Methods, apparatus, devices and storage media for recommending virtual network topologies

Abstract

This application relates to a method, apparatus, device, and storage medium for recommending virtual network topologies. The method includes: obtaining virtual network configuration information sent by a client, including the number of traffic path nodes to be deployed, one traffic ingress node, multiple traffic egress nodes, and service nodes; constructing a virtual network topology and a link cost model based on the virtual network configuration information; adding virtual endpoint nodes to the link cost model and adding virtual links between the virtual endpoint nodes and each traffic egress node to ensure that the costs of each virtual link are the same; iteratively executing a preset loop to update the pheromone concentration values ​​of the links in the link cost model; updating the pheromone concentration values ​​of the links in the link cost model again according to preset link importance weights and the latest link pheromone concentration values; and finding the optimal target virtual network topology from the updated link cost model. This method is beneficial for improving the transmission efficiency of virtual networks.

Description

Technical Field

[0001] This application relates to the field of cloud computing technology, and in particular to a method, apparatus, computer device, computer-readable storage medium, and computer program product for recommending virtual network topologies. Background Technology

[0002] In the current information technology field, the demand for cloud adoption among individual users, enterprises, and organizations is growing rapidly. Not only can cloud platforms be used to host traditional business applications, but innovative cloud-based solutions can also be explored. For example, the promotion of NFV (Network Functions Virtualization) transforms traditional hardware-dependent network functions (such as firewalls, load balancing, and network address translation) into software solutions that can run on standard virtual servers, enabling faster deployment and scaling of network services to respond to changing business needs.

[0003] However, when customers deploy various components based on a large number of cloud virtual machines, it is difficult to guarantee the transmission efficiency of the backbone link and the accessibility of network topology planning or expansion, resulting in low transmission efficiency of cloud virtual network topology. Summary of the Invention

[0004] Therefore, it is necessary to provide a method, apparatus, computer device, computer-readable storage medium, and computer program product that can improve the transmission efficiency of virtual networks, addressing the aforementioned technical problems.

[0005] Firstly, this application provides a method for recommending virtual network topologies, including:

[0006] Obtain the virtual network configuration information sent by the client. The virtual network configuration information includes the number of traffic path nodes to be deployed, as well as one traffic ingress node, multiple traffic egress nodes, and service nodes.

[0007] Based on the virtual network configuration information, a virtual network topology is constructed, and a link cost model is built based on the virtual network topology;

[0008] In the link cost model, a virtual endpoint node is added, and a virtual link is added between the virtual endpoint node and each traffic egress node to make the cost of each virtual link the same.

[0009] The preset loop body is executed iteratively to update the pheromone concentration value of the link in the link cost model until the preset number of iterations is reached, at which point the execution of the preset loop body stops.

[0010] Based on the preset link importance weights and the latest link pheromone concentration values, the pheromone concentration values ​​of the links in the link cost model are updated again.

[0011] Find the optimal target virtual network topology from the updated link cost model;

[0012] When the loop body is executed, the following steps are performed:

[0013] Multiple nodes are selected from multiple business nodes to construct a node set. The access probability of the traffic entry node and each node in the node set is determined. The node with the highest access probability is identified as the target node to be accessed. The target node is accessed, and the accessed target node is removed from the node set. The node set is updated, and the access probability of the target node and each node in the updated node set is determined. The steps of identifying the node with the highest access probability as the target node to be accessed are returned until the number of accessed nodes reaches a preset threshold. Finally, the pheromone concentration value of the link in the link cost model is updated.

[0014] In one embodiment, constructing a link cost model based on a virtual network topology includes:

[0015] Determine the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes;

[0016] Based on latency and bandwidth, determine the link cost between traffic ingress nodes, traffic egress nodes, and service nodes;

[0017] Based on the virtual network topology and link cost, construct a link cost model.

[0018] In one embodiment, determining the latency and bandwidth between the traffic ingress node, the traffic egress node, and the service node includes:

[0019] Based on virtual network configuration information, collect virtual machine information, network connection information and communication paths of traffic ingress nodes, traffic egress nodes and service nodes;

[0020] Based on the virtual machine information, network connection information, and communication paths of the traffic ingress node, traffic egress node, and business node, the communication between the traffic ingress node, traffic egress node, and business node is simulated, and communication data during the communication process is collected.

[0021] Based on communication data, determine the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes.

[0022] In one embodiment, finding the optimal target virtual network topology from the updated link cost model includes:

[0023] Based on the pheromone concentration and link cost of the links in the updated link cost model, a breadth-first search is used to find the target virtual network topology with the minimum sum of link costs for the path from the traffic ingress node to the traffic egress node.

[0024] In one embodiment, updating the pheromone concentration value of a link in the link cost model includes:

[0025] Determine the pheromone concentration increment on the visited links in the link cost model;

[0026] Update the pheromone concentration value of the visited links based on the pheromone concentration increment and the preset pheromone evaporation coefficient.

[0027] Based on the pheromone concentration values ​​on unvisited links in the link cost model and the preset pheromone evaporation coefficient, update the pheromone concentration values ​​of unvisited links.

[0028] In one embodiment, before determining the latency and bandwidth between the traffic ingress node, the traffic egress node, and the service node based on communication data, the method further includes:

[0029] Data preprocessing is performed on communication data, including at least one of dirty data removal and missing data processing.

