Multi-link adaptive scheduling and transmission method, device, equipment and storage medium

Abstract

The embodiments of the present disclosure provide a multi-link adaptive scheduling and transmission method, device, equipment and storage medium. Applied to the technical field of wireless communication technology, the scheduling method includes: evaluating the health of each link according to the real-time performance indicators of each link, predicting the quality status of each link, and calculating the expected quality score of each link; dynamically adjusting the transmission weight of each link at the next moment according to the expected quality score; adaptively scheduling each link according to the transmission weight using a hierarchical aggregation algorithm and a load balancing strategy; the transmission method includes: generating data packets corresponding to each link level based on a dynamic hierarchical aggregation algorithm and a load balancing strategy; allocating data transmission tasks at the next moment according to the transmission weight of each link at the next moment; and preparing for parallel transmission based on the data transmission tasks at the next moment based on a multiplexing method. In this way, the utilization rate of bandwidth resources and the stability of the communication network can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of wireless communication technology, and in particular to a multi-link adaptive scheduling and transmission method, device, equipment and storage medium. Background Art

[0002] With the rapid development of wireless communication technology, the reliability and efficiency of data transmission are increasingly required. However, traditional communication management methods make it difficult to monitor link quality in real time, intelligently allocate bandwidth resources, and dynamically adjust link loads, resulting in poor system performance and low transmission efficiency. Therefore, an intelligent management and control method is urgently needed to improve bandwidth resource utilization and network stability. Summary of the invention

[0003] The present disclosure provides a multi-link adaptive scheduling and transmission method, device, equipment and storage medium, aiming to improve transmission efficiency and the stability and reliability of communication networks.

[0004] According to a first aspect of the present disclosure, a multi-link adaptive scheduling method is provided. The scheduling method is applied to a scheduling server, and the scheduling server includes a multi-link adaptive scheduling model. The scheduling method includes:

[0005] According to the real-time performance indicators of each link, the health of each link is evaluated; according to the health, the quality status of each link is predicted using a pre-trained quality status prediction model, and the expected quality score of each link is calculated based on the prediction result; according to the expected quality score, the transmission weight of each link at the next moment is dynamically adjusted; according to the transmission weight of each link at the next moment, the hierarchical aggregation algorithm and load balancing strategy are used to adaptively schedule each link.

[0006] In some implementations of the first aspect, adaptive scheduling is performed on each link according to the transmission weight of each link at the next moment using a hierarchical aggregation algorithm and a load balancing strategy, including:

[0007] Preset link transmission weight level and level threshold;

[0008] According to the layer threshold, the layer to which the transmission weight of each link at the next moment belongs is determined, and the links whose transmission weights are within the same layer threshold range are dynamically hierarchically aggregated to obtain multiple link layers;

[0009] Based on the load balancing strategy, the expected load of each link at the next moment is judged. If a link at a certain level is expected to be severely congested, while links at other levels are expected to be relatively idle, a relatively idle aggregate link is dynamically decomposed to release resources to the severely congested link layer.

[0010] In some implementations of the first aspect, dynamically decomposing a relatively idle aggregate link to release resources to a severely congested link layer includes:

[0011] Dynamically decompose a relatively idle aggregate link, retain a single link for data flow transmission at this level, and merge the remaining links into the severely congested link layer.

[0012] According to a second aspect of the present disclosure, a multi-link adaptive scheduling model is provided. The model is a deep learning model that integrates supervised learning, reinforcement learning and decision-making mechanism, wherein:

[0013] A supervised learning unit is used to evaluate the health of each link based on its real-time performance indicators;

[0014] The supervised learning unit is also used to predict the quality status of each link according to the health level using a pre-trained quality status prediction model, and calculate the link quality score according to the prediction result;

[0015] A reinforcement learning unit, used to dynamically adjust the transmission weight of each link at the next moment according to the expected quality score;

[0016] The decision unit is used to adaptively schedule each link according to the transmission weight of each link at the next moment by using a hierarchical aggregation algorithm and a load balancing strategy.

[0017] According to a third aspect of the present disclosure, a multi-link adaptive transmission method based on the multi-link adaptive scheduling method according to any one of the first aspects of the present disclosure is provided, which is applied to a transmission server, and the transmission method includes:

[0018] Based on the dynamic hierarchical aggregation algorithm and load balancing strategy, the data stream to be transmitted is allocated to the corresponding link layer according to the link layer and the preset data stream priority;

[0019] Grouping and packaging the data streams in each link layer to generate data packets corresponding to each link layer;

[0020] According to the transmission weight of each link at the next moment, the transmission allocation ratio of the data packets in each layer link is determined, and the data transmission task at the next moment is allocated according to the transmission allocation ratio;

[0021] Based on the multiplexing method, corresponding parallel transmission preparation is carried out according to the data transmission task at the next moment.

[0022] In some implementations of the third aspect, based on a dynamic hierarchical aggregation algorithm and a load balancing strategy, the data stream to be transmitted is allocated to a corresponding link layer according to the link layer and a preset data stream priority, including:

[0023] Based on the dynamic hierarchical aggregation algorithm and load balancing strategy, the priority of the data stream to be transmitted is determined, and the data stream with higher priority is preferentially allocated to the aggregation link of the higher layer;

[0024] If there are free bits in the higher-level aggregate link, some medium-priority data flows are allocated to the higher-level aggregate link; if there are no free bits in the higher-level aggregate link, medium-priority data flows are allocated to the medium-level aggregate link;

[0025] If there are idle bits in the middle-level aggregation link, some low-priority data flows are allocated to the middle-level aggregation link; if there are no idle bits in the middle-level aggregation link, low-priority data flows are allocated to the low-level aggregation link.

