Link communication method and device, equipment, storage medium and program product

Abstract

The invention provides a link communication method and device, equipment, a storage medium and a program product, and relates to the technical field of data processing, and the method comprises the steps: obtaining an initial attribute vector of each replication node in a plurality of replication nodes in a replication communication network, the initial attribute vector is used for representing a communication condition and / or a resource use condition of a corresponding replication node, and each replication node is obtained by replication through an initial node in an initial communication network; based on the initial attribute vectors of the multiple replication nodes, determining an initial node corresponding to each replication node; and respectively updating the initial attribute vector of each replication node based on the first attribute vectors of the initial nodes corresponding to the plurality of replication nodes. According to the invention, the link segmentation multiplexing effect can be improved.

Description

Technical Field

[0001] This invention relates to the field of data processing technology, specifically to a link communication method, apparatus, device, storage medium, and program product. Background Technology

[0002] Link segmentation and multiplexing is a key technology in the Internet of Things (IoT) field, which can improve link utilization. In related technologies, nodes in the communication network are typically copied directly, and the copied nodes are configured with the same parameters as the original nodes to achieve link segmentation and multiplexing. However, in these technologies, different nodes have different service types and quality of service requirements, and different users have different needs for the links. For example, some users require high bandwidth and low latency, while others have higher requirements for link reliability. Simply copying the node parameters directly is insufficient to meet actual needs, resulting in poor link segmentation and multiplexing performance.

[0003] It is evident that the related technologies suffer from poor performance in link segmentation and multiplexing. Summary of the Invention

[0004] This invention provides a link communication method, apparatus, device, storage medium, and program product to solve the problem of poor link segmentation multiplexing in related technologies.

[0005] To solve the above problems, the present invention is implemented as follows:

[0006] In a first aspect, embodiments of the present invention provide a link communication method, including:

[0007] Obtain the initial attribute vector of each replica node in the replication communication network. The initial attribute vector is used to characterize the communication status and / or resource usage of the corresponding replica node. Each replica node is replicated from the initial node in the initial communication network.

[0008] The initial node corresponding to each of the multiple replica nodes is determined based on the initial attribute vectors of the replica nodes.

[0009] The initial attribute vector of each of the multiple replica nodes is updated based on the first attribute vector of the initial node corresponding to the initial node.

[0010] Secondly, embodiments of the present invention also provide a link communication device, comprising:

[0011] The first acquisition module is used to acquire the initial attribute vector of each replica node among multiple replica nodes in the replica communication network. The initial attribute vector is used to characterize the communication status and / or resource usage of the corresponding replica node. Each replica node is obtained by replicating an initial node in the initial communication network. The replica communication network is used for link segmentation and multiplexing.

[0012] The first determining module is used to determine the initial node corresponding to each of the plurality of replicated nodes based on the initial attribute vectors of the replicated nodes.

[0013] The update module is used to update the initial attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes, and each replica node is used to perform link communication based on the parameters corresponding to the updated attribute vector.

[0014] Thirdly, embodiments of the present invention also provide an electronic device, including a transceiver and a processor.

[0015] The transceiver is used to obtain the initial attribute vector of each of the multiple replica nodes in the replica communication network. The initial attribute vector is used to characterize the communication status and / or resource usage of the corresponding replica node. Each replica node is replicated from the initial node in the initial communication network. The replica communication network is used for link segmentation and multiplexing.

[0016] The processor is configured to determine the initial node corresponding to each of the plurality of replica nodes based on the initial attribute vectors of the replica nodes.

[0017] The processor is further configured to update the initial attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes, and each replica node is configured to perform link communication based on the parameters corresponding to the updated attribute vector.

[0018] Fourthly, embodiments of the present invention provide an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the steps of the link communication method described in the first aspect.

[0019] Fifthly, embodiments of the present invention provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the link communication method described in the first aspect.

[0020] In a sixth aspect, the present invention also provides a computer program product, including computer instructions, which, when executed by a processor, implement the steps of the link communication method described in the first aspect.

[0021] In this embodiment of the invention, an initial attribute vector is obtained for each of multiple replica nodes in a replica communication network. The initial attribute vector characterizes the communication status and / or resource usage of the corresponding replica node. Each replica node is replicated from an initial node in the initial communication network used for link segmentation multiplexing. An initial node corresponding to each replica node is determined based on its initial attribute vector. The initial attribute vector of each replica node is updated based on the first attribute vector of its corresponding initial node. Each replica node then performs link communication based on the parameters corresponding to the updated attribute vector. Thus, by updating the initial attribute vector of each replica node using the first attribute vector of its corresponding initial node, the utilization rate of link resources is improved when the replica node performs link communication based on the parameters corresponding to the updated attribute vector, thereby improving the effect of link segmentation multiplexing. Attached Figure Description

[0022] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0023] Figure 1 This is a flowchart of a link communication method provided in an embodiment of the present invention;

[0024] Figure 2 This is an optimized replication process provided in the embodiments of the present invention;

[0025] Figure 3 This is a structural diagram of a link communication device provided in an embodiment of the present invention;

[0026] Figure 4 This is a structural diagram of an electronic device provided in an embodiment of the present invention. Detailed Implementation

[0027] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0028] Please see Figure 1 , Figure 1This is a flowchart of a link communication method provided in an embodiment of the present invention, such as... Figure 1 As shown, it includes the following steps:

[0029] Step 101: Obtain the initial attribute vector of each replica node in the replica communication network. The initial attribute vector is used to characterize the communication status and / or resource usage of the corresponding replica node. Each replica node is replicated from the initial node in the initial communication network, which is used for link segmentation and multiplexing.

[0030] The aforementioned replication communication network is a communication network constructed by multiple replication nodes, which are nodes in the initial communication network. Data transmission is carried out through multiple replication nodes in the replication communication network and the links constructed between the replication nodes, thereby realizing link segmentation multiplexing relative to the initial communication network and effectively improving the utilization rate of link resources.

[0031] It should be noted that after replicating the initial nodes in the initial communication network to build a new replicated communication network, it is necessary to configure the parameters of the replicated nodes in the new replicated communication network so that each replicated node can maximize the utilization of link resources. Therefore, this invention proposes a link communication method to address this problem and enable parameter configuration of replicated nodes.

[0032] The aforementioned initial attribute vector is a vector corresponding to the communication parameters and / or resource usage parameters configured for the current replica node, such as bandwidth, latency, stability, and bit error rate during data transmission through the initial link. The initial attribute vector determines the current communication status and / or resource usage of the replica node. When the initial attribute vector is not adjusted, it can be based on the initial node settings; subsequent adjustments to the initial attributes adjust the current vector of the node.

[0033] In some implementations, setting the initial vector based on the initial node can be done by calculating the similarity between the replica node and the initial node based on the location and type of the current replica node, as well as the location and type of multiple initial nodes included in the initial network communication connection, selecting the initial node with the highest similarity from multiple initial nodes, and using its link multiplexing-related parameters as the initial vector of the replica node.

[0034] In other implementations, the initial vector of the replicated node can also be obtained without pre-configuration based on service type and node location.

[0035] Step 102: Determine the initial node corresponding to each of the multiple replica nodes based on the initial attribute vectors of the replica nodes.

[0036] The initial node corresponding to each of the above replica nodes is the closest initial node to that replica node, or the most relevant initial node to that replica node. Based on this, the initial attribute vector of the replica node can be adjusted according to the initial node corresponding to that replica node, so that the attribute vector of the replica node is closer to the actual data transmission requirements, thereby effectively improving the resource utilization of the link.

[0037] In some implementations, the initial node corresponding to the copied node can be determined by the similarity between the copied node and multiple initial nodes. This similarity can characterize the degree of similarity in attribute parameters between the copied node and the initial nodes, or it can represent the degree of similarity in type and / or position between the copied node and the initial nodes.

[0038] Step 103: Update the initial attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes. Each replica node is used for link communication based on the parameters corresponding to the updated attribute vector.

[0039] The aforementioned first attribute vector is a vector composed of the attribute parameters of the initial node in the initial network communication connection. The initial node can achieve communication through the initial link based on the attribute parameters corresponding to the first attribute vector. It should be noted that, since the replica node and the corresponding initial node have a high degree of similarity, the initial attribute vector of the replica node can be updated by the first attribute vector of the corresponding initial node to optimize the attribute vector of the replica node.

[0040] In this embodiment of the invention, an initial attribute vector is obtained for each of multiple replica nodes in a replica communication network. The initial attribute vector characterizes the communication status and / or resource usage of the corresponding replica node. Each replica node is replicated from an initial node in the initial communication network used for link segmentation multiplexing. An initial node corresponding to each replica node is determined based on its initial attribute vector. The initial attribute vector of each replica node is updated based on the first attribute vector of its corresponding initial node. Each replica node then performs link communication based on the parameters corresponding to the updated attribute vector. Thus, by updating the initial attribute vector of each replica node using the first attribute vector of its corresponding initial node, the utilization rate of link resources is improved when the replica node performs link communication based on the parameters corresponding to the updated attribute vector, thereby improving the effect of link segmentation multiplexing.