[0030] Secondly, this application also provides a virtual network topology device, comprising:

[0031] The configuration information acquisition module is used to acquire the virtual network configuration information sent by the client. The virtual network configuration information includes the number of traffic path nodes to be deployed, as well as one traffic inlet node, multiple traffic outlet nodes, and service nodes.

[0032] The model building module is used to construct a virtual network topology based on the virtual network configuration information, and to build a link cost model based on the virtual network topology. In the link cost model, virtual endpoint nodes are added, and virtual links are added between the virtual endpoint nodes and each traffic egress node so that the cost of each virtual link is the same.

[0033] The pheromone concentration adjustment module iteratively executes a preset loop to update the pheromone concentration value of the links in the link cost model until a preset number of iterations is reached, at which point the execution of the preset loop stops. Based on the preset link importance weights and the latest link pheromone concentration value, the module updates the link pheromone concentration value in the link cost model again. When the loop is executed, the following operations are performed: multiple nodes are selected from multiple business nodes to construct a node set; the access probability of the traffic entry node and each node in the node set is determined; the node with the highest access probability is identified as the target node to be accessed; the target node is accessed; the accessed target node is removed from the node set; the node set is updated; the access probability of the target node and each node in the updated node set is determined; the step of identifying the node with the highest access probability as the target node to be accessed is returned; this process continues until the number of accessed nodes reaches a preset threshold, at which point the pheromone concentration value of the links in the link cost model is updated.

[0034] The topology determination module is used to find the optimal target virtual network topology from the updated link cost model.

[0035] Thirdly, this application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in any of the above-described embodiments of the virtual network topology recommendation method.

[0036] Fourthly, this application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps in any of the above-described embodiments of the virtual network topology recommendation method.

[0037] Fifthly, this application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps in any of the above-described virtual network topology recommendation method embodiments.

[0038] The aforementioned virtual network topology recommendation method obtains virtual network configuration information sent by the client, constructs a link cost model based on this information, adds a virtual endpoint node to the model, and adds virtual links between the virtual endpoint node and each traffic exit node, with all virtual links having the same cost. Firstly, addressing the issue of multiple traffic exits in cloud users' virtual networks, this method, by creating a virtual endpoint node during the problem modeling phase and connecting each traffic exit to it with the same link cost, helps balance link costs and neutralize the impact of newly added spurious links on determining the optimal target virtual network topology, thus improving the accuracy of topology determination. Secondly, the improved algorithm loop constructs a set of nodes to be accessed by filtering from multiple business nodes, altering the node access mechanism. It determines the access node by determining the probability of the current access node accessing a node in the set of nodes to be accessed, updating the pheromone concentration value of the links in the link cost model. This promotes diversity in target network topology exploration, reduces the possibility of getting trapped in local optima, and improves the algorithm's convergence speed. Thirdly, by iteratively executing the loop, the pheromone concentration values ​​of the links in the link cost model are updated. Based on the preset link importance weights and the updated link pheromone concentration values, the pheromone concentration values ​​of the links in the link cost model are updated again. The optimal target virtual network topology is then found from the updated link cost model. Thus, determining the target virtual network topology through link pheromone concentration is beneficial for improving the transmission efficiency of the virtual network. Simultaneously, by combining preset link importance requirements, the solution difficulty is fully considered under the condition of extremely complex and large client networks. This allows the preset link importance to fully intervene in the topology optimization process, making the results more customer-oriented and improving the matching degree between the target network topology and client needs. Attached Figure Description

[0039] To more clearly illustrate the technical solutions in the embodiments of this application or related technologies, the drawings used in the description of the embodiments of this application or related technologies will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0040] Figure 1 This is a diagram illustrating the application environment of a virtual network topology recommendation method in one embodiment.

[0041] Figure 2 This is a flowchart illustrating a virtual network topology recommendation method in one embodiment;

[0042] Figure 3This is a schematic diagram of a link cost model in one embodiment;

[0043] Figure 4 This is a schematic diagram illustrating the important weights of a link in one embodiment;

[0044] Figure 5 This is a flowchart illustrating a virtual network topology recommendation method in another embodiment;

[0045] Figure 6 This is a flowchart illustrating the virtual network topology recommendation method in yet another embodiment;

[0046] Figure 7 This is a schematic diagram of a virtual network topology in one embodiment;

[0047] Figure 8 This is a schematic diagram of the target virtual network topology in one embodiment;

[0048] Figure 9 This is a flowchart illustrating the virtual network topology recommendation method in another embodiment;

[0049] Figure 10 This is a block diagram of a virtual network topology device in one embodiment;

[0050] Figure 11 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0051] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0052] The virtual network topology recommendation method provided in this application embodiment can be applied to, for example, Figure 1 In the application environment shown, client 102 communicates with server 104 via a network. Server 104 monitors and manages the virtual network in the cloud service platform. The virtual network consists of several cloud virtual devices, each corresponding to a node. Nodes can be traffic ingress nodes, traffic transit nodes, traffic egress nodes, or business nodes, etc. The data storage system can store the data that server 104 needs to process. The data storage system can be integrated on server 104 or placed on the cloud or other network servers.

