Data center optical interconnection resource migration method

Abstract

The invention relates to a data center optical interconnection resource migration method. The method comprises the following steps: acquiring first state information of each optical interconnection resource in a data center, and determining an optical interconnection resource to be migrated based on the first state information; formulating an optical interconnection resource migration scheme according to the first state information corresponding to the optical interconnection resource to be migrated; the optical interconnection resource migration scheme is used for indicating optical interconnection resource migration equipment to migrate optical interconnection resources needing to be migrated; and obtaining second state information of each migrated optical interconnection resource, and verifying the migrated optical interconnection resource needing to be migrated in the data center based on the second state information. By adopting the method, the migration scheme of the optical interconnection resources is formulated in an intelligent mode, the efficiency of migration of the optical interconnection resources can be improved, and stable operation of the data center is ensured.

Description

Technical Field

[0001] The invention belongs to the technical field of data center automation management, and in particular relates to a method for migrating optical interconnection resources in a data center. Background Art

[0002] In today's era of rapid digital development, data centers are the core hubs for information storage, processing and transmission, and their importance is self-evident. With the continuous increase in data center business volume and the continuous upgrading of technology, optical interconnection resources occupy a key position in the architecture of data centers.

[0003] However, as business needs change, on the one hand, data centers may need to adjust their architecture, upgrade equipment, or optimize their layout, which involves reconfiguring and migrating optical interconnection resources. On the other hand, optical interconnection resources may need to be migrated due to equipment failure, performance degradation, or network optimization needs.

[0004] Traditional optical interconnection resource migration often relies on manual experience for planning and execution, lacks systematicity and scientificity, and easily leads to low migration efficiency and high time cost.

[0005] Therefore, there is an urgent need for an intelligent data center optical interconnect resource migration method to improve the efficiency of migrating optical interconnect resources and ensure the stable operation of the data center. Summary of the invention

[0006] Based on this, it is necessary to provide a data center optical interconnection resource migration method, a data center optical interconnection resource migration device and a data center optical interconnection resource management system that can improve migration efficiency in response to the above technical problems.

[0007] In a first aspect, the present application provides a data center optical interconnection resource migration method, comprising:

[0008] Acquire first status information of each optical interconnection resource in the data center, and determine the optical interconnection resource to be migrated based on the first status information;

[0009] Formulate an optical interconnect resource migration plan according to the first state information corresponding to the optical interconnect resource to be migrated; the optical interconnect resource migration plan is used to instruct the optical interconnect resource migration device to migrate the optical interconnect resource to be migrated;

[0010] The second state information of each optical interconnection resource after migration is obtained, and the optical interconnection resources to be migrated after migration in the data center are verified based on the second state information.

[0011] In a second aspect, the present application also provides a data center optical interconnection resource migration device, comprising:

[0012] A resource monitoring module, used to monitor the status of each optical interconnection resource in the data center in real time, obtain first status information of each optical interconnection resource in the data center, and determine the optical interconnection resource to be migrated based on the first status information;

[0013] A migration planning module, used to formulate an optical interconnection resource migration plan according to the first state information corresponding to the optical interconnection resource to be migrated;

[0014] A migration execution module, used for migrating the optical interconnection resources to be migrated according to the optical interconnection resource migration plan;

[0015] The verification module obtains the second state information of each optical interconnection resource after migration, and verifies the optical interconnection resources to be migrated after migration in the data center based on the second state information.

[0016] In a third aspect, the present application also provides a data center optical interconnection resource management system, including: a data center, a sensor, a server, and an optical interconnection resource migration device;

[0017] The sensor is connected to the data center and the server, and is used to monitor the status of each optical interconnection resource in the data center in real time, and send first status information of each optical interconnection resource in the data center to the server;

[0018] The optical interconnection resource migration device is connected to the data center and the server, and is used to migrate the optical interconnection resources to be migrated according to the optical interconnection resource migration plan sent by the server;

[0019] The server is used to execute the method as described in the first aspect.

[0020] The above-mentioned data center optical interconnection resource migration method, device, and data center optical interconnection resource management system formulate a migration plan for optical interconnection resources in an intelligent manner, improve the efficiency of migrating optical interconnection resources, and ensure the stable operation of the data center. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 A schematic diagram of the process of the data center optical interconnection resource migration method provided by the present invention;

[0022] Figure 2 A schematic diagram of the steps for formulating an optical interconnection resource migration scheme in an optional embodiment of the present invention;

[0023] Figure 3 A schematic diagram of the steps of training a deep reinforcement learning model in an optional embodiment of the present invention;

[0024] Figure 4 A schematic diagram of the structure of a data center optical interconnection resource migration device provided by the present invention. DETAILED DESCRIPTION

[0025] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0026] refer to Figure 1 , which shows a flow chart of a data center optical interconnection resource migration method provided in an embodiment of the present application, the method is applied to a server, and includes the following steps:

[0027] S101: Acquire first status information of each optical interconnection resource in a data center, and determine optical interconnection resources that need to be migrated based on the first status information.

[0028] Specifically, the server obtains the first status information of each optical interconnection resource in the data center. Such status information may include the connection status, bandwidth usage, transmission efficiency, connection path, optical signal strength, optical link bit error rate, etc. of the optical interconnection resource. The server determines which optical interconnection resources need to be migrated by analyzing such status information. For example, if the bandwidth usage of a certain optical interconnection resource is too high or the transmission efficiency is low, the server may determine it as an optical interconnection resource that needs to be migrated.

