Optical communication distribution method for efficient network communication

Abstract

The invention discloses an optical communication distribution method for efficient network communication, which relates to the technical field of optical communication distribution, and comprises the following steps of: in the initial stage of network design, distributing an expected number of wavelengths to each communication link as initial expected wavelengths according to the expected load demand of each communication link; when data streams are transmitted by communication links, the load condition of each communication link is monitored in real time, and increment characteristics reflecting dynamic changes of network loads are extracted from network load information obtained in real time. According to the method, static and dynamic wavelength allocation is combined, the resource utilization efficiency and the network adaptability are improved, a high-load type link and a normal-load type link are dynamically divided through initial wavelength resource allocation, real-time monitoring, feature extraction and support vector machine prediction, the wavelength of the high-load link is adjusted in real time, the load pressure is relieved, and the service life of the high-load link is prolonged. The stable performance in the peak period is ensured, and the problems of resource waste and response lag are effectively solved.

Description

Technical Field

[0001] The present invention relates to the technical field of optical communication distribution, and in particular to an optical communication distribution method for efficient network communication. Background Art

[0002] Optical communication allocation refers to the reasonable allocation and management of bandwidth resources of optical signals in optical fiber communication networks to ensure that information can be transmitted efficiently and reliably. It usually involves resource allocation between multiple users or multiple data streams, especially on shared optical fibers or optical transmission links. Optical communication allocation schedules different wavelength, time or frequency resources to maximize the utilization of network capacity and ensure that each data stream obtains sufficient bandwidth to avoid congestion or resource conflicts. Common optical communication allocation methods include wavelength division multiplexing (WDM), time division multiplexing (TDM) and frequency division multiplexing (FDM). The application of these technologies enables optical communication networks to provide high-throughput and low-latency communication services in multi-user environments.

[0003] In existing optical communication networks, static wavelength allocation strategies are often used, which means that each wavelength has been fixedly allocated to a specific communication link during network design and will not change during the entire communication process. In other words, each wavelength is bound to a specific link from the beginning of network deployment, and the allocation of resources is pre-set and cannot be adjusted according to the actual network status. The main advantages of this static allocation method are that it is simple to implement, relatively stable network management, and can ensure reliable resource allocation. However, the disadvantages of this strategy are also obvious: when the network load changes, fixed wavelength resources may be inefficiently utilized. For example, some wavelengths may be idle during certain periods, while other wavelengths may be over-occupied, but because the allocation is fixed, it cannot be flexibly adjusted, resulting in a waste of bandwidth resources. In addition, the static allocation strategy cannot respond to congestion or failures in the network in a timely manner, and cannot flexibly respond to the dynamic needs of actual services, affecting the adaptability and efficiency of the network.

[0004] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0005] The purpose of the present invention is to provide an optical communication allocation method for efficient network communication, which allocates initial wavelength resources to cope with conventional loads in the early stage of design, and combines real-time monitoring and support vector machine models to predict network load dynamics, divides high-load and normal-load links, and adjusts the wavelength of high-load links in real time to alleviate load pressure and ensure stable performance during peak hours, thereby solving the problems of resource waste and response lag under static allocation strategies, making the network more efficient and stable in a dynamic load environment, so as to solve the problems in the above-mentioned background technology.

[0006] In order to achieve the above object, the present invention provides the following technical solution: an optical communication distribution method for efficient network communication, comprising the following steps:

[0007] In the early stage of network design, an expected number of wavelengths is allocated to each communication link as initial expected wavelengths according to the expected load demand of each communication link;

[0008] When the communication link transmits data stream, the load of each communication link is monitored in real time, and the incremental features reflecting the dynamic changes of network load are extracted from the real-time acquired network load information;

[0009] Under the monitoring window, the extracted dynamic incremental features are analyzed, and the analyzed feature data is input into the pre-trained support vector machine model to predict the communication link network load;

[0010] Based on the prediction results of the support vector machine model, the network load changes in the monitoring window are divided into two categories: "high load type" and "normal load type";

[0011] For communication links classified as normal load type, the preset initial wavelength resources continue to be used, that is, the initial expected wavelength continues to be allocated to each communication link to cope with normal network load fluctuations;

[0012] For communication links predicted to be high-loaded, wavelength resources are regulated according to the real-time load conditions, and more wavelengths are allocated to the communication links to ensure that the communication link network maintains stable performance during peak load periods.

[0013] Preferably, the initial expected wavelength refers to the initial wavelength resource allocated to each communication link according to the expected bandwidth requirement of the link during the network design phase, so as to meet the normal communication requirements of the link.

