An intelligent global hybrid cloud routing system based on SD-WAN architecture

Abstract

The application provides a kind of intelligent global hybrid cloud routing system based on SD-WAN architecture, belongs to network routing technical field, comprising: data acquisition module: based on the preset compression algorithm to the data stream to be transmitted is compressed, and then compressed data stream is obtained;Node determination module: obtain network performance related data, at the same time, determine the corresponding several dynamic distributed nodes of SD-WAN control layer;Node connection module: all distributed nodes are connected, at the same time, all distributed nodes are connected to SD-WAN control layer;Score determination module: the real-time network state data of all distributed nodes is collected and data analysis is carried out, and the path score of each available path is determined in combination with the preset user demand;Path determination module: determine the optimal path and redundant path;Path switching module: the network state data of all distributed nodes after the optimal path is obtained for data transmission, and then the redundant switching condition is determined and redundant path switching is carried out. The reliability and adaptability of network are improved.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of network routing, in particular to an intelligent global hybrid cloud routing system based on an SD-WAN architecture. BACKGROUND

[0002] With the rapid development of cloud computing and Internet technology, enterprises and individuals have increasing demand for global network connection. The traditional wide area network (WAN) architecture has been unable to meet the growing demand for data transmission, and there are problems such as limited network bandwidth, unstable connection and complex management.

[0003] SD-WAN (Software Defined Wide Area Network) technology has emerged as the times require, optimizing network performance through centralized control and intelligent path selection. However, existing SD-WAN solutions are based on static path configuration and cannot adjust paths in real time to respond to complex and changing network conditions, resulting in low transmission efficiency under different network conditions. Existing SD-WAN solutions often lack dynamic network performance evaluation and intelligent path selection capabilities, and cannot adjust paths in real time according to network conditions and user needs. Existing technologies rely more on manual configuration or simple rule-driven, and fail to fully utilize real-time network data to optimize routing strategies. Therefore, there are problems such as insufficient intelligent path selection, delayed response to network state changes, and untimely redundant path switching, which affect the performance and reliability of the network.

[0004] Therefore, the application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture. SUMMARY

[0005] The application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, which overcomes the shortcomings of inflexible path selection, delayed response and untimely redundant switching in existing SD-WAN technology by introducing dynamic distributed nodes and intelligent path scoring mechanisms. By collecting network performance data and user needs in real time, the system can automatically optimize path selection to ensure the stability and efficiency of data transmission. The node connection and path determination module works together to quickly adjust the path when the network state changes, improving the reliability and adaptability of the network. In addition, the redundant path switching module can switch to the standby path in time, effectively avoiding network interruption and significantly improving the overall network performance and user experience.

[0006] The application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, which includes:

[0007] Data acquisition module: acquires a data stream to be transmitted, compresses the data stream to be transmitted based on a preset compression algorithm, and then acquires a compressed data stream;

[0008] The node determination module obtains network performance related data based on a preset network performance monitoring tool, analyzes a preset network planning requirement, determines a plurality of dynamic distributed nodes corresponding to the SD-WAN control layer based on a requirement analysis result, the network performance related data and compressed data streams;

[0009] The node connection module obtains the node type of each distributed node, and then connects all the distributed nodes, and connects all the distributed nodes to the SD-WAN control layer.

[0010] The score determination module collects real-time network state data of all the distributed nodes based on the SD-WAN control layer and performs data analysis, and then determines the path score of each available path in combination with a preset user demand.

[0011] The path determination module determines the optimal path and the redundant path based on the path score of each available path and the real-time network state data of all the distributed nodes.

[0012] The path switching module obtains the network state data of all the distributed nodes after data transmission based on the SD-WAN control layer, and then determines the redundant switching condition and judges whether to switch the redundant path, and if it is judged to switch the redundant path, the redundant path is switched.

[0013] The edge routing cache module obtains the access characteristics of the data to be transmitted, combines a preset cache strategy, caches user high-frequency access data in the edge router, obtains network state data and dynamically adjusts the cache content, and then optimizes the data transmission process.

[0014] Preferably, the node determination module comprises:

[0015] The initial node determination unit analyzes a preset network planning requirement, and determines a plurality of initial distributed nodes based on a requirement analysis result.

[0016] The requirement acquisition unit performs compression analysis on the compressed data stream, determines a compression ratio, and obtains transmission requirements based on a preset compression ratio-demand database.

