SD-WAN traffic scheduling and collaborative management method for multi-cloud environment

Abstract

The invention discloses an SD-WAN traffic scheduling and collaborative management method for a multi-cloud environment, and the method comprises the steps: deploying an information collection agent in the multi-cloud environment, collecting the cloud node resource state and SD-WAN link transmission quality information in real time, enabling a constructed joint model to dynamically reflect the real-time state of a network, and breaking through the limitation of a conventional static model; the multi-objective optimization algorithm quickly recalculates different service flow optimal paths according to the update model, so that the flow is helped to avoid a resource shortage area, and the network flexibility and reliability are improved; traffic is classified carefully, a personalized optimization target is set, respective optimal paths are obtained through independent processing of an algorithm, differentiated service guarantee is provided for various services, and the service performance and the user experience are improved; the collaborative management node realizes effective collaboration of the cloud node and the network equipment, sends a scheduling instruction according to an optimal path, can quickly respond and update the instruction when a network state changes, realizes unified scheduling management, ensures dynamic optimization of flow, and improves network collaborative work and management efficiency.

Description

Technical Field

[0001] This invention relates to the field of SD-WAN traffic scheduling technology, and in particular to a method for SD-WAN traffic scheduling and collaborative management in a multi-cloud environment. Background Technology

[0002] With the rapid development of cloud computing technology, enterprises' demands for computing resources are becoming increasingly diverse and complex. A single cloud service provider often struggles to meet their needs in terms of cost, performance, and data compliance. Therefore, multi-cloud environments are gradually becoming the mainstream choice for enterprise IT architecture. Multi-cloud environments integrate resources from multiple different cloud service providers, offering enterprises greater flexibility and scalability. Software-defined wide area network (SD-WAN), as an innovative network technology, applies the concept of software-defined networking (SDN) to wide area network scenarios. By separating the network control plane from the data plane, it achieves flexible control and optimization of wide area network traffic. SD-WAN can dynamically select the optimal path for data transmission based on real-time network conditions and business needs, improving network performance and reliability while reducing network operating costs.

[0003] However, in a multi-cloud environment, traditional SD-WAN traffic scheduling and management methods face many challenges. On the one hand, multi-cloud environments are highly complex and dynamic. The resource status of different cloud nodes (such as CPU utilization and memory usage) fluctuates constantly with changes in business load. At the same time, the quality of network links between cloud nodes (such as latency, packet loss rate, and jitter) is also affected by various factors, such as network congestion and equipment failure. Traditional traffic scheduling methods are often based on static network models and fixed scheduling strategies, which make it difficult to perceive and adapt to the dynamic changes in the multi-cloud environment in real time. This results in inaccurate traffic scheduling, failure to fully utilize network resources, and even network congestion or business interruption.

[0004] On the other hand, traffic entering SD-WAN networks has diverse service types, such as real-time interactive services (e.g., video conferencing, voice calls), big data transmission services (e.g., file backup, data synchronization), and web browsing services. Different service types have very different requirements for network performance. For example, real-time interactive services are very sensitive to latency and jitter, while big data transmission services are more concerned with the bandwidth utilization of the link. Traditional traffic scheduling methods usually adopt a "one-size-fits-all" strategy, which does not fully consider the characteristics and needs of different service types, and cannot provide differentiated service quality assurance for various services, thus affecting the overall performance of the services and the user experience.

[0005] Furthermore, in a multi-cloud environment, there is a lack of effective collaborative management mechanisms between various cloud nodes and SD-WAN network devices. Different cloud service providers may use different management interfaces and protocols, which makes collaborative scheduling difficult. When the network status changes, it is difficult to quickly and accurately send scheduling instructions to relevant devices to achieve dynamic adjustment and optimization of traffic. Summary of the Invention

[0006] In view of this, the present invention proposes an SD-WAN traffic scheduling and collaborative management method for multi-cloud environments, which can effectively solve the shortcomings of existing technologies, such as difficulty in real-time perception and adaptation to the dynamic changes of multi-cloud environments, inability to provide differentiated service quality assurance for various services, and lack of effective collaborative management mechanisms between various cloud nodes and SD-WAN network devices.

