SDN (Software Defined Network) dynamic traffic adjustment method

Abstract

The invention discloses an SDN (Software Defined Network) dynamic traffic adjustment method, which comprises the following steps: firstly, acquiring a network traffic distribution state, analyzing the uniformity of traffic distribution, secondly, evaluating the availability rate of a core network link based on link information, evaluating the availability rate of nodes in a network cloud resource pool server cluster in combination with server node information and a network topology weight, and evaluating the availability rate of the nodes in the network cloud resource pool server cluster in combination with a service QoS (Quality of Service) index. And a final flow forwarding path is decided by taking time delay minimization as a target. By performing multi-dimensional dynamic evaluation on the network flow distribution, the link availability rate and the server node availability rate, and performing path decision by taking the time delay minimization as the target in combination with the service QoS requirement, the network resource utilization rate can be improved, the throughput can be enhanced, and the network congestion can be reduced.

Description

Technical Field

[0001] This application relates to the field of network communication technology, and more specifically, to an SDN dynamic traffic adjustment method. Background Technology

[0002] In the field of communication network technology, the convergence of various access technologies such as 5G, satellite communication, and WiFi has led to the formation of heterogeneous converged networks supporting diverse applications. While this network architecture improves connectivity flexibility and coverage, the diversity of access methods and the differences in service requirements pose significant challenges to the overall network resource scheduling. Software-defined networking (SDN) technology, as an emerging network management paradigm, separates the data plane from the control plane and utilizes a centralized controller to manage server cluster load balancing and link load balancing in software, providing opportunities for dynamic resource scheduling in converged networks. SDN technology enables efficient interconnection between large server clusters, thereby promoting the sharing and flexible allocation of network resources. Based on this, existing technologies have developed SDN traffic scheduling strategies such as dynamic routing path planning based on traffic classification, which have improved network throughput and user service quality to some extent.

[0003] However, existing SDN traffic scheduling strategies typically employ a minimum-cost approach when optimizing routing paths. This method often focuses on a single metric, lacking a comprehensive consideration of overall network traffic status, link availability, server performance availability, and current traffic bandwidth requirements. Therefore, existing strategies struggle to adapt to the dynamically changing service demands and multi-dimensional resource constraints in heterogeneous converged networks, potentially leading to uneven resource allocation, link congestion, or degraded service quality, thus limiting further improvements in overall network performance. Summary of the Invention

[0004] The purpose of this application is to overcome the shortcomings of existing technologies and provide an SDN dynamic traffic adjustment method. By dynamically evaluating network traffic distribution, link availability and server node availability from multiple dimensions, and combining business QoS requirements to make path decisions with the goal of minimizing latency, this method demonstrates the technical effects of improving network resource utilization, increasing throughput and reducing network congestion.

[0005] The objective of this application is achieved through the following technical solution: Firstly, this application proposes an SDN dynamic traffic adjustment method, the method comprising: Obtain network traffic distribution status and analyze the uniformity of traffic distribution; Assess the availability of core network links based on link information; The availability of nodes in the network cloud resource pool server cluster is evaluated by combining server node information and network topology weights. The availability of nodes includes the probability that the node is working normally and the current resource availability of the node. Based on link availability, node availability, and service QoS metrics, and with the goal of minimizing latency, the final traffic forwarding path is determined.

[0006] In one possible embodiment, the step of analyzing the uniformity of the flow distribution includes: Collect the flow direction of all data flows in the heterogeneous converged network, and construct the data flow matrix of switch nodes, server nodes, and link data flows. Covariance analysis is performed on the switch node data flow matrix, server node data flow matrix, and link data flow matrix to identify whether the current traffic distribution is uniform.

[0007] In one possible embodiment, the data flow matrix of the switch node is represented as X=F×S, the data flow matrix of the server node is represented as Y=F×H, and the data flow matrix of the link is represented as Z=F×E, where F represents the set of data flows, S represents the set of switches, H represents the set of server nodes, and E represents the set of links. Matrix element values Represents data stream Does the flow pass through the switch node? Matrix element values Represents data stream Does it flow through server nodes? Matrix element values Represents data stream Does it flow through the link? .