[0017] The node adjustment unit performs a first adjustment on the initial distributed nodes based on the transmission requirements, obtains network performance related data based on a preset network performance monitoring tool, performs a second adjustment on the initial distributed nodes after the first adjustment based on the network performance related data, and then obtains a plurality of dynamic distributed nodes.

[0018] Preferably, the initial node determination unit comprises:

[0019] The demand determination subunit determines a plurality of data centers and a plurality of service demands corresponding to each data center by performing demand analysis on the preset network planning demand.

[0020] The type determination subunit determines a plurality of node types corresponding to each service demand based on the type of each service demand and a preset type-node type database.

[0021] The demand analysis subunit determines the number of nodes of each node type corresponding to each service demand by performing demand analysis on each service demand corresponding to each data center, and further determines the node type corresponding to each data center and the number of nodes of each node type.

[0022] The location determination subunit determines the location of the nodes corresponding to each data center based on the location of each data center.

[0023] The node determination subunit determines a plurality of nodes corresponding to each data center based on the location of the nodes corresponding to each data center, the node type corresponding to each data center, and the number of nodes of each node type, and further determines the nodes of all data centers as initial distributed nodes.

[0024] Preferably, the node adjustment unit comprises:

[0025] The node evaluation subunit obtains network performance related data based on a preset network performance monitoring tool, and evaluates each initial distributed node based on the network performance related data.

[0026] The node determination subunit determines the first topology structure of the area within the preset range where each initial distributed node is located based on a preset routing table, searches the first topology structure of the area within the preset range where each initial distributed node is located based on a preset graph search algorithm, and further determines a plurality of associated nodes of each initial distributed node.

[0027] The node adjustment subunit disperses the load of each initial distributed node based on the associated nodes of each initial distributed node when the evaluation result of each initial distributed node does not meet the preset evaluation requirement.

[0028] Preferably, the node adjustment subunit comprises:

[0029] The number determination block determines the number of associated nodes for dispersing the load of each initial distributed node based on the evaluation result of each initial distributed node when the evaluation result of each initial distributed node does not meet the preset evaluation requirement.

[0030] Threshold determination block: obtain the distance between each initial distributed node and each associated node, and then determine the distance threshold corresponding to each initial distributed node by combining the number of associated nodes of the load of each initial distributed node:

[0031]

[0032] Wherein, T i is the distance threshold of the ith initial distributed node, M i is the number of all associated nodes of the ith initial distributed node, d ij is the distance between the ith initial distributed node and the jth associated node, N i is the number of associated nodes used to disperse the load of the ith initial distributed node, L i is the load of the ith initial distributed node, V i is the load variation coefficient of the ith initial distributed node, wherein, δ i is the distance fluctuation coefficient of the ith node.

[0033] Node determination block: the associated nodes whose distance from each initial distributed node is less than the distance threshold are determined as dispersion nodes;

[0034] Load dispersion block: disperse the load of each initial distributed node based on the dispersion nodes of each initial distributed node.

[0035] Preferably, the score determination module comprises:

[0036] Path determination unit: determine the source node and the target node based on the preset user demand, and then determine a plurality of available paths between the source node and the target node;

[0037] Score determination unit: collect real-time network state data of all distributed nodes and perform data analysis based on the SD-WAN control layer, and then determine the path score of each available path in combination with the preset user demand.

[0038] Preferably, the score determination subunit comprises:

[0039] Data acquisition block: collect real-time network state data of all distributed nodes using a preset network monitoring tool at the SD-WAN control layer, wherein the real-time network state data comprises a plurality of performance indicators;

[0040] Score calculation block: determine the weight of each performance indicator based on the preset user demand, and then determine the path score of each available path in combination with the preset score algorithm.

[0041] Preferably, the path determination module comprises:

[0042] An estimation determination unit determines an estimated path score of each available path based on real-time network state data of all distributed nodes and a preset algorithm;

[0043] A score determination unit determines a comprehensive path score of each available path based on the path score and the estimated path score of each available path;

[0044] An optimal determination unit determines an optimal path as the available path with the highest comprehensive path score and determines other available paths as candidate available paths except for the available path with the highest path score;

[0045] A redundancy determination unit determines a redundancy path as the candidate available path with a comprehensive path score greater than a preset score threshold.