[0029] S102, formulating an optical interconnection resource migration plan according to the first state information corresponding to the optical interconnection resource to be migrated; the optical interconnection resource migration plan is used to instruct the optical interconnection resource migration device to migrate the optical interconnection resource to be migrated.

[0030] Specifically, the server formulates an optical interconnect resource migration plan based on the first state information corresponding to the optical interconnect resource to be migrated. This optical interconnect resource migration plan is to instruct the optical interconnect resource migration device to perform a migration operation on the optical interconnect resource to be migrated. When formulating the plan, the server may consider the current state of the optical interconnect resource to be migrated, the target location of the migration, the time schedule of the migration, possible problems that may arise during the migration process, and the corresponding measures. For example, if the connection of an optical interconnect resource to be migrated is more complex, the server may arrange for the migration to be carried out during a low-traffic period and notify the relevant equipment in advance to be prepared.

[0031] S103: Acquire second status information of each migrated optical interconnection resource, and verify the migrated optical interconnection resources to be migrated in the data center based on the second status information.

[0032] Specifically, the server obtains the second state information of each optical interconnection resource after migration. This state information is similar to the first state information, but reflects the situation after migration. The server verifies the optical interconnection resources that need to be migrated in the data center after migration based on the second state information. The verification content may include whether the optical interconnection resources have been successfully migrated to the target location, whether the connection is normal, whether the bandwidth usage has been improved, whether the transmission efficiency has been improved, etc. If the verification finds a problem, the server will take corresponding measures to make adjustments to ensure the normal operation of the optical interconnection resources.

[0033] By utilizing the above method and using a server to intelligently formulate a migration plan for optical interconnection resources, the efficiency of migrating optical interconnection resources can be improved, and the stable operation of the data center can be ensured.

[0034] refer to Figure 2 In an optional embodiment, formulating an optical interconnection resource migration plan according to the first state information corresponding to the optical interconnection resource to be migrated includes the following steps:

[0035] S201: Determine a current connection path of the optical interconnection resource to be migrated in a data center according to first state information corresponding to the optical interconnection resource to be migrated.

[0036] Specifically, the first state information may include information related to the connection path, such as connected device nodes, port information, etc. The server analyzes this information to accurately sort out the specific connection direction of the optical interconnection resources to be migrated in the data center.

[0037] S202 . According to the current connection path of the optical interconnection resource to be migrated in the data center, an optimal migration path is planned with consideration of the path length, service impact, and optical signal loss.

[0038] Specifically, after determining the current connection path, the server plans the optimal migration path with path length, business impact and optical signal loss as important considerations. In terms of path length, the shorter the path length, the lower the migration cost and the smaller the signal transmission delay. In terms of optical signal loss, the server will select the path that minimizes the optical signal loss to ensure that the optical interconnection resources after migration can transmit data efficiently and stably. In terms of business impact, the server needs to evaluate the degree of interference of different paths on ongoing businesses and try to select the path with the least impact on the business; for example, different migration paths pass through different network devices and links in the data center, and these links themselves have different bandwidth allocations. If the link bandwidth passed by the migration path is limited, and the data traffic involved in the migration process is large, it will lead to an increase in the delay of business data transmission, thereby increasing the degree of business impact.

[0039] S203: Determine a migration time window according to first state information corresponding to the optical interconnection resource to be migrated.

[0040] Specifically, the first state information may include information such as the frequency of use of the optical interconnection resource, business peak period, historical network traffic data, etc. The server analyzes this information and selects a time period with relatively low business volume, relatively small network traffic, and minimal impact on the business as the migration time window. For example, if a certain optical interconnection resource is busy during the day and has relatively large network traffic, the server will choose to migrate at night.

[0041] S204: Use the migration time window and the migration path as an optical interconnection resource migration solution.

[0042] Specifically, the server will integrate the determined migration time window and the planned optimal migration path as the optical interconnect resource migration plan. This plan will provide clear guidance for the optical interconnect resource migration equipment, ensuring that the migration process is carried out efficiently and smoothly, while minimizing the impact on data center services.

[0043] In an optional embodiment, according to the current connection path of the optical interconnection resource to be migrated in the data center, taking the path length, service impact and optical signal loss as considerations, an optimal migration path is planned, including the following three steps:

[0044] Step 1: Set the parameters of the simulated annealing algorithm, including setting the initial temperature T0, the cooling coefficient λ and the maximum number of iterations N max , and establish the objective function E(x)=αL+γBI+βS; where x represents the current connection path; Represents the optical signal loss, P in is the input optical power of the optical interconnection resource, P out is the output optical power of the optical interconnection resource; Indicates the degree of business impact, n is the number of businesses, ω i is the weight of the ith business, t mi is the expected maximum delay time of the ith service during the migration process, t bi is the benchmark delay time for normal business operation; S is the path length; α is the weight coefficient of optical signal loss in the objective function; γ is the weight coefficient of business impact in the objective function; β is the weight coefficient of path length in the objective function;

[0045] Step 2: Perform iterative optimization, including: in each iteration, generate a new connection path x′ based on the current connection path x; calculate the optical signal loss corresponding to the new connection path x′ Business impact And the length S' of the new connection path, and then calculate the objective function value E(x') = αL' + γBI' + βS'; according to the state transition probability formula Determine whether to accept the new path x′. If accepted, update the current connection path to the new connection path; update the formula T according to the temperature k+1 =λT k , update temperature T; where T k is the temperature of the kth iteration, and k ≥ 0; T k+1 is the temperature of the k+1th iteration; λ is the cooling coefficient, and 0<λ<1;

[0046] Step 3: Perform termination judgment, including: checking whether the termination condition is met during each iteration. The termination condition is that the number of iterations reaches the predetermined maximum number of iterations N. max ; When the termination condition is met, the algorithm terminates and outputs the optimal new connection path x found as the optimal migration path.