[0007] The technical solution of this invention is implemented as follows:

[0008] A method for SD-WAN traffic scheduling and collaborative management in multi-cloud environments, comprising:

[0009] Deploy information collection agents on each cloud node in a multi-cloud environment, and build a joint model of the multi-cloud environment and SD-WAN network based on the information collection agents;

[0010] Traffic entering the SD-WAN network is classified into different service types.

[0011] The optimal scheduling path for traffic of different service types is obtained by processing traffic of different service types based on a multi-objective optimization algorithm.

[0012] The collaborative management node sends scheduling instructions to relevant cloud nodes and SD-WAN network devices based on the optimal scheduling path for traffic of different service types, thereby realizing SD-WAN traffic scheduling and collaborative management.

[0013] As a further optional solution to the aforementioned SD-WAN traffic scheduling and collaborative management method for multi-cloud environments, the deployment of information collection agents on each cloud node in the multi-cloud environment, and the construction of a joint model of the multi-cloud environment and the SD-WAN network based on the information collection agents, specifically includes:

[0014] Information collection agents are deployed on each cloud node in a multi-cloud environment to collect cloud node resource status information in real time, including CPU utilization, memory usage, and network bandwidth availability.

[0015] Collect transmission quality information of each link in the SD-WAN network, and calculate the overall link quality index based on the transmission quality information of each link;

[0016] Based on the collected cloud node resource status information and link comprehensive quality index, a joint model of multi-cloud environment and SD-WAN network is constructed. The joint model is represented by a graph structure, with cloud nodes as vertices and links as edges, and the weight of the edge is the link comprehensive quality index.

[0017] As a further optional solution to the aforementioned SD-WAN traffic scheduling and collaborative management method for multi-cloud environments, the calculation of the link comprehensive quality index based on the transmission quality information of each link is specifically calculated using the following formula:

[0018] ;

[0019] in, This is represented as the overall link quality index. This is represented as link delay. This is expressed as packet loss rate. This is represented as jitter. , , These are the weighting coefficients for link latency, packet loss rate, and jitter, respectively.

[0020] As a further optional solution to the aforementioned SD-WAN traffic scheduling and collaborative management method for multi-cloud environments, the step of classifying traffic entering the SD-WAN network into different service types specifically includes:

[0021] Deep packet inspection is performed on traffic entering the SD-WAN network to extract traffic feature information, including source IP address, destination IP address, source port number, destination port number, protocol type, and packet size distribution.

[0022] Based on traffic characteristics, clustering algorithms are used to categorize traffic into different business types.

[0023] As a further optional solution to the aforementioned SD-WAN traffic scheduling and collaborative management method for multi-cloud environments, the step of processing traffic of different service types based on a multi-objective optimization algorithm to obtain the optimal scheduling path for traffic of different service types specifically includes:

[0024] For traffic of different service types, multiple optimization objectives are set, including minimizing transmission latency, maximizing link utilization, and minimizing network cost;

[0025] Based on the joint model, and combined with traffic service types and optimization objectives, a multi-objective optimization model is constructed;

[0026] A multi-objective optimization algorithm is used to solve the multi-objective optimization model to obtain the optimal scheduling path for traffic of different service types.

[0027] As a further optional solution to the aforementioned SD-WAN traffic scheduling and collaborative management method for multi-cloud environments, the collaborative management node sends scheduling instructions to relevant cloud nodes and SD-WAN network devices according to the optimal scheduling path for traffic of different service types, specifically including:

[0028] The collaborative management node monitors the multi-cloud environment and SD-WAN network status in real time. When the resource status of the cloud node changes or the link transmission quality deteriorates, it sends scheduling instructions to the relevant cloud nodes and SD-WAN network devices according to the optimal scheduling path for traffic of different service types.