[0046] Compared with the prior art, the application has the following beneficial effects:

[0047] By introducing dynamic distributed nodes and an intelligent path score mechanism, the application overcomes the shortcomings of path selection inflexibility, response lag and untimely redundancy switching in the prior art SD-WAN technology. By collecting network performance data and user demand in real time, the system can automatically optimize path selection to ensure the stability and efficiency of data transmission. The node connection and path determination module work cooperatively to quickly adjust the path when the network state changes, thereby improving the reliability and adaptability of the network. In addition, the redundancy path switching module can switch to a backup path in time, effectively avoiding network interruption and significantly improving the overall network performance and user experience. BRIEF DESCRIPTION OF DRAWINGS

[0048] In order to more clearly illustrate the technical solutions in the application or prior art, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are some embodiments of the application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of these drawings.

[0049] Figure 1 is a structural schematic diagram of an intelligent global hybrid cloud routing system based on an SD-WAN architecture provided by an embodiment of the application;

[0050] Figure 2 is a structural schematic diagram of a node determination module of an intelligent global hybrid cloud routing system based on an SD-WAN architecture provided by an embodiment of the application. DETAILED DESCRIPTION

[0051] In order to make the objects, technical solutions and advantages of the present application clearer, the technical solutions in the present application will be described clearly and completely below in conjunction with the drawings in the present application. Obviously, the described embodiments are part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0052] Embodiment 1

[0053] The embodiments of the present application provide an intelligent global hybrid cloud routing system based on an SD-WAN architecture, as shown in the accompanying drawings, comprising: Figure 1

[0054] The data acquisition module acquires a to-be-transmitted data stream, compresses the to-be-transmitted data stream based on a preset compression algorithm, and then acquires a compressed data stream.

[0055] The node determination module acquires network performance related data based on a preset network performance monitoring tool, simultaneously performs demand analysis on a preset network planning requirement, determines a plurality of dynamic distributed nodes corresponding to an SD-WAN control layer based on a demand analysis result, network performance related data and the compressed data stream.

[0056] The node connection module acquires the node type of each distributed node, and then connects all the distributed nodes, and simultaneously connects all the distributed nodes to the SD-WAN control layer.

[0057] The score determination module collects real-time network state data of all the distributed nodes based on the SD-WAN control layer and performs data analysis, and then determines the path score of each available path in combination with a preset user demand.

[0058] The path determination module determines an optimal path and a redundant path based on the path score of each available path and the real-time network state data of all the distributed nodes.

[0059] The path switching module acquires the network state data of all the distributed nodes after data transmission based on the optimal path of the SD-WAN control layer, and then determines a redundant switching condition and judges whether to perform redundant path switching. If it is judged to perform redundant path switching, redundant path switching is performed.

[0060] The edge routing cache module acquires the access features of the to-be-transmitted data, combines a preset cache strategy, caches user high-frequency access data in an edge router, acquires network state data and then dynamically adjusts cache content, and then optimizes the data transmission process.

[0061] ​In this embodiment, the preset compression algorithm is based on H.265 / HEVC optimization: by dynamically adjusting the H.265 encoding parameters, the encoding efficiency of high motion pictures (such as product display) and static pictures (such as anchor pictures) in e-commerce live streaming scenarios is optimized: dynamic bit rate adjustment: real-time adjustment of bit rate (CBR / VBR mixed mode) according to user bandwidth, scene adaptive optimization: by analyzing the live streaming picture content, using region of interest encoding (ROI Encoding) to improve product area clarity and reduce background area encoding resource occupation, realizing H.265 / HEVC optimization in e-commerce live streaming scenarios, improving encoding efficiency and video quality, and simultaneously performing adaptive adjustment according to user bandwidth and picture content to realize effective data stream compression. Further, data stream compression is realized.

[0062] In this embodiment, the preset network performance monitoring tool refers to a tool or device configured in the system in advance for real-time monitoring and analysis of network performance. These tools can collect network bandwidth, delay, packet loss rate, network congestion and other performance data to help the system understand the health status of the network. For example, network management software (such as SolarWinds, PRTG Network Monitor) can be used to monitor the bandwidth usage and delay of each node in the enterprise network in real time, and collect these data for subsequent path selection and redundancy switching decisions.