[0008] In one possible implementation, the link availability is calculated using the following formula: ,in It is a link The remaining bandwidth, It is a coefficient that takes a value between 0 and 1. Used to measure the maximum load on the link. It is a link Bandwidth during construction.

[0009] In one possible implementation, link criticality The calculation formula is: ,in, It is the average load, representing all flows through the link. The ratio of the sum of bandwidth to the number of data streams.

[0010] In one possible embodiment, the remaining bandwidth The calculation formula is: .

[0011] In one possible embodiment, the probability of a node functioning correctly is calculated using the following formula: ,in It is a node Normal working hours, It is a node Failure time.

[0012] In one possible embodiment, the formula for calculating the current resource availability of a node is: ,in , , These are weighting coefficients. , , These represent the maximum computing resources, storage resources, and bandwidth resources of server node h, respectively. , , This indicates the computing, storage, and bandwidth resources that have been used by server node h.

[0013] In one possible embodiment, the formula for calculating the service node availability is: ,in, It represents the current resource availability of the node. It represents the probability that the node is functioning normally.

[0014] In one possible embodiment, the formula for calculating network topology weights is: ,in Indicates server node and The path between them and These are server nodes and The node availability rate.

[0015] The main solution and its various further alternatives described above can be freely combined to form multiple solutions, all of which are solutions that can be adopted and are claimed in this application; furthermore, the (non-conflicting alternatives) can also be freely combined with each other and with other alternatives. Those skilled in the art, after understanding the solution of this application, will realize from the prior art and common general knowledge that there are many combinations, all of which are technical solutions to be protected in this application, and will not be exhaustively listed here.

[0016] This application discloses an SDN dynamic traffic adjustment method. First, it acquires the network traffic distribution status and analyzes the uniformity of the traffic distribution. Second, it evaluates the availability of core network links based on link information. Third, it evaluates the availability of nodes in the network cloud resource pool server cluster by combining server node information and network topology weights. Finally, it combines service QoS indicators and, with the goal of minimizing latency, determines the final traffic forwarding path. By dynamically evaluating network traffic distribution, link availability, and server node availability from multiple dimensions, and by making path decisions with the goal of minimizing latency in conjunction with service QoS requirements, it can improve network resource utilization, enhance throughput, and reduce network congestion. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart illustrating an SDN dynamic traffic adjustment method proposed in an embodiment of this application is shown.

[0019] Figure 2 This illustration shows a flowchart of an SDN dynamic traffic adjustment strategy based on traffic distribution and remaining resource availability proposed in an embodiment of this application. Detailed Implementation

[0020] The following specific examples illustrate the implementation of this application. Those skilled in the art can easily understand other advantages and effects of this application from the content disclosed in this specification. This application can also be implemented or applied through other different specific embodiments, and various details in this specification can also be modified or changed based on different viewpoints and applications without departing from the spirit of this application. It should be noted that, unless otherwise specified, the following embodiments and features in the embodiments can be combined with each other.

[0021] Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0022] In existing technologies, heterogeneous converged networks integrating multiple access methods such as 5G, satellite, and WiFi support diverse applications, posing greater challenges to overall network resource scheduling. Software-defined networking (SDN) technology offers opportunities for dynamic resource scheduling in converged networks. It separates the data plane and control plane, and uses a centralized controller to manage server cluster load balancing and link load balancing in software. This provides efficient interconnection methods among large server clusters, enabling the sharing and flexible allocation of network resources. Based on this, SDN traffic scheduling strategies such as dynamic routing path planning based on traffic classification have emerged, improving network throughput and user service quality. However, current SDN traffic scheduling strategies typically use a minimum-cost approach to optimize routing paths, rarely considering the overall network traffic status, link availability, server performance availability, and current traffic bandwidth requirements to optimize routing paths.