[0041] In one embodiment, the plurality of replica nodes includes a first replica node, and determining the initial node corresponding to each replica node based on the initial attribute vectors of the plurality of replica nodes includes:

[0042] Obtain the first attribute vectors of multiple initial nodes in the initial communication network;

[0043] Calculate the first similarity distance between the initial attribute vector of the first replicated node and the first attribute vectors of the plurality of initial nodes;

[0044] The initial node with the highest first similarity distance among the plurality of initial nodes is determined as the first initial node corresponding to the first copied node.

[0045] In this embodiment of the invention, for any one of the replicated nodes (i.e., the first replicated node), the first attribute vectors of multiple initial nodes in the initial communication network are obtained; the first similarity distance between the initial attribute vector of the first replicated node and the first attribute vectors of the multiple initial nodes is calculated; and the initial node with the highest first similarity distance among the multiple initial nodes is determined as the first initial node corresponding to the first replicated node. Thus, the first initial node corresponding to the first replicated node is determined through the first similarity distance, and the initial attribute vector of the first replicated node can be optimized based on the initial attribute vector corresponding to the first initial node.

[0046] In some implementations, the first similarity distance between the initial attribute vector of the first replicated node and the first attribute vector of the plurality of initial nodes can be calculated by directly calculating the similarity between the initial attribute vector and the first attribute vector, and then obtaining the first similarity distance based on the similarity conversion.

[0047] In other implementations, the initial attribute vector and the first attribute vector can be divided into two parts: the first part is the node attribute vector of the replicated node and the initial node, and the second part is the communication attribute vector of the replicated node and the initial node. The first distance between the node attribute vectors is measured by Euclidean distance, and the second distance between the communication attribute vector values ​​is measured by Value Difference Metric (VDM). The first similarity distance is obtained by weighting the first distance and the second distance.

[0048] The node attribute vector includes parameters such as node identifier, node type, node location, and node resources, which can be collected using network topology discovery tools such as SNMP and Telnet. The communication attribute vector includes parameters related to communication between different nodes, such as CPU utilization, memory utilization, and link bandwidth.

[0049] It should be noted that since the parameters corresponding to the communication attribute vectors are discrete, VDM calculates the distance by statistically analyzing the distribution probability of attribute values, which is more suitable for measuring the similarity of discrete attributes. Therefore, VDM is used to measure and obtain the second distance.

[0050] Furthermore, considering that multiple parameters will be obtained during data collection, but not every parameter will affect communication, it is necessary to filter different parameters to obtain the parameters that need to be managed and updated, and then generate the initial attribute vector and the first attribute vector through the parameter values ​​of these parameters.

[0051] The algorithm employs an information gain-based feature selection algorithm to calculate the information gain of each parameter. The parameters are sorted from largest to smallest according to their gain values, and the top k parameters are selected as key attribute parameters. The value of k is determined through cross-validation. The initial attribute vector of the replicated node is generated by using the parameter values ​​of the k parameters in the replicated node, and the first attribute vector of the first node is generated by using the parameter values ​​of the k parameters in the initial node.

[0052] In some implementations, the values ​​of key attribute parameters can be normalized to map the range of different attribute parameters to [0,1], so as to avoid the difference in units between different attributes affecting the accuracy of distance calculation, thereby improving the accuracy of attribute vector optimization for replicated nodes.

[0053] In one embodiment, updating the initial attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes includes:

[0054] Update the initial attribute vector of the first replicated node to the first attribute vector of the first initial node;

[0055] A second copy node is determined, wherein the similarity distance between the initial attribute vector of the second copy node and the first attribute vector of the first initial node is less than the first similarity distance;

[0056] The initial attribute vector of the second replicated node is updated to a weighted vector, which is the weighted value of the initial attribute vector of the second replicated node and the first attribute vector of the first initial node.

[0057] In this embodiment of the invention, the initial attribute vector of the first replica node is updated to the first attribute vector of the first initial node to update the first replica node; then, a second replica node is determined, wherein the similarity distance between the initial attribute vector of the second replica node and the first attribute vector of the first initial node is less than the first similarity distance; the initial attribute vector of the second replica node is updated to a weighted vector, wherein the weighted vector is the weighted value of the initial attribute vector of the second replica node and the first attribute vector of the first initial node. In this way, by adjusting the attribute vector of the second replica node through the attribute vector of the first replica node, the parameters of the first replica node are transferred to the second replica node, so that the attribute vector of the second replica node is updated simultaneously. This improves link utilization when link communication occurs between the first and second replica nodes.

[0058] Furthermore, after the second copy node is updated, other copy nodes can continue to be updated. Specifically, a third copy node is determined. The similarity between the attribute vector of the third copy node and the first attribute vector of the first initial node is less than the similarity between the attribute vector of the second copy node and the first attribute vector of the first initial node. Then, the attribute vector of the third copy node is updated according to the attribute vector of the second copy node. At this time, the attribute vector of the third copy node is the weighted value of the initial attribute vector and the current attribute vector of the second copy node, thus realizing the transfer of parameters from the second copy node to the third copy node.

[0059] It should be noted that this invention employs a random walk algorithm to transfer attribute parameters between the initial communication network and the replicated communication network. Specifically, the starting node of the initial communication network is determined, the similarity distance between this node and each replicated node is calculated, and the Softmax function is used to transform the similarity distance into a probability distribution, i.e.:

[0060] P(j|i)=exp(-d(i,j)) / ∑exp(-d(i,k));

[0061] Here, d(i,j) represents the similarity distance between the initial node i and the replicated node j, P(j|i) represents the probability of transitioning from the initial node i to the replicated node j, and exp is the natural exponential function. The Softmax function ensures that the sum of all transition probabilities is 1, making the result a valid probability distribution. Thus, the similarity between the replicated node and the initial node can be determined through probability.

[0062] Furthermore, by using probability, the next replica node of the current replica node can be selected sequentially, the attribute vector of the current replica node can be passed to the next replica node, and the attribute vector of the next node can be updated, until the terminal node of the replication link is reached, forming an attribute transfer path, thus realizing the updating of replica nodes in the replication communication network.

[0063] In one embodiment, the method further includes:

[0064] The process involves repeatedly determining the initial node corresponding to each replica node based on the updated attribute vectors of the multiple replica nodes, and updating the attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the multiple replica nodes, until the first preset condition is met and the update of the attribute vector of each replica node is stopped.

[0065] The first preset condition includes the following:

[0066] The number of updates has reached a set threshold;

[0067] The average similarity distance is less than a set distance threshold. The average similarity distance is the average of the similarity distances of the multiple replicated nodes. The similarity distance of the first replicated node is the similarity distance between the attribute vector of the first replicated node and the first attribute vector of the corresponding initial node.

[0068] In this embodiment of the invention, the initial node corresponding to each replica node is repeatedly determined based on the updated attribute vectors of the plurality of replica nodes. The attribute vector of each replica node is then updated based on the first attribute vector of the initial node corresponding to the plurality of replica nodes, until a first preset condition is met, at which point the updating of the attribute vector of each replica node stops. In this way, the attribute vector of each replica node can be continuously optimized through repeated updates, thereby effectively improving the utilization rate of link resources when each replica node communicates via the link.

[0069] The first preset condition is that the number of updates reaches a set threshold. At this point, multiple updates have been performed, and the attribute vector of the current copied node can be considered to have reached its optimal value. Alternatively, the first preset condition is that the average similarity distance is less than a set distance threshold. At this point, the overall value of the multiple copied nodes is close to that of the initial node, and the attribute vector of the current copied node can be considered to have reached its optimal value, so no further updates are needed.

[0070] In some implementations, the similarity distance of the first replicated node can be the cosine similarity distance between the attribute vector of the first replicated node and the first attribute vector of the corresponding initial node.

[0071] For example, when analyzing the initial communication network, 15 attribute parameters, such as CPU utilization, memory utilization, and link bandwidth, of the initial nodes are collected. The attribute values ​​are normalized, and the value range is mapped to [0,1]. Then, an information gain-based feature selection algorithm is used to calculate the information gain value of each attribute parameter. The attributes are sorted from largest to smallest according to their information gain values, and the top 5 attributes are selected as key attributes. The optimal k value is determined through 10-fold cross-validation.

[0072] Next, the attributes of the replicated nodes are initialized. Based on the node type (e.g., switch, router) and location (e.g., access layer, aggregation layer), the initial node with the highest similarity is selected from the initial communication network. Cosine similarity is used to calculate the similarity between the node attribute vectors, and the key attributes of the node with the highest similarity are copied to the corresponding replicated node. When calculating the attribute distance between the original node and the replicated node, for continuous attributes such as bandwidth, Euclidean distance is used to calculate the L2 norm, and for discrete attributes such as device model, the VDM algorithm is used to calculate the probability distribution difference, thus obtaining the distance matrix between the nodes.