[0063] In this embodiment, the SD-WAN control layer is the central management layer in the entire SD-WAN architecture, responsible for monitoring and controlling the path selection of network traffic, processing user demands, and issuing control instructions to each distributed node. It realizes intelligent routing optimization through data exchange with each node. For example, an SD-WAN control layer may include a centralized SD-WAN management platform that monitors the network status of each distributed node and dynamically selects the optimal path according to real-time data. Common SD-WAN control platforms include Viptela, VeloCloud, etc.

[0064] In this embodiment, the node type of each distributed node refers to the role or function of the node in the network. For example, some nodes may be data center nodes, some nodes may be branch office routers, or they may be redundant nodes for backup. For example, one distributed node may be the data center of the enterprise headquarters, and the other node may be the network router of the remote office. The node type can affect the priority and path selection of data transmission.

[0065] In this embodiment, all distributed nodes are connected, and connecting all distributed nodes to the SD-WAN control layer means that the system connects all distributed nodes through the network to form a unified network architecture. At the same time, these nodes also need to be connected to the SD-WAN control layer, so that the control layer can monitor and manage the network status and path selection of all nodes in real time. For example, assume a multinational enterprise has multiple data centers and branch offices in different regions, and each location has deployed distributed nodes. These nodes (such as routers in data centers or network devices in remote offices) are connected through the Internet or dedicated lines and connected to the control layer of the SD-WAN (such as a central management platform) for network traffic management and routing.

[0066] In this embodiment, the optimal path refers to the data transmission path that can provide the best performance (such as bandwidth, low latency, and low packet loss rate) under given network conditions. The redundant path is a backup path that is usually activated when the optimal path fails or performance decreases to ensure the continuity and reliability of network connections. Assuming that the path from A to B is the main path, which has the best bandwidth and latency, it is called the optimal path. If this path becomes unavailable for some reason, the system will automatically switch to the alternative path C, which is the redundant path, to ensure that data can continue to be transmitted.

[0067] In this embodiment, the redundant switching condition refers to the condition or standard for the system to determine whether to switch to the redundant path. For example, if the latency of a path exceeds a certain threshold or the data packet loss rate is too high, the system will consider that the redundant path needs to be activated. If the packet loss rate of a path exceeds 5% or the latency increases to more than 300ms, the system will trigger the redundant switching condition and activate the backup path.

[0068] In this embodiment, redundant path switching refers to automatically switching to a redundant path when the main path has problems or performance decreases to ensure the stability of network connections and the continuity of data transmission. For example, when the bandwidth of the optimal path decreases or fails, the system will automatically switch the traffic to the redundant path to ensure that network services do not interrupt. This redundant switching is usually performed in the network architecture of cloud service providers to improve fault tolerance.

[0069] In this embodiment, the edge routing cache module includes the following units:

[0070] Traffic detection unit: monitors the type of data requested by users (such as video streaming, API requests, static resources, etc.), counts data access frequency, identifies high-frequency access content, and determines data cache priority.

[0071] Cache strategy judgment unit: based on user access behavior, network load condition, decide whether to cache data, adopt machine learning algorithm to carry out flow prediction, improve cache hit rate, combine data deduplication compression technology, optimize storage space, improve cache efficiency.

[0072] Edge storage unit: store high-frequency access data in edge router, reduce remote server access; support P2P cache sharing, cooperative storage between multiple edge devices, reduce bandwidth consumption. Adopt LRU (Least Recently Used) algorithm, regularly eliminate low-frequency access data, improve cache utilization.

[0073] Intelligent scheduling unit: through intelligent scheduling, when user requests data, preferentially acquire data from local cache or P2P node, reduce remote server access pressure; combine zero packet loss mechanism, ensure data transmission stability, improve transmission quality.

[0074] Cache update and invalidation processing unit: adopt TTL (Time To Live) control, ensure that cache data does not expire; combine dynamic cache synchronization, ensure that cache data is consistent with remote server data, avoid cache pollution, support real-time data update mechanism, automatically synchronize high-change-rate data, improve system adaptability.

[0075] The beneficial effects of the above technical solutions are: by introducing the intelligent global hybrid cloud routing system based on the SD-WAN architecture, the problem of static and inflexible network path selection in the prior art is solved, by dynamically obtaining network performance data and user demand, the system can intelligently evaluate and optimize the data transmission path in real time, ensure that the optimal path is selected, in addition, the redundant path switching module automatically judges and switches to the standby path when the network state changes, effectively improves the reliability and fault tolerance of the network, reduces 30%-50% of repeated flow requests, reduces enterprise operating costs, and improves data access speed: API request response time is shortened by 50%, live video stall rate is reduced by 40%, optimizes network flow distribution: utilizes P2P cache sharing to improve data availability, reduces server pressure, and improves system stability: combines intelligent scheduling and zero packet loss mechanism to ensure stable and reliable data transmission. This scheme significantly improves the intelligence, real-time and adaptability of path selection, optimizes the data transmission efficiency and user experience in the global hybrid cloud environment.