[0047] Specifically, the simulated annealing algorithm used here is a random optimization algorithm based on the Monte Carlo iterative solution strategy, which is used to find the global optimal solution or the approximate optimal solution in a large search space. Its core idea is to simulate the solid annealing process and accept a solution worse than the current solution with a certain probability during the search process to avoid falling into the local optimal solution. The basic formula mainly involves the state transition probability and the objective function.

[0048] First, set the current state (which can be understood as the current migration path in path optimization) to x, the new state (new migration path) to x′, the objective function to E(x) (in the optimization of optical interconnection resource migration paths, it can be a function including factors such as optical signal loss L, service impact BI and path length S, such as E(x)=αL+γBI+βS, where α, γ and β are weight coefficients of optical signal loss L, service impact BI and path length S in the objective function, respectively), the temperature parameter to T, and the state transition probability formula is: Cooling coefficient λ and maximum number of iterations N max .in, P in is the input optical power of the optical interconnection resource, P out is the output optical power of the optical interconnection resource; n is the number of services, ω i is the weight of the ith business, t mi is the expected maximum delay time of the ith service during the migration process, t bi This is the baseline latency for normal business operation.

[0049] Then, iterative optimization is performed. In each iteration, a new connection path x′ is generated based on the current connection path x; the optical signal loss corresponding to the new connection path x′ is calculated Degree of service impact and the length S′ of the new connection path, and then the objective function value E(x′) = αL′ + γBI′ + βS′ is calculated; according to the state transition probability formula judge whether to accept the new path x′. If accepted, update the current connection path to the new connection path; update the temperature T according to the temperature update formula T k+1 = λT k , update the temperature T; where T k is the temperature of the k-th iteration, and k ≥ 0; T k+1 is the temperature of the (k + 1)-th iteration; λ is the cooling coefficient, and 0 < λ < 1.

[0050] Finally, check whether the termination condition is satisfied in each iteration process. The termination condition is that the number of iterations reaches the predetermined maximum number of iterations N max ; when the termination condition is satisfied, the algorithm terminates and outputs the found optimal new connection path x as the optimal migration path.

[0051] Among them, the above state transition probability formula means that when E(x′) < E(x), the formula is P(x→x′) = 1. Here, E(x) is the objective function value corresponding to the current state (current connection path), and E(x′) is the objective function value corresponding to the new state (new connection path). When the objective function value of the new connection path is less than the objective function value of the current connection path, it means that the new connection path is better in terms of the optimization objective (comprehensively considering factors such as optical signal loss, degree of service impact, path length, etc.). Therefore, in this case, the probability of accepting the new path is 1, that is, the new connection path will definitely be accepted as the current path for the next iteration. For example, if the objective function value of the current path is 10 and the objective function value of the new path is 8, since the new connection path is better, the algorithm will毫不犹豫地 accept the new connection path.

[0052] When E(x′) ≥ E(x), the formula is At this time, the objective function value of the new connection path is greater than or equal to the objective function value of the current connection path, which indicates that the new connection path may not be as good as the current connection path in terms of the comprehensive performance measured by the objective function. However, to avoid falling into local optimality, the algorithm does not directly reject this new connection path, but decides whether to accept it with a probability. This probability is related to the difference in the objective function values E(x′) - E(x) and the current temperature parameter T. The exponential part The smaller the value, the greater the probability P(x→x′) of accepting a new connection path. When the temperature T is high, even if the difference in the objective function value is large, the probability of accepting a new connection path is relatively high. This is like when the energy state of the system is relatively chaotic at high temperature, and it is easier to accept some states with slightly higher energy (analogous to the objective function value in simulated annealing). As the temperature T decreases, the probability of accepting a worse new path will also decrease, and the algorithm will gradually converge to a better solution. For example, assuming the current temperature T = 10, the objective function value difference E(x′)-E(x) = 2, the calculated acceptance probability is There is a certain probability of accepting this new connection path; when temperature T = 1, under the same objective function value difference, the acceptance probability The probability of acceptance is reduced.

[0053] In the process of optimizing the connection paths of optical interconnection resources, the state transition probability formula enables the algorithm to be more flexible in exploring new connection paths. Sometimes, although a new connection path may cause a slight increase in optical signal loss or a slight increase in business impact or a slight decrease in path length (making the objective function value larger), the algorithm still has a chance to accept this new path through this probability formula. This mechanism helps the algorithm escape the trap of local optimal solutions and find the global optimal or near-optimal connection path in a wider path space, thereby better balancing multiple optimization goals such as optical signal loss, business impact, and path length.

[0054] In an optional embodiment, determining the current connection path of the optical interconnection resource to be migrated in the data center according to the first state information corresponding to the optical interconnection resource to be migrated includes the following steps:

[0055] Acquire topological structure data of the data center from first state information corresponding to the optical interconnection resource to be migrated;

[0056] According to the topological structure data of the data center, the current connection path of the optical interconnection resources to be migrated in the data center is determined.