[0029] As a further optional solution for the SD-WAN traffic scheduling and collaborative management method for multi-cloud environments, the collaborative management node adopts a distributed architecture, and multiple collaborative management nodes communicate and collaborate on decision-making through message queues.

[0030] An SD-WAN traffic scheduling and collaborative management system for multi-cloud environments includes:

[0031] The information collection agent deployment module is used to deploy information collection agents on various cloud nodes in a multi-cloud environment.

[0032] The joint model building module is used to build joint models of multi-cloud environments and SD-WAN networks based on information gathering agents.

[0033] The traffic service classification module is used to classify traffic entering the SD-WAN network into different service types.

[0034] The multi-objective optimization processing module is used to process traffic of different service types based on multi-objective optimization algorithms to obtain the optimal scheduling path for traffic of different service types.

[0035] The collaborative management and scheduling module is used by the collaborative management node to send scheduling instructions to relevant cloud nodes and SD-WAN network devices according to the optimal scheduling path of traffic of different service types, so as to realize SD-WAN traffic scheduling and collaborative management.

[0036] A computing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the above-described SD-WAN traffic scheduling and collaborative management methods for multi-cloud environments.

[0037] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described SD-WAN traffic scheduling and collaborative management methods for multi-cloud environments.

[0038] The beneficial effects of this invention are as follows: Deploying information collection agents on various cloud nodes in a multi-cloud environment enables real-time and comprehensive collection of resource status information of cloud nodes and transmission quality information of each link in the SD-WAN network. The joint model of the multi-cloud environment and SD-WAN network constructed based on this real-time information accurately reflects the real-time state of the network. Compared to the limitations of traditional methods based on static models and fixed strategies, the joint model can dynamically capture the complex changes in the multi-cloud environment. The multi-objective optimization algorithm, based on the updated model information, quickly recalculates the optimal scheduling path for different service types of traffic, ensuring that traffic can avoid resource-scarce areas in a timely manner and fully utilize other available resources in the network. This effectively adapts to the dynamic changes in the multi-cloud environment, improving the network's flexibility and reliability. Simultaneously, it performs detailed service classification on traffic entering the SD-WAN network, setting personalized optimization objectives based on the characteristics and needs of different service types. The multi-objective optimization algorithm independently processes traffic of different service types to obtain their respective optimal scheduling paths. This differentiated processing approach ensures that each service type obtains the most suitable transmission path and resource allocation for its characteristics within the network, thereby providing precise and differentiated quality of service assurance for various services and significantly improving overall service performance and user experience. Furthermore, effective collaborative management between various cloud nodes and SD-WAN network devices is achieved through collaborative management nodes. Based on the optimal scheduling path for different service types obtained through a multi-objective optimization algorithm, the collaborative management node sends scheduling instructions to relevant cloud nodes and SD-WAN network devices. When network conditions change, such as a decrease in link transmission quality or a change in cloud node resource status, the collaborative management node can respond quickly, re-obtain the optimal scheduling path, and promptly send updated scheduling instructions to relevant devices. This collaborative management mechanism breaks down the differences in management interfaces and protocols between different cloud service providers, achieving unified scheduling and management of network devices in a multi-cloud environment. This ensures that traffic can be dynamically adjusted and optimized according to the optimal path, improving the overall network's collaborative working capability and management efficiency. Attached Figure Description

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

[0040] Figure 1 This is a flowchart of an SD-WAN traffic scheduling and collaborative management method for multi-cloud environments according to the present invention;

[0041] Figure 2This is a schematic diagram illustrating the composition of an SD-WAN traffic scheduling and collaborative management system for multi-cloud environments according to the present invention.