[0014] Preferably, incremental features reflecting dynamic changes in network load are extracted from network load information acquired in real time, wherein the extracted features include the number of data packets waiting to be processed in the communication link queue and the rate of change of the arrival timing of data packets in the communication link. After acquisition, the number of data packets waiting to be processed in the communication link queue and the rate of change of the arrival timing of data packets in the communication link are analyzed under the monitoring window, and a queue length index and a jitter rate index are generated respectively. The queue length index is used to quantify the situation of data packets queuing for processing on the communication link, reflecting the degree of buffer congestion caused when the link is overloaded. The jitter rate index is used to quantify the fluctuation of the arrival timing of data packets, reflecting the volatility and instability of network delay.

[0015] Preferably, after obtaining the queue length index and jitter rate index generated by analyzing the extracted incremental features that reflect the dynamic changes in network load, the queue length index and jitter rate index are input into a pre-trained support vector machine model to generate a network load index, and the network load of the communication link is predicted through the network load index.

[0016] Preferably, the network load index generated when predicting the network load of the communication link by the support vector machine model under the monitoring window is compared and analyzed with the preset network load index reference threshold, and the network load change under the monitoring window is divided. The specific division steps are as follows:

[0017] If the network load index generated in the monitoring window is greater than or equal to the network load index reference threshold, the network load change in the monitoring window is classified as "high load type";

[0018] If the network load index generated in the monitoring window is less than the network load index reference threshold, the network load change in the monitoring window is classified as a "normal load type".

[0019] Preferably, the specific steps of generating a queue length index after analyzing the number of data packets waiting to be processed in the communication link queue under the monitoring window are as follows:

[0020] In the monitoring window, the number of data packets waiting to be processed in the link queue is collected in real time. First, the cumulative queue load on the link is calculated. The cumulative queue load refers to the difference between the number of backlogged data packets in the link queue and the actual transmission capacity of the network within a period of time. An exponential function is defined to calculate the cumulative queue load in each monitoring window. The calculation expression of the cumulative queue load is:

[0021]

[0022] Among them, Q index (t) represents the cumulative queue load at time t, L queue(t) represents the number of packets waiting to be processed in the queue at time t, S rate is the processing rate of the link, that is, the number of data packets that the link can process per unit time, t1 and t2 represent the start time and end time of the monitoring window respectively;

[0023] The generated cumulative queue load is normalized to generate a queue length index. Specifically, the cumulative queue load generated in each time period is mapped into a unified range. The normalized calculation expression is:

[0024]

[0025] In the formula, Q normalized represents the normalized queue length index, α and β are parameters that control the weight increase of queue length under high load conditions, and C max is the maximum processing capability of the link.

[0026] Preferably, the specific steps of generating a jitter rate index after analyzing the rate of change of the arrival timing of data packets in the communication link under the monitoring window are as follows:

[0027] Analyze the changes in the arrival timing of data packets, calculate the time difference between the arrival times of adjacent data packets, and capture the interval fluctuations of data packet arrival. For a given communication link, let the arrival times of data packets be t1, t2, t3, …, t n , where n is the total number of data packets, then the calculation expression for the delay difference between data packets is: Δt i =t i+1 -t i , i = 1, 2, 3, ..., n-1, where Δt i It represents the delay difference between the ith and i+1th data packets, reflecting the change in the arrival timing of adjacent data packets;

[0028] Analyze the delay difference Δt i The incremental changes of the delay difference are captured to capture the fluctuation characteristics of finer granularity and quantify the dynamic changes of the delay difference. Specifically, the incremental difference is introduced and recorded as ΔΔt i , calculate the incremental difference ΔΔt i , the calculation expression is: ΔΔt i =Δt i+1 -Δt i , incremental differential ΔΔt i Reflects the acceleration of the delay difference change, Δt i+1 represents the delay difference between the i+1th and i+2th data packets;

[0029] Calculate the jitter rate index. The jitter rate index quantifies the volatility of the packet arrival timing through the maximum value of the incremental difference and the variance of the overall delay difference of the communication link, and measures the severity of the jitter. The calculation expression of the jitter rate index is:

[0030]

[0031] Among them, max(ΔΔt i ) represents the incremental difference ΔΔt i The maximum value, i.e. the maximum incremental change, reflects the degree of sudden fluctuation; is the variance of the delay difference, which is used to represent the overall fluctuation characteristics of the link. rate Indicates the jitter rate index.

