment time index. The wavelength tuning and stabilization time and the frame / packet alignment time are both greater than or equal to the corresponding expected values.

[0121] Based on the wavelength tuning and settling time exponents and the alignment time exponent, the function establishment response coefficients are obtained through a function establishment response model, which is expressed as follows:

[0122] ;

[0123] in, This indicates that the function establishes response coefficients. Indicates the wavelength tuning and settling time index. The alignment time index is represented by the following. The larger the value, the faster the function is established.

[0124] Among them, wavelength tuning and settling time refers to the time required for the output wavelength of a tunable laser to reach the target value and stabilize within a preset tolerance range after receiving a wavelength switching command. This time directly reflects the agility and stability of the laser's wavelength switching. It can be obtained by integrating a wavelength monitoring unit into the laser module to track the output wavelength in real time and record the time required from the issuance of the tuning command to wavelength stabilization. Frame / packet alignment time refers to the time required to adjust data streams from different channels to the same reference frame start point during optical channel aggregation. This is crucial for ensuring the timing consistency and data integrity of the aggregated data streams. This time can be obtained by setting a timing detection circuit in the aggregation module to measure the time required from receiving the synchronization signal of each data stream to the completion of alignment of all data streams. The function establishment response model adopts an exponential decay function form. In this model, the wavelength tuning and settling time is exponentially... Alignment Time Index As inputs to the model, the larger their values, the greater the negativity of the exponential term, leading to a decrease in the functional response coefficient. The smaller the value, the slower the function is established. Conversely, the larger the value, the slower the function is established. The larger the value, the faster the feature is established. This model is sensitive to the negative impact of any increase in the time exponent on the feature establishment response coefficient, thus providing a standardized coefficient between 0 and 1 to assess the speed of feature establishment.

[0125] The following is a concrete example. In a CPO switching matrix, when a tunable laser needs to be switched from the current wavelength to the target wavelength, the system's internal monitoring module measures the wavelength tuning and stabilization time to be 600 microseconds. Simultaneously, when aggregating data streams from multiple optical channels, the aggregation unit measures the frame / packet alignment time to be 250 nanoseconds. Assume the preset expected value for wavelength tuning and stabilization time is 450 microseconds, with an allowable deviation of 150 microseconds; the preset expected value for frame / packet alignment time is 180 nanoseconds, with an allowable deviation of 70 nanoseconds. Based on the above calculations, the wavelength tuning and stabilization time exponent... Alignment time index Substituting these two indices into the function to build a response model, we obtain... At this point, the function establishes the response coefficient. The value of 0.135 indicates that the speed of function establishment is relatively low.

[0126] Through the above technical solution, this application provides an effective method for quantifying the rapid response of optical channel function establishment in a CPO switching matrix. By accurately measuring wavelength tuning and settling time and frame / packet alignment time, and standardizing them to an exponential form, and then calculating the function establishment response coefficient using an exponential decay model, the efficiency of optical channel function establishment can be objectively and accurately evaluated. This allows the system to more finely perceive and quantify the time overhead during function establishment, thus providing a more reliable basis for calculating the physical response coefficient. Ultimately, this helps improve the adaptability and response speed of the CPO switching matrix to changes in service load when dynamically aggregating optical channels, ensuring that optical channels can quickly and stably complete configuration and data transmission in complex and ever-changing service environments, thereby optimizing overall network performance.

[0127] Preferably, the physical response coefficients are obtained as follows:

[0128] Obtain the fast photoelectric response coefficient and the functional establishment response coefficient;

[0129] Based on the fast photoelectric response coefficient and functional establishment of the response coefficient, the physical response coefficient is obtained through a physical response model, which is expressed as follows:

[0130] ;

[0131] in, Represents the physical response coefficient. Represents the fast photoelectric response coefficient. This indicates that the function establishes response coefficients. Represents the mismatch penalty factor, the Furthermore, the larger the value, the higher the overall response agility of physical reconfiguration.