[0026] In some implementations of the third aspect, allocating the data transmission task at the next moment according to the transmission allocation ratio includes:

[0027] Preliminarily allocate the data transmission task at the next moment according to the transmission allocation ratio, and determine the expected load of each link at the next moment;

[0028] If the data transmission task of a certain layer link is heavier, while the links of other layers are relatively idle, a relatively idle aggregate link will be dynamically decomposed, the corresponding transmission allocation ratio will be readjusted, and resources will be coordinated to the link layer with heavier data transmission tasks.

[0029] According to a fourth aspect of the present disclosure, a multi-link adaptive transmission device based on the multi-link adaptive transmission method according to any one of the third aspects of the present disclosure is provided, which is applied to a transmission server, and the transmission device includes:

[0030] The first processing module is used to allocate the data stream to be transmitted to the corresponding link layer according to the link layer and the preset data stream priority based on the dynamic hierarchical aggregation algorithm and the load balancing strategy;

[0031] The first processing module is further used to group and pack the data streams in each link layer to generate data packets corresponding to each link layer;

[0032] The first processing module is further used to determine the transmission allocation ratio of data packets in each layer link according to the transmission weight of each link at the next moment, and allocate the data transmission task at the next moment according to the transmission allocation ratio;

[0033] The second processing module is used to perform corresponding parallel transmission preparation based on the multiplexing method according to the data transmission task at the next moment.

[0034] According to a fifth aspect of the present disclosure, an electronic device is provided. The electronic device includes: at least one processor; and a memory connected to the at least one processor in communication; the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the method described in the first aspect or the third aspect above.

[0035] According to a sixth aspect of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, where the computer instructions are used to cause a computer to execute the method described in the first aspect or the third aspect above.

[0036] In this disclosure:

[0037] Through the real-time health of each link, the quality status of each link is predicted and the quality score is further calculated. The transmission weight of each link in the future is adjusted according to the quality score, which helps to reasonably reallocate resources before link congestion or failure occurs to prevent potential congestion or failure, and make the scheduling method more proactive and flexible;

[0038] According to the transmission weight of each link at the future moment, on the basis of using the multiplexing method to realize parallel transmission, combined with the hierarchical aggregation algorithm and load balancing strategy, each link is adaptively scheduled, so that the data flow in each link is adaptively transmitted; through hierarchical aggregation, links with similar transmission weights are combined together to facilitate unified management and scheduling, and the load balancing strategy ensures the reasonable allocation of resources between the various levels, avoiding the situation where some links are overloaded while other links are idle. In this way, through dynamic aggregation and decomposition, reallocation of resources, and parallel transmission, bandwidth resources can be intelligently allocated, link load can be dynamically adjusted, data transmission efficiency can be improved, and the communication network can be ensured to operate stably in a good state.

[0039] It should be understood that the contents described in the summary of the invention are not intended to limit the key or important features of the embodiments of the present disclosure, nor are they intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] The above and other features, advantages and aspects of the embodiments of the present disclosure will become more apparent with reference to the following detailed description in conjunction with the accompanying drawings. The accompanying drawings are used to better understand the present solution and do not constitute a limitation of the present disclosure. In the accompanying drawings, the same or similar reference numerals represent the same or similar elements, among which:

[0041] Figure 1 A flowchart of a multi-link adaptive scheduling method provided by an embodiment of the present disclosure is shown;

[0042] Figure 2A schematic diagram of a multi-link adaptive scheduling method provided by an embodiment of the present disclosure is shown;

[0043] Figure 3 A schematic diagram of a multi-link adaptive scheduling model provided by an embodiment of the present disclosure is shown;

[0044] Figure 4 A flowchart of a multi-link adaptive transmission method provided by an embodiment of the present disclosure is shown;

[0045] Figure 5 A schematic diagram of a multi-link adaptive transmission device provided by an embodiment of the present disclosure is shown;

[0046] Figure 6 A block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the embodiments of the present disclosure clearer, the technical solution in the embodiments of the present disclosure will be clearly and completely described below in conjunction with the drawings in the embodiments of the present disclosure. Obviously, the described embodiments are part of the embodiments of the present disclosure, not all of the embodiments. Based on the embodiments in the present disclosure, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present disclosure.

[0048] In addition, the term "and / or" in this article is only a description of the association relationship between the associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship.

[0049] In response to the problems mentioned in the background technology, the present disclosure provides a multi-link adaptive scheduling and transmission method, apparatus, device and storage medium.

[0050] Specifically, the performance indicators of each link are monitored in real time, and the health of each link is evaluated based on the performance indicators; based on the health, the quality status of each link is predicted using a pre-trained quality status prediction model, and the expected quality score of each link is calculated based on the prediction result; the transmission weight of each link at the next moment is dynamically adjusted based on the expected quality score; based on the transmission weight of each link at the next moment, a hierarchical aggregation algorithm and a load balancing strategy are used to adaptively schedule each link.

[0051] Furthermore, based on the dynamic hierarchical aggregation algorithm and load balancing strategy, the data streams to be transmitted are allocated to the corresponding link layers according to the link layers and the preset data stream priorities; the data streams in each link layer are grouped and packaged to generate data packets corresponding to each link layer; according to the transmission weight of each link at the next moment, the transmission allocation ratio of the data packets in each level link is determined, and the data transmission tasks at the next moment are allocated according to the transmission allocation ratio; based on the multiplexing method, corresponding parallel transmission preparations are performed according to the data transmission tasks at the next moment.