[0073] Furthermore, using Dijkstra's algorithm, starting from the replica node, the nearest initial node is found, and its key attributes are copied to the replica node. Then, a random walk algorithm is employed, starting from the initial node of the original link. The Softmax function is used to convert distance into transition probability, i.e., P(i,j)=exp(-d(i,j)) / ∑exp(-d(i,k)), where d(i,j) represents the distance from the initial node i to the replica node j. The next-hop replica node is selected through probability sampling. The attribute vector of the current replica node is weighted and averaged with the initial attribute vector of the next-hop replica node to update the attribute vector of the next-hop replica node. For example, the weights of the two attribute vectors can be 0.8 and 0.2 respectively, and they are passed sequentially until the terminal node of the replica link is reached, thus updating all replica nodes.

[0074] The iteration can be stopped when the average distance between the attribute vectors of the copied node and the original node is less than 0.01 or the number of iterations reaches 100. In each iteration, the cosine similarity of the node attribute vectors is calculated. If the similarity increase is less than 0.001 after five consecutive iterations, the iteration terminates early. The resulting node attribute vectors of the copied link have an average cosine similarity of over 0.95 with the attribute vectors of the original link nodes, achieving effective replication and transfer of unique attributes.

[0075] In one embodiment, after updating the initial attribute vector of each replicated node based on the first attribute vector of the initial node corresponding to the plurality of replicated nodes, the method further includes:

[0076] Obtain multiple initial replication links corresponding to the multiple replication nodes, where each initial replication link is used for communication between replication nodes;

[0077] Multiple intermediate links are generated based on the ant colony optimization algorithm, and these intermediate links are used to replicate communication between nodes.

[0078] Calculate the fitness value of each of the plurality of initial replication links and the plurality of intermediate links, the fitness value being used to characterize the communication status of the link;

[0079] Multiple update links are obtained through filtering. The fitness values ​​of these multiple update links are greater than the fitness values ​​of other links. The other links are the links other than the multiple update links among the multiple initial replication links and the multiple intermediate links.

[0080] In this embodiment of the invention, multiple initial replication links corresponding to the multiple replication nodes are obtained, each of which is used for communication between replication nodes. Multiple intermediate links are generated based on an ant colony optimization algorithm, and these intermediate links are also used for communication between replication nodes. The fitness value of each of the multiple initial replication links and the multiple intermediate links is calculated, and the fitness value characterizes the communication status of the link. Multiple updated links are then selected, where the fitness value of the updated links is greater than the fitness values ​​of other links, which are the links other than the updated links among the multiple initial replication links and the multiple intermediate links. Thus, by updating the multiple initial replication links with multiple intermediate links generated by the ant colony optimization algorithm, the updated links better meet the actual data transmission requirements of link communication compared to the initial replication links, thereby further improving the utilization rate of link resources.

[0081] The fitness value of a link is used to characterize the effectiveness of data transmission. A higher fitness value indicates a better transmission effect, while a lower fitness value indicates a worse transmission effect. By calculating the fitness value of each link, multiple update links with better transmission effects can be obtained.

[0082] In one embodiment, the plurality of intermediate links includes a first intermediate link, and the generation of the plurality of intermediate links based on the ant colony optimization algorithm includes:

[0083] Randomly generate the first path corresponding to the first replicated node;

[0084] The next node of the first path is randomly selected based on the transition probability, wherein the transition probability is calculated using global pheromones, and the next node is one of the plurality of replicated nodes;

[0085] The first intermediate link is generated based on the first path and the next node of the first path;

[0086] The global pheromone is updated based on the length of the first intermediate link.

[0087] In this embodiment of the invention, the next node of the first path is randomly selected based on the transition probability, which is calculated using global pheromones. The next node is one of the plurality of replicated nodes. A first intermediate link is generated based on the first path and its next node. The global pheromones are updated based on the length of the first intermediate link. Thus, the first path is generated using global pheromones, and the first intermediate link is obtained through the first path. Simultaneously, the global pheromones are adjusted to ensure that different intermediate links are distinguished when multiple intermediate links are generated, thereby obtaining different intermediate links and enabling the updating of the initial replicated link through multiple intermediate links.

[0088] For example, when optimizing the initial replication link in a replication communication network, the initial replication link is first abstracted into a weighted bidirectional graph model. The replication nodes in the model represent network devices, such as routers and switches, the edges represent replication link connections, and the weights of the edges represent attribute indicators such as link latency and bandwidth. For example, the link latency between replication node 1 and replication node 2 is 5ms and the bandwidth is 100Mbps.

[0089] Furthermore, minimizing latency, maximizing bandwidth, and maximizing reuse rate are set as optimization objectives. A multi-objective optimization model is constructed, and a weighted summation method is used to normalize multiple objectives and then weight them together to form a fitness value f = 0.5 × latency + 0.3 / bandwidth + 0.2 × reuse rate. The fitness value of the link is calculated through the objective optimization function.

[0090] Furthermore, initialize the parameters of the ant colony algorithm, such as setting the ant colony size m=50, the initial value of global pheromone τ=0.1, the pheromone evaporation coefficient ρ=0.1, the maximum number of iterations Nmax=100, and design the heuristic information η(i,j)=1 / (d(i,j)+1), where d(i,j) represents the time delay between nodes i and j.

[0091] When generating intermediate links, an initial solution is randomly generated, i.e., 20 replica nodes are randomly selected as starting points to construct 20 initial paths, and the fitness value of each path is calculated. Entering the iterative optimization phase, for each ant, according to the state transition probability formula P(i,j)=[τ(i,j)]^α×[η(i,j)]^β / ∑[τ(i,k)]^α×[η(i,k)]^β, where α=1 and β=2, a roulette wheel selection strategy is used to select the next replica node until the destination is reached, resulting in a complete intermediate path.

[0092] Then, the global pheromone is updated using the global pheromone update formula τ(i,j)=(1-ρ)×τ(i,j)+∑Δτ(i,j). Here, Δτ(i,j)=Q / Lk, Q=100, and Lk is the path length of the k-th ant. The pheromone concentration on the path is updated using the path length.

[0093] Repeat the above process until all 50 ants have completed one iteration, resulting in 50 new paths (including the initial copy path and intermediate paths). Calculate the fitness value of each path and compare it with the current best solution. If the fitness value of a new solution is better, then it is adopted as the new best solution.

[0094] In some implementations, a second preset condition can be set to determine whether to stop updating and replicating the link. This second preset condition could be that the number of updates reaches a preset number, or that the change in the fitness value of the link before and after the update is less than a threshold. For example, if the change in fitness value of the optimal solution over 20 consecutive iterations is less than 0.001, or the number of iterations reaches 100, then the current optimal solution is output, i.e., the link scheme with the minimum latency, maximum bandwidth, and highest reuse rate. For example, the optimal path is {1-3-5-7-9}, with a latency of 8ms, a bandwidth of 500Mbps, a reuse rate of 0.9, and a fitness value of 0.85; otherwise, iterative optimization continues until the second preset condition is met.

[0095] In one embodiment, after updating the initial attribute vector of each replicated node based on the first attribute vector of the initial node corresponding to the plurality of replicated nodes, the method further includes:

[0096] Obtain the communication parameters of each replication link in the multiple replication links included in the replication communication network;

[0097] The link health index of each of the multiple replicated links is calculated based on the communication parameters of each replicated link.

[0098] The first optimization strategy corresponding to the first replication link is determined based on the first mapping relationship. The first replication link is the replication link whose link health index is less than a set health threshold among the multiple replication links. The first mapping relationship includes multiple communication parameters and the optimization strategy corresponding to each communication parameter.

[0099] The communication parameters of the first replicated link are adjusted based on the first optimization strategy.

[0100] In this embodiment of the invention, communication parameters of each of the multiple replication links in the replication communication network are obtained; a link health index is calculated for each replication link based on the communication parameters; a first optimization strategy is determined for a first replication link based on a first mapping relationship, wherein the first replication link is a replication link whose link health index is less than a set health threshold, and the first mapping relationship includes multiple communication parameters and an optimization strategy corresponding to each communication parameter; the communication parameters of the first replication link are adjusted based on the first optimization strategy. Thus, by calculating the link health index to determine the communication effect of each link, it can be determined whether the replication link needs adjustment; when adjustment is needed, the first optimization strategy is determined based on the link health index to adjust the communication parameters of the replication link.

[0101] For example, network traffic analysis tools can be used to collect communication parameters on the optimized communication network in real time, such as traffic size, traffic direction, and traffic rate, to generate traffic statistics reports; network performance monitoring platforms, such as Cacti and Nagios, can be used to periodically probe the communication quality parameters of the links, including link bandwidth utilization, latency, jitter, and packet loss rate, and obtain performance indicator data from network devices through the Simple Network Management Protocol (SNMP) to generate performance monitoring reports; the traffic statistics reports and performance monitoring reports can be correlated and analyzed to calculate the link health index.

[0102] The link health index is calculated as 1 - (0.5 × bandwidth utilization + 0.3 × latency / threshold + 0.2 × packet loss rate). When the overall health index is lower than the preset threshold (e.g., 0.8), a link alarm is triggered, generating an alarm event.