[0132] The physical response coefficient is obtained through a physical response model based on the fast photoelectric response coefficient and the function establishment response coefficient. This step aims to comprehensively consider both the fast photoelectric response capability and the function establishment response capability to obtain a physical response coefficient that can fully reflect the overall agility of the physical reconfiguration of the CPO exchange matrix. The physical response model not only considers the independent contributions of the two response capabilities but also introduces a mismatch penalty factor to quantify and penalize the degree of mismatch between the two response capabilities, thereby more accurately assessing the performance bottlenecks that the system may encounter in actual operation. The physical response model is expressed as: This model specifically defines how to obtain the fast photoelectric response coefficient. Establish response coefficients with functions Calculate the physical response coefficient .in, and The product term reflects the fundamental contribution resulting from the synergistic effect of the two response capabilities. And... The term is a penalty factor used to measure and The degree of difference between them. When and The closer the two are, the closer the penalty factor is to 1, indicating a high degree of matching and a small negative impact on the overall physical response coefficient; conversely, the greater the difference between the two, the smaller the penalty factor, indicating a low degree of matching and a large negative impact on the overall physical response coefficient. Mismatch Penalty Factor This is an adjustable parameter used to control the intensity of the penalty. The model ensures the physical response coefficient... It considers not only the response speed of each component but also their coordination, thus more realistically reflecting the overall agility of physical reconfiguration. Physical Response Coefficient This is the final output, representing the overall agility of the CPO exchange matrix in physical reconfiguration. Its value ranges from 0 to 1, with a higher value indicating more agile physical reconfiguration. Fast photoelectric response coefficient. Establish response coefficients with functions These are the model inputs, quantifying the optoelectronic device response speed and the link establishment speed, respectively. Mismatch penalty factor. It is a positive real number used to adjust when and When there is a mismatch, for The degree of punishment. For example, The settings can be empirically determined or optimized through simulation based on the sensitivity of the actual system to different response characteristic mismatches.

[0133] As a specific implementation method, when obtaining the physical response coefficients in the CPO exchange matrix, the fast photoelectric response coefficients can first be obtained through measurement or simulation. Establish response coefficients with functions For example, fast photoelectric response coefficient The response coefficient can be obtained by normalizing and weighting the laser on / off time, modulator drive response time, and photodetector response time; its value can be 0.8. Functional establishment of the response coefficient. This can be obtained by similarly processing the wavelength tuning and settling time, as well as the frame / packet alignment time; its value can be 0.7. Assume a mismatch penalty factor. Based on the system characteristics being set to 5, then, This calculation result This refers to the physical response coefficient in that specific scenario, which comprehensively reflects the synergistic effect of fast photoelectric response and function build-up response, as well as the degree of mismatch between them. This physical response coefficient This can then be used as input to calculate the overall capability coefficient, which in turn affects the assessment of the dynamic permission level for link switching performance.

[0134] The above technical solutions avoid errors caused by evaluating physical response capabilities from a single dimension or simple combination, resulting in a more refined and accurate assessment of physical response coefficients. This facilitates more accurate calculation of comprehensive capability coefficients in subsequent comprehensive capability models, thereby improving the accuracy of dynamic permission assessment for link switching performance. Ultimately, this enables the load-based dynamic aggregation method for optical channels in the CPO switching matrix to make decisions based on a more reliable physical system state, thereby optimizing the dynamic aggregation strategy of optical channels, improving system resource utilization and service carrying capacity, while ensuring the stability and reliability of link switching.

[0135] Preferably, the dynamic permission level for link switching performance is determined as follows:

[0136] Based on the decision agility coefficient and the physical response coefficient, the comprehensive capability coefficient is obtained through a comprehensive capability model, which is expressed as follows:

[0137] ;

[0138] in, This represents the overall capability coefficient. Indicates decision-making agility. Represents the physical response coefficient;

[0139] Based on the basic clock phase synchronization error tolerance and the overall capability coefficient, the clock phase synchronization error threshold is obtained through a clock phase synchronization error model, which is expressed as follows:

[0140] ;

[0141] in, This indicates the clock phase synchronization error threshold. This indicates the tolerance for the phase synchronization error of the base clock. Indicates the threshold tightening coefficient ( );