[0076] Embodiment 2:

[0077] The embodiment of the application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, a node determination module, as shown in Figure 2 , comprising:

[0078] Initial node determination unit: demand analysis is performed on preset network planning requirements, and a plurality of initial distributed nodes are determined based on the demand analysis result.

[0079] Demand acquisition unit: perform compression analysis on the compressed data stream, determine the compression ratio, and acquire the transmission demand based on the preset compression ratio-demand database;

[0080] Node adjustment unit: based on the transmission demand, first adjust the initial distributed nodes, acquire network performance related data based on the preset network performance monitoring tool, and based on the network performance related data, second adjust the initial distributed nodes after the first adjustment, and further acquire a plurality of dynamic distributed nodes.

[0081] In this embodiment, the initial distributed nodes refer to network nodes determined according to the preset network planning demand and demand analysis result, which are data centers, remote branch offices, or other necessary infrastructure nodes. They are the basic units for realizing data routing and transmission in the system. For example, in the network design of a certain global enterprise, the initial distributed nodes may include the data center of the headquarters, branch offices in various regions, and important edge nodes (such as cloud access points in certain regions). These initial nodes are set in the design stage according to the network demand of the enterprise (such as high bandwidth and large traffic), and they form the preliminary architecture of the network. In the initial stage of network deployment, network traffic may be transmitted through these nodes until the network path and node structure are optimized through real-time monitoring and adjustment.

[0082] In this embodiment, the first adjustment of the initial distributed nodes based on the transmission demand includes: according to the compression ratio, querying the corresponding transmission demand from the database. Assuming that the compression ratio is 3, after querying the database, it may be found that the transmission demand requires a certain bandwidth range (such as 10-20 Mbps), a delay of no more than 200 milliseconds, and a packet loss rate of less than 0.5%. Evaluate the load capacity of the existing nodes: check the existing load capacity of the initial distributed nodes, including the bandwidth, processing capacity (such as CPU, memory and storage resources) of the nodes and the current network connection state. For example, view the bandwidth utilization rate, CPU usage rate and memory occupancy rate of each node through the network performance monitoring tool. Determine the available resources of the nodes: calculate how much additional load each node can still carry. For an initial distributed node, if its current bandwidth utilization rate is 60% and its total bandwidth is 100 Mbps, there is still 40 Mbps of available bandwidth. According to the bandwidth demand acquired from the database, compare it with the available bandwidth of the node. If the required bandwidth exceeds the available bandwidth of the node, it may be necessary to consider adding new links or upgrading existing links, or transferring part of the business to other nodes with more available bandwidth. For example, if the transmission demand requires 30 Mbps of bandwidth and the current node only has 20 Mbps of available bandwidth, it can be considered to transfer part of the data traffic to other nodes or add new links to the node to increase the bandwidth.

[0083] The beneficial effects of the above technical solution are: by introducing the initial node determination unit and the node adjustment unit, intelligent dynamic node management based on the SD-WAN architecture is realized. First, the system determines the initial distributed nodes according to the preset network planning requirement analysis, providing a basic framework for network construction. Then, through real-time data obtained based on network performance monitoring tools, the position and configuration of the initial nodes are adjusted in combination with compressed data flow, the network performance is optimized, and the distributed nodes are generated. Compared with the prior art, this intelligent dynamic adjustment mechanism can respond to network load changes in real time, ensure the stability, flexibility and efficiency of the network, greatly improve the reliability and user experience of the system, and meet the changing network requirements.

[0084] Embodiment 3:

[0085] The embodiment of the application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, an initial node determination unit, comprising:

[0086] The demand determination subunit performs demand analysis on the preset network planning requirements, and then determines a plurality of data centers and a plurality of business requirements corresponding to each data center;

[0087] The type determination subunit determines a plurality of node types corresponding to each business requirement based on the type of each business requirement and a preset type-node type database;

[0088] The demand analysis subunit performs demand analysis on each business requirement corresponding to each data center, and then determines the number of nodes of each node type corresponding to each business requirement, and then determines the node type corresponding to each data center and the number of nodes of each node type;

[0089] The location determination subunit determines the location of the nodes corresponding to each data center based on the location of each data center;

[0090] The node determination subunit determines a plurality of nodes corresponding to each data center based on the location of the nodes corresponding to each data center, the node type corresponding to each data center, and the number of nodes of each node type, and then determines the nodes of all data centers as initial distributed nodes.