[0052] In this way, potential congestion or failure can be prevented, and link scheduling and data transmission can be made more proactive and flexible; through dynamic aggregation and decomposition, reallocation of resources, and parallel transmission, bandwidth resources can be intelligently allocated, link load can be dynamically adjusted, and transmission efficiency can be improved, ensuring that the communication network operates stably in a good state.

[0053] The above contents are described in more detail below with reference to the accompanying drawings and specific embodiments.

[0054] According to a first aspect of the present disclosure, a multi-link adaptive scheduling method is provided. The scheduling method is applied to a scheduling server, and the scheduling server includes a multi-link adaptive scheduling model.

[0055] Figure 1 A flowchart of a multi-link adaptive scheduling method provided by an embodiment of the present disclosure is shown; Figure 1 As shown, the multi-link adaptive scheduling method 100 may include the following steps:

[0056] S110, evaluating the health of each link according to the real-time performance indicators of each link.

[0057] The performance indicators of the link include, but are not limited to: uplink bandwidth utilization, downlink bandwidth utilization, packet loss rate, delay, delay jitter, signal-to-noise ratio, and bit error rate.

[0058] The performance indicators of the link can be collected through distributed sensors and monitoring nodes and transmitted to the scheduling server. Based on the acquired data, the server calculates the health of the link by summing or weighted summing the data, and combines multiple performance indicators into a health value. The health value is used as the basis for training the quality status prediction model, which effectively reduces the number of original performance indicators that the model needs to process, reduces the computational complexity of the model, and improves the efficiency of the model.

[0059] S120, predicting the quality status of each link according to the health level using a pre-trained quality status prediction model, and calculating an estimated quality score of each link according to the prediction result.

[0060] Among them, the training steps of the quality status prediction model can be as follows:

[0061] 1. Obtain the link health at each historical moment, use the link health from tn to t as the sample set, and use the link health at t+n as the sample label to train the quality status prediction model; where n is a positive integer greater than or equal to 1.

[0062] 2. Use evaluation indicators to evaluate the model and optimize the model according to the evaluation results.

[0063] According to the data characteristics, the above quality status prediction model can be implemented by using a recurrent neural network or a long short-term memory network, etc. Here we take a simple recurrent neural network model as an example.

[0064] Select a recurrent neural network model with a multi-layer LSTM network structure and initialize the weights and bias items of the model.

[0065] Assuming that the link health data from time tn to time t has been obtained, the historical time window size n=10 is determined, and the data of the past 10 time points are used to predict the link health at the next time point.

[0066] Take the link health data from time t-10 to time t from the data set and construct the input sequence x of the sample set t -10,x t -11, ..., x t ; The link health at time t+1 is taken as the output, that is, the sample label is y = {x t+1}; Repeat this process until the end of the data set, generate multiple input-output pairs (i.e., samples), and further generate a sample set X = {x t-1 ,x t}, where x t Indicates the link health at time t.

[0067] Using the sample set X = {x t-1 , x t}Train the model and update the model's weights and biases through back propagation.

[0068] Based on historical data, use gradient descent and choose mean squared error (MSE) or other evaluation methods to estimate the parameters of the model.

[0069] If the model performance does not meet the requirements after evaluation, the model can be tuned in the following ways:

[0070] Adjust model parameters and change parameters such as regularization strength and learning rate in the linear regression model.

[0071] Increase the input parameters and input the health values ​​of more historical moments into the model as features.

[0072] Repeat the above training and evaluation process until a model that meets the standards is obtained.

[0073] As another example, using the above sample set and sample labels, a linear regression quality status prediction model can also be trained. The model can be expressed as:

[0074] ;

[0075] Among them, θ0, θ1, and θ2 are model parameters.

[0076] Based on historical data, the gradient descent method is used to estimate the parameters of the model by minimizing the mean square error (MSE) between the predicted value and the actual value. If the performance of the model does not meet the requirements after evaluation, the model is tuned in the following ways:

[0077] Adjust model parameters and change parameters such as regularization strength and learning rate in the linear regression model.

[0078] Increase the input parameters and input the health values ​​of more historical moments into the model as features.

[0079] Repeat the above training and evaluation process until a model that meets the standards is obtained.

[0080] S130, dynamically adjusting the transmission weight of each link at the next moment according to the estimated quality score.

[0081] Specifically, a weight function is defined, and the predicted quality score is mapped to a specific transmission weight using the weight function, so that the transmission weight of each link at the next moment is dynamically adjusted according to the predicted quality score.

[0082] According to the embodiments of the present disclosure, the link weight is adjusted in advance using the real-time link performance indicator so as to further adjust the resource allocation, thereby ensuring the efficient use of bandwidth resources.

[0083] Exemplarily, the expected quality score is defined to range from 0-1 and the transmission weight is defined to range from 0-10.