[0142] Based on the basic link state detection delay tolerance and the comprehensive capability coefficient, the link state detection delay threshold is obtained through a link state detection delay model, which is expressed as follows:

[0143] ;

[0144] in, Indicates the link state detection latency threshold. Indicates the latency tolerance for basic link state detection;

[0145] Based on the baseline inter-channel crosstalk level (maximum allowable crosstalk value), the ideal inter-channel crosstalk level, and the overall capability coefficient, the channel crosstalk threshold is obtained through a channel crosstalk model, which is expressed as follows:

[0146] ;

[0147] in, Indicates the channel crosstalk threshold. Indicates the level of crosstalk between reference channels. This indicates the ideal level of crosstalk between channels;

[0148] The current inter-channel crosstalk level is compared with the channel crosstalk threshold to obtain the channel crosstalk level compliance.

[0149] The clock phase synchronization error threshold and the link status detection delay threshold are compared with the current clock phase synchronization error and the current link status detection delay to obtain the clock phase synchronization error compliance and the link status detection delay compliance.

[0150] Based on clock phase synchronization error compliance, link state detection delay compliance, and channel crosstalk level compliance, a dynamic link handover performance compliance model is used to obtain the dynamic compliance of link handover performance. This dynamic compliance model is expressed as follows:

[0151] ;

[0152] in, Indicates the dynamic permission level for link switching performance. Indicates the clock phase synchronization error compliance. Indicates the compliance of link state detection delay. Indicates the channel crosstalk level compliance. Represents the weight coefficient and The Furthermore, the larger the value, the higher the link state security margin (the safer the channel switching).

[0153] Among them, the comprehensive ability coefficient It is an indicator that measures the overall performance and responsiveness of a system, taking into account decision-making agility. and physical response coefficient This coefficient reflects the system's overall capabilities in rapid decision-making and physical reconfiguration; it is derived from the comprehensive capability model. Calculations show that the overall capability of a system is the product of decision agility and the physical response coefficient, meaning that both must be maintained at a high level to achieve a high overall capability. Decision agility The ability of a system to make quick and accurate decisions in the face of change is measured by constructing a decision agility model based on SDN controller decision latency, southbound transmission latency of control signaling, and local switching chip latency. Physical response coefficient. The response speed and efficiency of optical devices and optical links in the CPO switching matrix during physical reconfiguration are measured by obtaining them based on the fast photoelectric response coefficient and the functionally established response coefficient, and then through a physical response model.

[0154] Basic clock phase synchronization error tolerance The clock phase synchronization error threshold represents the maximum clock phase synchronization error that the system can tolerate under ideal or reference conditions, signifying the system's inherent tolerance to clock synchronization errors. This is the maximum permissible clock phase synchronization error under the current system's overall capabilities. It is a dynamically adjusted threshold determined by a clock phase synchronization error model. Calculations show that when the system's overall capabilities are strong, this threshold will tighten, requiring higher synchronization accuracy. Threshold tightening coefficient. It is a coefficient between 0 and 1, used to control the overall capability coefficient. The impact of threshold adjustment can be configured, for example, based on the system's requirements for reliability and performance (assigned based on expert experience or obtained through fitting historical data).

[0155] Basic link state detection delay tolerance This represents the maximum link state detection delay that the system can tolerate under ideal or baseline conditions, signifying the system's inherent tolerance for link state detection speed. Link state detection delay threshold. This is the maximum allowable link state detection delay under the current overall system capabilities. It is a dynamically adjusted threshold determined by the link state detection delay model. The calculations were performed to adapt to the current overall capabilities of the system.

[0156] Reference channel crosstalk level This is the maximum inter-channel crosstalk value that the system can tolerate under worst-case or baseline conditions, and it is usually a negative value. Ideal inter-channel crosstalk level. This is the expected inter-channel crosstalk value under optimal or ideal conditions, typically a smaller negative value. Channel crosstalk threshold. This is the permissible inter-channel crosstalk threshold under the current overall system capabilities. It is a dynamically adjusted threshold determined by the channel crosstalk model. Calculations show that when the system has strong overall capabilities, the threshold will be closer to the ideal value, requiring lower crosstalk.