[0042] Figure 3 This is a schematic diagram of the composition of a computing device according to the present invention. Detailed Implementation

[0043] The technical solutions in the embodiments of the present invention will be clearly and completely described below. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] refer to Figures 1 to 3 A method for SD-WAN traffic scheduling and collaborative management in multi-cloud environments, comprising:

[0045] Information collection agents are deployed on various cloud nodes in a multi-cloud environment. A joint model of the multi-cloud environment and the SD-WAN network is then constructed based on these information collection agents. Specifically, this includes:

[0046] Information collection agents are deployed on various cloud nodes in a multi-cloud environment to collect real-time cloud node resource status information, including CPU utilization, memory usage, and network bandwidth availability. The collection period is [not specified]. ,in The system is dynamically adjusted based on the type of cloud node service. For services with high real-time requirements, the adjustment is made accordingly. The value range is 1-5 seconds; for services with low real-time requirements, The value range is 10-30 seconds;

[0047] Collect transmission quality information of each link in the SD-WAN network, and calculate the overall link quality index based on the transmission quality information of each link;

[0048] Based on the collected cloud node resource status information and link comprehensive quality index, a joint model of multi-cloud environment and SD-WAN network is constructed. The joint model is represented by a graph structure, with cloud nodes as vertices and links as edges, and the weight of the edge is the link comprehensive quality index.

[0049] Specifically, information collection agents are deployed on various cloud nodes in a multi-cloud environment to collect real-time and accurate cloud node resource status information, covering key indicators such as CPU utilization, memory usage, and network bandwidth availability. The collection cycle is dynamically adjusted according to the cloud node's service type. For services with high real-time requirements, the collection cycle is shortened to 1-5 seconds to ensure timely capture of subtle changes in service status. For services with low real-time requirements, the collection cycle is appropriately extended to 10-30 seconds to ensure information timeliness while reducing unnecessary resource consumption. At the same time, the transmission quality information of each link in the SD-WAN network is collected, and the link comprehensive quality index is calculated. Based on this rich information, a graph structure joint model is constructed with cloud nodes as vertices and links as edges with edge weights equal to the link comprehensive quality index, comprehensively and accurately reflecting the real-time status of the multi-cloud environment and the SD-WAN network.

[0050] Based on the constructed joint model, the resource load of each cloud node in the network and the transmission quality of the links can be understood in real time. This allows for dynamic selection of the optimal path based on the real-time network status during traffic scheduling. For example, when the CPU utilization of a certain cloud node is too high, new traffic can bypass that node and select a cloud node with relatively sufficient resources for transmission. When the transmission quality of a certain link deteriorates (such as increased latency or increased packet loss rate), traffic can be switched to other links with better quality in a timely manner. This dynamic traffic scheduling method effectively avoids network congestion, improves link utilization, and thus significantly improves the performance and stability of the entire network.

[0051] The joint model provides a clear view for the collaborative management of multi-cloud environments and SD-WAN networks. Based on this model, the collaborative management mechanism can accurately grasp the operating status of each cloud node and link, thereby formulating more reasonable collaborative strategies. For example, when performing traffic scheduling, the collaborative management can coordinate the actions of different cloud nodes and SD-WAN network devices according to the information in the model, ensuring that traffic is transmitted along the optimal path. At the same time, when network anomalies occur, the problem can be quickly located and corresponding measures can be taken to solve it, improving the network's operational efficiency and manageability.

[0052] It should be noted that the information collection agent adopts a lightweight containerized deployment method to reduce the occupation of cloud node resources.

[0053] In some embodiments, the calculation of the link comprehensive quality index based on the transmission quality information of each link is specifically calculated using the following formula:

[0054] ;

[0055] in, This is represented as the overall link quality index. This is represented as link delay. This is expressed as packet loss rate. This is represented as jitter. , , These are the weighting coefficients for link latency, packet loss rate, and jitter, respectively. The weight values ​​are determined based on the business's sensitivity to latency, packet loss rate, and jitter.

[0056] Specifically, by comprehensively considering link latency Packet loss rate and shaking Three key indicators, and using weighting coefficients , , They are combined into a comprehensive quality index This method can comprehensively and objectively evaluate the overall transmission quality of a link. Latency reflects the time required for data to travel from the sender to the receiver, packet loss rate reflects the proportion of data lost during transmission, and jitter indicates the fluctuation in the arrival time of data packets. These three indicators characterize the performance of the link from different perspectives. Together, they can more accurately reflect whether the link is suitable for transmitting data for a specific service. For example, for real-time interactive services such as video conferencing, low latency and low jitter are crucial. This calculation method can take these key factors into consideration and provide a basis for selecting a suitable link for the service.