[0084] Next, we define a weight function using the Sigmoid function to map the expected quality score to the transmission weight. Here, we adjust the output of the Sigmoid function by a scaling factor (e.g. 10) and a possible offset to match the range of the transmission weight, then:

[0085] def sigmoid(x, scale=10, shift=0):

[0086] # Sigmoid function

[0087] # x: Estimated quality score of the input

[0088] # scale: scaling factor, adjust the output range of the Sigmoid function

[0089] # shift: offset, adjust the center position of the Sigmoid function

[0090] return scale / (1 + np.exp(-x + shift))

[0091] def map_quality_to_weight(quality_score):

[0092] # Weight function, mapping the expected quality score to the transmission weight

[0093] # quality_score: Estimated quality score, ranging from 0 to 1

[0094] # Returns the transmission weight, ranging from 0 to 10

[0095] return sigmoid(quality_score, scale=10)

[0096] # Example: Using the weight function to calculate the transmission weight

[0097] predicted_quality_scores = np.array([0.1, 0.3, 0.5, 0.7, 0.9]) # predicted quality scores for the examples

[0098] transmission_weights = [map_quality_to_weight(score) for score inpredicted_quality_scores]

[0099] # Print results

[0100] for quality, weight in zip(predicted_quality_scores, transmission_weights):

[0101] print(f"Expected quality score: {quality:.2f}, transfer weight: {weight:.2f}")

[0102] In the above code example, the expected quality score is mapped to a transmission weight between 0 and 10 through the sigmoid function. In this code example, the range of the transmission weight can be changed by adjusting the scale parameter, or the center position of the function curve can be changed by adjusting the shift parameter.

[0103] For the sake of simplicity, this example does not adjust the shift parameter to change the center position of the function curve. In actual applications, the center position of the function curve can be further adjusted by adjusting the shift parameter to achieve refined control of the weight mapping.

[0104] Furthermore, a weight table can be defined to dynamically adjust the transmission weight of each link at the next moment. Assuming that the range of the expected quality score is 0-1 and the range of the transmission weight is 0-10, then:

[0105]

[0106] Calculate the corresponding transmission weight according to the weight table.

[0107] S140, adaptively schedule each link according to the transmission weight of each link at the next moment using a layered aggregation algorithm and a load balancing strategy. Specifically:

[0108] Preset link transmission weight levels and level thresholds; determine the level to which the transmission weight of each link at the next moment belongs based on the level threshold, dynamically aggregate links whose transmission weights are within the same level threshold range, and obtain multiple link levels.

[0109] Based on the load balancing strategy, the expected load of each link at the next moment is determined. If a link at a certain level is expected to be severely congested, while links at other levels are expected to be relatively idle, a relatively idle aggregate link is dynamically decomposed to release resources to the severely congested link layer. The load balancing strategy includes but is not limited to: minimum load priority, polling, and weighted polling strategy.

[0110] Furthermore, a relatively idle aggregate link is dynamically decomposed to release resources to the severely congested link layer, that is, a relatively idle aggregate link is dynamically decomposed, a single link is reserved for data flow transmission at this level, and the remaining links are merged into the severely congested link layer.

[0111] If the links at each level are expected to be load balanced, the data stream is transmitted normally according to the link level.

[0112] For example:

[0113] 1. Define link transmission weight levels and level thresholds

[0114] Suppose there are N links, and the transmission weight of each link is w i , where i is the index of the link, i = 1, 2, ..., N, and w i The range is [0, X]; M levels are preset, and the threshold range of each level is [T min_j , T max_j ], where j is the index of the level, j = 1, 2, ..., M.

[0115] 2. Dynamic hierarchical aggregation

[0116] Determine the link level: For each link i, according to its transmission weight w i Determine the level to which it belongs. If T min_j ≤w i < T max_j , then link i belongs to level j.

[0117] Among them, the layer where the link with higher transmission weight is located is the high layer, which is used to transmit data streams with higher priority; the layer where the link with average transmission weight is located is the medium layer, which is used to transmit data streams with medium priority; the layer where the link with lower transmission weight is located is the low layer, which is used to transmit data streams with lower priority.

[0118] Aggregate links: Aggregate links whose transmission weights are within the same layer threshold range to form multiple link layers, each of which contains a group of links with similar weights.

[0119] 3. Load balancing strategy

[0120] Estimated load situation: Get the data stream to be transmitted, determine the corresponding aggregate link level according to the priority of the data stream to be transmitted, and further determine the data stream to be transmitted that can be allocated to each link in each level, so as to determine the estimated load situation of each level and each link in the level at the next moment. That is, according to the priority of the data stream to be transmitted and the corresponding aggregate link level, determine the estimated load situation of each level and each link in each level at the next moment.

[0121] Determine congestion: For each link level, calculate its average expected load L avg_j , if L of a certain level j avg_j Exceeding the preset congestion threshold L threshold , then the level is considered to be severely congested.

[0122] Dynamic resource adjustment: If a link at a certain level is expected to be severely congested, while links at other levels are expected to be relatively idle, then an idle level with the lowest expected load is selected, assuming it is k.

[0123] Select a link i from level K as a reserved link and distribute the other links equally to one or more congested link layers; or remove link i from the current level and make it a new level separately, merging level K into the congested link layer.

[0124] Figure 2 A schematic diagram of a multi-link adaptive scheduling method provided by an embodiment of the present disclosure is shown; Figure 2 As shown in the figure, in a complex network environment, the performance indicators of each link are monitored and analyzed in real time to comprehensively evaluate the health of the link; once sufficient link health data is collected, the pre-trained quality status prediction model is used to accurately predict the quality status of each link, and based on the prediction results, the expected quality score of each link is calculated; after obtaining the expected quality score, the transmission weight of each link at the next moment is dynamically adjusted; this step is crucial because it directly affects the efficiency of data transmission and load balancing; further, the hierarchical aggregation algorithm and load balancing strategy are used to achieve adaptive scheduling of each link.

[0125] According to a second aspect of the present disclosure, a multi-link adaptive scheduling model is provided, which is a deep learning model that integrates supervised learning, reinforcement learning, and decision-making mechanisms.