[0091] In this embodiment, the business requirement refers to the network function, performance or service required by an enterprise or organization in network architecture design according to the actual application scenario and target. For example, a department may need high-bandwidth, low-latency network connection to support video conferencing or big data analysis and other key tasks. For example, the business requirements of a financial institution may include fast real-time data transmission and highly secure data encryption. The business requirements of an e-commerce platform may include high availability and large-scale concurrent user access support.

[0092] In this embodiment, the data center refers to a physical facility for centralized storage, management, processing of large amounts of data and running various network application services. It usually contains servers, storage devices, network hardware, and power and cooling systems, etc. Infrastructure to support the daily operations of enterprises, for example: Amazon's AWS service or Google's Google Cloud data center, which provides cloud computing services and data storage for global users.

[0093] In this embodiment, the type of business requirement refers to the way of classifying business requirements according to different business scenarios or tasks. Each type of business requirement may involve specific technical requirements, performance indicators or network configuration requirements. The type of business requirement usually affects the configuration, node type and number of the data center, for example: the type of business requirement includes: high bandwidth requirement: for large data transmission or high-resolution video streaming service, low latency requirement: such as real-time voice communication, online games and other applications sensitive to delay, high security requirement: such as financial services, medical data transmission and other applications that require high encryption and protection, high availability requirement: such as critical applications that require 24 / 7 operation, require no single point of failure, and ensure continuous service.

[0094] The beneficial effects of the above technical solution are: through the intelligent global hybrid cloud routing system based on the SD-WAN architecture, a precise and dynamic node determination mechanism is provided. The system conducts comprehensive demand analysis according to the preset network planning requirements, combines the business requirements and node types of each data center, intelligently determines the required node type, number and specific location of each data center, and this method is more flexible than traditional static configuration, which can dynamically adjust the node layout according to actual requirements, optimize the network architecture, improve the network performance, reliability and scalability of the system through precise node selection and configuration, and ensure efficient and stable data transmission in the global hybrid cloud environment, effectively supporting the business requirements and technical development of enterprises.

[0095] Embodiment 4:

[0096] The embodiment of the application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, a node adjustment unit, comprising:

[0097] The node evaluation subunit: based on the preset network performance monitoring tool, the network performance related data is obtained, and each initial distributed node is evaluated based on the network performance related data;

[0098] The node determination subunit: determines the first topology structure of the area within the preset range of each initial distributed node based on the preset routing table, searches the first topology structure of the area within the preset range of each initial distributed node based on the preset graph search algorithm, and then determines a plurality of associated nodes of each initial distributed node.

[0099] The node adjustment subunit: when the evaluation result of each initial distributed node does not meet the preset evaluation requirement, the load of each initial distributed node is dispersed based on the associated nodes of each initial distributed node.

[0100] In this embodiment, the preset range in which each initial distributed node is located refers to a specific area or network range in which each initial distributed node is located. This range is preset and used for evaluating the performance of the node or performing topology search. This range is determined based on geographical location, network connectivity, bandwidth limitation, or other network design factors. For example, assuming that a distributed node is located in the Asian region in a global enterprise network. The preset range of this node may be a network area covering part of the Asian region, including multiple access points, routers, and other infrastructure. During evaluation, only the network topology structure within the range in which this node is located is considered, and other regions around the world are not included.

[0101] In this embodiment, the first topology structure refers to a preliminary network topology formed according to the network design and the connection mode of the node within the preset range. This topology structure describes the connection relationship between nodes, including the arrangement of nodes, connection paths, and network performance of each node. For example, in a regional network of a large enterprise, there may be multiple distributed nodes (such as routers of multiple branch offices). These nodes are connected to each other through dedicated lines or virtual private networks (VPNs) to form a local network topology. This local topology can be used as the “first topology structure” for subsequent performance evaluation and adjustment.