[0053] Specifically, client 102 can send virtual network configuration information to server 104. Server 104 obtains the virtual network configuration information sent by the client, including the number of traffic path nodes to be deployed, one traffic ingress node, multiple traffic egress nodes, and service nodes. Next, based on the virtual network configuration information, a virtual network topology is constructed, and a link cost model is built based on the virtual network topology. Virtual endpoint nodes are added to the link cost model, and virtual links are added between the virtual endpoint nodes and each traffic egress node to ensure that the cost of each virtual link is the same. Then, a preset loop is iteratively executed to update the pheromone concentration values ​​of the links in the link cost model until a preset number of iterations is reached, at which point the execution of the preset loop stops. When the loop is executed, the following steps are implemented: from multiple... Multiple nodes are selected from the business nodes to construct a node set. The access probability of the traffic entry node and each node in the node set is determined. The node with the highest access probability is identified as the target node to be accessed. The target node is accessed, and the accessed target node is removed from the node set. The node set is updated, and the access probability of the target node and each node in the updated node set is determined. The step of identifying the node with the highest access probability as the target node to be accessed is returned until the number of accessed nodes reaches a preset threshold. The pheromone concentration value of the link in the link cost model is updated. Finally, based on the preset link importance weight and the latest link pheromone concentration value, the pheromone concentration value of the link in the link cost model is updated again. The optimal target virtual network topology is found from the updated link cost model.

[0054] Client 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. IoT devices can include smart speakers, smart TVs, smart air conditioners, smart in-vehicle devices, and projection devices. Portable wearable devices can include smartwatches, smart bracelets, and head-mounted displays. Head-mounted displays can be virtual reality (VR) devices, augmented reality (AR) devices, and smart glasses. Server 104 can be a standalone physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing cloud computing services.

[0055] In one exemplary embodiment, such as Figure 2 As shown, a virtual network topology recommendation method is provided, which can be applied to... Figure 1 Taking server 104 as an example, the explanation includes the following steps S100 to S700. Among them:

[0056] S100: Obtain the virtual network configuration information sent by the client. The virtual network configuration information includes the number of nodes through which the traffic to be deployed passes, as well as one traffic ingress node, multiple traffic egress nodes, and service nodes.

[0057] Virtual network configuration information refers to the device information within a user's virtual network deployed on a public cloud. This information characterizes the client's network topology optimization needs. Virtual network configuration information includes the IDs of the virtual machines corresponding to the traffic ingress nodes, traffic egress nodes, and service nodes in the user's virtual network, as well as the number of nodes through which the traffic to be deployed will pass. A virtual network is a network resource in a public cloud environment. Traditional network functions (such as firewalls, load balancers, and intrusion detection systems) can be implemented in software using NFV technology and run as virtual machines on standard server hardware.

[0058] In practical applications, cloud service providers can offer a graphical virtual network configuration interface, allowing users to create, configure, and manage virtual networks. This involves obtaining the virtual machine IDs of the traffic ingress nodes, traffic egress nodes, and service nodes configured by the client through the virtual network configuration interface, as well as the number of traffic transit nodes to be deployed. For example, the virtual network configuration information includes 2 traffic transit nodes, 1 traffic ingress node, 10 service nodes, and 2 traffic egress nodes to be deployed. The client's requirement is to select 2 target service nodes from the 10 service nodes to perform the function of traffic transit nodes, and to provide a network topology including the traffic ingress node, traffic transit nodes, and traffic egress nodes.

[0059] S200 constructs a virtual network topology based on virtual network configuration information and builds a link cost model based on the virtual network topology.

[0060] Among them, the link cost model is used to understand the structure of virtual networks and optimize network performance.

[0061] In practical applications, based on the information of virtual network devices in the virtual network configuration, the network information (IP addresses, etc.) of the devices in the virtual network can be determined, the configuration files of the devices in the virtual network can be obtained, the connection relationships between devices can be determined, and the virtual network topology can be constructed using a network topology drawing tool. Metrics for evaluating link costs (such as bandwidth, latency, packet loss rate, and link stability) can be determined, and the link costs can be determined using preset cost evaluation standards. Cost annotations can then be added to the virtual network topology to construct a link cost model.

[0062] S300 adds a virtual endpoint node to the link cost model and adds a virtual link between the virtual endpoint node and each traffic egress node to make the cost of each virtual link the same.

[0063] In practical applications, to respond to users' topology optimization needs (such as users needing to select target business nodes from business nodes to undertake traffic transit node functions, providing a topology including traffic ingress nodes, traffic egress nodes, and traffic transit nodes), and addressing the issue of multiple traffic egress points in the user's virtual network, a virtual endpoint can be hypothesized. Each traffic egress node is connected to the virtual endpoint, and the virtual link costs are the same. This allows for better balancing of the cost of each link when the optimized topology is determined, reducing the impact of newly added links on the optimization topology solution. For example... Figure 3 As shown, the link cost model includes one traffic ingress node, ten service nodes, and two traffic egress nodes. The links between the nodes are represented by the dashed lines in the figure. The boxes represent the default service nodes and links. `timeN` represents the communication cost between nodes, such as the cost, and `N` is used to distinguish the communication cost between different devices. A virtual endpoint node is added to the link cost model. Virtual links are added between the virtual endpoint node and each traffic egress node. The link cost of the virtual links is configured to be the same, such as a preset link cost value. This preset link cost value can be set based on experience.