[0057] Specifically, the first status information may include a description of the overall network architecture of the data center, such as the connection relationship between various devices, the hierarchical structure, etc. The server accurately obtains the topological structure data of the data center by carefully screening and extracting this information. This step provides a basis for determining the specific location and connection status of the optical interconnection resources to be migrated.

[0058] After obtaining the topological structure data of the data center, the server will gradually track and analyze the connection direction of the optical interconnection resources to be migrated based on the device nodes and connection relationships in the topological structure. For example, by finding the devices connected to the resource and the connection links between these devices, the server can clearly determine the current connection path of the optical interconnection resources to be migrated in the data center and prepare for subsequent migration planning.

[0059] In an optional embodiment, determining the migration time window according to the first state information corresponding to the optical interconnection resource to be migrated includes the following steps:

[0060] Acquiring historical network traffic data from first state information corresponding to the optical interconnection resource to be migrated;

[0061] Based on historical network traffic data, use the trained deep reinforcement learning model to predict network traffic and obtain network traffic prediction data;

[0062] According to the network traffic forecast data, the time period with the minimum network traffic is determined as the migration time window.

[0063] Specifically, the server obtains historical network traffic data from the first state information corresponding to the optical interconnection resource to be migrated. The first state information may include traffic records of the network environment in which the optical interconnection resource is located over a period of time in the past, including information such as the traffic size in different time periods and the time when the peak traffic occurs. By extracting these historical network traffic data, the server provides a basis for subsequent traffic prediction and determination of the migration time window.

[0064] The server uses the trained deep reinforcement learning model to predict network traffic based on the historical network traffic data obtained to obtain network traffic prediction data. The deep reinforcement learning model can grasp the changing rules and trends of network traffic by learning from historical data. The server inputs the historical network traffic data into the model, and the model outputs the network traffic prediction results for a period of time in the future after calculation and analysis. This process uses advanced machine learning technology to improve the accuracy and reliability of traffic prediction.

[0065] Based on the network traffic forecast data, the server determines the time period with the lowest network traffic as the migration time window. By analyzing the network traffic forecast data, the server can find the time period with lower network traffic. Migrating optical interconnection resources during this time period can minimize the impact on network services. Because lower network traffic means fewer ongoing data transmission tasks, the interference to existing services during the migration process will also be reduced accordingly. After careful screening and comparison, the server finally determines the time period with the lowest network traffic as the best time window for migration to ensure the smooth progress of the migration process.

[0066] refer to Figure 3 In an optional embodiment, the training process of the deep reinforcement learning model includes the following steps:

[0067] S301, obtaining historical network traffic data 30 days ago from historical network traffic data, wherein the historical network traffic data 30 days ago is used to train a deep reinforcement learning model, and the historical network traffic data within 30 days is used to input the trained deep reinforcement learning model to obtain network traffic prediction data.

[0068] S302, preprocessing the historical network traffic data from 30 days ago to obtain a network traffic data set; dividing the network traffic data set into a training set and a validation set.

[0069] S303. Build a deep reinforcement learning model.

[0070] S304: Input the training set into the deep reinforcement learning model to train the deep reinforcement learning model.

[0071] S305: Input the verification set into the trained deep reinforcement learning model to optimize the parameters of the deep reinforcement learning model, and finally obtain the trained deep reinforcement learning model.

[0072] Specifically, the server obtains the historical network traffic data from 30 days ago from the historical network traffic data, and specifies that this part of the data is used to train the deep reinforcement learning model, while the historical network traffic data within 30 days is used to input the trained model to obtain network traffic prediction data. This division ensures that model training and prediction use data from different time periods, avoiding overfitting and improving the model's generalization ability.

[0073] Clean the historical network traffic data from 30 days ago to remove abnormal values. The network traffic data is N = {n1, n2, ..., n t}, where n t represents the network traffic at time t. By calculating the mean and standard deviation Data points outside the range [μ-k1σ,μ+k1σ] are considered outliers and removed, where k1 is a constant. Normalization is performed using the formula The data is made to be in the range of [0,1]. After preprocessing, the network traffic data set is obtained, and then it is divided into a training set (80%), a validation set (10%), and a test set (10%) in chronological order to ensure the temporal continuity of the data.

[0074] DQN (Deep Q-Network) can be used to build a deep reinforcement learning model. The state space is defined as the network traffic at the current moment and the previous moments, that is, where n t-i represents the network flow at time ti, and k2 represents the number of historical moments considered. The action space can be simplified in this model that focuses on network flow prediction to not involve specific action selection, but mainly focus on the predicted output of network flow. The reward function is defined as R t =-(Predicted t -Actual t ) 2 , here Predicted t is the network traffic predicted by the model at time t, Actual t is the actual network traffic at time t, so that the model gets higher rewards when the predicted value is closer to the actual value.