[0102] In this embodiment, the preset evaluation requirement refers to some standards or thresholds set by a network performance monitoring tool during node evaluation. These requirements usually cover network performance indicators (such as bandwidth, delay, packet loss rate, etc.) and are used to judge whether a node meets the performance requirements of the system. If a node does not meet the preset requirements, corresponding load adjustment or node reconfiguration will be triggered. For example, the preset evaluation requirements of a distributed node may include: delay less than 100 ms, bandwidth greater than 1 Gbps, and packet loss rate less than 0.1%. If the evaluation result of a node does not meet these requirements (for example, the delay is too high or the packet loss rate is too large), the system will take measures to adjust the load of the node or replace it with other nodes.

[0103] The beneficial effects of the above technical solutions are: the network performance is optimized by dynamically adjusting the distributed nodes, the network performance data is obtained and analyzed in real time by the node evaluation subunit, the performance of each initial distributed node is evaluated, when the node does not meet the preset evaluation requirement, the node adjustment subunit disperses the load based on the associated nodes to ensure the network load balance and stability, in addition, the system accurately determines the topology structure and associated nodes between nodes through the routing table and graph search algorithm, and the flexibility and expansibility of the network are improved, the intelligent and automatic dynamic adjustment mechanism effectively avoids the network performance bottleneck that may be caused by the traditional static node configuration, and improves the overall efficiency and reliability of the system.

[0104] Embodiment 5:

[0105] The embodiment of the application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, a node adjustment subunit, and the like.

[0106] The number determination block is configured to determine the number of associated nodes for dispersing the load of each initial distributed node based on the evaluation result of each initial distributed node when the evaluation result of each initial distributed node does not meet the preset evaluation requirement.

[0107] The threshold value determination block is configured to obtain the distance between each initial distributed node and each associated node, and then determine the distance threshold value corresponding to each initial distributed node in combination with the number of associated nodes of the load of each initial distributed node.

[0108]

[0109]

[0110] wherein T i is the distance threshold value of the i th initial distributed node, M i is the number of all associated nodes of the i th initial distributed node, d ij is the distance between the i th initial distributed node and the j th associated node, N i is the number of associated nodes for dispersing the load of the i th initial distributed node, L i is the load of the i th initial distributed node, V i is the load variation coefficient of the i th initial distributed node, wherein δ i is the distance fluctuation coefficient of the i th node.

[0111] The node determination block is configured to determine the associated nodes with a distance less than the distance threshold value between each initial distributed node and each associated node as the dispersion nodes.

[0112] The load dispersion block is configured to disperse the load of each initial distributed node based on the dispersion nodes of each initial distributed node.

[0113] In this embodiment, the number of associated nodes is calculated by the number determination block according to the evaluation result of each initial distributed node that does not meet the preset evaluation requirement, and the number of associated nodes for load dispersion is calculated. This number is usually determined by a certain algorithm or rule according to the evaluation result of the initial distributed node with excessive load or poor performance, for example, if the evaluation result shows that the load of the node exceeds a certain percentage (such as 80%), n associated nodes may be needed to disperse the load, and the value of n may be dynamically adjusted according to the severity of the load and the capacity of the system.

[0114] In this embodiment, the distance fluctuation coefficient is an index for measuring the change of the distance between the node and the associated node. It reflects the connection stability of a node and the flexibility of network topology change. The higher the distance fluctuation coefficient, the greater the change of the distance between the node and its associated node, which may lead to unstable network connection; and the lower the distance fluctuation coefficient, the more stable the connection distance between the nodes.

[0115] The beneficial effects of the above technical solution are: through the number determination block and the threshold determination block, the system intelligently calculates the number of associated nodes required for load dispersion according to the performance evaluation result of each initial distributed node and the distance between nodes, and sets a reasonable distance threshold, effectively selects the dispersion nodes with close distance to the initial node and load bearing capacity, reduces network delay and bandwidth pressure, at the same time, the load dispersion block can ensure the balanced allocation of load between nodes, improve the stability and performance of the network, compared with the traditional static load allocation method, the present application has higher flexibility and adaptability, can dynamically optimize according to the real-time network status, and improves the efficiency and reliability of the system.

[0116] Embodiment 6:

[0117] The embodiment of the present application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, a score determination module, which comprises:

[0118] The path determination unit determines the source node and the target node based on the preset user demand, and then determines a plurality of available paths between the source node and the target node;

[0119] The score determination unit collects real-time network state data of all distributed nodes based on the SD-WAN control layer and performs data analysis, and then determines the path score of each available path in combination with the preset user demand.