[0032] Preferably, for a communication link predicted to be of high load type, wavelength resources are regulated according to the real-time load situation, and more wavelengths are allocated to the communication link. The specific steps are as follows:

[0033] After being predicted as a high-load type, the additional wavelength demand increment is first determined to meet the high-load requirements of the link. The wavelength demand increment is based on the network load index Network Load and network load index reference threshold Threshold Load The difference is calculated by the expression:

[0034]

[0035] , where Δλ represents the wavelength demand increment, λ initial is the number of initially expected wavelengths of the communication link, representing the preset number of wavelengths under normal load;

[0036] Based on the calculated wavelength demand increment Δλ, the wavelength resource adjustment coefficient is determined to accurately control the increment of wavelength resources to avoid over-allocation of resources. The calculation expression of the wavelength resource adjustment coefficient is:

[0037]

[0038] Among them, α′ represents the wavelength resource adjustment coefficient, Some of them introduced the logarithmic variation of network load index;

[0039] The final number of wavelengths allocated under high load is calculated based on the wavelength resource adjustment coefficient and the wavelength demand increment. The formula is as follows:

[0040] λ final =λ initial +α′·Δλ

[0041] Among them, λ finalrepresents the final number of wavelengths allocated to the link under high load conditions, and α′·Δλ is the wavelength increment adjusted according to the load conditions.

[0042] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0043] The present invention achieves a significant improvement in resource utilization efficiency and enhanced network adaptability by combining the advantages of static and dynamic wavelength allocation. Initial wavelength resources are preset at the beginning of network design to ensure that the link can cope with conventional loads. At the same time, through real-time monitoring and incremental feature extraction, the support vector machine model is used to predict network load changes and dynamically divide high-load and normal-load links. For high-load links, by adjusting wavelength resources in real time, load pressure is effectively alleviated, congestion is avoided, and stable link performance during peak hours is ensured. The overall solution takes into account both resource stability and flexibility, solves the resource waste and response lag problems caused by static allocation strategies, and enables the network to operate more efficiently and stably in complex and dynamic load environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0045] Figure 1 The present invention is a method flow chart of an optical communication allocation method for efficient network communication. DETAILED DESCRIPTION

[0046] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0047] The present invention provides Figure 1 An optical communication distribution method for efficient network communication shown includes the following steps:

[0048] In the early stage of network design, an expected number of wavelengths is allocated to each communication link as initial expected wavelengths according to the expected load demand of each communication link;

[0049] The initial expected wavelength refers to the initial wavelength resource allocated to each communication link during the network design phase based on the expected bandwidth requirements of the link to meet the normal communication needs of the link. These initial expected wavelengths are reasonably allocated resources based on estimates of network service types, link lengths, data transmission rates, and other communication parameters. The number of initial expected wavelengths is usually fixed to ensure that the link can smoothly transmit data. Its main function is to provide stable bandwidth support for the basic communications of the network to ensure the reliable operation of the network.

[0050] The setting of the initial expected wavelength should be determined based on the physical characteristics of each communication link and the expected communication needs. For example, when the link distance is long or the data transmission rate requirement is high, more wavelength resources are needed to provide sufficient bandwidth. For short-distance links or when the transmission demand is low, fewer wavelengths can be allocated. When setting the initial expected wavelength, the designer can refer to the historical data of similar networks, standard communication demand models, and expected quality of service requirements to ensure that the initial wavelength resources of each link can meet conventional needs without causing waste of resources.

[0051] When the communication link transmits data stream, the load of each communication link is monitored in real time, and the incremental features reflecting the dynamic changes of network load are extracted from the real-time acquired network load information;

[0052] Under the monitoring window, the extracted dynamic incremental features are analyzed, and the analyzed feature data is input into the pre-trained support vector machine model to predict the communication link network load;

[0053] When the link transmits data streams, real-time monitoring of the load of each communication link can be achieved in a variety of ways. Common methods include using network traffic monitoring tools (such as SNMP, NetFlow, or sFlow) to collect real-time link bandwidth utilization, traffic rate, delay, packet loss rate and other data. In addition, network detection equipment or sensors can also be deployed to collect data packet information on the link in real time, analyze the flow direction, size and transmission delay of the data packet, and further obtain the real-time load status of the link.

[0054] Incremental features reflecting dynamic changes in network load are extracted from network load information acquired in real time, wherein the extracted features include the number of data packets waiting to be processed in the communication link queue and the rate of change of the arrival timing of data packets in the communication link. After acquisition, the number of data packets waiting to be processed in the communication link queue and the rate of change of the arrival timing of data packets in the communication link are analyzed under the monitoring window, and then a queue length index and a jitter rate index are generated respectively. The queue length index is used to quantify the situation of data packets queuing for processing on the communication link, reflecting the degree of buffer congestion caused when the link is overloaded. The jitter rate index is used to quantify the fluctuation of the arrival timing of data packets, reflecting the volatility and instability of network delay. It is usually used to evaluate the impact of delay imbalance and sudden changes in traffic on the performance of the network when the load increases.

[0055] After obtaining the queue length index and jitter rate index generated by analyzing the extracted incremental features reflecting the dynamic changes in network load, the queue length index and jitter rate index are input into a pre-trained support vector machine model to generate a network load index, and the network load of the communication link is predicted through the network load index.