[0126] Figure 3 A schematic diagram of a multi-link adaptive scheduling model provided by an embodiment of the present disclosure is shown; Figure 3 As shown, the multi-link adaptive scheduling model 300 may include the following units:

[0127] The supervised learning unit 310 is used to evaluate the health of each link according to the real-time performance indicators of each link.

[0128] The supervised learning unit 310 is further configured to predict the quality status of each link according to the health level using a pre-trained quality status prediction model, and calculate a link quality score according to the prediction result.

[0129] The reinforcement learning unit 320 is used to dynamically adjust the transmission weight of each link at the next moment according to the expected quality score.

[0130] The decision unit 330 is used to perform adaptive scheduling on each link according to the transmission weight of each link at the next moment by using a layered aggregation algorithm and a load balancing strategy.

[0131] Among them, the decision unit 330 is also used to preset the link transmission weight level and level threshold; according to the level threshold, the level to which the transmission weight of each link at the next moment belongs is determined, and the links whose transmission weights are within the same level threshold range are dynamically layered and aggregated to obtain multiple link levels.

[0132] Based on the load balancing strategy, the expected load of each link at the next moment is judged. If a link at a certain level is expected to be severely congested, while links at other levels are expected to be relatively idle, a relatively idle aggregate link is dynamically decomposed to release resources to the severely congested link layer.

[0133] Furthermore, the decision unit 330 is also used to dynamically decompose a relatively idle aggregate link to release resources to the severely congested link layer, that is, to dynamically decompose a relatively idle aggregate link, retain a single link therein for data flow transmission at this level, and merge the remaining links into the severely congested link layer.

[0134] Understandably, Figure 3 Each unit in the multi-link adaptive scheduling model 300 shown has the function of implementing each step in the multi-link adaptive transmission method 100 provided in the embodiment of the present disclosure, and can achieve its corresponding technical effect. The unit working process of the described module can refer to the corresponding process in the aforementioned method embodiment. For the convenience and conciseness of the description, it will not be repeated here.

[0135] According to a third aspect of the present disclosure, a multi-link adaptive transmission method based on the multi-link adaptive scheduling method described in any one of the first aspects of the present disclosure is provided, which is applied to a transmission server.

[0136] Figure 4 A flowchart of a multi-link adaptive transmission method provided by an embodiment of the present disclosure is shown; Figure 4 As shown, the multi-link adaptive transmission method 400 may include the following steps:

[0137] S410, based on a dynamic hierarchical aggregation algorithm and a load balancing strategy, the data streams to be transmitted are allocated to corresponding link levels according to the link levels and preset data stream priorities.

[0138] Specifically, based on the dynamic hierarchical aggregation algorithm and load balancing strategy, the priority of the data stream to be transmitted is determined, and the data stream with higher priority is preferentially allocated to the aggregation link of the higher layer;

[0139] If there are free bits in the higher-level aggregate link, some medium-priority data flows are allocated to the higher-level aggregate link; if there are no free bits in the higher-level aggregate link, medium-priority data flows are allocated to the medium-level aggregate link;

[0140] If there are idle bits in the middle-level aggregation link, some low-priority data flows are allocated to the middle-level aggregation link; if there are no idle bits in the middle-level aggregation link, low-priority data flows are allocated to the low-level aggregation link.

[0141] Furthermore, after the data packets are distributed to each level, the number of data flows to be transmitted in each link can be further divided according to the weight of the link in each level.

[0142] S420, grouping and packaging the data streams in each link layer to generate data packets corresponding to each link layer.

[0143] S430, determining the transmission allocation ratio of data packets in each layer link according to the transmission weight of each link at the next moment, and allocating the data transmission task at the next moment according to the transmission allocation ratio.

[0144] Specifically, according to the transmission weight of each link at the next moment, the transmission allocation ratio of the data packets in each level of link is determined, the data transmission task at the next moment is preliminarily allocated according to the transmission allocation ratio, and the expected load of each link at the next moment is judged; if the data transmission task of a certain level of link is heavier, while other levels of links are relatively idle, a relatively idle aggregate link is dynamically decomposed, the corresponding transmission allocation ratio is readjusted, and resources are coordinated to the link layer with the heavier data transmission task.

[0145] S440, based on the multiplexing method, according to the data transmission task at the next moment, corresponding parallel transmission preparation is performed.

[0146] Among them, the multiplexing method can be implemented using multiplexing technology or other network protocols. The following are some network protocols that can implement multiplexing:

[0147] 1. TCP protocol: TCP connections can be multiplexed by multiple application layer processes. By using sockets and port numbers, multiple application layer processes can communicate with the remote host at the same time on the same host without having to establish a separate physical connection for each process. This is actually a logical multiplexing.

[0148] 2. UDP protocol: The UDP protocol can assign a unique port number to each application layer process, and use this port number to implement data multiplexing and sharing. Since UDP is connectionless, it allows multiple application layer processes to send and receive data at the same time without having to establish and maintain a connection status like TCP.

[0149] 3. IP protocol: The IP protocol can use addresses and routing mechanisms to enable multiple different data streams to be transmitted on the same physical network, thereby achieving multiplexing at the network layer.

[0150] 4. MPLS protocol: MPLS protocol is a protocol for making forwarding decisions on IP networks. It uses short labels to identify data streams, allowing fast data forwarding between routers. MPLS supports the simultaneous transmission of multiple different data streams on a single physical link through the label stack mechanism, realizing network layer multiplexing.

[0151] According to the embodiments of the present disclosure, on the basis of realizing parallel transmission through multiplexing, combined with the layered aggregation algorithm and load balancing strategy, according to the link characteristics and the importance of the data flow, the transmission quality of the key data flow is prioritized while taking into account the overall resource utilization; the overall resource utilization is improved to ensure the efficient operation and stable transmission of the communication network.