[0103] Furthermore, a Bayesian network-based inference model can be used to locate the cause of link failures and generate a fault diagnosis report based on the priority and combination of preset rules. Based on the fault diagnosis report, a rule-based policy engine technology can be used to match predefined link optimization strategies.

[0104] In some implementations, the first mapping relationship is pre-configured, and optimization strategies corresponding to different parameters can be determined through this first mapping relationship. For example, when bandwidth utilization > 0.9 and latency > 50ms, the link is determined to be congested, and the optimization strategy is Strategy 1; or when packet loss rate > 0.1 and jitter > 30ms, the link is determined to be unstable, and the optimization strategy is Strategy 2, and so on. Automated configuration management tools, such as Ansible and Puppet, can also be used to issue configuration commands to network devices to dynamically adjust link parameters.

[0105] Furthermore, after adjusting the link parameters, traffic and performance monitoring continue. Reinforcement learning algorithms are used, with the improvement in link quality indicators as a reward signal. Through algorithms such as Q-learning, the strategy is continuously optimized and adjusted, the optimization effect is evaluated, and the link quality improvement rate is calculated through before-and-after comparative analysis.

[0106] For example, the link quality improvement rate = 0.4 × (1 - latency improvement rate) + 0.3 × (1 - packet loss rate improvement rate) + 0.2 × (1 - jitter improvement rate) + 0.1 × bandwidth utilization improvement rate. The evaluation results are used as the baseline for the next round of monitoring. The health threshold and alarm rules are dynamically adjusted and continuously iterated and optimized until the link quality stabilizes at the expected level.

[0107] In one embodiment, before obtaining the initial attribute vector of each of the multiple replica nodes in the replica communication network, the method further includes:

[0108] The multiplexing communication parameters of each initial link in the multiple initial links included in the initial communication network are obtained. The multiplexing communication parameters are used to characterize the parameters that can be reused in the corresponding initial links.

[0109] A multiplexing matrix corresponding to each initial link is generated based on the multiplexing communication parameters of the multiple initial links;

[0110] Clustering is performed on the multiple reuse matrices corresponding to the multiple initial links to obtain multiple clusters and a multiple reuse granularity corresponding to each cluster, wherein each cluster includes at least one multiple reuse matrix;

[0111] The number of copies of each initial node in the initial communication network is determined based on the reuse granularity corresponding to the multiple clusters.

[0112] Each initial node is replicated based on the number of replications of the multiple initial nodes to obtain the multiple replicated nodes.

[0113] In this embodiment of the invention, multiplexing communication parameters of each initial link in the plurality of initial links included in the initial communication network are obtained. These multiplexing communication parameters characterize the parameters that can be reused in the corresponding initial links. A multiplexing matrix corresponding to each initial link is generated based on the multiplexing communication parameters. The multiplexing matrices corresponding to the plurality of initial links are clustered to obtain multiple clusters and a multiplexing granularity corresponding to each cluster. Each cluster includes at least one multiplexing matrix. The replication quantity of each initial node in the initial communication network is determined based on the multiplexing granularity corresponding to the multiple clusters. Each initial node is replicated based on its replication quantity to obtain the plurality of replicated nodes. Thus, by generating a multiplexing matrix corresponding to each initial link, a multiplexing granularity is obtained through clustering. The replication quantity of each initial node can then be determined based on the multiplexing granularity, thereby enabling the replication of initial nodes in the initial communication network to obtain multiple replicated nodes in the replicated communication network.

[0114] In some implementations, multiplexed communication parameters may include parameters such as bandwidth, latency, stability, and bit error rate. These parameters can be preprocessed to convert the data into a standardized format. Further, principal component analysis is used to analyze the standardized multiplexed communication parameters, obtaining a correlation coefficient matrix between different multiplexed communication parameters. Based on the correlation coefficient matrix, a subset of key attributes with strong correlations is selected according to a preset threshold, serving as the basis for determining whether a link supports multiplexing. According to a preset fuzzy rule, such as "if the bandwidth is greater than 100Mbps and the latency is less than 10ms, then the link supports multiplexing," the attribute values ​​in the subset of key attributes are matched with the rule conditions to obtain the judgment result of whether the link supports multiplexing, thereby selecting a set of key attributes that can support link multiplexing.

[0115] Furthermore, after obtaining the correlation coefficient matrix of the initial links, the initial links in the initial communication network can be clustered using the fuzzy C-means clustering algorithm based on the multiplexed communication parameters in the key attribute set, and the membership matrix of each multiplexed communication parameter can be calculated. Then, the objective function is minimized using the least squares method, and the cluster centers and memberships are iteratively updated until the objective function converges, thereby obtaining the multiplexing matrix corresponding to the initial links.

[0116] After obtaining the multiplexing matrix corresponding to the initial links, a bottom-up Agglomerative Nesting (AGNES) hierarchical clustering algorithm is used to calculate the average connection distance between clusters of different multiplexing matrices to achieve the merging of the initial links. Merging stops when the inter-cluster distance exceeds a preset threshold, thus obtaining multiple clusters and the multiplexing granularity corresponding to each cluster, and consequently, the multiplexing granularity corresponding to each initial link.

[0117] In other implementations, besides the scheme described above that calculates the multiplexing granularity of the initial link using a clustering algorithm, the multiplexing granularity of each initial link can also be determined based on the service type and Quality of Service (QoS) of the initial link. For example, a mapping relationship between link multiplexing granularity and metrics such as latency and packet loss rate can be pre-created. By obtaining metrics such as latency and packet loss rate of the initial link and combining them with the mapping relationship, the multiplexing granularity of the initial link can be determined.

[0118] For example, for voice services with high real-time requirements, a smaller multiplexing granularity can be set to ensure low latency; while for services with high reliability requirements, a larger multiplexing granularity needs to be set to ensure high availability.

[0119] In some implementations, a genetic algorithm can be used to optimize the granularity of link replication. Specifically, the reuse granularity is encoded as a chromosome, and the fitness function is a weighted sum of indicators such as latency, packet loss rate, and utilization rate. Individuals with higher fitness values ​​are selected for crossover and mutation, with a crossover probability of 0.8 and a mutation probability of 0.1. New individuals are generated using uniform crossover and random mutation, and the optimal link reuse granularity value is obtained after 500 generations.

[0120] For example, the sampling fuzzy C-means clustering method clusters the reuse characteristics of links. This process involves calculating a membership matrix of key metrics such as latency, packet loss rate, and utilization rate. An objective function is optimized, which is the weighted sum of squared distances from all samples to their respective cluster centers (the weights are determined by the fuzzy membership degrees). This process iteratively updates the sample cluster centers (based on the current weighted average membership degree) and recalculates the membership degrees (based on the ratio of distances from samples to each cluster center). This process repeats 100 times, ultimately deriving a reuse matrix that reflects the similarity of link segment reuse.

[0121] Furthermore, hierarchical clustering is performed using the AGNES algorithm. This algorithm measures the average connection distance between different categories as the merging criterion. The specific calculation formula is the average inter-class distance d(Ci,Cj), which is obtained by summing the distances between each pair of samples and then dividing by the product of the number of samples in the two categories. Clustering starts from the bottom layer and gradually merges the closest categories until the average inter-class distance exceeds a preset threshold of 0.5, thus determining the reuse granularity.

[0122] Alternatively, based on the link service type and QoS requirements, an association rule mining algorithm can be used. The Apriori algorithm is used to calculate frequent itemsets and association rules for service types and QoS indicators, with support ≥ 0.2 and confidence ≥ 0.8. Strong association rules are obtained to show that voice services require latency ≤ 30ms and packet loss rate ≤ 0.1%, while services with high reliability requirements require latency ≤ 50ms and packet loss rate ≤ 0.05%. Based on this, the link multiplexing granularity for voice services and services with high reliability requirements is set to 2 and 3, respectively.

[0123] Furthermore, to optimize the granularity of data replication, a genetic algorithm was employed, where the reuse granularity was encoded as an 8-bit binary string serving as chromosomes in the genetic algorithm. Individual fitness was evaluated using a fitness function f(x), which comprehensively considered three metrics: delay, packet loss, and utilization, with weights of w1=0.5, w2=0.3, and w3=0.2, respectively. During the evolutionary process, a roulette wheel selection method was used, employing uniform crossover (crossover probability set to 0.8) and random mutation (mutation probability 0.1) to generate new generations of individuals. After 500 iterations, the optimal reuse granularity value of 3 was obtained.

[0124] In one embodiment, the initial communication network includes a second initial node, and determining the number of replications for each initial node in the initial communication network based on the reuse granularity corresponding to the plurality of clusters includes:

[0125] Determine the reuse granularity corresponding to the second initial node. The reuse granularity of the second initial node includes the reuse granularity corresponding to the cluster of the initial link where the second initial node is located.

[0126] Obtain the service type and / or required parameters of the second initial node;

[0127] The number of replications corresponding to the second initial node is determined based on the second mapping relationship, which includes multiple replication numbers, the service type and / or requirement parameters corresponding to each replication number, and the reuse granularity.