[0056] A pre-trained support vector machine (SVM) model is a support vector machine model that is trained in advance using a large amount of historical data in the network load prediction task, and can automatically make predictions based on the input features (in this case, the queue length index and the jitter rate index). Support vector machine is a powerful supervised learning algorithm that divides data points into different categories or performs regression analysis by finding an optimal hyperplane in the feature space. During the training phase, the SVM model uses a labeled data set (that is, known network load conditions and corresponding queue length, jitter rate, etc.) to optimize a loss function to determine the optimal hyperplane, thereby learning a generalizable classification or regression function. This process includes feature extraction of the data, adjustment of model parameters, and selection of appropriate kernel functions to deal with nonlinear features. After training, the SVM model can generate the corresponding network load index based on new input data, such as queue length index and jitter rate index.

[0057] In practical applications, the role of pre-trained support vector machine models lies in their efficient classification and regression capabilities, especially when faced with high-dimensional, complex, and nonlinear data. SVM can map data to high-dimensional space through appropriate kernel techniques in order to find the classification boundary or regression curve of the data. Specifically for the application in this article, features such as queue length index and jitter rate index reflect the dynamic changes of network load. By inputting these features into the SVM model, the system can derive a network load index to predict the future load of the communication link. The output of the SVM model, namely the network load index, reflects the current load status of the network link and can help the network management system make more timely decisions. The pre-trained SVM model can not only make efficient predictions based on existing data, but also be gradually optimized as more data is accumulated, improving the accuracy of predictions and the level of intelligent network management.

[0058] A significant increase in the number of packets waiting to be processed in a communication link queue is usually an indication that the network load on that communication link is becoming high. When the load on a link increases, the number of packets transmitted exceeds the link's processing power or bandwidth capacity, causing packets to accumulate in the queue awaiting further processing. An increase in queue length reflects that the link is experiencing a bottleneck in processing data and is unable to forward all packets in a timely manner. This is usually caused by a surge in network traffic, insufficient bandwidth, or link congestion. A backlog of packets in the queue indicates that the link is overloaded, which can lead to increased latency, data loss, and overall performance degradation, further confirming that the network is becoming more loaded. Therefore, a significant increase in queue length is an important indicator of increased network load.

[0059] The specific steps for generating a queue length index after analyzing the number of data packets waiting to be processed in the communication link queue under the monitoring window are as follows:

[0060] In the monitoring window, the number of data packets waiting to be processed in the link queue is collected in real time. First, the cumulative queue load on the link is calculated. The cumulative queue load refers to the difference between the number of data packets accumulated in the link queue and the actual transmission capacity of the network within a period of time (such as per second, per minute, etc.). In order to reflect the dynamic changes of the queue process, an exponential function is defined to calculate the cumulative queue load in each monitoring window. This process takes into account the arrival rate, processing rate and queue backlog of data packets. The calculation expression of the cumulative queue load is:

[0061]

[0062] Among them, Q index (t) represents the cumulative queue load at time t, L queue (t) represents the number of packets waiting to be processed in the queue at time t (i.e., link load), S rateis the processing rate of the link, that is, the number of data packets that the link can process per unit time, t1 and t2 represent the start time and end time of the monitoring window respectively;

[0063] Through this integral function, the cumulative change of link load in the entire monitoring window can be obtained, further quantifying the difference between network load and link processing capacity. The cumulative queue load reflects the increase in queues caused by insufficient link processing capacity during the monitoring period, reflecting the degree of buffer congestion.

[0064] The generated cumulative queue load is normalized to generate a queue length index. Specifically, the cumulative queue load generated in each time period is mapped to a unified range (for example, 0 to 1) to facilitate subsequent analysis and decision-making. In the normalization process, a weighting function based on load changes is used to dynamically adjust the contribution of the cumulative queue load in different time periods, especially to enhance the weight of high-load periods to highlight the bottleneck points of the link. The normalized calculation expression is:

[0065]

[0066] In the formula, Q normalized represents the normalized queue length index. α and β are parameters that control the weight of the queue length under high load conditions. They are usually set through experimental tuning and are not specifically limited here. C max is the maximum processing capability of the link (i.e., maximum throughput or maximum packet processing rate);

[0067] α generally controls the overall sensitivity of the queue length index. When the network is highly loaded, the value of α is increased to increase the weight of the queue length index under high load conditions and strengthen the responsiveness of network bottlenecks under high load. Conversely, when the load is low, the value of α is reduced to reduce the impact of the queue length index under low load. This value can be adjusted dynamically based on real-time traffic, link performance, and historical load patterns. The parameter β controls the amplification effect of the queue length index under high load conditions. It defines the acceleration of the queue length index growth when the link load exceeds a certain threshold. When the network load increases, the value of β increases, making the queue length index grow more rapidly during high load periods, helping to better identify and respond to congestion problems. Similarly, when the network load is light, the value of β is reduced so that the queue length index does not overreact to slight changes in load.