[0152] It should be noted that the dynamic link layer aggregation based on the multiplexing method can be implemented on the basis of the above multiplexing method in combination with the LACP protocol (Link Aggregation Control Protocol) or other similar protocols.

[0153] The multi-link adaptive scheduling method 100 and the multi-link adaptive transmission method 400 are described in more detail below using a specific embodiment.

[0154] Suppose there are N = 10 links and the transmission weight of each link is w i (where i is the index of the link, i = 1, 2, ..., 10) is in the range of [0, 10]; M = 3 levels are preset, and the threshold range of each level is as follows:

[0155] First level (j = 1): T min1 = 0, T max1 = 3 (including links with weights in the range [0, 3)); Second level (j = 2): T min2 = 3, T max2 = 7 (including links with weights in the range [3, 7)); Third level (j = 3): T min3 = 7, T max3 = 10 (including links with weights in the range [7, 10]).

[0156] Among them, the third layer is a high layer, which is used to transmit data streams with higher priorities; the second layer is a medium layer, which is used to transmit data streams with medium priorities; and the third layer is a low layer, which is used to transmit data streams with lower priorities.

[0157] For each link i, check its transmission weight w i , and assign them to the corresponding levels:

[0158] Assume w1 = 1, w2 = 2, w3 = 1, then links 1, 2, and 3 belong to the first level (j = 1); Assume w4 = 6, w5 = 4, w6 = 5, and w7 = 6, then links 4, 5, 6, and 7 belong to the second level (j = 2); Assume w8 = 8, w9 = 8, w10 = 11, w11 = 12, w12 = 13, w13 = 14, w14 = 15, w15 = 16, w16 = 17, w17 = 18, w18 = 19, w19 = 20, w20 = 21, w20 = 22, w20 = 23, w20 = 24, w20 = 25 10 = 9, then links 8, 9, and 10 belong to the third level (j = 3).

[0159] Aggregate the links within the same level threshold range and calculate the average expected load L for each link level. avg_j , let the congestion threshold L threshold = 85%.

[0160] Obtain the data stream to be transmitted, determine the corresponding aggregation link layer according to the priority of the data stream to be transmitted, and further determine the data stream to be transmitted that can be allocated to each link in each layer, so as to judge the expected load situation of each layer and each link in the layer at the next moment; among which, the priority of the data stream to be transmitted can be set according to actual business needs.

[0161] Assume that there are currently three groups of data streams to be transmitted, with priorities of high, medium, and low, corresponding to the third, second, and first layers respectively.

[0162] High-level data flow D high : The size is 100GB and needs to be transmitted through the third-level link;

[0163] Medium-level data flow D medium : The size is 60GB and needs to be transmitted through the second-level link;

[0164] Low-level data flow D low : The size is 30GB and needs to be transmitted through the first-level link.

[0165] These data streams are grouped and packaged:

[0166] D high Divided into 10 data packets, each of size 10GB;

[0167] D medium Divided into 6 data packets, each of size 10GB;

[0168] D low Divided into 3 data packs, each 10GB in size.

[0169] According to the transmission weight of each link at the next moment, the transmission allocation ratio of the data packets in each layer link is determined, which can be specifically as follows:

[0170] D highIt is divided into 10 data packets and transmitted through the third-level link. The third-level link weights are: w8 = 8, w9 = 8, w 10 = 9, then w8 and w9 are allocated 3 packets each, which is w 10 Allocate 4 packets.

[0171] D medium It is divided into 6 data packets and transmitted through the second-level link. The second-level link weights are: w4 = 6, w5 = 4, w6 = 5, w7 = 6, then 2 data packets are allocated to w4 and w7 respectively, and 1 data packet is allocated to w5 and w6 respectively.

[0172] D low It is divided into 3 data packets and transmitted through the first-level link. The link weights of the first level are: w1 = 1, w2 = 2, w3 = 1, then 1 data packet is allocated to each of w1, w2, and w3.

[0173] As another example, according to the transmission weight of each link at the next moment, the transmission allocation ratio of the data packets in each layer link is determined, which may be specifically as follows:

[0174] The total number of data packets to be transmitted is 19.

[0175] The sum of the link weights at the third level is 25 (8+8+9), at the second level it is 21 (6+4+5+6), and at the first level it is 4 (1+2+1).

[0176] The allocation ratio of the third layer = 25 / (25 + 21 + 4) = 0.5; 19 × 0.5 = 9.5. After rounding, the third layer can be allocated 10 data packets, and D high The 10 packets are assigned to the third layer. The link weights of the third layer are: w8 = 8, w9 = 8, w 10 = 9, then 3 packets are allocated to w8 and w9 respectively, which is w 10 Allocate 4 packets.

[0177] The allocation ratio of the second level = 21 / (25 + 21 + 4) = 0.42; 19 × 0.42 = 7.98. After rounding, the second level can be allocated 8 data packets, so D medium 6 packets and D low The 2 data packets are allocated to the second layer. The link weights of the second layer are: w4 = 6, w5 = 4, w6 = 5, w7 = 6, so 3 data packets are allocated to w4 and w7 respectively, and 1 data packet is allocated to w5 and w6 respectively.

[0178] The allocation ratio of the first level = 4 / (25 + 21 + 4) = 0.08; 19 × 0.08 = 1.52. After rounding, the first level can be allocated 2 data packets, so D low The 1 data packet is allocated to w2 of the first layer, and w1 and w3 of the first layer are idle and not transmitted.