[0075] The server inputs the training set into the deep reinforcement learning model to train the model. The network traffic data samples in the training set are defined according to the state space For each training sample time t, the model input is a vector consisting of the network traffic values ​​at k2 times before the current time t. Setting the model parameter to θ, the model prediction function can be expressed as in The model is in state S t The predicted value of network traffic at time t is given below. During the training process, the model minimizes the loss function To adjust the parameter θ, here y i is the actual network traffic value, is the model prediction value, and n is the number of training samples. During the training process, the model makes predictions based on the input state space data (historical network traffic sequence in the training set) and calculates the current predicted reward value in combination with the reward function. Although the loss function is L(θ), the reward function measures the quality of the model prediction from another perspective. During the training process, the model is essentially trying to find a set of parameters θ that make the prediction result As close as possible to the actual value y i , to maximize the reward R t From the perspective of gradient descent, the gradient information of the reward function can provide a reference direction for adjusting model parameters. Through the back-propagation algorithm, the model parameters are adjusted according to the loss function to make the model's prediction results as close as possible to the actual network traffic, that is, to optimize in the direction of maximizing the reward.

[0076] The validation set data is also organized according to the definition of the state space. When the validation set is input into the model, for each sample time t in the validation set, the input received by the model is By calculating the prediction error on the validation set, such as the mean square error Where m is the number of samples in the validation set, is the actual network traffic value in the validation set, is the model's prediction value on the validation set. During the validation phase, when calculating the prediction results on the validation set, the reward function can be used to intuitively evaluate the prediction quality of the model on the validation set. For example, the average reward for all samples in the validation set can be calculated as in is the reward value calculated based on the validation set samples. If the mean square error of the validation set does not achieve the expected decrease or overfitting occurs, some regularization methods can be used to optimize the model. For example, add the L2 regularization term to the loss function, and the new loss function is Where λ is the regularization coefficient, θ j is a parameter of the model. By adjusting the value of λ, the complexity of the model can be controlled to prevent overfitting, thereby improving the prediction performance of the model on unseen data (validation set).

[0077] The test set is input into the trained model, and the prediction ability of the model in the actual application scenario is evaluated by calculating the mean square error and other indicators of the model on the test set. The data of the test set is also input according to the state space definition. Assume that the number of test set samples is p, and the mean square error on the test set is calculated. To evaluate the model performance, is the actual network traffic value in the test set, is the predicted value of the model on the test set. For the test set, the average reward can also be calculated The prediction performance of the model in actual application scenarios is evaluated by observing the size of the reward value. At the same time, combined with indicators such as the mean square error on the test set, it is possible to more comprehensively evaluate whether the model can accurately predict network traffic and whether it can make effective predictions based on the goal embodied by the reward function (reducing the gap between the prediction and the actual value). If the model performs well on the test set, it means that the model has good generalization ability and can be used for actual network traffic prediction; if the performance is poor, it is necessary to further adjust the model structure and parameters or retrain and optimize.

[0078] In an optional embodiment, determining that the optical interconnection resource needs to be migrated based on the first state information includes the following steps:

[0079] Acquire the optical link bit error rate in each optical interconnection resource and the optical signal strength corresponding to the connection path from the first state information;

[0080] Marking the optical interconnection resources whose optical link bit error rate is higher than a preset value or whose optical signal strength is not within a preset range;

[0081] The marked optical interconnection resources are used as optical interconnection resources that need to be migrated.

[0082] Specifically, the server obtains the optical link bit error rate in each optical interconnection resource and the optical signal strength corresponding to the connection path from the first state information of the optical interconnection resource. The first state information may include various performance parameters and state indicators of the optical interconnection resource. The server accurately obtains the two key parameters of the optical link bit error rate and the optical signal strength through a specific data extraction method. The optical link bit error rate reflects the probability of an error in the optical signal during transmission, while the optical signal strength indicates the strength of the optical signal.

[0083] The server marks the optical interconnection resources whose optical link bit error rate is higher than the preset value or whose optical signal strength is not within the preset range. The preset value and preset range are determined based on the data center's performance requirements and stability standards for optical interconnection resources. If the optical link bit error rate is too high, it means that the frequency of optical signal errors during transmission is high, which may affect the accuracy and reliability of data transmission. If the optical signal strength is not within the preset range, it may cause the signal to attenuate too much or too little, affecting the transmission quality and distance of the signal. The server compares the actual optical link bit error rate and optical signal strength with the preset value and preset range, and marks the optical interconnection resources that do not meet the requirements.

[0084] The server identifies the marked optical interconnect resources as those that need to be migrated. After the marking step, those optical interconnect resources that are considered to have poor performance or do not meet the requirements are identified as objects that need to be migrated. The server organizes and records these marked optical interconnect resources to provide clear goals for the subsequent formulation of optical interconnect resource migration plans.

[0085] In an optional embodiment, verifying the migrated optical interconnection resources to be migrated in the data center based on the second state information includes the following steps:

[0086] Step 1: obtaining the second state information corresponding to the optical interconnection resource to be migrated after migration from the second state information; and obtaining the optical link bit error rate and optical signal strength of the optical interconnection resource to be migrated after migration as a verification result according to the second state information corresponding to the optical interconnection resource to be migrated after migration;

[0087] Step 2: If the bit error rate in the verification result is higher than a preset value and / or the optical signal strength is not within a preset range, the fault location is determined according to the verification result, and the fault is repaired;

[0088] Step 3: Repeat the above steps until the verification result meets the preset conditions; wherein the preset conditions are that the bit error rate in the verification result is lower than a preset value and the optical signal strength is within a preset range.