[0057] The overall link quality index provides an important basis for traffic scheduling decisions in SD-WAN networks. In a multi-cloud environment, there are multiple possible links to choose from. By calculating the overall quality index of each link, the advantages and disadvantages of each link can be clearly compared. The traffic scheduling algorithm can guide traffic to links with higher overall quality based on these indices, thereby achieving reasonable allocation and utilization of resources. For example, when the overall quality index of a link drops, the traffic scheduling system can switch traffic to other links with better quality in a timely manner, avoiding the impact of poor link quality on services and improving the reliability and stability of the network.

[0058] From the perspective of overall network performance, traffic scheduling based on the link comprehensive quality index can effectively reduce data transmission latency, packet loss, and jitter, and improve data transmission efficiency and success rate. At the same time, this quantitative evaluation method allows network administrators to have a more intuitive understanding of the link performance status, which facilitates network monitoring, troubleshooting, and performance optimization. For example, by regularly calculating and analyzing the changing trend of the link comprehensive quality index, potential problems with the links can be identified in advance, and corresponding measures can be taken for prevention and maintenance, thereby improving the management efficiency and operational quality of the entire network.

[0059] Traffic entering the SD-WAN network is classified into different service types, including:

[0060] Deep packet inspection is performed on traffic entering the SD-WAN network to extract traffic feature information, including source IP address, destination IP address, source port number, destination port number, protocol type, and packet size distribution.

[0061] Based on traffic characteristics, clustering algorithms are used to categorize traffic into different service types, such as real-time interactive services, big data transmission services, and web browsing services. The similarity metric for the clustering algorithm uses the following formula:

[0062] ;

[0063] in, and For two traffic streams, Similarity between source IP and destination IP For the similarity between the source port and the destination port, For protocol type similarity, For the similarity of data packet size distribution, , , , These are the weighting coefficients for the similarity of each feature.

[0064] Specifically, deep packet inspection technology extracts detailed traffic characteristics, including source IP address, destination IP address, source port number, destination port number, protocol type, and packet size distribution. This provides a comprehensive and detailed data foundation for business classification. These rich characteristics can characterize traffic attributes from multiple dimensions, enabling more accurate identification of different business types, such as real-time interactive services, big data transmission services, and web browsing services, when using clustering algorithms for classification. Accurate business type identification is a prerequisite for subsequent differentiated traffic scheduling and quality of service assurance, ensuring that network resources can be rationally allocated according to business needs.

[0065] A specific similarity metric formula is used to calculate the similarity between traffic flows, comprehensively considering multiple factors such as source and destination IP similarity, source and destination port similarity, protocol type similarity, and packet size distribution similarity, and weighting coefficients are applied. , , , Adjusting the importance of each factor. This scientific similarity measurement method enables clustering algorithms to more reasonably classify traffic into different business categories, avoiding the inaccurate classification problems that may be caused by a single feature measurement. For example, for some traffic using the same protocol but with different source / destination IPs and ports, it can be accurately classified based on comprehensive similarity, improving the accuracy and reliability of business classification;

[0066] In multi-cloud SD-WAN networks, traffic types are diverse and dynamically changing. This service classification method can detect and classify incoming traffic in real time, adapting promptly to changes in network service traffic. Whether a new service type emerges or the traffic characteristics of existing service types change, it can quickly identify and classify them through deep packet inspection and clustering algorithms. This ensures that the network can always schedule and manage traffic according to the latest service situation. For example, when an enterprise adds a new real-time collaboration service, this method can quickly identify it as a real-time interactive service and schedule traffic according to the corresponding strategy to ensure the normal operation of the service.

[0067] It should be noted that the clustering algorithm is the K-means algorithm, and the cluster centers are dynamically adjusted according to the real-time changes in traffic during each iteration.