[0179] Furthermore, suppose the prediction result shows that there are 12 packets to be transmitted at the second level, and each link is expected to be severely congested, while there is only 1 packet to be transmitted at the first level, which is relatively idle. At this time, dynamic adjustment is performed:

[0180] Select link w1 from the first level as the reserved link for transmission D low The other links are incorporated into the second layer, and the links in the second layer are updated to: w2 = 2, w3 = 1, w4 = 6, w5 = 4, w6 = 5, w7 = 6, which are used to transmit 12 D medium Data packet.

[0181] The recalculated second-level transmission allocation ratio is as follows:

[0182] w2 and w3 are each allocated 1 data packet, w5 and w6 are each allocated 2 data packets, and w4 and w7 are each allocated 3 data packets.

[0183] After the data packets to be transmitted are allocated, further, based on the multiplexing method, parallel transmission preparation is performed according to the adjusted data transmission tasks.

[0184] According to the embodiments of the present disclosure, on the basis of parallel transmission using the multiplexing method, combined with the hierarchical aggregation algorithm and the load balancing strategy, the routing strategy is dynamically adjusted according to the real-time load situation and resource allocation situation of the link to ensure that the data flow can be more evenly distributed on the links at each level, thereby realizing multi-link adaptive scheduling and transmission and optimizing the performance of the communication network.

[0185] According to the embodiments of the present disclosure, the following technical effects are achieved:

[0186] 1. Based on the link characteristics and the importance of data flows, priority is given to ensuring the transmission quality of key data flows while taking into account overall resource utilization; improving overall resource utilization to ensure efficient operation and stable transmission of the network.

[0187] 2. Prevent potential congestion or failures, and make link scheduling and data transmission more proactive and flexible; through dynamic aggregation and decomposition, reallocation of resources, and parallel transmission, it is possible to intelligently allocate bandwidth resources, dynamically adjust link loads, and improve transmission efficiency, ensuring that the communication network operates stably in a good state.

[0188] 3. Under the framework of the deep learning model, multi-link adaptive scheduling and transmission are performed based on the multiplexing method combined with the layered aggregation algorithm and load balancing strategy to maximize the use of multiplexing technology, improve network throughput and reliability, and meet the needs of efficient data transmission in complex network environments.

[0189] 4. Use deep reinforcement learning technology to adjust the multi-link aggregation strategy according to the dynamic changes of the environment and optimize data transmission efficiency; improve the adaptability and effectiveness of the system to ensure stable operation and efficient data transmission in complex network environments.

[0190] It should be noted that, for the aforementioned method embodiments, for the sake of simplicity, they are all described as a series of action combinations, but those skilled in the art should be aware that the present disclosure is not limited by the order of the actions described, because according to the present disclosure, certain steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all optional embodiments, and the actions and modules involved are not necessarily required by the present disclosure.

[0191] The above is an introduction to a method embodiment. The following is a further explanation of the solution described in the third aspect of the present disclosure through an apparatus embodiment.

[0192] Figure 5 A schematic diagram of a multi-link adaptive transmission device provided by an embodiment of the present disclosure is shown; Figure 5 As shown, the multi-link adaptive transmission device 500 may include:

[0193] The first processing module 510 is used to allocate the data stream to be transmitted to the corresponding link layer according to the link layer and the preset data stream priority based on the dynamic hierarchical aggregation algorithm and the load balancing strategy;

[0194] The first processing module 510 is further used to group and package the data streams in each link layer to generate data packets corresponding to each link layer;

[0195] The first processing module 510 is further used to determine the transmission allocation ratio of the data packets in each layer link according to the transmission weight of each link at the next moment, and allocate the data transmission task at the next moment according to the transmission allocation ratio;

[0196] The second processing module 520 is used to perform corresponding parallel transmission preparation based on the data transmission task at the next moment based on the multiplexing method.

[0197] Understandably, Figure 5Each module in the multi-link adaptive transmission device 500 shown has the function of implementing each step in the multi-link adaptive transmission method 300 provided in the embodiment of the present disclosure, and can achieve its corresponding technical effect. The specific working process of the described module can refer to the corresponding process in the aforementioned method embodiment. For the convenience and conciseness of the description, it will not be repeated here.

[0198] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device and a readable storage medium.

[0199] Figure 6 A block diagram of an exemplary electronic device capable of implementing embodiments of the present disclosure is shown.

[0200] The electronic device 600 is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or claimed herein.

[0201] The electronic device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a ROM 602 or a computer program loaded from a storage unit 608 into a RAM 603. In the RAM 603, various programs and data required for the operation of the electronic device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An I / O interface 605 is also connected to the bus 604.

[0202] Multiple components in the electronic device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the electronic device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0203] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as method 100 or method 300. For example, in some embodiments, method 100 or method 300 may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the electronic device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the method 100 or method 300 described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to execute the method 100 or the method 300 in any other appropriate manner (eg, by means of firmware).

[0204] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0205] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0206] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or device, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0207] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0208] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., a scheduling server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0209] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship of client and server is generated by computer programs running on respective computers and having a client-server relationship with each other. The server may be a cloud server, a server of a distributed system, or a server combined with a blockchain.