[0089] Specifically, the server obtains the second state information corresponding to the optical interconnection resources to be migrated after the migration from the second state information. After the migration of the optical interconnection resources is completed, the server will collect new state data to evaluate the migration effect. Then, based on these second state information, the server obtains the optical link bit error rate and optical signal strength of the optical interconnection resources to be migrated after the migration. These two parameters, as verification results, directly reflect the performance status of the optical interconnection resources after migration. The optical link bit error rate reflects the accuracy of data transmission. The lower the bit error rate, the higher the reliability of data transmission; the optical signal strength determines the transmission quality and distance of the signal. The appropriate optical signal strength can ensure stable signal transmission.

[0090] If the bit error rate in the verification result is higher than the preset value and / or the optical signal strength is not within the preset range, it means that there are still problems with the migrated optical interconnection resources. The server needs to determine the location of the fault based on the verification results. This involves checking the connection paths, device interfaces, etc. of the optical interconnection resources one by one, and determining the specific location where the fault may occur by analyzing the abnormalities of the bit error rate and optical signal strength, combined with the topology and device configuration information of the data center. Once the fault location is determined, the server will generate a fault report, which is for maintenance personnel to review and take appropriate measures to repair the fault based on this fault report. For example, if the problem is caused by a poor connection, you may need to reconnect or replace the connection line; if it is a device failure, you may need to replace or repair the relevant equipment.

[0091] The server will cycle through the above steps, continuously verifying and repairing the migrated optical interconnection resources until the verification results meet the preset conditions. The preset conditions are that the bit error rate in the verification results is lower than the preset value and the optical signal strength is within the preset range. This process ensures that the optical interconnection resources can work normally after migration and meet the performance requirements of the data center. Through continuous cycle verification and repair, the server can gradually optimize the status of the optical interconnection resources and improve the overall stability and reliability of the data center.

[0092] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0093] refer to Figure 4 , is a data center optical interconnection resource migration device 400 provided in an embodiment of the present application, the device comprising:

[0094] The resource monitoring module 401 is used to monitor the status of each optical interconnection resource in the data center in real time, obtain first status information of each optical interconnection resource in the data center, and determine the optical interconnection resource to be migrated based on the first status information.

[0095] The migration planning module 402 is used to formulate an optical interconnection resource migration plan according to the first state information corresponding to the optical interconnection resource to be migrated.

[0096] The migration execution module 403 is used to migrate the optical interconnection resources to be migrated according to the optical interconnection resource migration plan.

[0097] The verification module 404 obtains the second status information of each optical interconnection resource after migration, and verifies the optical interconnection resources to be migrated after migration in the data center based on the second status information.

[0098] Optionally, the migration planning module 402 includes:

[0099] The connection path acquisition unit is used to determine the current connection path of the optical interconnection resource to be migrated in the data center according to the first state information corresponding to the optical interconnection resource to be migrated.

[0100] The migration path planning unit is used to plan the optimal migration path according to the current connection path of the optical interconnection resources to be migrated in the data center, taking the path length, service impact and optical signal loss as considerations.

[0101] The migration time window determining unit is used to determine the migration time window according to the first state information corresponding to the optical interconnection resource to be migrated.

[0102] The migration scheme determining unit is used to use the migration time window and the migration path as the optical interconnection resource migration scheme.

[0103] Optionally, the migration path planning unit is specifically used to perform the following operations, including:

[0104] Step 1: Set the parameters of the simulated annealing algorithm, including setting the initial temperature T0, the cooling coefficient λ and the maximum number of iterations N max , and establish the objective function E(x) = αL + γBI + βS;

[0105] Among them, x represents the current connection path; Represents the optical signal loss, P in is the input optical power of the optical interconnection resource, P out is the output optical power of the optical interconnection resource; Indicates the degree of business impact, n is the number of businesses, ω i is the weight of the ith business, t mi is the expected maximum delay time of the ith service during the migration process, t bi is the benchmark delay time for normal business operation; S is the path length; α is the weight coefficient of optical signal loss in the objective function; γ is the weight coefficient of business impact in the objective function; β is the weight coefficient of path length in the objective function;

[0106] Step 2: Perform iterative optimization, including:

[0107] In each iteration, a new connection path x′ is generated based on the current connection path x;

[0108] Calculate the optical signal loss corresponding to the new connection path x′ Business impact And the length S′ of the new connection path, and then calculate the objective function value E(x′)=αL′+γBI′+βS′;

[0109] According to the state transition probability formula Determine whether to accept the new path x′, if accepted, update the current connection path to the new connection path;

[0110] Update the formula according to the temperature T k+1 =λT k , update temperature T;

[0111] Among them, T k is the temperature of the kth iteration, and k ≥ 0; T k+1 is the temperature of the k+1th iteration; λ is the cooling coefficient, and 0<λ<1;

[0112] Step 3: Make termination judgment, including:

[0113] During each iteration, check whether the termination condition is met. The termination condition is that the number of iterations reaches the predetermined maximum number of iterations N. max ;

[0114] When the termination condition is met, the algorithm terminates and outputs the optimal new connection path x found as the optimal migration path.

[0115] Optionally, the connection path acquisition unit includes:

[0116] The topology data acquisition component is used to acquire the topology data of the data center from the first state information corresponding to the optical interconnection resources to be migrated.

[0117] The topology data analysis component is used to determine the current connection path of the optical interconnection resources to be migrated in the data center according to the topology data of the data center.

[0118] Optionally, the migration time window determining unit includes:

[0119] The historical network traffic data acquisition component is used to acquire the historical network traffic data from the first state information corresponding to the optical interconnection resource to be migrated.