[0064] S400: Iteratively execute the preset loop body to update the pheromone concentration value of the link in the link cost model until the preset number of iterations is reached, then stop executing the preset loop body.

[0065] Among them, the pheromone concentration value is used to characterize the quality of the link and is a key factor affecting the performance of the ant colony algorithm.

[0066] In practical applications, a solution loop can be pre-set to determine the optimal path with the highest transmission efficiency, which includes traffic ingress nodes, traffic egress nodes, and target service nodes (traffic path nodes). In this embodiment, the optimal path solution loop is designed based on the ant colony algorithm. This loop is executed iteratively until a preset number of iterations (e.g., 200) is reached, at which point execution stops. This updates the pheromone concentration value of the link in the link cost model using the ant colony algorithm, thus facilitating the determination of the optimal path.

[0067] S500 updates the pheromone concentration value of the link in the link cost model again based on the preset link importance weight and the latest link pheromone concentration value.

[0068] The preset link importance weights are the importance weights of the links between the traffic ingress node, the traffic path nodes to be deployed, and the traffic egress node, used to characterize the importance of the service, such as... Figure 4The diagram illustrates the key weights of the links between the traffic ingress node, the traffic path nodes to be deployed, and the traffic egress node. The preset link weights can be configured by the client and obtained through virtual network configuration information. That is, the virtual network configuration information sent by the client also includes the preset link weights. `timeN` represents the communication cost between nodes, and `N` is used to distinguish the communication cost between different devices.

[0069] In practical applications, the preset link importance weights are obtained from the virtual network configuration information. For each link, the pheromone concentration value of the link is multiplied by the corresponding link weight to obtain the updated link pheromone concentration value.

[0070] S600 finds the optimal target virtual network topology from the updated link cost model.

[0071] The optimal target virtual network topology can be the topology corresponding to the path with the highest transmission efficiency from the traffic ingress node to the traffic egress node.

[0072] In practical applications, after updating the link cost model, the optimal target virtual network topology can be found from the link cost model using a search algorithm. For example, the Dijkstra algorithm or A* search algorithm can be used to find the path with the minimum sum of link costs from the traffic ingress node to the traffic egress node, and the target virtual network topology can be drawn based on this path.

[0073] S700, when the loop body is executed, the following steps are implemented: multiple nodes are selected from multiple business nodes, a node set is constructed, the access probability of the traffic entry node and each node in the node set is determined, the node with the highest access probability is determined as the target node to be accessed, the target node is accessed, the accessed target node is removed from the node set, the node set is updated, the access probability of the target node and each node in the updated node set is determined, the step of determining the node with the highest access probability as the target node to be accessed is returned, until the number of accessed nodes reaches a preset number threshold, and the pheromone concentration value of the link in the link cost model is updated.

[0074] In this embodiment, the mechanism for ants visiting nodes in the ant colony algorithm is improved. The ant colony algorithm is a heuristic optimization algorithm that simulates the foraging behavior of ants in nature. The access paths of ants to nodes represent feasible solutions to the target virtual network topology recommendation problem. Through the positive feedback of the ant colony, the pheromone concentration values ​​of links in the link cost model are updated, resulting in higher pheromone concentration values ​​for links in better paths. In practical applications, algorithm initialization is performed: initializing the number of ants, the pheromone concentration values ​​of links in the link cost model, and the pheromone evaporation coefficient of the ant colony algorithm, and defining a heuristic function based on link cost. This helps improve the diversity of ant exploration, reduces the possibility of getting trapped in local optima, and improves the convergence speed of the ant colony algorithm. When the loop body is executed, the following processing is performed:

[0075] Multiple nodes are randomly selected from the business nodes to construct a node set, representing the set of nodes to be visited by the ant colony in this loop. The access probability of the traffic entry node and each node in the node set is determined. Specifically, for each ant, the access probability of the traffic entry node and any node in the node set is determined, which can be calculated by applying the state transition equation of the ant colony algorithm based on the link cost between the traffic entry node and that node. The node with the highest access probability is identified as the target node to be visited. This target node is visited, and the visited target node is removed from the node set, updating the node set. The access probability of the target node and each node in the updated node set is determined again. Specifically, for each ant, the access probability of the target node and any node in the node set is determined, which can be calculated by applying the state transition equation of the ant colony algorithm based on the link cost between the target node and that node. The process of identifying the node with the highest access probability as the target node to be visited is repeated until the number of visited nodes reaches a preset threshold. The preset threshold is determined based on the number of nodes the traffic to be deployed will pass through. Specifically, the preset threshold can be the sum of the number of traffic entry nodes (1 node), traffic entry nodes, and traffic exit nodes (counted as 1 node). For each ant, if the number of nodes visited by the ant meets the preset threshold, it indicates that the ant has reached the traffic exit node, and the next node to be visited is the virtual endpoint node. The ant stops searching.