[0128] In this embodiment of the invention, the reuse granularity corresponding to the second initial node is determined. The reuse granularity of the second initial node includes the reuse granularity corresponding to the cluster of the initial link where the second initial node is located. The service type and / or requirement parameters of the second initial node are obtained. The replication quantity corresponding to the second initial node is determined according to a second mapping relationship. The second mapping relationship includes multiple replication quantities, and the service type and / or requirement parameters corresponding to each replication quantity, as well as the reuse granularity. In this way, by using the reuse granularity, service type, and / or requirement parameters of the initial node, the replication quantity corresponding to the initial node is determined, thereby realizing the replication of the initial node to obtain a replication node.

[0129] The second mapping relationship is pre-configured and used to determine the number of replications. Specifically, different multiplexing granularity thresholds are set according to the service type and QoS requirements of the link to construct the second mapping relationship. For example, when the key attributes of a node, such as bandwidth and latency, meet the threshold requirements, replication is performed, and the number of replications can be dynamically adjusted according to the second mapping relationship.

[0130] In some implementations, the topology of the initial communication network can be modeled according to the link reuse granularity. Configuration information of network devices can be collected, and node identifiers, types, and key attributes can be extracted and stored in a topology database. An adjacency list is used to represent the connections between link nodes, with each node containing node representation, node type, and key attributes. A depth-first search algorithm is used to recursively traverse all initial nodes, starting from the initial node of the initial link. During the traversal, nodes matching the key attributes and their connections are copied, and the traversal path and node access status are recorded to avoid repeated traversals. This yields the service type and / or required parameters of each initial node, as well as the reuse granularity of the initial node.

[0131] Furthermore, when replicating nodes, the number of nodes to be replicated is determined based on the link reuse granularity. If the key attributes of the current node do not match the target granularity, its child nodes are traversed to ensure that nodes are replicated as much as possible.

[0132] In one embodiment, after replicating each initial node based on the replication number of the plurality of initial nodes to obtain the plurality of replicated nodes, the method further includes:

[0133] Multiple initial replication links are constructed based on the aforementioned multiple replication nodes.

[0134] It should be noted that after obtaining multiple replication nodes, it is necessary to establish an initial replication link between the multiple replication nodes in order to enable communication through the initial replication link.

[0135] When replicating connections, the system retrieves all neighboring nodes of the current node from the adjacency list of the initial link and determines whether these neighboring initial nodes have already been replicated to the new link. If they have been replicated, a connection is established between the replicated node of the current node and the replicated node of the neighboring initial node in the new link; otherwise, the neighboring node is added to the stack of the node to be replicated, awaiting further processing. Through recursive traversal and replication of nodes and their connections, multiple initial replicated links with similar topology, matching node attributes, and consistent reuse granularity are ultimately obtained.

[0136] Furthermore, after traversing all replicated nodes and creating the initial replicated link, it is determined whether there are any replicated nodes without established links. If not, the data in the stack of the node to be replicated is discarded; if it exists, the corresponding node in the stack of the node to be replicated is determined, the node in the stack of the node to be replicated is replicated, and the replicated link is established.

[0137] It should be noted that during the replication process, a hash table is used to record the access and replication status of nodes. A hash function maps node identifiers to buckets in the hash table. Each bucket contains a list of node statuses, which is used to quickly determine whether a node has been accessed or replicated, thereby improving the time efficiency of the algorithm.

[0138] In some implementations, the replicated links can be verified by comparing the traffic data on the replicated links with the initial links to calculate the similarity of traffic distribution. If the similarity reaches 95% or higher, the replication is considered successful; if the similarity is low, further analysis of the reasons for the differences is needed, and link optimization should be performed.

[0139] In some implementations, the topology of multiple replicated links can be simplified. Specifically, using Kruskal's algorithm, all edges of the replicated links are first sorted by weight in ascending order. Then, starting with the edge with the smallest weight, if the two nodes on that edge do not form a cycle in the minimum spanning tree, the edge is added to the minimum spanning tree, and so on, until all nodes are added.

[0140] In addition to the above methods, Prim's algorithm can also be used. First, select a starting node, and then each time select the edge with the smallest weight from the edges connected to the current minimum spanning tree, add the other node connected by that edge to the minimum spanning tree, until all nodes have been added.

[0141] In this way, optimization of multiple replicated links is achieved through the two methods described above. Furthermore, the optimized topology can be evaluated by calculating the total weight of the new replicated links and comparing it with the original replicated links. If the total weight is reduced by more than 20%, the optimization is considered effective, resulting in a link topology with minimum cost and reducing the overhead of link reuse.

[0142] For example, when replicating the initial communication network, the configuration information of network devices, such as device name, IP address, and interface bandwidth, is first collected via the SNMP protocol of the network management system. The identifier, type, and key attributes of the initial nodes are then extracted and stored in a database. Next, the NetworkX library in Python is used to model the link topology as an undirected weighted graph. The identifier of the initial node is used as a node in the graph, and the bandwidth of the initial link is used as the edge weight, to generate an adjacency list representing the connections between nodes.

[0143] Then, using a depth-first search algorithm, all nodes in the graph are recursively traversed starting from the initial node. A stack is used to store the nodes to be traversed, and a hash table is used to record the access status of each node to avoid repeated traversals. During the traversal, different multiplexing granularity thresholds are set according to the service type and QoS requirements of the link. For example, the bandwidth threshold for voice links is 1Mbps, and the latency threshold is 50ms; the bandwidth threshold for data links is 10Mbps, and the latency threshold is 100ms. If the key attributes of the current node meet the threshold requirements, the number of nodes to be replicated is determined according to the threshold level (i.e., the replication number is determined through a second mapping relationship).

[0144] Furthermore, the current node is copied to the new link, inheriting the connection relationships of the original node. If the key attributes of the current node do not meet the threshold requirements, its child nodes are added to the stack to be traversed to continue traversal. When copying connection relationships, all neighboring nodes of the current node are obtained through the adjacency list. It is determined whether these neighboring nodes have been copied. If they have been copied, the connection relationships between them are established in the new link; if they have not been copied, the neighboring nodes are added to the stack to be copied.

[0145] In this way, through recursive traversal, a replicated link with consistent reuse granularity and matching node attributes is finally obtained. In some implementations, the Netflow protocol of the network management system can be used to collect traffic data of the replicated link and the initial link, and calculate the JS divergence of the traffic distribution. If the JS divergence is less than 0.05, the replication is considered successful; otherwise, the reasons for the traffic differences are analyzed, such as mismatched node attributes or inconsistent link topology, and link optimization is performed, such as adjusting the reuse granularity and deleting redundant nodes.

[0146] In one embodiment, after constructing multiple initial replication links based on the plurality of replication nodes, the method further includes:

[0147] Obtain the attribute parameters and business parameters for each initial replication link;

[0148] Based on the attribute parameters and service parameters of each initial replication link, an evaluation result is generated for the corresponding initial replication link. The evaluation result is used to characterize the replication effect of the initial replication link.

[0149] The reuse granularity of each initial replication link is adjusted based on the evaluation results of each initial replication link.

[0150] In this embodiment of the invention, attribute parameters and service parameters of each initial replication link are obtained; an evaluation result corresponding to each initial replication link is generated based on the attribute parameters and service parameters; and the multiplexing granularity corresponding to each initial replication link is adjusted based on the evaluation result. Thus, by generating an evaluation result corresponding to each initial replication link based on its attribute parameters and service parameters, the accuracy of the multiplexing granularity corresponding to the initial replication link can be determined through the evaluation result, thereby adjusting the multiplexing granularity to improve the accuracy of the final replication link.

[0151] The attribute parameters for each initial replication link include peak traffic and response latency, while the service parameters include service type and traffic volume. These attribute and service parameters can be obtained using network monitoring tools such as Nagios and Zabbix.

[0152] Furthermore, association rule mining algorithms, such as the Apriori algorithm or the FP-Growth algorithm, can be used to perform association analysis on the link attribute and business attribute datasets. By setting a minimum support threshold of 0.1 and a minimum confidence threshold of 0.8, strong association rules between link attributes and business attributes can be filtered out. For example, "link bandwidth > 100Mbps" corresponds to "business response latency < 10ms", with a support of 0.8 and a confidence of 0.9. These parameters can be adjusted according to specific business scenarios and data characteristics.

[0153] Based on the discovered association rules, an adaptation matrix is ​​constructed between link attributes and business attributes. The matrix elements represent the degree of adaptation between link attribute values ​​and business attribute values. The degree of adaptation can be measured by the confidence of the association rules. The higher the confidence, the higher the degree of adaptation.

[0154] The fitting matrix can be clustered using the K-means clustering algorithm to divide the link attributes and business attributes into several fitting clusters. The fitting degree between link attributes and business attributes within each cluster is high, while the fitting degree between clusters is low. The silhouette coefficient is used to evaluate the clustering effect. The formula for calculating the silhouette coefficient is S=(ba) / max(a,b), where a represents the average distance between a sample and other samples in the same cluster, and b represents the average distance between a sample and the nearest cluster. The silhouette coefficient ranges from [-1,1], with a larger value indicating a better clustering effect. The maximum silhouette coefficient is selected as the clustering result.