[0023] Therefore, to address the aforementioned technical issues, this application proposes an SDN dynamic traffic adjustment method. Based on a measurement of network traffic distribution and considering the uneven distribution characteristics, it selects available network paths by combining existing link availability and service node availability. Furthermore, by incorporating user service QoS requirements, it optimizes the network routing strategy with the goal of minimizing path latency, significantly improving the average link utilization and throughput of the heterogeneous converged network and effectively reducing network congestion.

[0024] Please refer to Figure 1 , Figure 1 This paper presents a flowchart illustrating an SDN dynamic traffic adjustment method according to an embodiment of this application. The method includes: Step S1: Obtain the network traffic distribution status and analyze the uniformity of the traffic distribution.

[0025] The SDN controller collects real-time flow information of all data flows in a multi-access heterogeneous converged network through its southbound interface. To quantitatively analyze the distribution of network traffic, three key data flow matrices are constructed: a switch node data flow matrix X, composed of a data flow set F and a switch set S; a server node data flow matrix Y, composed of a data flow set F and a server node set H; and a link data flow matrix Z, composed of a data flow set F and a link set E. The elements in the matrices are binary states (0 or 1), used to identify whether a specific data flow passes through a particular switch, server node, or link. By performing covariance analysis on these three matrices, the uniformity of current network traffic distribution can be accurately identified. If the analysis results indicate uneven traffic distribution, it means that some devices or links in the network are at risk of overload, while other resources may be idle.

[0026] The steps for analyzing the uniformity of flow distribution include: Collect the flow direction of all data flows in the heterogeneous converged network, and construct the data flow matrix of switch nodes, server nodes, and link data flows. Covariance analysis is performed on the switch node data flow matrix, server node data flow matrix, and link data flow matrix to identify whether the current traffic distribution is uniform.

[0027] First, the SDN controller periodically or triggeredly collects real-time flow information of all active data flows in the heterogeneous converged network. This information records in detail the network elements traversed by each data flow. Based on the collected raw data, the system constructs three core data flow correlation matrices to represent the mapping relationship between traffic and network infrastructure using mathematical models: The data flow matrix of a switch node is represented as X=F×S, the data flow matrix of a server node is represented as Y=F×H, and the data flow matrix of a link is represented as Z=F×E, where F represents the set of data flows, S represents the set of switches, H represents the set of server nodes, and E represents the set of links. Matrix element values Represents data stream Does the flow pass through the switch node? Matrix element values Represents data stream Does it flow through server nodes? Matrix element values Represents data stream Does it flow through the link? .

[0028] After constructing the aforementioned matrices, covariance analysis is performed on these three matrices to comprehensively assess the uniformity of traffic distribution. By calculating the covariance of specific dimensions within or between these matrices, the concentration or dispersion of data flow on network devices and links can be effectively measured. If the analysis results show a high positive correlation in the covariance, it means that the current network traffic distribution is uneven, with a few switches, servers, or links carrying the vast majority of the data flow. This can easily lead to performance bottlenecks and single points of failure, while other network resources may be inefficiently idle. Conversely, if the covariance results approach zero or show a low correlation, it indicates that the traffic distribution is relatively uniform, and network resources are being utilized relatively evenly.

[0029] Step S2: Evaluate the availability of core network links based on link information.

[0030] The formula for calculating link availability is: ,in It is a link The remaining bandwidth, It is a coefficient that takes a value between 0 and 1. Used to measure the maximum load on the link. It is a link Bandwidth during construction.

[0031] Link criticality The calculation formula is: ,in, It is the average load, representing all flows through the link. The ratio of the sum of bandwidth to the number of data streams.

[0032] Remaining bandwidth The calculation formula is: , It is the criticality of the link. It is an indicator variable relating data flow and links.