[0076] Once the number of nodes visited by all ants in the ant colony meets the preset threshold, the single ant path exploration ends. Based on the access paths of all ants, the pheromone concentration value of the link in the link cost model is updated according to the pheromone evaporation coefficient of the ant colony algorithm, so that the pheromone concentration value of the links visited by multiple ants is higher.

[0077] In the aforementioned virtual network topology recommendation method, virtual network configuration information sent by the client is obtained. A link cost model is constructed based on this information, and a virtual endpoint node is added to the model. Virtual links are then added between the virtual endpoint node and each traffic exit node, with all virtual links having the same cost. Firstly, addressing the issue of multiple traffic exits in cloud users' virtual networks, a virtual endpoint node is created during the problem modeling phase. Connecting each traffic exit to this virtual endpoint node with the same link cost helps balance link costs and mitigate the impact of newly added spurious links on determining the optimal target virtual network topology, thus improving the accuracy of topology determination. Secondly, an improved algorithm loop is designed. By selecting nodes to be accessed from multiple business nodes and constructing a set of nodes to be accessed, the node access mechanism is altered. The probability of the current accessing node accessing a node in the set of nodes to be accessed is determined to identify the accessing node and update the pheromone concentration value of the link in the link cost model. This promotes diversity in target network topology exploration, reduces the possibility of getting trapped in local optima, and improves the algorithm's convergence speed. Thirdly, by iteratively executing the loop, the pheromone concentration values ​​of the links in the link cost model are updated. Based on the preset link importance weights and the updated link pheromone concentration values, the pheromone concentration values ​​of the links in the link cost model are updated again. The optimal target virtual network topology is then found from the updated link cost model. Thus, determining the target virtual network topology through link pheromone concentration is beneficial for improving the transmission efficiency of the virtual network. Simultaneously, by combining preset link importance requirements, the solution difficulty is fully considered under the condition of extremely complex and large client networks. This allows the preset link importance to fully intervene in the topology optimization process, making the results more customer-oriented and improving the matching degree between the target network topology and client needs.

[0078] In an exemplary embodiment, S200 includes S220 to S260. Wherein:

[0079] S220 determines the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes.

[0080] In practical applications, based on network configuration information, the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and business nodes can be queried in the management console of the cloud service.

[0081] S240 determines the link cost between traffic ingress nodes, traffic egress nodes, and service nodes based on latency and bandwidth.

[0082] In practical applications, weights can be pre-assigned to latency and bandwidth, and the link cost between traffic ingress nodes, traffic egress nodes, and service nodes can be determined by weighted summation of latency and bandwidth.

[0083] In other embodiments, the link cost of a fake link can be determined by taking the average link cost between the traffic ingress node, the traffic egress node, and the service node as the link cost of the fake link.

[0084] S260 constructs a link cost model based on the virtual network topology and link cost.

[0085] In practical applications, a network model can be constructed based on the virtual network topology using a topology model building tool, and link cost labels can be added to the links in the network model to obtain a link cost model.

[0086] In this embodiment, by determining the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes, the link cost between these nodes is calculated, and a link cost model is constructed. This is beneficial for optimizing the topology based on the link cost to reduce the possibility of network congestion, reduce data transmission latency, and improve data transmission efficiency.

[0087] To improve the accuracy of topology determination, in one exemplary embodiment, such as Figure 5 As shown, S240 also includes S242 to S246. Wherein:

[0088] S242, based on virtual network configuration information, collects virtual machine information, network connection information and communication paths of traffic ingress nodes, traffic egress nodes and service nodes respectively.

[0089] The virtual machine information may include the virtual machine's ID, IP address, MAC address, and host machine ID. Network connection information may include the virtual machine's network port.

[0090] In practical applications, the virtual machine information, network connection information, and communication paths of traffic ingress nodes, traffic egress nodes, and business nodes can be collected based on the virtual configuration of the virtual machine, such as security groups and network access control lists.

[0091] S244 simulates the communication between traffic ingress nodes, traffic egress nodes, and service nodes based on their respective virtual machine information, network connection information, and communication paths, and collects communication data during the communication process.

[0092] The communication data may include traffic data received by the virtual machine, traffic data sent, data packet size, data packet sending frequency, etc.

[0093] In practical applications, virtual switching software can be used to simulate business communication between traffic ingress nodes, traffic egress nodes, and business nodes, based on their respective virtual machine information, network connection information, and communication paths. For example, the packet in and packet out functions of OpenVSwitch (Open Virtual Switching Standard) can be used to simulate communication between devices on the business network plane and collect communication data during the communication process.

[0094] S246, based on communication data, determines the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes.

[0095] In practical applications, the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes are calculated based on the time difference between sending and receiving data packets in the communication data and the amount of data transmitted per unit time.

[0096] In this embodiment, based on virtual network configuration information, the communication between traffic ingress nodes, traffic egress nodes, and service nodes is simulated, communication data is collected, and the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes are determined based on the communication data. This is beneficial for topology optimization based on latency and bandwidth data, thereby improving the accuracy of determining the optimal topology.

[0097] In one exemplary embodiment, such as Figure 6 As shown, S600 includes S620. Wherein:

[0098] S620, based on the pheromone concentration value and link cost of the links in the updated link cost model, uses breadth-first search to find the target virtual network topology with the minimum sum of link costs for the path from the traffic ingress node to the traffic egress node in the updated link cost model.