[0155] Furthermore, evaluation results are generated based on the clustering results to determine whether the attributes of the replicated links are located in the same adaptation cluster as the target business attributes. If they are in the same cluster, it means that the attributes of the replicated links meet the business reuse requirements. These links are added to the reuse link set, and their key attributes and reuse granularity are recorded as the basis for subsequent link scheduling and optimization. An evaluation result indicating that the replication effect of the initial replicated links is good is generated. If they are not in the same cluster, it means that the attributes of the replicated links do not match the business reuse requirements, and link optimization is required. An evaluation result indicating that the replication effect of the initial replicated links is poor is generated.

[0156] In some implementations, the reuse granularity of each initial replication link is adjusted based on the evaluation results of each initial replication link. This can be achieved by using methods such as Principal Component Analysis (PCA) based on variance or Factor Analysis (FA) based on correlation to evaluate and rank the importance of link attributes and business attributes, and select the top-ranked key attributes as optimization targets.

[0157] Based on the compatibility relationship between link attributes and service attributes, a heuristic search algorithm, such as a genetic algorithm or simulated annealing algorithm, is used to search for the optimal combination of link attributes. The optimization effect of the attribute combination is evaluated through a fitness function, which can comprehensively consider multiple optimization objectives such as link reuse rate, service response latency, and link utilization. For example, the fitness function f = w1 × reuse rate + w2 × (1 / latency) + w3 × utilization rate, where w1, w2, and w3 are the weight coefficients of link reuse rate, service response latency, and link utilization, respectively. The comprehensive fitness value is obtained by weighted summation.

[0158] Furthermore, parameters such as population size, crossover probability, and mutation probability in the genetic algorithm, and initial temperature, cooling rate, and termination temperature in the simulated annealing algorithm, can be set according to the scale and complexity of the specific problem. The optimal link attribute combination obtained from the search is mapped to the reuse granularity as the new reuse granularity, and the original links are copied and their attributes mapped again to obtain optimized copied links.

[0159] After obtaining the optimized replication link, the reusability of the replication link is re-evaluated. When the fitness improvement of n consecutive iterations is less than a certain threshold, or the total number of iterations reaches the upper limit, the optimization stops and the optimal solution is output, which is the set of replication links that meet the business reuse requirements and their key attributes and reuse granularity.

[0160] For example, using the Zabbix network monitoring platform, business data of the original links can be collected via the SNMP protocol, including the number of requests and response latency, the amount of file transfers and the transfer rate, as well as indicators such as the bandwidth utilization and packet loss rate of the links, to form a business attribute dataset.

[0161] Then, using the FP-Growth association rule mining algorithm, with a minimum support of 0.05 and a minimum confidence of 0.8, we obtained the association rule "link packet loss rate < 0.01" corresponding to "response latency < 50ms", with a support of 0.12 and a confidence of 0.95; and "link bandwidth utilization > 0.8" corresponding to "transmission rate < 1MB / s", with a support of 0.08 and a confidence of 0.9.

[0162] Based on the association rules, a 10×10 adaptation matrix is ​​constructed, where rows represent link attributes and columns represent business attributes. For example, the (1,1)th element indicates that "packet loss rate < 0.01" corresponds to an adaptation degree of 0.95 for "response latency < 50ms". K-means clustering is performed on the adaptation matrix, with K=3, resulting in 3 adaptation clusters. The silhouette coefficient of each link is then calculated.

[0163] Furthermore, links with a profile coefficient greater than 0.6 are grouped into the set that meets the multiplexing requirements. For links with a profile coefficient less than 0.4, the PCA algorithm is used to perform attribute importance analysis, selecting the top three principal components as key attributes, such as bandwidth utilization, packet loss rate, and latency jitter. Then, a genetic algorithm is used to search for the optimal combination of link attributes. The population size is set to 50, the crossover probability is 0.8, the mutation probability is 0.1, and the fitness function is f = 0.5 × multiplexing rate + 0.3 × (1 / average latency) + 0.2 × utilization rate. The optimal attribute combination is obtained after 500 iterations and mapped to a new multiplexing granularity.

[0164] The original links are copied and their attributes are remapped to obtain optimized copied links. The evaluation and optimization process is repeated until the fitness improvement is less than 0.01 after 10 consecutive iterations, or the total number of iterations reaches 1000. The optimal set of links that meet the reuse requirements and their reuse granularity are output.

[0165] Furthermore, such as Figure 2 As shown, in this invention, the attribute parameters of the initial communication network are obtained; it is determined whether the attribute parameters meet the conditions for link multiplexing; the correlation between the attribute parameters is analyzed to obtain key attributes; the multiplexing granularity is determined based on the key attributes; the replication granularity is determined based on the replication granularity, and a replication link is generated; the attribute parameters of the replication link are updated; it is determined whether the replication link meets the multiplexing requirements; and the replication link is optimized and continuously monitored.

[0166] Please see Figure 3 , Figure 3 This is a structural diagram of a link communication device provided in an embodiment of the present invention, such as... Figure 3 As shown, the link communication device 300 includes:

[0167] The first acquisition module 301 is used to acquire the initial attribute vector of each replica node among multiple replica nodes in the replica communication network. The initial attribute vector is used to characterize the communication status and / or resource usage of the corresponding replica node. Each replica node is obtained by replicating an initial node in the initial communication network. The replica communication network is used for link segmentation and multiplexing.

[0168] The first determining module 302 is used to determine the initial node corresponding to each of the plurality of replicated nodes based on the initial attribute vectors of the replicated nodes.

[0169] The update module 303 is used to update the initial attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes, and each replica node is used to perform link communication based on the parameters corresponding to the updated attribute vector.

[0170] In one embodiment, the plurality of replication nodes includes a first replication node, and the first determining module 302 includes:

[0171] The first acquisition unit is used to acquire the first attribute vectors of multiple initial nodes in the initial communication network;

[0172] The calculation unit is used to calculate the first similarity distance between the initial attribute vector of the first replicated node and the first attribute vectors of the plurality of initial nodes;

[0173] The first determining unit is used to determine the initial node with the highest first similarity distance among the plurality of initial nodes as the first initial node corresponding to the first copied node.

[0174] In one embodiment, the update module 303 includes:

[0175] The first update unit is used to update the initial attribute vector of the first replicated node to the first attribute vector of the first initial node;

[0176] The second determining unit is used to determine the second copy node, wherein the similarity distance between the initial attribute vector of the second copy node and the first attribute vector of the first initial node is less than the first similarity distance;

[0177] The second update unit is used to update the initial attribute vector of the second replicated node to a weighted vector, wherein the weighted vector is the weighted value of the initial attribute vector of the second replicated node and the first attribute vector of the first initial node.

[0178] In one embodiment, the link communication device 300 further includes:

[0179] The processing module is used to repeatedly determine the initial node corresponding to each copy node based on the updated attribute vectors of the plurality of copy nodes, and update the attribute vector of each copy node based on the first attribute vector of the initial node corresponding to the plurality of copy nodes, until the first preset condition is met and the update of the attribute vector of each copy node is stopped.

[0180] The first preset condition includes the following:

[0181] The number of updates has reached a set threshold;

[0182] The average similarity distance is less than a set distance threshold. The average similarity distance is the average of the similarity distances of the multiple replicated nodes. The similarity distance of the first replicated node is the similarity distance between the attribute vector of the first replicated node and the first attribute vector of the corresponding initial node.

[0183] In one embodiment, the link communication device 300 further includes:

[0184] The second acquisition module is used to acquire multiple initial replication links corresponding to the multiple replication nodes, wherein each initial replication link is used for communication between replication nodes.

[0185] The first generation module is used to generate multiple intermediate links based on the ant colony optimization algorithm, and the multiple intermediate links are used to replicate communication between nodes.

[0186] The first calculation module is used to calculate the fitness value of each of the plurality of initial replication links and the plurality of intermediate links, wherein the fitness value is used to characterize the communication status of the link;

[0187] A filtering module is used to filter and obtain multiple update links, wherein the fitness values ​​of the multiple update links are greater than the fitness values ​​of other links, and the other links are links other than the multiple update links among the multiple initial replication links and the multiple intermediate links.

[0188] In one embodiment, the plurality of intermediate links includes a first intermediate link, and the first generation module includes:

[0189] The first generation unit is used to randomly generate the first path corresponding to the first replication node;

[0190] The selection unit is used to randomly select the next node of the first path according to the transition probability, wherein the transition probability is calculated by global pheromone, and the next node is one of the plurality of replicated nodes;

[0191] The second generation unit is used to generate the first intermediate link based on the first path and the next node of the first path;

[0192] The third update unit is used to update the global pheromone based on the length of the first intermediate link.