[0157] Channel crosstalk level compliance This is an indicator obtained by comparing the current inter-channel crosstalk level with the channel crosstalk threshold. It reflects the degree of compliance of the current crosstalk level with the allowable threshold. For example, this ratio can be obtained by directly comparing the current crosstalk value with the threshold, or by normalizing the difference between the two to a specific range using a function to quantify the compliance. Clock phase synchronization error compliance. This is an indicator obtained by comparing the current clock phase synchronization error with a clock phase synchronization error threshold, reflecting the degree of compliance of the current clock synchronization error with the allowable threshold. For example, this ratio could be calculated by comparing the threshold with the current error value, or by mapping the relationship between the error and the threshold to a compliance score using a function. (Link state detection delay compliance) It is an index obtained by comparing the current link state detection delay with the link state detection delay threshold. It reflects the degree of compliance of the current detection delay with the allowable threshold. For example, the ratio can be a ratio of the threshold to the current delay value, or a function can be used to map the relationship between the delay and the threshold to a compliance score.

[0158] Link switching performance dynamic permission It is a comprehensive indicator, based on a dynamic licensing model for link switching performance. The calculation shows that it integrates clock phase synchronization error compliance, link state detection delay compliance, and channel crosstalk level compliance, taking into account their respective weighting coefficients. This can be obtained through expert experience assignment or through the analytic hierarchy process. This permission level... The value ranges from 0 to 1. A larger value indicates a higher link state security margin, meaning a safer channel switching.

[0159] The solution proposed in this application introduces a comprehensive capability coefficient. This allows for dynamic adjustment of thresholds for various performance indicators, thereby more accurately reflecting the system's actual capabilities under different operating conditions. Specifically, the system first adjusts the thresholds based on decision agility. and physical response coefficient Through a comprehensive capability model Calculate the comprehensive ability coefficient This coefficient reflects the system's overall capability in rapid decision-making and physical reconfiguration. Next, this comprehensive capability coefficient is used... The system will dynamically calculate the clock phase synchronization error threshold. Link state detection delay threshold and channel crosstalk threshold These thresholds will be based on the comprehensive capability coefficient. and threshold tightening coefficient Adjustments will be made to improve the overall system capabilities. Higher dynamic thresholds require higher clock synchronization accuracy, link status detection speed, and channel crosstalk levels to ensure the quality of aggregated data transmission and signal integrity. After obtaining these dynamic thresholds, the system compares the current inter-channel crosstalk level, clock phase synchronization error, and link status detection delay with the corresponding dynamic thresholds to obtain the channel crosstalk level compliance. Clock phase synchronization error compliance and link state detection delay compliance These compliance metrics quantify the gap between the current system performance and the dynamic threshold. Finally, based on these compliance metrics, a dynamic licensing model for link switching performance is implemented. Calculate the final dynamic permittivity for link switching performance. This model weights and combines the compliance scores of each item, and uses the tanh function to normalize the results to the range (0,1), making... The value can intuitively reflect the security margin of link switching. The solution in this application introduces a comprehensive capability coefficient. To dynamically adjust various performance thresholds, enabling dynamic permission levels for link switching performance. The evaluation is more refined and adaptive. It not only considers the system's own decision-making and physical response capabilities, but also combines real-time evaluation of key performance indicators such as clock synchronization, link detection, and channel crosstalk. This allows for a more accurate assessment of the suitability of CPO switching matrix for dynamic aggregation of optical channels, avoiding performance degradation or service interruption caused by blind aggregation.