[0068] Based on a multi-objective optimization algorithm, traffic of different service types is processed to obtain the optimal scheduling path for traffic of different service types, specifically including:

[0069] For traffic of different service types, multiple optimization objectives are set, including minimizing transmission latency, maximizing link utilization, and minimizing network cost;

[0070] Based on the joint model, and combined with traffic service types and optimization objectives, a multi-objective optimization model is constructed;

[0071] A multi-objective optimization algorithm is used to solve the multi-objective optimization model to obtain the optimal scheduling path for traffic of different service types.

[0072] Specifically, by setting multiple optimization objectives for different types of traffic, including minimizing transmission latency, maximizing link utilization, and minimizing network costs, this solution comprehensively considers multiple key aspects of network performance. For real-time interactive services, such as video conferencing and online games, minimizing transmission latency ensures fast data transmission, reduces lag and delays, and provides a smooth user experience. For big data transmission services, maximizing link utilization improves data transmission efficiency and shortens transmission time. Minimizing network costs helps enterprises reduce network operating costs while ensuring service quality. This multi-objective optimization approach results in a comprehensive improvement and balance in network performance.

[0073] Based on the joint model, a multi-objective optimization model is constructed by combining traffic service types and optimization objectives. This model can fully consider the characteristics and needs of different service types. Different services have very different requirements for network performance. For example, real-time interactive services are sensitive to latency, while big data transmission services are more concerned with bandwidth and link utilization. By closely combining service types with optimization objectives, this solution can tailor the optimal scheduling path for each service type, ensuring that traffic is processed and transmitted in the network in the most appropriate way, thereby accurately adapting to the personalized needs of various services.

[0074] Multi-cloud environments are highly dynamic, with cloud node resource status and network link quality constantly changing. Scheduling methods based on multi-objective optimization algorithms can respond to these changes in real time. When the link quality in the network deteriorates or cloud node resources become scarce, the optimization algorithm can recalculate and adjust the scheduling paths for traffic of different service types to ensure that traffic is always transmitted along the optimal path. This dynamic adaptability enables the network to remain stable and efficient in complex and ever-changing environments, improving the reliability and availability of the network.

[0075] By solving the model using a multi-objective optimization algorithm to obtain the optimal scheduling path, network resources can be allocated more rationally. The algorithm will allocate traffic to the most suitable links based on various optimization objectives and real-time network status, avoiding situations where some links are overloaded while others are idle. For example, during periods of low link utilization, more traffic can be directed to these links to improve overall link utilization. At the same time, while ensuring service quality, lower-cost links are selected for transmission, thereby optimizing the allocation of network resources and improving resource utilization efficiency.

[0076] It should be noted that the multi-objective optimization model, based on the joint model and combined with traffic service types and optimization objectives, specifically includes:

[0077] Based on the constructed joint model encompassing the resource status of each cloud node in a multi-cloud environment and the transmission quality information of SD-WAN network links, and combined with different traffic service types categorized according to traffic characteristics, such as real-time interaction, big data transmission, and web browsing, and with optimization objectives including minimizing transmission latency, maximizing link utilization, and minimizing network cost, the following multi-objective optimization model is constructed:

[0078] ;

[0079] ;

[0080] ;

[0081] Where n is the number of selectable links, Let be the delay of the i-th link. This is a binary variable representing whether traffic selects the i-th link ( =1 indicates selection. =0 indicates no selection) The bandwidth already used by the i-th link. Let be the total bandwidth of the i-th link. Let $\frac{i}{i}$ be the unit bandwidth cost of the $i$ link. This model is used to determine the optimal scheduling path for traffic of different service types in multi-cloud environments and SD-WAN networks.

[0082] Furthermore, the multi-objective optimization algorithm is a non-dominated sorting genetic algorithm (NSGA-II), which continuously evolves the population through genetic operations (selection, crossover, mutation) to approximate the Pareto optimal solution set of the multi-objective optimization problem.