[0193] In one embodiment, the link communication device 300 further includes:

[0194] The third acquisition module is used to acquire the communication parameters of each of the multiple replication links included in the replication communication network;

[0195] The second calculation module is used to calculate the link health index of each of the multiple replicated links based on the communication parameters of the multiple replicated links.

[0196] The second determining module is used to determine the first optimization strategy corresponding to the first replication link based on the first mapping relationship. The first replication link is the replication link whose link health index is less than a set health threshold among the plurality of replication links. The first mapping relationship includes a plurality of communication parameters and an optimization strategy corresponding to each communication parameter.

[0197] The adjustment module is used to adjust the communication parameters of the first replicated link based on the first optimization strategy.

[0198] In one embodiment, the link communication device 300 further includes:

[0199] The fourth acquisition module is used to acquire the multiplexing communication parameters of each initial link in the multiple initial links included in the initial communication network. The multiplexing communication parameters are used to characterize the parameters that can be reused for the corresponding initial link.

[0200] The second generation module is used to generate a multiplexing matrix corresponding to each initial link based on the multiplexing communication parameters of the multiple initial links;

[0201] The clustering module is used to cluster the multiple reuse matrices corresponding to the multiple initial links to obtain multiple clusters and the multiple reuse granularity corresponding to each cluster, wherein each cluster includes at least one multiple reuse matrix;

[0202] The third determining module is used to determine the number of copies of each initial node in the initial communication network based on the reuse granularity corresponding to the multiple clusters;

[0203] The replication module is used to replicate each of the multiple initial nodes based on the replication number of the multiple initial nodes, thereby obtaining the multiple replicated nodes.

[0204] In one embodiment, the initial communication network includes a second initial node, and the third determining module includes:

[0205] The third determining unit is used to determine the multiplexing granularity corresponding to the second initial node, wherein the multiplexing granularity of the second initial node includes the multiplexing granularity corresponding to the cluster of the initial link where the second initial node is located;

[0206] The second acquisition unit is used to acquire the service type and / or requirement parameters of the second initial node;

[0207] The fourth determining unit is used to determine the number of replications corresponding to the second initial node according to the second mapping relationship. The second mapping relationship includes multiple replication numbers, as well as the service type and / or requirement parameters corresponding to each replication number, and the reuse granularity.

[0208] The link communication device provided in this embodiment of the invention can realize the various processes of the above-described link communication method, with one-to-one correspondence of technical features and the same technical effect. To avoid repetition, it will not be described again here.

[0209] It should be noted that the link communication device in the embodiments of the present invention can be a device, or it can be a component, integrated circuit, or chip in an electronic device.

[0210] This invention also provides an electronic device, including: a processor, a memory, and a program stored in the memory and executable on the processor, wherein the program, when executed by the processor, implements the above-described functionality. Figure 1The various processes of the link communication method embodiment shown can achieve the same technical effect, and will not be described again here to avoid repetition.

[0211] For details, see Figure 4 As shown, this embodiment of the invention also provides an electronic device, including a bus 401, a transceiver 402, an antenna 403, a bus interface 404, a processor 405, and a memory 406.

[0212] The transceiver 402 is used to obtain the initial attribute vector of each replica node among multiple replica nodes in the replica communication network. The initial attribute vector is used to characterize the communication status and / or resource usage of the corresponding replica node. Each replica node is replicated through the initial node in the initial communication network. The replica communication network is used for link segmentation multiplexing.

[0213] The processor 405 is configured to determine the initial node corresponding to each of the plurality of replica nodes based on the initial attribute vectors of the replica nodes.

[0214] The processor 405 is further configured to update the initial attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes, and each replica node is configured to perform link communication based on the parameters corresponding to the updated attribute vector.

[0215] In one embodiment, the plurality of replica nodes includes a first replica node, and determining the initial node corresponding to each replica node based on the initial attribute vectors of the plurality of replica nodes includes:

[0216] Obtain the first attribute vectors of multiple initial nodes in the initial communication network;

[0217] Calculate the first similarity distance between the initial attribute vector of the first replicated node and the first attribute vectors of the plurality of initial nodes;

[0218] The initial node with the highest first similarity distance among the plurality of initial nodes is determined as the first initial node corresponding to the first copied node.

[0219] In one embodiment, updating the initial attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes includes:

[0220] Update the initial attribute vector of the first replicated node to the first attribute vector of the first initial node;

[0221] A second copy node is determined, wherein the similarity distance between the initial attribute vector of the second copy node and the first attribute vector of the first initial node is less than the first similarity distance;

[0222] The initial attribute vector of the second replicated node is updated to a weighted vector, which is the weighted value of the initial attribute vector of the second replicated node and the first attribute vector of the first initial node.

[0223] In one embodiment, the processor 405 is further configured to repeatedly determine the initial node corresponding to each replica node based on the updated attribute vectors of the plurality of replica nodes, and update the attribute vector of each replica node based on the first attribute vector of the initial node corresponding to the plurality of replica nodes, until the attribute vector of each replica node is stopped when a first preset condition is met.

[0224] The first preset condition includes the following:

[0225] The number of updates has reached a set threshold;

[0226] The average similarity distance is less than a set distance threshold. The average similarity distance is the average of the similarity distances of the multiple replicated nodes. The similarity distance of the first replicated node is the similarity distance between the attribute vector of the first replicated node and the first attribute vector of the corresponding initial node.

[0227] In one embodiment, the transceiver 402 is further configured to acquire multiple initial replication links corresponding to the multiple replication nodes, wherein each initial replication link is used for communication between replication nodes;

[0228] The processor 405 is also used to generate multiple intermediate links based on the ant colony optimization algorithm, and the multiple intermediate links are used for communication between replication nodes.

[0229] The processor 405 is also configured to calculate the fitness value of each of the plurality of initial replication links and the plurality of intermediate links, the fitness value being used to characterize the communication status of the link;

[0230] The processor 405 is further configured to filter and obtain multiple update links, wherein the fitness values ​​of the multiple update links are greater than the fitness values ​​of other links, and the other links are links other than the multiple update links among the multiple initial replication links and the multiple intermediate links.

[0231] In one embodiment, the plurality of intermediate links includes a first intermediate link, and the generation of the plurality of intermediate links based on the ant colony optimization algorithm includes:

[0232] Randomly generate the first path corresponding to the first replicated node;

[0233] The next node of the first path is randomly selected based on the transition probability, wherein the transition probability is calculated using global pheromones, and the next node is one of the plurality of replicated nodes;

[0234] The first intermediate link is generated based on the first path and the next node of the first path;

[0235] The global pheromone is updated based on the length of the first intermediate link.

[0236] In one embodiment, the transceiver 402 is further configured to acquire communication parameters of each of the plurality of replication links included in the replication communication network;

[0237] The processor 405 is also configured to calculate the link health index of each of the multiple replicated links based on the communication parameters of the multiple replicated links.

[0238] The processor 405 is further configured to determine a first optimization strategy corresponding to the first replication link based on the first mapping relationship. The first replication link is a replication link among the plurality of replication links whose link health index is less than a set health threshold. The first mapping relationship includes a plurality of communication parameters and an optimization strategy corresponding to each communication parameter.

[0239] The processor 405 is further configured to adjust the communication parameters of the first replicated link based on the first optimization strategy.

[0240] In one embodiment, the transceiver 402 is further configured to acquire multiplexing communication parameters of each initial link among the plurality of initial links included in the initial communication network, wherein the multiplexing communication parameters are used to characterize the parameter situation that the corresponding initial link can be multiplexed.

[0241] The processor 405 is further configured to generate a multiplexing matrix corresponding to each initial link based on the multiplexing communication parameters of the plurality of initial links;

[0242] The processor 405 is further configured to cluster the multiplexing matrices corresponding to the plurality of initial links to obtain a plurality of clusters and a multiplexing granularity corresponding to each cluster, wherein each cluster includes at least one multiplexing matrix;

[0243] The processor 405 is further configured to determine the number of copies of each initial node in the initial communication network based on the reuse granularity corresponding to the plurality of clusters;

[0244] The processor 405 is further configured to replicate each initial node based on the replication number of the multiple initial nodes, thereby obtaining the multiple replicated nodes.

[0245] In one embodiment, the initial communication network includes a second initial node, and determining the number of replications for each initial node in the initial communication network based on the reuse granularity corresponding to the plurality of clusters includes:

[0246] Determine the reuse granularity corresponding to the second initial node. The reuse granularity of the second initial node includes the reuse granularity corresponding to the cluster of the initial link where the second initial node is located.

[0247] Obtain the service type and / or required parameters of the second initial node;

[0248] The number of replications corresponding to the second initial node is determined based on the second mapping relationship, which includes multiple replication numbers, the service type and / or requirement parameters corresponding to each replication number, and the reuse granularity.

[0249] exist Figure 4 In this document, a bus architecture (represented by bus 401) is used. Bus 401 may include any number of interconnected buses and bridges, linking various circuits including one or more processors represented by processor 405 and memory represented by memory 406. Bus 401 may also link various other circuits such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. Bus interface 404 provides an interface between bus 401 and transceiver 402. Transceiver 402 may be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 405 is transmitted over a wireless medium via antenna 403, which further receives data and transmits data to processor 405.