[0068] The normalization formula is used to stabilize the queue length index when the network load is low; when the load is too heavy, the queue length index will increase rapidly due to the weighting coefficient, thereby significantly amplifying the high load situation and reflecting the degree of congestion and load increase of the link. This process ensures that the queue length index can not only quantify the normal load, but also provide sensitive feedback when the network load surges.

[0069] It can be seen from the queue length index that the larger the performance value of the queue length index generated after analyzing the number of data packets waiting to be processed in the communication link queue under the monitoring window, it usually means that the network load of the communication link is more serious. When the queue length on the communication link increases, it means that more data packets are waiting to be processed, reflecting the insufficient processing capacity of the link, resulting in an aggravation of the queuing phenomenon. As the number of data packets in the queue increases, the degree to which the link load exceeds its processing capacity is also increasing, which makes the queue length index larger, representing the severity of the network load. On the contrary, when the queue length index is small, it means that there are fewer data packets on the link and the queuing pressure of the queue is small, indicating that the network load is light or in a normal state. Therefore, the queue length index can effectively quantify the degree of link load. The larger the queue length index value, the more serious the network load.

[0070] Poor stability in the rate of change of packet arrival timing in a communication link usually means that the network load of the communication link is becoming higher. When the network load increases, the data traffic on the link may increase or fluctuate suddenly, resulting in uneven changes in the packet arrival timing. Specifically, when the load increases, the link may experience a large amount of data transmission in a short period of time, which will cause large fluctuations in the arrival interval of the data packets, that is, increased delay jitter. Poor rate of change stability indicates that when the link handles these bursts of traffic, it may generate queuing, delay fluctuations, and congestion, further affecting the data transmission quality and latency performance. Therefore, poor stability in the rate of change of packet arrival timing is a clear sign of increased network load.

[0071] The specific steps of generating the jitter rate index after analyzing the rate of change of the arrival timing of data packets in the communication link under the monitoring window are as follows:

[0072] Analyze the changes in the arrival timing of data packets and calculate the time difference between the arrival times of adjacent data packets (i.e., delay). The purpose of this process is to capture the interval fluctuations of data packet arrival. For a given communication link, let the arrival times of data packets be t1, t2, t3, …, t n , where n is the total number of data packets, then the calculation expression for the delay difference between data packets is: Δt i =t i+1 -t i, i = 1, 2, 3, ..., n-1, where Δt i It represents the delay difference between the ith and i+1th data packets, reflecting the change in the arrival timing of adjacent data packets;

[0073] This step will provide basic data for the subsequent calculation of the jitter rate index, that is, the delay difference Δt i .

[0074] Analyze the delay difference Δt i The incremental changes of the delay difference are used to capture the fluctuation characteristics of finer granularity. The purpose of this step is to quantify the dynamic changes of the delay difference and identify whether there is a significant increase in fluctuation. To this end, the incremental difference (i.e., the difference between adjacent delay differences) is introduced, denoted as ΔΔt i , the calculation expression is: ΔΔt i =Δt i+1 -Δt i , incremental differential ΔΔt i It reflects the acceleration of the delay difference change and can capture the drastic change or "jitter" of the delay fluctuation. i+1 represents the delay difference between the i+1th and i+2th data packets;

[0075] If the incremental difference ΔΔt i A large value indicates that the delay difference between adjacent data packets fluctuates suddenly during the change process, suggesting that the network load fluctuates significantly.

[0076] Calculate the jitter rate index. The jitter rate index quantifies the volatility of the packet arrival timing through the maximum value of the incremental difference and the variance of the overall delay difference of the communication link, and measures the severity of the jitter. The calculation expression of the jitter rate index is:

[0077]

[0078] Among them, max(ΔΔt i ) represents the incremental difference ΔΔt i The maximum value, i.e. the maximum incremental change, reflects the degree of sudden fluctuation; is the variance of the delay difference, which is used to represent the overall fluctuation characteristics of the link. rate represents the jitter rate index;

[0079] This formula combines the amplitude of delay difference fluctuations and their cumulative effects to provide a more refined and complex jitter rate index that can accurately reflect the delay imbalance caused by increased network load and its impact on network performance.

[0080] It can be seen from the jitter rate index that the larger the performance value of the jitter rate index generated by analyzing the rate of change of the arrival timing of data packets in the communication link under the monitoring window, the more serious the network load of the communication link is. The jitter rate index reflects the volatility and instability of the arrival timing of data packets. Especially when the network load increases, the traffic on the link may fluctuate suddenly, resulting in a sharp change in the arrival interval of data packets. When the network load increases, the link may face greater transmission delay and delay fluctuation, resulting in an increase in the change of the delay difference and its increment, thereby increasing the jitter rate index. On the contrary, if the network load is light, the delay change of the link tends to be stable, the fluctuation of the arrival timing of data packets is small, and the jitter rate index is low. Therefore, the increase of the jitter rate index is an obvious indicator of the increase of network load, which can effectively reflect the congestion of the link and the instability of network performance.