[0099] In practical applications, based on the pheromone concentration and link cost of the links in the updated link cost model, and using a breadth-first search algorithm, starting from the traffic ingress node, the optimal next node is searched. When selecting the next node, the probability of selecting each unvisited neighbor is determined based on the pheromone concentration and link cost of the links between the current node and its unvisited neighboring nodes. The neighboring node with the highest probability is selected as the next node. The sum of link costs from the starting point to this node is calculated and recorded. The priority queue is updated based on the sum of link costs, iterating to the virtual endpoint node, and outputting the path with the minimum sum of link costs between the traffic ingress node and the traffic egress node. Figure 7As shown, the sum of link costs for the path from the traffic ingress node -> service node 4 -> service node 5 to the traffic egress node is minimized. This determines that the traffic passes through service nodes 4 and 5. `timeN` represents the communication cost between nodes, and `N` distinguishes the communication cost between different devices. Based on this path, the target virtual network topology is determined, as follows: Figure 8 As shown, timeN represents the communication cost between nodes, and N is used to distinguish the communication cost between different devices.

[0100] In this embodiment, a breadth-first search is performed on the link cost model based on the pheromone concentration value and link cost of the links. This increases the exploratory nature of node selection in the breadth-first search, improves search efficiency, enhances the robustness and flexibility of the algorithm, and improves its adaptability to optimizing complex user network models. This, in turn, helps to improve the accuracy of determining the target virtual network model.

[0101] To accelerate the convergence of path solving, in one example implementation, such as Figure 9 As shown, S400 includes S420 to S460. Wherein:

[0102] S420, based on the pheromone concentration value of the link in the link cost model and the preset pheromone evaporation coefficient, update the pheromone concentration value of the link in the link cost model.

[0103] S440, determine the increment of pheromone concentration value on the visited links in the link cost model.

[0104] S460, update the pheromone concentration value of the visited links based on the pheromone concentration value increment.

[0105] After a single ant colony exploration is completed, the pheromone concentration value of each link in the link cost model is decayed according to the preset pheromone volatilization coefficient, and the pheromone concentration value of the link in the link cost model is updated to simulate the natural volatilization of pheromones over time.

[0106] Based on the positive feedback mechanism of the ant colony algorithm, the pheromone concentration value on links visited by more ants in the link cost model is increased. For each ant, the sum of link costs in the ant's node access path is determined based on the ant's node access path. The incremental pheromone concentration value on the links already visited by the ant is then determined based on the sum of link costs in the ant's node access path.

[0107] After determining the pheromone concentration increment value of the visited links, the pheromone concentration value of the visited links is updated according to the pheromone concentration increment value.

[0108] After updating the pheromone concentration values ​​of visited links based on the pheromone concentration increment, the current optimal path with the minimum sum of link costs between the current traffic ingress node and the traffic egress node is found. Based on the sum of link costs of the current optimal path, the pheromone concentration increment value of the current optimal path is determined (e.g., 1 / sum of link costs). The pheromone concentration values ​​of the links on the current optimal path are then updated based on this increment value to enhance the pheromone concentration of the current optimal solution.

[0109] In this embodiment, for the visited links in the link cost model, the pheromone concentration value increment is determined and the visited link value is updated to strengthen the pheromone concentration value on the visited links. For unvisited links, the pheromone concentration value of the unvisited links is updated according to the pheromone evaporation coefficient to reduce the pheromone concentration value of the unvisited links, thereby improving the speed of determining the optimal path.

[0110] To improve the accuracy of processing communication data, in an exemplary embodiment, S245 is included before S246. Wherein:

[0111] S245, perform data preprocessing on the communication data, the data preprocessing including at least one of dirty data removal and missing data processing.

[0112] Dirty data can be inaccurate, incomplete, or duplicate data in communication data.

[0113] In practical applications, network fluctuations (such as packet loss and latency fluctuations) can affect the quality of communication data during simulated communication and data acquisition. To improve the reliability and accuracy of communication data, hash tables or sets are used to store information of processed data packets to avoid duplicate recording. The Z-score method is used to detect and remove outliers in the communication data. Missing data is filled using interpolation methods, such as weighted filling based on the values ​​of the sampling points before and after the missing data.

[0114] In this embodiment, data cleaning preprocessing, which involves removing dirty data and processing missing data, helps to reduce the impact of network vibration on the collected communication data. By processing missing data to fill in the missing data, the vibration of the original network is reflected in the overall stability of the link, which helps to improve the accuracy of determining the latency and bandwidth between nodes.

[0115] To provide a clearer explanation of the virtual network topology recommendation method provided in this application, a specific embodiment is described below, which includes the following steps:

[0116] S1, obtain the virtual network configuration information sent by the client. The virtual network configuration information includes the number of nodes through which the traffic to be deployed passes, as well as one traffic ingress node, multiple traffic egress nodes, and service nodes.

[0117] S2, based on virtual network configuration information, collects the virtual machine information, network connection information, and communication path of each traffic ingress node, traffic egress node, and service node. Based on the virtual machine information, network connection information, and communication path of each traffic ingress node, traffic egress node, and service node, it simulates the communication between the traffic ingress node, traffic egress node, and service node, collects the communication data during the communication process, and performs data preprocessing on the communication data. The data preprocessing includes at least one of dirty data removal and missing data processing.