[0160] The following is a specific example illustrating how, in a CPO switching matrix, to obtain dynamic permission levels for link switching performance, the following steps can be taken:

[0161] First, the system continuously monitors the SDN controller decision latency, the southbound transmission latency of control signaling, and the local switching chip latency, and calculates the decision agility based on a preset decision agility model. For example, decision-making agility. It can be obtained through the exponential decay function Obtain, among which These are the decision delay index, transmission delay index, and chip delay index, respectively. Simultaneously, the system also monitors the laser turn-on / off time, modulator drive response time, photodetector response time, wavelength tuning and stabilization time, and frame / packet alignment time, and calculates the physical response coefficients based on a preset physical response model. For example, physical response coefficient It can be done Obtain, among which For fast photoelectric response coefficient, Establish response coefficients for the function. This is the mismatch penalty factor. Subsequently, the calculated decision agility... and physical response coefficient Input into the comprehensive capability model In the process, the comprehensive ability coefficient is obtained. For example, if It is 0.9. If the value is 0.8, then the comprehensive ability coefficient is... The value is 0.72. Next, we will use this comprehensive capability coefficient... This allows for dynamic adjustment of various performance thresholds. For example, assuming a basic clock phase synchronization error tolerance... The threshold tightening factor is 10 ps. The value is 0.5. Therefore, the clock phase synchronization error threshold is... Similarly, assuming a basic link state detection delay tolerance... The threshold tightening factor is 100ns. It remains at 0.5. Therefore, the link state detection delay threshold... For the channel crosstalk threshold, assume a baseline inter-channel crosstalk level. -20dB, ideal inter-channel crosstalk level The value is -30dB. Therefore, the channel crosstalk threshold is... After acquiring these dynamic thresholds, the system monitors the current inter-channel crosstalk level, clock phase synchronization error, and link status detection delay in real time. For example, if the current inter-channel crosstalk level is -25dB, the clock phase synchronization error is 5ps, and the link status detection delay is 50ns, then ratio processing is performed to obtain the compliance rate. For example, through appropriate normalization, the clock phase synchronization error compliance rate can be obtained. The link state detection delay compliance is 0.9. The channel crosstalk level compliance is 0.8. The compliance rate is 0.95. Finally, these compliance scores are... and preset weighting coefficients (For example Substituting these values ​​into the dynamic licensing model for link handover performance, the final dynamic licensing level for link handover performance is calculated. .

[0162] Through the above technical solution, this application provides a more refined and adaptive dynamic permissiveness evaluation mechanism for link switching performance. By introducing a comprehensive capability coefficient and dynamically adjusting the thresholds for clock phase synchronization error, link status detection delay, and inter-channel crosstalk level based on this coefficient, the system can fully consider its real-time operating status and capabilities when evaluating link switching security. This avoids misjudgments that may result from using fixed thresholds, i.e., being overly conservative when the system capability is strong, or having excessively high risks when the system capability is weak. Therefore, the solution of this application can more accurately reflect the actual suitability of CPO switching matrix for dynamic aggregation of optical channels, thereby maximizing resource utilization efficiency and service response speed while ensuring communication quality and system stability, and effectively reducing the risk of service interruption or performance degradation caused by improper link switching.

[0163] Preferably, the extinction ratio optimization model is expressed as:

[0164] ;

[0165] in, Indicates the extinction ratio of the target signal. Indicates the current signal extinction ratio. Indicates the dynamic permission level for target link switching performance. Indicates the dynamic permission level for link switching performance. This represents the environmental stability coefficient.

[0166] in, The extinction ratio represents the target signal extinction ratio, which is the expected extinction ratio of the aggregated optical signal. It reflects the ratio of the optical power in the "on" state to the optical power in the "off" state of the laser. Generally, a higher extinction ratio means better signal quality and lower inter-symbol interference. This value is the output of the optimization model and is used to configure the optical transceiver or modulator, or as a parameter of the feedback control loop to adjust the operating point of the optical components. This represents the current signal extinction ratio, which is the current optical signal extinction ratio measured or estimated in real time. It reflects the optical signal quality before optimization. This value can be obtained in real time by monitoring the optical power meter and signal analyzer integrated in the CPO module, or it can be estimated based on historical data and the current operating conditions of the optical components. This represents the target link switching performance dynamic permission level, which is the expected or ideal dynamic link switching performance permission level. It serves as a reference point for the optimization model and guides the adjustment of the target extinction ratio to achieve specific performance goals. This value can be pre-configured by the network operator according to the Service Level Agreement (SLA) or network design requirements, or it can be dynamically adjusted by a higher-level network management system based on the overall network load and performance goals. The dynamic permissibility of link switching performance is a dynamically calculated permissibility that reflects the current capability of the system to perform reliable and agile optical channel switching. Its value is derived from various factors such as decision agility, physical response, clock synchronization, link detection, and crosstalk. This value is used as input to the extinction ratio optimization model and is calculated by the preceding steps of the method. It can be continuously updated based on the real-time measurement and calculation of the underlying parameters. The environmental stability coefficient quantifies the stability of the operating environment, taking into account factors such as buffer emptying time, internal temperature change rate of CPO package, and power supply voltage fluctuation. The higher the value, the more stable the environment. This value is also used as the input of the extinction ratio optimization model, which comes from the previous steps of the method and can reflect the current environmental conditions based on sensor readings and system monitoring data.