[0083] The collaborative management node sends scheduling instructions to relevant cloud nodes and SD-WAN network devices based on the optimal scheduling path for traffic of different service types. Specifically, these instructions include:

[0084] The collaborative management node monitors the multi-cloud environment and SD-WAN network status in real time. When the resource status of the cloud node changes or the link transmission quality deteriorates, it sends scheduling instructions to the relevant cloud nodes and SD-WAN network devices according to the optimal scheduling path for traffic of different service types.

[0085] In some embodiments, the collaborative management node adopts a distributed architecture, and multiple collaborative management nodes communicate and collaborate on decision-making through message queues.

[0086] Specifically, the collaborative management node sends scheduling instructions to relevant cloud nodes and SD-WAN network devices based on the optimal scheduling path for traffic of different service types, ensuring that traffic can be transmitted strictly according to the calculated optimal path. This enables network resources to be accurately allocated according to service needs and real-time network status. For example, for real-time interactive services, scheduling instructions can guide them to avoid congested links and choose low-latency paths, thereby ensuring the smooth operation of services and effectively improving user experience.

[0087] The collaborative management node monitors the multi-cloud environment and SD-WAN network status in real time. When the resource status of the cloud node changes (such as a sudden increase or decrease in CPU utilization) or the link transmission quality deteriorates (such as increased link latency or increased packet loss rate), it can promptly send scheduling instructions according to the new optimal scheduling path. This dynamic adjustment mechanism enables the network to respond quickly to environmental changes, avoid service interruption or performance degradation caused by changes in network status, and enhance the adaptability and stability of the network in complex and ever-changing multi-cloud environments.

[0088] The collaborative management nodes adopt a distributed architecture, with multiple nodes communicating and coordinating decisions through message queues. This improves the system's reliability and processing capacity. As an asynchronous communication mechanism, message queues decouple the collaborative management nodes, allowing them to handle tasks independently while maintaining information consistency and coordination through message passing. When one collaborative management node fails, other nodes can continue to work, ensuring the continuous operation of the system and achieving efficient collaborative management, thereby improving the overall network management efficiency and reliability.

[0089] An SD-WAN traffic scheduling and collaborative management system for multi-cloud environments includes:

[0090] The information collection agent deployment module is used to deploy information collection agents on various cloud nodes in a multi-cloud environment.

[0091] The joint model building module is used to build joint models of multi-cloud environments and SD-WAN networks based on information gathering agents.

[0092] The traffic service classification module is used to classify traffic entering the SD-WAN network into different service types.

[0093] The multi-objective optimization processing module is used to process traffic of different service types based on multi-objective optimization algorithms to obtain the optimal scheduling path for traffic of different service types.

[0094] The collaborative management and scheduling module is used by the collaborative management node to send scheduling instructions to relevant cloud nodes and SD-WAN network devices according to the optimal scheduling path of traffic of different service types, so as to realize SD-WAN traffic scheduling and collaborative management.

[0095] A computing device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of any of the above-described SD-WAN traffic scheduling and collaborative management methods for multi-cloud environments.

[0096] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of any of the above-described SD-WAN traffic scheduling and collaborative management methods for multi-cloud environments.

[0097] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for SD-WAN traffic scheduling and collaborative management in a multi-cloud environment, characterized in that, include: Deploy information collection agents on each cloud node in a multi-cloud environment, and build a joint model of the multi-cloud environment and SD-WAN network based on the information collection agents; Traffic entering the SD-WAN network is classified into different service types. The optimal scheduling path for traffic of different service types is obtained by processing traffic of different service types based on a multi-objective optimization algorithm. The collaborative management node sends scheduling instructions to relevant cloud nodes and SD-WAN network devices based on the optimal scheduling path for traffic of different service types, thereby realizing SD-WAN traffic scheduling and collaborative management.