[0033] Step S2 dynamically assesses the real-time carrying capacity and health status of each link in the network. First, it defines the link criticality. Its value is the historical average load of the link. With link construction bandwidth The ratio reflects the inherent load level of the link. Subsequently, a configurable coefficient is introduced. (0< ≤1) to define a reasonable maximum load threshold for the link, i.e. Remaining bandwidth of the link This is calculated by subtracting the total bandwidth of all traffic currently flowing through the link from the maximum load threshold. Ultimately, the link availability is... Defined as remaining bandwidth With maximum available bandwidth The ratio. This indicator It can sensitively reflect the real-time congestion level of a link; the higher the value, the more abundant the resources available for new data streams on that link.

[0034] Step S3: Combine server node information and network topology weights to evaluate the availability of nodes in the network cloud resource pool server cluster. The availability of a node includes the probability that the node is working normally and the current resource availability of the node.

[0035] Probability of a node working properly The calculation formula is: ,in It is a node Normal working hours, It is a node Failure time.

[0036] Current resource availability of the node The calculation formula is: ,in , , These are weighting coefficients. , , These represent the maximum computing resources, storage resources, and bandwidth resources of server node h, respectively. , , This indicates the computing, storage, and bandwidth resources that are already occupied by server node h. This represents the summation of all data streams.

[0037] The formula for calculating service node availability is: ,in, It represents the current resource availability of the node. It represents the probability that the node is functioning normally.

[0038] The formula for calculating network topology weights is: ,in Indicates server node and The path between them and These are server nodes and The node availability rate.

[0039] Step S3 performs a comprehensive evaluation from two dimensions: server node reliability and real-time resource load. Node availability consists of two parts: one is the probability that the node is functioning normally. The first is a reliability indicator, obtained by statistically analyzing the ratio of a node's historical uptime to its total uptime; the second is the node's current resource availability, which is calculated using a weighted model to determine the node's overall utilization rate across the three core resources of computing, storage, and bandwidth. This model comprehensively considers the computing resources already occupied by tasks. Storage resources and bandwidth resources and respectively with the total resources corresponding to the nodes. , , Compare, and then use weighting coefficients , , (satisfy The weighted sum is then applied. Ultimately, the service node availability rate is calculated. Depend on Therefore, when evaluating the global weight of a path between nodes, not only the link itself is considered, but also the node availability of the servers at both ends of the path, i.e., the path weight. This enables a unified assessment of resources along the entire "server-link-server" path.

[0040] Step S4: Based on the availability of links, the availability of nodes, and service QoS indicators, and with the goal of minimizing latency, determine the final traffic forwarding path.

[0041] Step S4 is the final decision-making step. The SDN controller integrates the network-wide information obtained in the preceding steps, including link availability (representing link status), node availability (representing server health and load), and the constructed integrated path weights. Based on this, the service requests received by the controller will include their specific QoS (Quality of Service) level requirements. The controller first filters all potential paths from the network topology that can meet the basic QoS requirements of the service, forming a set of available paths. Then, within this set, with minimizing end-to-end latency as the core optimization objective, and comprehensively considering the congestion risk implied by the path weights and the server processing capacity, it determines the optimal traffic forwarding path. Subsequently, the SDN controller implements the decision result to the corresponding network devices by distributing flow tables, thereby completing dynamic and optimized traffic scheduling, effectively improving the overall network throughput and reducing congestion risk.

[0042] Figure 2 This diagram illustrates a flowchart of an SDN dynamic traffic adjustment strategy based on traffic distribution and remaining resource availability proposed in this application. The entire process begins with a comprehensive understanding of the network status, with inputs including key parameters such as data flow size, network topology, link bandwidth, service node normal operation probability, and resource utilization. First, the system constructs three data flow matrices—switch, link, and server—based on the input raw data. Covariance analysis is then performed on these matrices to assess the uniformity of traffic distribution, thereby identifying the risk of load imbalance in the network. Second, the system performs link availability assessment and service node availability assessment in parallel. Link assessment quantifies available bandwidth resources by calculating link criticality and maximum load, while node assessment determines service capability by combining node normal operation probability and real-time resource utilization.