[0081] The support vector machine model is not specifically limited here, and can realize the queue length index Q normalized and Jittre rate Perform comprehensive analysis to generate network load index Network Load The support vector machine model can be used. In order to realize the technical solution of the present invention, the present invention provides a specific implementation method; Network load index Network Load The resulting calculation formula is:

[0082]

[0083] Where x1 and x2 are queue length index Q respectively. normalized and Jittre rate The preset proportional coefficient, and x1 and x2 are both greater than 0.

[0084] In this formula, the preset proportional coefficients (i.e., x1 and x2) refer to the artificially set weights used to adjust the queue length index Q normalized and Jittre rate Calculating the network load index Network Load By setting different preset proportional coefficients, network administrators or system designers can control the influence of these two parameters on the final load index. For example, if the x1 value is large and x2 is small, the calculation result will focus more on the influence of the queue length index, and vice versa. This adjustment can make the model more flexible to adapt to different network environments and business needs.

[0085] It can be seen from the network load index that the larger the performance value of the queue length index generated after analyzing the number of data packets waiting to be processed in the communication link queue under the monitoring window, the larger the performance value of the jitter rate index generated after analyzing the rate of change of the arrival timing of data packets in the communication link under the monitoring window. That is, the larger the performance value of the network load index generated when predicting the network load situation of the communication link through the support vector machine model under the monitoring window, the more serious the network load of the communication link is, and vice versa.

[0086] Based on the prediction results of the support vector machine model, the network load changes in the monitoring window are divided into two categories: "high load type" and "normal load type";

[0087] The network load index generated when predicting the network load of the communication link through the support vector machine model under the monitoring window is compared and analyzed with the pre-set network load index reference threshold, and the network load change under the monitoring window is divided. The specific division steps are as follows:

[0088] If the network load index generated in the monitoring window is greater than or equal to the network load index reference threshold, the network load change in the monitoring window is classified as "high load type";

[0089] High load type means that the current network load of the communication link has reached a high level, which usually means that the link is close to or has exceeded its design capacity, and problems such as insufficient bandwidth, increased latency, and data packet loss may occur. At this time, the traffic demand in the network exceeds the normal processing range of the system, and the link load is close to saturation or has obvious congestion. Classifying the network load as high load type helps to identify bottlenecks on the link and remind the system to perform load management or resource allocation to ensure communication stability and service quality.

[0090] If the network load index generated in the monitoring window is less than the network load index reference threshold, the network load change in the monitoring window is classified as "normal load type";

[0091] Normal load type means that the current network load of the communication link is within the controllable range, and the link's bandwidth, delay, packet loss rate and other performance indicators are at normal levels. At this time, the link's resources can meet the current traffic demand, and load fluctuations will not have a significant impact on link performance. Classifying the network load as normal load type indicates that the link is operating well, no additional resource allocation is required, and the network can operate stably with the expected efficiency.

[0092] For communication links classified as normal load type, the preset initial wavelength resources continue to be used, that is, the initial expected wavelength continues to be allocated to each communication link to cope with normal network load fluctuations;

[0093] For those communication links that are predicted or classified as normal load type, that is, links where the current network load is within the normal range, no additional wavelength resource adjustment is required. The system will continue to use the "initial expected wavelength" pre-allocated during the design phase to support data transmission. These initial wavelength resources are set according to the basic communication needs of the link, which is sufficient to cope with normal traffic fluctuations and daily load changes without the need to temporarily increase or decrease wavelength resources. Therefore, under normal load conditions, the system does not need to reconfigure or dynamically adjust the wavelength, which simplifies resource management and reduces unnecessary operational overhead. This strategy ensures the stability of resources and avoids the complexity caused by over-allocation, while maintaining efficient operation of the link and low management costs. When the link load is within the normal fluctuation range, the initial wavelength resources are sufficient to meet the transmission needs and ensure the continuity and stability of communication.

[0094] For communication links predicted to be high-loaded, wavelength resources are regulated according to the real-time load conditions, and more wavelengths are allocated to the communication links to ensure that the communication link network maintains stable performance during peak load periods.

[0095] For a communication link predicted to be high-load, wavelength resources are regulated according to the real-time load situation, and more wavelengths are allocated to the communication link. The specific steps are as follows:

[0096] After being predicted as a high-load type, the additional wavelength demand increment is first determined to meet the high-load requirements of the link. The wavelength demand increment is based on the network load index Network Load and network load index reference threshold Threshold Load The difference is calculated by the expression:

[0097]

[0098] Where Δλ represents the wavelength demand increment, λ initial is the number of initially expected wavelengths of the communication link, representing the preset number of wavelengths under normal load;

[0099] By calculating the wavelength demand increment Δλ, the system is able to quantify the additional wavelengths needed to cope with high load conditions.