[0210] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0211] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A multi-link adaptive scheduling method, characterized in that: The scheduling method is applied to a scheduling server, the scheduling server includes a multi-link adaptive scheduling model, and the scheduling method includes: Evaluate the health of each link based on its real-time performance indicators; According to the health, a pre-trained quality status prediction model is used to predict the quality status of each link, and an estimated quality score of each link is calculated according to the prediction result; Dynamically adjust the transmission weight of each link at the next moment according to the expected quality score; According to the transmission weight of each link at the next moment, each link is adaptively scheduled using the hierarchical aggregation algorithm and load balancing strategy; including: Preset link transmission weight level and level threshold; According to the layer threshold, the layer to which the transmission weight of each link at the next moment belongs is determined, and the links whose transmission weights are within the same layer threshold range are dynamically hierarchically aggregated to obtain multiple link layers; Based on the load balancing strategy, the expected load of each link at the next moment is judged. If a link at a certain level is expected to be severely congested, while links at other levels are expected to be relatively idle, a relatively idle aggregate link is dynamically decomposed to release resources to the severely congested link layer.

2. The method according to claim 1, characterized in that The method of dynamically decomposing a relatively idle aggregate link to release resources to a severely congested link layer includes: Dynamically decompose a relatively idle aggregate link, retain a single link for data flow transmission at this level, and merge the remaining links into the severely congested link layer.

3. A multi-link adaptive scheduling device, characterized in that: Applied to a scheduling server, the scheduling server includes a multi-link adaptive scheduling model, the device includes a supervised learning unit, a reinforcement learning unit and a decision unit, wherein: A supervised learning unit, used to monitor the performance indicators of each link in real time and evaluate the health of each link based on the performance indicators; The supervised learning unit is further used to predict the quality status of each link according to the health level using a pre-trained quality status prediction model, and calculate the estimated quality score of each link according to the prediction result; A reinforcement learning unit, used for dynamically adjusting the transmission weight of each link at the next moment according to the expected quality score; The decision unit is used to adaptively schedule each link according to the transmission weight of each link at the next moment by using the hierarchical aggregation algorithm and load balancing strategy; it includes: Preset link transmission weight level and level threshold; According to the layer threshold, the layer to which the transmission weight of each link at the next moment belongs is determined, and the links whose transmission weights are within the same layer threshold range are dynamically hierarchically aggregated to obtain multiple link layers; Based on the load balancing strategy, the expected load of each link at the next moment is judged. If a link at a certain level is expected to be severely congested, while links at other levels are expected to be relatively idle, a relatively idle aggregate link is dynamically decomposed to release resources to the severely congested link layer.

4. A multi-link adaptive transmission method based on the multi-link adaptive scheduling method according to any one of claims 1 to 2, applied to a transmission server, characterized in that: The transmission method comprises: Based on the dynamic hierarchical aggregation algorithm and load balancing strategy, the data stream to be transmitted is allocated to the corresponding link layer according to the link layer and the preset data stream priority; including: Determine the priority of the data stream to be transmitted, and preferentially allocate the data stream with higher priority to the aggregated link of higher layer; if the aggregated link of higher layer has idle bits, allocate some data streams with medium priority to the aggregated link of higher layer; if the aggregated link of higher layer has no idle bits, allocate the data streams with medium priority to the aggregated link of medium layer; if the aggregated link of medium layer has idle bits, allocate some data streams with lower priority to the aggregated link of medium layer; if the aggregated link of medium layer has no idle bits, allocate the data streams with lower priority to the aggregated link of lower layer; Grouping and packaging the data streams in each link layer to generate data packets corresponding to each link layer; Determine the transmission allocation ratio of data packets in each layer link according to the transmission weight of each link at the next moment, and allocate the data transmission task at the next moment according to the transmission allocation ratio; Based on the multiplexing method, corresponding parallel transmission preparation is carried out according to the data transmission task at the next moment.

5. The method according to claim 4, characterized in that The allocating the data transmission task at the next moment according to the transmission allocation ratio includes: Preliminarily allocate the data transmission task at the next moment according to the transmission allocation ratio, and determine the expected load of each link at the next moment; If the data transmission task of a certain layer link is heavier, while the links of other layers are relatively idle, a relatively idle aggregate link will be dynamically decomposed, the corresponding transmission allocation ratio will be readjusted, and resources will be coordinated to the link layer with heavier data transmission tasks.

6. A multi-link adaptive transmission device based on the multi-link adaptive transmission method according to any one of claims 4 to 5, applied to a transmission server, characterized in that: The transmission device comprises: The first processing module is used to allocate the data stream to be transmitted to the corresponding link layer according to the link layer and the preset data stream priority based on the dynamic hierarchical aggregation algorithm and the load balancing strategy; including: Determine the priority of the data stream to be transmitted, and preferentially allocate the data stream with higher priority to the aggregated link of higher layer; if the aggregated link of higher layer has idle bits, allocate some data streams with medium priority to the aggregated link of higher layer; if the aggregated link of higher layer has no idle bits, allocate the data streams with medium priority to the aggregated link of medium layer; if the aggregated link of medium layer has idle bits, allocate some data streams with lower priority to the aggregated link of medium layer; if the aggregated link of medium layer has no idle bits, allocate the data streams with lower priority to the aggregated link of lower layer; The first processing module is further used to group and pack the data streams in each link layer to generate data packets corresponding to each link layer; The first processing module is further used to determine the transmission allocation ratio of the data packets in each layer link according to the transmission weight of each link at the next moment, and allocate the data transmission task at the next moment according to the transmission allocation ratio; The second processing module is used to perform corresponding parallel transmission preparation based on the multiplexing method according to the data transmission task at the next moment.

7. An electronic device, characterized in that: include: at least one processor; as well as a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the method of any one of claims 1-2 or claims 4-5.

8. A non-transitory computer-readable storage medium storing computer instructions, characterized in that: The computer instructions are used to cause the computer to execute the method according to any one of claims 1-2 or claims 4-5.

Images