[0043] Ultimately, the above analysis results—including the identified set of available paths and their resource status, as well as the QoS level requirements of user services—are jointly input into the path decision module. This module, with latency minimization as its core optimization objective, determines the final traffic forwarding path from all available paths and outputs the result. This closed-loop process enables dynamic and precise traffic scheduling based on real-time network conditions and service requirements, effectively improving network resource utilization and transmission performance.

[0044] Compared with the prior art, the embodiments of this application have the following beneficial effects: First, traditional path planning algorithms rely solely on the shortest path, which can easily lead to link overload and fail to dynamically distribute traffic. This application analyzes traffic distribution, remaining link bandwidth, and node resource availability in real time, and selects paths with the goal of minimizing latency, avoiding local congestion and thus significantly improving overall bandwidth utilization efficiency.

[0045] Second, traditional methods limit the range of available paths due to over-reliance on shared links, resulting in limited throughput. This application fully utilizes all available links and achieves reasonable traffic allocation through covariance analysis of traffic distribution and resource availability, effectively improving the average throughput of network data flows.

[0046] Third, by dynamically assessing traffic uniformity, link availability, and server node availability, this invention can promptly identify and avoid busy devices or links. By optimizing path selection in conjunction with service QoS requirements, it reduces localized overload and effectively lowers the probability of network congestion.

[0047] The above description is merely a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of this application should be included within the protection scope of this application.

Claims

1. A method for SDN dynamic traffic adjustment, the method comprising: The method comprises: acquiring network traffic distribution state, analyzing uniformity of traffic distribution; evaluating available rate of core network link based on link information; combining server node information and network topology weight to evaluate available rate of node in network cloud resource pool server cluster, the available rate of node including probability of normal work of node and current resource available rate of node; based on available rate of link, available rate of node and service QoS index, and taking minimum delay as target, deciding final traffic forwarding path.

2. The method of claim 1, wherein, The step of analyzing uniformity of traffic distribution comprises: collecting flow direction of all data flows in heterogeneous fusion network, constructing switch node data flow matrix, server node data flow matrix and link data flow matrix; performing covariance analysis on the switch node data flow matrix, server node data flow matrix and link data flow matrix to identify whether current traffic distribution is uniform.

3. The method of claim 2, wherein, The switch node data flow matrix is represented as X=F×S, the server node data flow matrix is represented as Y=F×H, and the link data flow matrix is represented as Z=F×E, wherein F represents data flow set, S represents switch set, H represents server node set, and E represents link set; Matrix element value Representative data flow Whether or not to flow through a switch node Matrix element value Representative data flow Whether or not to flow through a server node Matrix element value Representative data flow Whether or not to flow through a link .

4. The method of claim 3, wherein, The formula for calculating the availability of the link is: wherein is the residual bandwidth of the link , is a coefficient with a value between 0 and 1, used to measure the maximum load of the link, is the bandwidth of the link when it is built, is the criticality of the link.

5. The method of claim 4, wherein, Link criticality The formula for calculating the link criticality is: wherein, is the average load, representing the ratio of the sum of the bandwidth of the traffic flowing through the link to the number of data flows.

6. The method of claim 5, wherein, Residual bandwidth The calculation formula is: , is the association indication variable of data flow and link.

7. The method of claim 1, wherein, The formula for calculating the probability of the node working normally is: wherein is the time of the node working normally, is the time of the node failure.

8. The method of claim 1, wherein, The calculation formula of the current resource availability of the node is: wherein , , is a weight coefficient, , , respectively represent the maximum computing resource, storage resource and bandwidth resource of the server node h, , , represent the computing resource, storage resource and bandwidth resource of the server node h that have been occupied, represent the summation of all data streams.

9. The method of claim 1, wherein, The calculation formula of the service node availability is: wherein, is the current resource availability of the node, is the probability of normal operation of the node.

10. The method of claim 4, wherein, The calculation formula of the network topology weight is: wherein represents the path between the server nodes and , and and are the node availability of the server nodes and respectively.

Images