[0100] Based on the calculated wavelength demand increment Δλ, the wavelength resource adjustment coefficient is determined to accurately control the increment of wavelength resources to avoid over-allocation of resources. The calculation expression of the wavelength resource adjustment coefficient is:

[0101]

[0102] Among them, α' represents the wavelength resource adjustment coefficient, The logarithmic variation of the network load index is partially introduced to ensure that the wavelength resource adjustment coefficient is smoother when the network load index is close to the network load index reference threshold;

[0103] The calculation of wavelength resource adjustment coefficients can help accurately control wavelength resource allocation to cope with traffic demands during high-load periods.

[0104] The final number of wavelengths allocated under high load is calculated based on the wavelength resource adjustment coefficient and the wavelength demand increment. The formula is as follows:

[0105] λ final =λ initial +α′·Δλ

[0106] Among them, λ final represents the final number of wavelengths allocated to the link under high load conditions, α′·Δλ is the wavelength increment adjusted according to the load conditions;

[0107] Through this calculation, the system can flexibly adjust wavelength allocation so that the link can cope with the current high load demand while preventing resource waste.

[0108] Dynamically allocate wavelength resources according to the real-time load of the communication link, and allocate more wavelength resources to the links with higher loads to avoid network congestion and ensure smooth transmission of data streams. Through this resource regulation, it is possible to effectively cope with the traffic demand of the link during peak load periods, thereby maintaining the stability of link performance and avoiding bandwidth bottlenecks, increased latency, or data packet loss caused by insufficient resources. The key to this step is to allocate network resources intelligently and flexibly, so that the system can still provide reliable and high-quality communication services under high load conditions, ensuring the overall efficiency and service quality of the network.

[0109] This optical communication allocation method combines the advantages of static and dynamic wavelength allocation to achieve a significant improvement in resource utilization efficiency and enhanced network adaptability. Initial wavelength resources are preset at the beginning of network design to ensure that the link can cope with regular loads. At the same time, through real-time monitoring and incremental feature extraction, the support vector machine model is used to predict network load changes and dynamically divide high-load and normal-load links. For high-load links, by adjusting wavelength resources in real time, load pressure can be effectively alleviated, congestion can be avoided, and stable link performance during peak hours can be ensured. The overall solution takes into account both resource stability and flexibility, solves the problems of resource waste and response lag caused by static allocation strategies, and enables the network to operate more efficiently and stably in complex and dynamic load environments.

[0110] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0111] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

[0112] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

Claims

1. An optical communication distribution method for efficient network communication, characterized in that: The following steps are involved: In the early stage of network design, an expected number of wavelengths is allocated to each communication link as initial expected wavelengths according to the expected load demand of each communication link; When the communication link transmits data stream, the load of each communication link is monitored in real time, and the incremental features reflecting the dynamic changes of network load are extracted from the real-time acquired network load information; Under the monitoring window, the extracted dynamic incremental features are analyzed, and the analyzed feature data is input into the pre-trained support vector machine model to predict the communication link network load; Based on the prediction results of the support vector machine model, the network load changes in the monitoring window are divided into two categories: "high load type" and "normal load type"; For communication links classified as normal load type, the preset initial wavelength resources continue to be used, that is, the initial expected wavelength continues to be allocated to each communication link to cope with normal network load fluctuations; For communication links predicted to be high-loaded, wavelength resources are regulated according to the real-time load conditions, and more wavelengths are allocated to the communication links to ensure that the communication link network maintains stable performance during peak load periods.

2. The optical communication distribution method for efficient network communication according to claim 1, characterized in that: The initial expected wavelength refers to the initial wavelength resource allocated to each communication link during the network design phase based on the expected bandwidth requirements of the link to meet the normal communication requirements of the link.

3. The optical communication distribution method for efficient network communication according to claim 1, characterized in that: Incremental features reflecting dynamic changes in network load are extracted from network load information acquired in real time, wherein the extracted features include the number of data packets waiting to be processed in the communication link queue and the rate of change of the arrival timing of data packets in the communication link. After acquisition, the number of data packets waiting to be processed in the communication link queue and the rate of change of the arrival timing of data packets in the communication link are analyzed under the monitoring window, and a queue length index and a jitter rate index are generated respectively. The queue length index is used to quantify the situation of data packets queuing for processing on the communication link, reflecting the degree of buffer congestion caused when the link is overloaded. The jitter rate index is used to quantify the fluctuation of the arrival timing of data packets, reflecting the volatility and instability of network delay.

4. The optical communication distribution method for efficient network communication according to claim 3, characterized in that: After obtaining the queue length index and jitter rate index generated by analyzing the extracted incremental features reflecting the dynamic changes in network load, the queue length index and jitter rate index are input into a pre-trained support vector machine model to generate a network load index, and the network load of the communication link is predicted through the network load index.