[0167] The extinction ratio optimization model mathematically calculates the extinction ratio of the target signal. Extinction ratio of current signal Dynamically calculated link switching performance and dynamic permission level Predefined target link switching performance dynamic permission and environmental stability coefficient Connecting them. Specifically, the model utilizes the hyperbolic tangent function. To adjust the current signal extinction ratio. Dynamic permission level for current dynamic link switching performance. Approaching target permission level hour, The term approaches zero, and the hyperbolic tangent term also approaches zero, which means the target extinction ratio will be close to the current extinction ratio. If Significantly lower than (Indicating poor switching performance) A positive hyperbolic tangent term leads to an increase in the target extinction ratio. Conversely, if... Higher than (Indicating that the switching performance is better than the target), then A negative term, especially a negative hyperbolic tangent term, may allow for a reduction in the target extinction ratio. Environmental stability coefficient. As a proportionality factor in the denominator, it affects the adjustment of the extinction ratio. and Sensitivity to differences between them. Larger. A more stable environment will make adjustments less sensitive, and smaller adjustments will be less likely to occur. Environmental instability makes it more sensitive, enabling more proactive adjustments when the environment is unpredictable. This mathematical formula provides a precise and adaptive mechanism to determine the optimal target signal extinction ratio. By explicitly incorporating dynamic link switching performance permittivity and environmental stability, this method ensures that optical signal quality (represented by the extinction ratio) can be dynamically adjusted to match current system capabilities and environmental conditions. This makes optical channel aggregation more robust and efficient, preventing performance degradation under different operating scenarios and optimizing resource utilization. The model's structure ensures that adjustments are not arbitrary but guided by explicit performance targets and regulated by environmental factors, thereby enhancing the reliability and adaptability of the CPO switching matrix.

[0168] For example, in the operating scenario of a CPO switching matrix, assuming the currently measured signal extinction ratio The value is 10dB. Based on system calculations, the current link switching performance dynamic permissibility is... The value is 0.6, while the system dynamically sets the target link switching performance permissibility based on business needs and network policies. The value is 0.8. Simultaneously, by monitoring and calculating environmental parameters such as the internal temperature and power supply voltage fluctuations of the CPO package, the environmental stability coefficient was obtained. The value is 0.7. Substituting these parameters into the above extinction ratio optimization model, we get: The CPO switching matrix will adjust the bias current of the optical transmitter or the driving voltage of the modulator to increase the actual signal extinction ratio to 12.77dB based on this target value, in order to cope with the current low link switching performance allowance and ensure the signal quality and stability of optical channel aggregation.

[0169] In another scenario, if the link switching performance has dynamic permissibility... It is 0.9, higher than the target. The value is 0.8, and the environmental stability coefficient is... If it is 0.7, then This indicates that when system performance exceeds expectations, the extinction ratio requirement can be appropriately reduced, potentially leading to lower power consumption or extended device lifespan.