2. The SD-WAN traffic scheduling and collaborative management method for multi-cloud environments according to claim 1, characterized in that, The deployment of information collection agents on each cloud node in the multi-cloud environment, and the construction of a joint model of the multi-cloud environment and SD-WAN network based on the information collection agents, specifically includes: Information collection agents are deployed on each cloud node in a multi-cloud environment to collect cloud node resource status information in real time. The cloud node resource status information includes CPU utilization, memory usage, and network bandwidth availability. Collect transmission quality information of each link in the SD-WAN network, and calculate the overall link quality index based on the transmission quality information of each link; Based on the collected cloud node resource status information and link comprehensive quality index, a joint model of multi-cloud environment and SD-WAN network is constructed. The joint model is represented by a graph structure, with cloud nodes as vertices and links as edges, and the weight of the edge is the link comprehensive quality index.

3. The SD-WAN traffic scheduling and collaborative management method for multi-cloud environments according to claim 2, characterized in that, The link comprehensive quality index is calculated based on the transmission quality information of each link, and the specific calculation formula is as follows: ； in, This is represented as the overall link quality index. This is represented as link delay. This is expressed as packet loss rate. This is represented as jitter. , , These are the weighting coefficients for link latency, packet loss rate, and jitter, respectively.

4. The SD-WAN traffic scheduling and collaborative management method for multi-cloud environments according to claim 3, characterized in that, The process of classifying traffic entering the SD-WAN network into different service types specifically includes: Deep packet inspection is performed on traffic entering the SD-WAN network to extract traffic feature information, including source IP address, destination IP address, source port number, destination port number, protocol type, and packet size distribution. Based on traffic characteristics, clustering algorithms are used to categorize traffic into different business types.

5. The SD-WAN traffic scheduling and collaborative management method for multi-cloud environments according to claim 4, characterized in that, The process of processing traffic of different service types based on a multi-objective optimization algorithm to obtain the optimal scheduling path for traffic of different service types specifically includes: For traffic of different service types, multiple optimization objectives are set, including minimizing transmission latency, maximizing link utilization, and minimizing network cost; Based on the joint model, and combined with traffic service types and optimization objectives, a multi-objective optimization model is constructed; A multi-objective optimization algorithm is used to solve the multi-objective optimization model to obtain the optimal scheduling path for traffic of different service types.

6. The SD-WAN traffic scheduling and collaborative management method for multi-cloud environments according to claim 5, characterized in that, The collaborative management node sends scheduling instructions to relevant cloud nodes and SD-WAN network devices based on the optimal scheduling path for traffic of different service types, specifically including: The collaborative management node monitors the multi-cloud environment and SD-WAN network status in real time. When the resource status of the cloud node changes or the link transmission quality deteriorates, it sends scheduling instructions to the relevant cloud nodes and SD-WAN network devices according to the optimal scheduling path for traffic of different service types.

7. The SD-WAN traffic scheduling and collaborative management method for multi-cloud environments according to claim 6, characterized in that, The collaborative management node adopts a distributed architecture, and multiple collaborative management nodes communicate and collaborate on decision-making through message queues.

8. An SD-WAN traffic scheduling and collaborative management system for multi-cloud environments, characterized in that, include: The information collection agent deployment module is used to deploy information collection agents on various cloud nodes in a multi-cloud environment. The joint model building module is used to build joint models of multi-cloud environments and SD-WAN networks based on information gathering agents. The traffic service classification module is used to classify traffic entering the SD-WAN network into different service types. The multi-objective optimization processing module is used to process traffic of different service types based on multi-objective optimization algorithms to obtain the optimal scheduling path for traffic of different service types. The collaborative management and scheduling module is used by the collaborative management node to send scheduling instructions to relevant cloud nodes and SD-WAN network devices according to the optimal scheduling path of traffic of different service types, so as to realize SD-WAN traffic scheduling and collaborative management.

9. A computing device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the SD-WAN traffic scheduling and collaborative management method for multi-cloud environments as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The storage medium stores a computer program, which, when executed by a processor, implements the steps of the SD-WAN traffic scheduling and collaborative management method for multi-cloud environments as described in any one of claims 1-7.

Images