[0118] S3, based on communication data, determines the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes. Based on the latency and bandwidth, it determines the link cost between traffic ingress nodes, traffic egress nodes, and service nodes. Based on the virtual network topology and link cost, it constructs a link cost model.

[0119] S4. Add a virtual endpoint node to the link cost model, and add a virtual link between the virtual endpoint node and each traffic exit node so that the cost of each virtual link is the same.

[0120] S5, iteratively execute the preset loop body to update the pheromone concentration value of the link in the link cost model until the preset number of iterations is reached, then stop executing the preset loop body.

[0121] When the loop is executed, it performs the following steps: select multiple nodes from multiple business nodes, construct a node set, determine the access probability of the traffic entry node and each node in the node set, determine the node with the highest access probability as the target node to be accessed, access the target node, remove the accessed target node from the node set, update the node set, determine the access probability of the target node and each node in the updated node set, return to the step of determining the node with the highest access probability as the target node to be accessed, until the number of accessed nodes reaches a preset threshold, and update the pheromone concentration value of the link in the link cost model.

[0122] S6. Based on the preset link importance weights and the latest link pheromone concentration values, update the link pheromone concentration values ​​in the link cost model again.

[0123] S7. Based on the pheromone concentration value and link cost of the links in the updated link cost model, a breadth-first search is used to find the target virtual network topology with the minimum sum of link costs for the path from the traffic ingress node to the traffic egress node in the updated link cost model.

[0124] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0125] In one exemplary embodiment, such as Figure 10 As shown, a virtual network topology device 600 is provided, including: a configuration information acquisition module 610, a model construction module 620, a pheromone concentration value adjustment module 630, and a topology determination module 640, wherein:

[0126] The configuration information acquisition module 610 is used to acquire the virtual network configuration information sent by the client. The virtual network configuration information includes the number of nodes through which the traffic to be deployed passes, as well as one traffic ingress node, multiple traffic egress nodes, and service nodes.

[0127] The model building module 620 is used to build a virtual network topology based on the virtual network configuration information, and to build a link cost model based on the virtual network topology; a virtual endpoint node is added to the link cost model, and a virtual link is added between the virtual endpoint node and each traffic exit node so that the cost of each virtual link is the same.

[0128] The pheromone concentration adjustment module 630 is used to iteratively execute a preset loop to update the pheromone concentration value of the link in the link cost model until a preset number of iterations is reached, at which point the execution of the preset loop stops. Based on the preset link importance weight and the latest link pheromone concentration value, the pheromone concentration value of the link in the link cost model is updated again. When the loop is executed, the following operations are performed: multiple nodes are selected from multiple business nodes, a node set is constructed, the access probability of the traffic entry node and each node in the node set is determined, the node with the highest access probability is determined as the target node to be accessed, the target node is accessed, the accessed target node is removed from the node set, the node set is updated, the access probability of the target node and each node in the updated node set is determined, and the step of determining the node with the highest access probability as the target node to be accessed is returned, until the number of accessed nodes reaches a preset number threshold, and the pheromone concentration value of the link in the link cost model is updated.

[0129] The topology determination module 640 is used to find the optimal target virtual network topology from the updated link cost model.

[0130] In an exemplary embodiment, the model building module 620 is further configured to determine the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes; determine the link cost between traffic ingress nodes, traffic egress nodes, and service nodes based on the latency and bandwidth; and construct a link cost model based on the virtual network topology and link cost.

[0131] In an exemplary embodiment, the model building module 620 is further configured to: collect virtual machine information, network connection information, and communication paths of traffic ingress nodes, traffic egress nodes, and service nodes based on virtual network configuration information; simulate communication between traffic ingress nodes, traffic egress nodes, and service nodes based on the virtual machine information, network connection information, and communication paths of traffic ingress nodes, traffic egress nodes, and service nodes, and collect communication data during the communication process; and determine the latency and bandwidth between traffic ingress nodes, traffic egress nodes, and service nodes based on the communication data.

[0132] In an exemplary embodiment, the pheromone concentration adjustment module 630 is further configured to, based on the pheromone concentration values ​​and link costs of the links in the updated link cost model, use breadth-first search to find the target virtual network topology with the minimum sum of link costs for the path from the traffic ingress node to the traffic egress node in the updated link cost model.

[0133] In an exemplary embodiment, the virtual network topology recommendation device 600 further includes a data preprocessing module 650 for preprocessing communication data, the data preprocessing including at least one of dirty data removal and missing data processing.

[0134] Each module in the aforementioned virtual network topology device 600 can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in or independent of the processor in a computer device, or stored in the memory of a computer device as software, so that the processor can call and execute the operations corresponding to each module.

[0135] In one exemplary embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as follows: Figure 11As shown, this computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operating system and computer programs stored in the non-volatile storage media. The database stores data. The I / O interfaces are used for exchanging information between the processor and external devices. The communication interface is used for communicating with external terminals via a network connection. When executed by the processor, the computer program implements a virtual network topology recommendation method.

[0136] Those skilled in the art will understand that Figure 11 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0137] In one exemplary embodiment, a computer device is provided, including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps in any of the above embodiments of the virtual network topology recommendation method.