[0170] By introducing a specific extinction ratio optimization model, this application can accurately quantify and dynamically adjust the target signal extinction ratio during optical channel aggregation. This model comprehensively considers the current signal extinction ratio, dynamic link switching performance allowance, target link switching performance dynamic allowance, and environmental stability coefficient. This makes the determination of the target extinction ratio no longer static or empirical, but adaptively adjustable based on the real-time operating status of the CPO switching matrix and environmental changes. When link switching performance is lower than expected, the model can guide the system to increase the target extinction ratio to compensate for potential performance degradation and ensure the signal quality and reliability of the aggregation channel. Conversely, when system performance is good and the environment is stable, the model also allows for an appropriate reduction in the extinction ratio requirement, thereby optimizing system power consumption or extending device lifespan. This dynamic and quantitative optimization mechanism significantly improves the adaptability and robustness of the CPO switching matrix under complex service loads and changing environments, effectively avoiding signal quality degradation or resource waste caused by improper extinction ratio settings, thus ensuring the stability and efficiency of optical channel aggregation.

[0171] 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.

[0172] 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 dynamic aggregation of optical channels based on service load in a CPO switching matrix, characterized in that, include: A decision agility model is constructed based on the SDN controller decision delay, the southbound transmission delay of control signaling, and the local switching chip delay to obtain decision agility. The environmental stability coefficient is obtained based on the buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation. A fast photoelectric response coefficient is obtained based on the laser turn-on / turn-off time, the modulator drive response time, and the photodetector response time. Based on wavelength tuning and settling time, as well as frame / packet alignment time, the function establishes response coefficients; Based on the fast photoelectric response coefficient and functional establishment of the response coefficient, the physical response coefficient is obtained; Based on decision agility and physical response coefficient, clock phase synchronization error, link status detection delay and inter-channel crosstalk level are used to obtain dynamic permission for link switching performance. Based on the dynamic permission of link switching performance, environmental stability coefficient, and current signal extinction ratio, the extinction ratio of the target signal is obtained through an extinction ratio optimization model.

2. The method for dynamic aggregation of optical channels based on service load in a CPO switching matrix according to claim 1, characterized in that, The extinction ratio optimization model is expressed as follows: ； in, Indicates the extinction ratio of the target signal. Indicates the current signal extinction ratio. Indicates the dynamic permission level for target link switching performance. Indicates the dynamic permission level for link switching performance. This represents the environmental stability coefficient.

3. The method for dynamic aggregation of optical channels based on service load in the CPO switching matrix according to claim 2, characterized in that, The method for dynamic permission allocation of link switching performance is as follows: Based on the decision agility coefficient and the physical response coefficient, the comprehensive capability coefficient is obtained through a comprehensive capability model, which is expressed as follows: ； in, This represents the overall capability coefficient. Indicates decision-making agility. Represents the physical response coefficient; Based on the basic clock phase synchronization error tolerance and the overall capability coefficient, the clock phase synchronization error threshold is obtained through a clock phase synchronization error model, which is expressed as follows: ； in, This indicates the clock phase synchronization error threshold. This indicates the tolerance for the phase synchronization error of the base clock. Indicates the threshold tightening factor; Based on the basic link state detection delay tolerance and the comprehensive capability coefficient, the link state detection delay threshold is obtained through a link state detection delay model, which is expressed as follows: ； in, Indicates the link state detection latency threshold. Indicates the latency tolerance for basic link state detection; Based on the baseline inter-channel crosstalk level, the ideal inter-channel crosstalk level, and the overall capability coefficient, the channel crosstalk threshold is obtained through a channel crosstalk model, which is expressed as follows: ； in, Indicates the channel crosstalk threshold. Indicates the level of crosstalk between reference channels. This indicates the ideal level of crosstalk between channels; The current inter-channel crosstalk level is compared with the channel crosstalk threshold to obtain the channel crosstalk level compliance. The clock phase synchronization error threshold and the link status detection delay threshold are compared with the current clock phase synchronization error and the current link status detection delay to obtain the clock phase synchronization error compliance and the link status detection delay compliance. Based on clock phase synchronization error compliance, link state detection delay compliance, and channel crosstalk level compliance, a dynamic link handover performance compliance model is used to obtain the dynamic compliance of link handover performance. This dynamic compliance model is expressed as follows: ； in, Indicates the dynamic permission level for link switching performance. Indicates the clock phase synchronization error compliance. Indicates the compliance of link state detection delay. Indicates the channel crosstalk level compliance. Represents the weight coefficient and The Furthermore, the larger the value, the higher the link state security margin.