[0120] The network traffic data prediction component is used to predict network traffic based on historical network traffic data using a trained deep reinforcement learning model to obtain network traffic prediction data.

[0121] The migration time window selection component is used to determine the time period with the minimum network traffic as the migration time window according to the network traffic prediction data.

[0122] Optionally, the network traffic data prediction component is also used to train a deep reinforcement learning model. The training process includes the following steps:

[0123] Obtain historical network traffic data from 30 days ago from historical network traffic data; the historical network traffic data from 30 days ago is used to train the deep reinforcement learning model, and the historical network traffic data within 30 days is used to input the trained deep reinforcement learning model to obtain network traffic prediction data;

[0124] Preprocess the historical network traffic data from 30 days ago to obtain a network traffic data set; divide the network traffic data set into a training set and a validation set;

[0125] Build deep reinforcement learning models;

[0126] Inputting the training set into the deep reinforcement learning model to train the deep reinforcement learning model;

[0127] The validation set is input into the trained deep reinforcement learning model to optimize the parameters of the deep reinforcement learning model, and finally a trained deep reinforcement learning model is obtained.

[0128] Optionally, the resource monitoring module 401 includes:

[0129] The optical interconnection resource state parameter acquisition component is used to acquire the optical link bit error rate in each optical interconnection resource and the optical signal strength corresponding to the connection path from the first state information.

[0130] The marking component is used to mark the optical interconnection resources whose optical link bit error rate is higher than a preset value or whose optical signal strength is not within a preset range.

[0131] The optical interconnection resource to be migrated determining component is used to take the marked optical interconnection resource as the optical interconnection resource to be migrated.

[0132] Optionally, the verification module 404 is specifically configured to perform the following operations, including:

[0133] Step 1: obtaining the second state information corresponding to the optical interconnection resource to be migrated after migration from the second state information; and obtaining the optical link bit error rate and optical signal strength of the optical interconnection resource to be migrated after migration as a verification result according to the second state information corresponding to the optical interconnection resource to be migrated after migration;

[0134] Step 2: If the bit error rate in the verification result is higher than a preset value and / or the optical signal strength is not within a preset range, the fault location is determined according to the verification result, and the fault is repaired;

[0135] Step 3: Repeat the above steps until the verification result meets the preset conditions; wherein the preset conditions are that the bit error rate in the verification result is lower than a preset value and the optical signal strength is within a preset range.

[0136] The embodiment of the present application also provides a data center optical interconnection resource management system, the system comprising: a data center, a sensor, a server, and an optical interconnection resource migration device;

[0137] The sensor is connected to the data center and the server, and is used to monitor the status of each optical interconnection resource in the data center in real time, and send first status information of each optical interconnection resource in the data center to the server;

[0138] The optical interconnection resource migration device is connected to the data center and the server, and is used to migrate the optical interconnection resources to be migrated according to the optical interconnection resource migration plan sent by the server;

[0139] The server is used to execute the data center optical interconnection resource migration method as described in any one of the above items.

[0140] Specifically, a data center is composed of a large number of data center servers, storage devices, network devices, etc. Optical interconnection resources include optical fiber cables, optical modules, optical switches, etc. Sensors can be optical fiber sensors or photoelectric detectors. Servers can be physical servers, virtual servers, or cloud servers. Optical interconnection resource migration devices can be optical switches, optical fiber plug-in devices, or optical module hot-swappable devices.

[0141] For the device embodiments, since they basically correspond to the method embodiments, the relevant parts can refer to the partial description of the method embodiments. The device embodiments described above are only schematic, wherein the components described as separate parts may or may not be physically separated, and the parts displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the disclosed solution. A person of ordinary skill in the art can understand and implement it without paying any creative work.

[0142] The above-mentioned embodiments only express several implementation methods of the embodiments of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the patent of the embodiments of the present application. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the embodiments of the present application, and these all belong to the protection scope of the embodiments of the present application.

Claims

1. A data center optical interconnection resource migration method, characterized in that: The method comprises: Acquire first status information of each optical interconnection resource in the data center, and determine the optical interconnection resource to be migrated based on the first status information; Formulate an optical interconnect resource migration plan according to the first state information corresponding to the optical interconnect resource to be migrated; the optical interconnect resource migration plan is used to instruct the optical interconnect resource migration device to migrate the optical interconnect resource to be migrated; The second state information of each of the optical interconnection resources after migration is acquired, and the optical interconnection resources to be migrated after migration in the data center are verified based on the second state information.

2. The method according to claim 1, characterized in that The step of formulating an optical interconnection resource migration plan according to the first state information corresponding to the optical interconnection resource to be migrated includes: Determining a current connection path of the optical interconnection resource to be migrated in the data center according to the first state information corresponding to the optical interconnection resource to be migrated; According to the current connection path of the optical interconnection resource to be migrated in the data center, taking the path length, service impact and optical signal loss as considerations, planning an optimal migration path; Determining a migration time window according to the first state information corresponding to the optical interconnection resource to be migrated; The migration time window and the migration path are used as an optical interconnection resource migration solution.