5. The optical communication distribution method for efficient network communication according to claim 4, characterized in that: The network load index generated when predicting the network load of the communication link through the support vector machine model under the monitoring window is compared and analyzed with the pre-set network load index reference threshold, and the network load change under the monitoring window is divided. The specific division steps are as follows: If the network load index generated in the monitoring window is greater than or equal to the network load index reference threshold, the network load change in the monitoring window is classified as "high load type"; If the network load index generated under the monitoring window is less than the network load index reference threshold, the network load change under the monitoring window is classified as "normal load type".

6. The optical communication distribution method for efficient network communication according to claim 3, characterized in that: The specific steps for generating a queue length index after analyzing the number of data packets waiting to be processed in the communication link queue under the monitoring window are as follows: In the monitoring window, the number of data packets waiting to be processed in the link queue is collected in real time. First, the cumulative queue load on the link is calculated. The cumulative queue load refers to the difference between the number of backlogged data packets in the link queue and the actual transmission capacity of the network within a period of time. An exponential function is defined to calculate the cumulative queue load in each monitoring window. The calculation expression of the cumulative queue load is: Among them, Q index (t) represents the cumulative queue load at time t, L queue (t) represents the number of packets waiting to be processed in the queue at time t, S rate is the processing rate of the link, that is, the number of data packets that the link can process per unit time, t1 and t2 represent the start time and end time of the monitoring window respectively; The generated cumulative queue load is normalized to generate a queue length index. Specifically, the cumulative queue load generated in each time period is mapped into a unified range. The normalized calculation expression is: In the formula, Q normalized represents the normalized queue length index, α and β are parameters that control the weight increase of queue length under high load conditions, and C max is the maximum processing capability of the link.

7. The optical communication distribution method for efficient network communication according to claim 3, characterized in that: The specific steps of generating the jitter rate index after analyzing the rate of change of the arrival timing of data packets in the communication link under the monitoring window are as follows: Analyze the changes in the arrival timing of data packets, calculate the time difference between the arrival times of adjacent data packets, and capture the interval fluctuations of data packet arrival. For a given communication link, let the arrival times of data packets be t1, t2, t3, …, t n , where n is the total number of data packets, then the calculation expression for the delay difference between data packets is: Δt i =t i+1 -t i , i = 1, 2, 3, ..., n-1, where Δt i It represents the delay difference between the ith and i+1th data packets, reflecting the change in the arrival timing of adjacent data packets; Analyze the delay difference Δt i The incremental changes of the delay difference are captured to capture the fluctuation characteristics of finer granularity and quantify the dynamic changes of the delay difference. Specifically, the incremental difference is introduced and recorded as ΔΔt i , calculate the incremental difference ΔΔt i , the calculation expression is: ΔΔt i =Δt i+1 -Δt i , incremental differential ΔΔt i Reflects the acceleration of the delay difference change, Δt i+1 represents the delay difference between the i+1th and i+2th data packets; Calculate the jitter rate index. The jitter rate index quantifies the volatility of the packet arrival timing through the maximum value of the incremental difference and the variance of the overall delay difference of the communication link, and measures the severity of the jitter. The calculation expression of the jitter rate index is: Among them, max(ΔΔt i ) represents the incremental difference ΔΔt i The maximum value, i.e. the maximum incremental change, reflects the degree of sudden fluctuation; is the variance of the delay difference, which is used to represent the overall fluctuation characteristics of the link. rate Indicates the jitter rate index.

8. The optical communication distribution method for efficient network communication according to claim 5, characterized in that: For a communication link predicted to be high-load, wavelength resources are regulated according to the real-time load situation, and more wavelengths are allocated to the communication link. The specific steps are as follows: After being predicted as a high-load type, the additional wavelength demand increment is first determined to meet the high-load requirements of the link. The wavelength demand increment is based on the network load index Network Load and network load index reference threshold Threshold Load The difference is calculated by the expression: Where Δλ represents the wavelength demand increment, λ initial is the number of initially expected wavelengths of the communication link, representing the preset number of wavelengths under normal load; Based on the calculated wavelength demand increment Δλ, the wavelength resource adjustment coefficient is determined to accurately control the increment of wavelength resources to avoid over-allocation of resources. The calculation expression of the wavelength resource adjustment coefficient is: Among them, α' represents the wavelength resource adjustment coefficient, Some of them introduced the logarithmic variation of network load index; The final number of wavelengths allocated under high load is calculated based on the wavelength resource adjustment coefficient and the wavelength demand increment. The formula is as follows: l final =λ initial +α′·Dl Among them, λ final represents the final number of wavelengths allocated to the link under high load conditions, and α′·Δλ is the wavelength increment adjusted according to the load conditions.

Images