[0250] Processor 405 is responsible for managing bus 401 and general processing, and can also provide various functions, including timing, peripheral interface, voltage regulation, power management, and other control functions. Memory 406 can be used to store data used by processor 405 during operation.

[0251] Optionally, the processor 405 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or a graphics processing unit (GPU).

[0252] This invention also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the above-described functions. Figure 1 The various processes of the corresponding link communication method embodiments, which can achieve the same technical effect, will not be described again here to avoid repetition. The computer-readable storage medium mentioned includes, for example, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0253] The present invention also provides a computer program product, including computer instructions that, when executed by a processor, implement the above-described... Figure 1 The various processes of the corresponding link communication method embodiments can achieve the same technical effect, and will not be described again here to avoid repetition.

[0254] In the embodiments of this invention, the terms "first," "second," etc., are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to these processes, methods, products, or devices. Additionally, the use of "and / or" in this application indicates at least one of the connected objects, such as A and / or B and / or C, representing eight possibilities: A alone, B alone, C alone, both A and B present, both B and C present, both A and C present, and A, B, and C present.

[0255] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Unless otherwise specified, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element.

[0256] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, air conditioner, or second terminal device, etc.) to execute the methods of the various embodiments of this application.

[0257] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method of link communication, characterized by, The method comprises: obtaining an initial attribute vector of each of a plurality of replicated nodes in a replicated communication network, the initial attribute vector being used to represent a communication condition and / or a resource usage condition of the corresponding replicated node, each of the replicated nodes being replicated by an initial node in an initial communication network; determining a corresponding initial node of each of the replicated nodes based on the initial attribute vector of the plurality of replicated nodes; updating the initial attribute vector of each of the replicated nodes based on a first attribute vector of the corresponding initial node of the plurality of replicated nodes.

2. The method of claim 1, wherein, The plurality of replicated nodes comprises a first replicated node, and the determining of the corresponding initial node of each of the replicated nodes based on the initial attribute vector of the plurality of replicated nodes comprises: obtaining a first attribute vector of a plurality of initial nodes in the initial communication network; calculating a first similarity distance between the initial attribute vector of the first replicated node and the first attribute vector of the plurality of initial nodes; determining that an initial node with the highest first similarity distance in the plurality of initial nodes is a first initial node corresponding to the first replicated node.

3. The method of claim 2, wherein, The updating of the initial attribute vector of each of the replicated nodes based on the first attribute vector of the corresponding initial node of the plurality of replicated nodes comprises: updating the initial attribute vector of the first replicated node to the first attribute vector of the first initial node; determining a second replicated node, a similarity distance between the initial attribute vector of the second replicated node and the first attribute vector of the first initial node being less than the first similarity distance; updating the initial attribute vector of the second replicated node to a weighted vector, the weighted vector being a weighted value of the initial attribute vector of the second replicated node and the first attribute vector of the first initial node.

4. The method of claim 3, wherein, The method further comprises: repeating the determining of the corresponding initial node of each of the replicated nodes based on the updated attribute vector of the plurality of replicated nodes and the updating of the attribute vector of each of the replicated nodes based on the first attribute vector of the corresponding initial node of the plurality of replicated nodes until the attribute vector of each of the replicated nodes is stopped from being updated when a first preset condition is met; wherein the first preset condition comprises one of: a number of updates reaching a set number threshold; an average similarity distance being less than a set distance threshold, the average similarity distance being an average value of the similarity distances of the plurality of replicated nodes, and the similarity distance of the first replicated node being a similarity distance between the attribute vector of the first replicated node and the first attribute vector of the corresponding initial node of the first replicated node.

5. The method of any one of claims 1 to 4, wherein, After the updating of the initial attribute vector of each of the replicated nodes based on the first attribute vector of the corresponding initial node of the plurality of replicated nodes, the method further comprises: obtaining a plurality of initial replication links corresponding to the plurality of replicated nodes, each of the initial replication links being used for communication between the replicated nodes; generating a plurality of intermediate links based on an ant colony optimization algorithm, the plurality of intermediate links being used for communication between the replicated nodes; calculating a fitness value of each link in the plurality of initial replication links and the plurality of intermediate links, the fitness value being used to represent a communication condition of the link; screening a plurality of updated links, the fitness value of the plurality of updated links being greater than the fitness value of other links, the other links being links other than the plurality of updated links in the plurality of initial replication links and the plurality of intermediate links.

6. The method of claim 5, wherein, The plurality of intermediate links includes a first intermediate link, and the plurality of intermediate links is generated based on an ant colony optimization algorithm, including: randomly generating a first path corresponding to a first replication node; randomly selecting a next node of the first path according to a transition probability, the transition probability being calculated by a global pheromone, the next node being one of the plurality of replication nodes; generating the first intermediate link based on the first path and the next node of the first path; updating the global pheromone based on the length of the first intermediate link.

7. The method of any one of claims 1 to 4, wherein, After the first attribute vector of the initial node corresponding to the plurality of replication nodes is updated to obtain the initial attribute vector of each replication node, the method further includes: obtaining a communication parameter of each replication link included in the replication communication network; calculating a link health index of each replication link based on the communication parameter of the plurality of replication links; determining a first optimization strategy corresponding to a first replication link based on a first mapping relationship, the first replication link being a replication link with a link health index less than a set health threshold in the plurality of replication links, the first mapping relationship including a plurality of communication parameters and an optimization strategy corresponding to each communication parameter; adjusting the communication parameter of the first replication link based on the first optimization strategy.

8. The method of any one of claims 1 to 4, wherein, Before the initial attribute vector of each replication node in the replication communication network is obtained, the method further includes: obtaining a multiplexing communication parameter of each initial link included in the initial communication network, the multiplexing communication parameter being used to represent a parameter condition that the corresponding initial link can multiplex; generating a multiplexing matrix corresponding to each initial link based on the multiplexing communication parameter of the plurality of initial links; clustering the multiplexing matrix corresponding to the plurality of initial links to obtain a plurality of clustering clusters and a multiplexing granularity corresponding to each clustering cluster, each clustering cluster including at least one multiplexing matrix; determining a replication number of each initial node in the initial communication network based on the multiplexing granularity corresponding to the plurality of clustering clusters; replicating each initial node based on the replication number of the plurality of initial nodes to obtain the plurality of replication nodes.

9. The method of claim 8, wherein, The initial communication network includes a second initial node, and the replication number of each initial node in the initial communication network is determined based on the multiplexing granularity corresponding to the plurality of clustering clusters, including: determining a multiplexing granularity corresponding to the second initial node, the multiplexing granularity of the second initial node being a multiplexing granularity corresponding to a clustering cluster including an initial link where the second initial node is located; obtaining a service type and / or requirement parameter of the second initial node; The second initial node corresponds to a copy quantity according to a second mapping relationship. The second mapping relationship includes a plurality of copy quantities, and a service type and / or a requirement parameter corresponding to each copy quantity, and a multiplexing granularity.

10. A link communication device, characterized by Comprise: The first acquisition module is used for acquiring an initial attribute vector of each of a plurality of copy nodes in a copy communication network, the initial attribute vector is used for representing a communication condition and / or a resource use condition of the corresponding copy node, each of the copy nodes is obtained by copying an initial node in an initial communication network, and the copy communication network is used for link segment multiplexing. The first determination module is used for determining an initial node corresponding to each of the copy nodes based on the initial attribute vectors of the plurality of copy nodes. The update module is used for updating the initial attribute vector of each of the copy nodes based on the first attribute vector of the initial node corresponding to each of the copy nodes, and each of the copy nodes is used for link communication based on a parameter corresponding to the updated attribute vector.

11. An electronic device, comprising: Comprise a transceiver and a processor, The transceiver is used for acquiring an initial attribute vector of each of a plurality of copy nodes in a copy communication network, the initial attribute vector is used for representing a communication condition and / or a resource use condition of the corresponding copy node, each of the copy nodes is obtained by copying an initial node in an initial communication network, and the copy communication network is used for link segment multiplexing. The processor is used for determining an initial node corresponding to each of the copy nodes based on the initial attribute vectors of the plurality of copy nodes. The processor is also used for updating the initial attribute vector of each of the copy nodes based on the first attribute vector of the initial node corresponding to each of the copy nodes, and each of the copy nodes is used for link communication based on a parameter corresponding to the updated attribute vector.

12. An electronic device, comprising: Comprise: The processor, the memory and the program stored on the memory and executable on the processor, when the program is executed by the processor, the steps of the link communication method in any one of claims 1 to 9 are implemented.

13. A computer-readable storage medium, characterized in that, The computer program is stored on the computer readable storage medium, and when the computer program is executed by the processor, the steps of the link communication method in any one of claims 1 to 9 are implemented.

14. A computer program product, characterised in that, Comprise computer instructions, when the computer instructions are executed by the processor, the steps of the link communication method in any one of claims 1 to 9 are implemented.

Images