3. The method according to claim 2, characterized in that The planning of the optimal migration path according to the current connection path of the optical interconnection resource to be migrated in the data center, taking the path length, the service impact degree and the optical signal loss as considerations, includes: Step 1: Set the parameters of the simulated annealing algorithm, including setting the initial temperature T0, the cooling coefficient λ and the maximum number of iterations N max , and establish the objective function E(x) = αL + γBI + βS; Wherein, x represents the current connection path; Represents the optical signal loss, P in is the input optical power of the optical interconnection resource, P out is the output optical power of the optical interconnection resource; Indicates the degree of business impact, n is the number of businesses, ω i is the weight of the ith business, t mi is the expected maximum delay time of the ith service during the migration process, t bi is the benchmark delay time for normal operation of the service; S is the path length; α is the weight coefficient of optical signal loss in the objective function; γ is the weight coefficient of service impact in the objective function; β is the weight coefficient of path length in the objective function; Step 2: Perform iterative optimization, including: In each iteration, a new connection path x′ is generated based on the current connection path x; Calculate the optical signal loss corresponding to the new connection path x′ Business impact And the length S′ of the new connection path, and then calculate the objective function value E(x′)=αL′+γBI′+βS′; According to the state transition probability formula Determine whether to accept the new path x′, if accepted, update the current connection path to the new connection path; Update the formula according to the temperature T k+1 =λT k , update temperature T; Among them, T k is the temperature of the kth iteration, and k ≥ 0; T k+1 is the temperature of the k+1th iteration; λ is the cooling coefficient, and 0<λ<1; Step 3: Make termination judgment, including: During each iteration, check whether the termination condition is met, and the termination condition is that the number of iterations reaches the predetermined maximum number of iterations N. max ; When the termination condition is met, the algorithm terminates and outputs the found optimal new connection path x as the optimal migration path.

4. The method according to claim 2, characterized in that: The step of determining the current connection path of the optical interconnection resource to be migrated in the data center according to the first state information corresponding to the optical interconnection resource to be migrated includes: Acquire topological structure data of the data center from first state information corresponding to the optical interconnection resource to be migrated; According to the topological structure data of the data center, a current connection path of the optical interconnection resource to be migrated in the data center is determined.

5. The method according to claim 2, characterized in that: The step of determining the migration time window according to the first state information corresponding to the optical interconnection resource to be migrated includes: Acquiring historical network traffic data from first state information corresponding to the optical interconnection resource to be migrated; According to the historical network traffic data, the network traffic is predicted using the trained deep reinforcement learning model to obtain network traffic prediction data; According to the network traffic prediction data, a time period with minimum network traffic is determined as the migration time window.

6. The method according to claim 5, characterized in that The training process of the deep reinforcement learning model includes: Obtaining historical network traffic data 30 days ago from the historical network traffic data; wherein the historical network traffic data 30 days ago is used to train the deep reinforcement learning model, and the historical network traffic data within 30 days is used to input the trained deep reinforcement learning model to obtain the network traffic prediction data; Preprocessing the historical network traffic data from 30 days ago to obtain a network traffic data set; dividing the network traffic data set into a training set and a validation set; Constructing the deep reinforcement learning model; Inputting the training set into the deep reinforcement learning model to train the deep reinforcement learning model; The validation set is input into the trained deep reinforcement learning model to optimize the parameters of the deep reinforcement learning model, and finally obtain the trained deep reinforcement learning model.

7. The method according to claim 1, characterized in that The determining, based on the first state information, that optical interconnection resources need to be migrated includes: Acquire the optical link bit error rate and the optical signal strength corresponding to the connection path in each optical interconnection resource from the first state information; Marking the optical interconnection resources whose optical link bit error rate is higher than a preset value or whose optical signal strength is not within a preset range; The marked optical interconnection resources are used as the optical interconnection resources to be migrated.

8. The method according to claim 1, characterized in that The verifying the migrated optical interconnection resources to be migrated in the data center based on the second state information includes: Step 1: acquiring the second state information corresponding to the migrated optical interconnection resource to be migrated from the second state information; and acquiring the optical link bit error rate and optical signal strength of the migrated optical interconnection resource to be migrated as a verification result according to the second state information corresponding to the migrated optical interconnection resource to be migrated; Step 2: If the bit error rate in the verification result is higher than a preset value and / or the optical signal strength is not within a preset range, the fault location is determined according to the verification result, and the fault is repaired; Step 3: Repeat the above steps until the verification result meets the preset conditions; wherein the preset conditions are that the bit error rate in the verification result is lower than a preset value and the optical signal strength is within a preset range.

9. A data center optical interconnection resource migration device, characterized in that: The device comprises: A resource monitoring module, used for real-time monitoring of the status of each optical interconnection resource in the data center, obtaining first status information of each optical interconnection resource in the data center, and determining the optical interconnection resource to be migrated based on the first status information; A migration planning module, configured to formulate an optical interconnection resource migration plan according to the first state information corresponding to the optical interconnection resource to be migrated; A migration execution module, configured to migrate the optical interconnection resources to be migrated according to the optical interconnection resource migration plan; The verification module obtains the second status information of each of the optical interconnection resources after migration, and verifies the optical interconnection resources to be migrated after migration in the data center based on the second status information.

10. A data center optical interconnection resource management system, characterized in that: The system includes: a data center, a sensor, a server, and an optical interconnection resource migration device; The sensor is connected to the data center and the server, and is used for real-time monitoring of the status of each optical interconnection resource in the data center, and sending first status information of each optical interconnection resource in the data center to the server; The optical interconnection resource migration device is connected to the data center and the server, and is used to migrate the optical interconnection resources to be migrated according to the optical interconnection resource migration plan sent by the server; The server is used to execute the method according to any one of claims 1 to 8.

Images