[0138] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the steps in any of the above embodiments of the virtual network topology recommendation method.

[0139] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in any of the above embodiments of the virtual network topology recommendation method.

[0140] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of the relevant data must comply with relevant regulations.

[0141] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, artificial intelligence (AI) processors, etc., and are not limited to these.

[0142] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0143] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A virtual network topology recommendation method, characterized in that, The method includes: Obtain the virtual network configuration information sent by the client. The virtual network configuration information includes the number of traffic path nodes to be deployed, as well as one traffic inlet node, multiple traffic outlet nodes, and service nodes. Based on the virtual network configuration information, a virtual network topology is constructed, and a link cost model is built based on the virtual network topology; A virtual endpoint node is added to the link cost model, and a virtual link is added between the virtual endpoint node and each of the traffic exit nodes so that the cost of each virtual link is the same. The preset loop body is executed iteratively to update the pheromone concentration value of the link in the link cost model until the preset number of iterations is reached, at which point the execution of the preset loop body stops. Based on the preset link importance weights and the latest pheromone concentration values ​​of the links, the pheromone concentration values ​​of the links in the link cost model are updated again. Find the optimal target virtual network topology from the updated link cost model; When the loop body is executed, the following steps are performed: Multiple nodes are selected from the multiple business nodes to construct a node set. The access probability of the traffic entry node and each node in the node set is determined. The node with the highest access probability is identified as the target node to be accessed. The target node is accessed. The accessed target node is removed from the node set. The node set is updated. The access probability of the target node and each node in the updated node set is determined. The step of identifying the node with the highest access probability as the target node to be accessed is returned. This process continues until the number of accessed nodes reaches a preset threshold. Finally, the pheromone concentration value of the link in the link cost model is updated.

2. The method according to claim 1, characterized in that, The link cost model constructed based on the virtual network topology includes: Determine the latency and bandwidth between the traffic ingress node, the traffic egress node, and the service node; Based on the latency and the bandwidth, determine the link cost between the traffic ingress node, the traffic egress node, and the service node; Based on the virtual network topology and the link cost, a link cost model is constructed.

3. The method according to claim 2, characterized in that, Determining the latency and bandwidth between the traffic ingress node, the traffic egress node, and the service node includes: Based on the virtual network configuration information, collect the virtual machine information, network connection information and communication path of the traffic ingress node, the traffic egress node and the service node respectively; Based on the virtual machine information, network connection information, and communication path of the traffic ingress node, the traffic egress node, and the service node, the communication between the traffic ingress node, the traffic egress node, and the service node is simulated, and communication data during the communication process is collected. Based on the communication data, the latency and bandwidth between the traffic ingress node, the traffic egress node, and the service node are determined.

4. The method according to any one of claims 1 to 3, characterized in that, The process of finding the optimal target virtual network topology from the updated link cost model includes: Based on the pheromone concentration value and link cost of the links in the updated link cost model, a target virtual network topology with the minimum sum of link costs for the path from the traffic ingress node to the traffic egress node is found through a breadth-first search in the updated link cost model.

5. The method according to claim 3, characterized in that, The updating of the pheromone concentration value of the link in the link cost model includes: The pheromone concentration value of the link in the link cost model is updated based on the pheromone concentration value of the link in the link cost model and the preset pheromone evaporation coefficient. Determine the pheromone concentration increment on the visited links in the link cost model; The pheromone concentration value of the visited links is updated based on the increment of the pheromone concentration value.

6. The method according to claim 5, characterized in that, Before determining the latency and bandwidth between the traffic ingress node, the traffic egress node, and the service node based on the communication data, the method further includes: The communication data is preprocessed, and the preprocessing includes at least one of dirty data removal and missing data processing.

7. A virtual network topology recommendation device, characterized in that, The device includes: The configuration information acquisition module is used to acquire the virtual network configuration information sent by the client. The virtual network configuration information includes the number of traffic path nodes to be deployed, as well as one traffic inlet node, multiple traffic outlet nodes, and service nodes. The model building module is used to build a virtual network topology based on the virtual network configuration information, and to build a link cost model based on the virtual network topology; to add virtual endpoint nodes to the link cost model, and to add virtual links between the virtual endpoint nodes and each traffic exit node, so that the cost of each virtual link is the same. The pheromone concentration adjustment module is used to iteratively execute a preset loop to update the pheromone concentration value of the links in the link cost model until a preset number of iterations is reached, at which point the execution of the preset loop stops. Based on the preset link importance weights and the latest pheromone concentration value of the links, the pheromone concentration value of the links in the link cost model is updated again. When the loop is executed, the following operations are performed: multiple nodes are selected from multiple service nodes to construct a node set; the access probability of the traffic entry node and each node in the node set is determined; the node with the highest access probability is identified as the target node to be accessed; the target node is accessed; the accessed target node is removed from the node set; the node set is updated; the access probability of the target node and each node in the updated node set is determined; the step of identifying the node with the highest access probability as the target node to be accessed is returned; this process continues until the number of accessed nodes reaches a preset threshold, at which point the pheromone concentration value of the links in the link cost model is updated. The topology determination module is used to find the optimal target virtual network topology from the updated link cost model.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images