4. The method for dynamic aggregation of optical channels based on service load in a CPO switching matrix according to claim 3, characterized in that, The method for obtaining the physical response coefficients is as follows: Obtain the fast photoelectric response coefficient and the functional establishment response coefficient; Based on the fast photoelectric response coefficient and functional establishment of the response coefficient, the physical response coefficient is obtained through a physical response model, which is expressed as follows: ； in, Represents the physical response coefficient. Represents the fast photoelectric response coefficient. This indicates that the function establishes response coefficients. Represents the mismatch penalty factor, the Furthermore, the larger the value, the higher the overall response agility of physical reconfiguration.

5. The method for dynamic aggregation of optical channels based on service load in a CPO switching matrix according to claim 2, characterized in that, The method for obtaining the environmental stability coefficient is as follows: Acquire buffer emptying time, CPO package internal temperature change rate, and power supply voltage fluctuation; The buffer emptying time, the internal temperature change rate of the CPO package, and the power supply voltage fluctuation are compared with the corresponding maximum allowable values ​​to obtain the buffer emptying time index, the temperature change rate index, and the voltage fluctuation index. Based on the buffer emptying time exponent, temperature change rate exponent, and voltage fluctuation exponent, the environmental stability coefficient is obtained through an environmental stability model, which is expressed as follows: ； in, Indicates the environmental stability coefficient. This indicates the buffer emptying time index. Indicates the rate of temperature change index. Indicates the voltage fluctuation index. Represents the weight coefficient and The Furthermore, the larger the value, the more stable the environment.

6. The method for dynamic aggregation of optical channels based on service load in a CPO switching matrix according to claim 3, characterized in that, The way to obtain decision agility is: Acquire SDN controller decision latency, control signaling southbound transmission latency, and local switching chip latency; The decision delay index, transmission delay index, and chip delay index of the SDN controller, the southbound transmission delay of the control signaling, and the local switching chip delay are compared with the corresponding reference values ​​to obtain the decision delay index, transmission delay index, and chip delay index. Decision agility is obtained through a decision agility model based on the decision delay index, transmission delay index, and chip delay index. The decision agility model is expressed as follows: ； in, Indicates decision-making agility. Indicates the decision delay index, Indicates the transmission delay index. Indicates the chip delay index, the Furthermore, the larger the value, the more agile the decision-making.

7. The method for dynamic aggregation of optical channels based on service load in a CPO switching matrix according to claim 4, characterized in that, The method for obtaining the fast photoelectric response coefficient is as follows: Acquire the laser turn-on / turn-off time, modulator drive response time, and photodetector response time; After performing maximum-minimum normalization on the laser turn-on / off time, modulator drive response time, and photodetector response time, and taking their complements, the laser stabilization time index, modulator drive response time index, and photodetector response time index are obtained. Based on the laser stabilization time index, the modulator drive response time index, and the photodetector response time index, the fast photoelectric response coefficient is obtained through a fast photoelectric response model, which is expressed as follows: ； in, Represents the fast photoelectric response coefficient. Indicates the laser settling time index. Indicates the modulator drive response time exponent. This represents the response time index of a photodetector. Represents the weight coefficient and The The larger the value, the faster the device response.

8. The method for dynamic aggregation of optical channels based on service load in a CPO switching matrix according to claim 4, characterized in that, The method for obtaining the response coefficient is as follows: Obtain wavelength tuning and settling time, as well as frame / packet alignment time; The wavelength tuning and stabilization time and the frame / packet alignment time are compared with the corresponding expected values ​​after the difference is processed and then compared with the allowable deviation from the expected value to obtain the wavelength tuning and stabilization time index and the alignment time index. The wavelength tuning and stabilization time and the frame / packet alignment time are both greater than or equal to the corresponding expected values. Based on the wavelength tuning and settling time exponents and the alignment time exponent, the function establishment response coefficients are obtained through a function establishment response model, which is expressed as follows: ； in, This indicates that the function establishes response coefficients. Indicates the wavelength tuning and settling time index. The alignment time index is represented by the following. The larger the value, the faster the function is established.

Images