[0120] In this embodiment, the source node and the target node are determined based on the preset user demand, and then a plurality of available paths between the source node and the target node are determined as follows: first, the preset user demand needs to be analyzed in depth. The user demand can include multiple aspects, such as service type (file transfer, video conference, database access, etc.), quality of service requirement (acceptable delay range, required bandwidth, packet loss threshold, etc.), source location and target location information, etc. Service mapping: according to the service type, the node where the service is located is determined as the target node. For example, if a user needs to access a file storage service, the system will find the distributed node where the server storing the service is located and take it as the target node. This can be completed through a service discovery protocol (such as DNS query, service registry center, etc.) or a pre-configured service mapping table. Source node determination: the source node is usually the network node connected by the user equipment initiating the service request. For enterprise networks, the source node can be the access device of a branch office; for user access networks, the source node can be the access point connected by the user's computer, mobile device or Internet of Things device. For example, for a video conference request initiated by a mobile user, the source node can be the wireless network access point connected by the user's mobile phone. Determining a plurality of available paths between the source node and the target node: topology information query: network topology map: use the network topology map stored in the SD-WAN control layer, which contains all distributed nodes and their connection relationships. By querying the topology map, all possible connection links and combinations of intermediate nodes from the source node to the target node can be found, which constitute available paths. Routing table reference: check the routing table, which contains reachability information between network nodes. For each source node and target node pair, the routing table indicates which next hop node is reachable, so that possible paths can be constructed. For example, starting from the source node, according to the next hop information of the routing table, gradually extend to find multiple paths leading to the target node.

[0121] The above technical solution has the beneficial effect that the scoring determination module dynamically optimizes the network path. Through the path determination unit, the system intelligently determines a plurality of available paths between the source node and the target node according to the preset user demand, providing multiple network options. The scoring determination unit assigns a path score to each path based on real-time network state data and the user's specific demand, ensuring the selection of the optimal network path. This dynamic scoring and path selection mechanism is more flexible than traditional static routing, can provide more efficient and stable connections in complex and changing network environments, and improves network performance and user experience.

[0122] Embodiment 7:

[0123] The embodiment of the application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, and a scoring determination subunit, which comprises:

[0124] Data collection block: Real-time network status data of all distributed nodes is collected using preset network monitoring tools at the SD-WAN control layer, wherein the real-time network status data contains several performance indicators;

[0125] Score calculation block: The weight of each performance indicator is determined based on preset user requirements, and then the path score of each available path is determined by combining a preset scoring algorithm.

[0126] In this embodiment, performance indicators are various data or standards that measure network performance, which are usually used to evaluate network quality, stability and efficiency. These indicators may include but are not limited to bandwidth, delay, packet loss rate, response time, throughput, etc. By monitoring and analyzing these indicators, the system can better understand the current network status and make appropriate adjustments. For example, in an SD-WAN network, common performance indicators may include: bandwidth: the maximum rate of network data transmission, such as 100Mbps or 1Gbps. Delay: the data transmission delay from the source node to the target node, such as 50 milliseconds. Packet loss rate: the proportion of lost packets during data transmission, such as 1% packet loss rate. These performance indicators help determine the quality of network paths, which in turn affect path selection.

[0127] In this embodiment, determining the weight of each performance indicator based on preset user requirements means that different importance weights are assigned to different network performance indicators according to the specific needs and application scenarios of users. Different applications or businesses have greater dependence on certain performance indicators, so higher weights need to be given in the scoring process. The weight setting depends on the specific business requirements, for example, some applications may have very high requirements for delay, but lower requirements for bandwidth. Assuming that a company's core business is video conferencing, then delay (low delay is crucial) may be more important than bandwidth (transmission rate). Based on this requirement, the system may assign a higher weight (such as 0.7) to delay and a lower weight (such as 0.3) to bandwidth. In this way, in the path scoring, the impact of delay will be greater than that of bandwidth, so a path with lower delay will be selected.

[0128] In this embodiment, the path score of each available path is determined by combining a preset scoring algorithm, which uses a weighted average formula to calculate the path score. Assuming that W = [w1, w2, w3,..., wn] is the weight vector of each performance indicator, and P = [p1, p2, p3,..., pn] is the normalized performance indicator vector, the path score Score can be calculated by the formula Score = w1*p1 + w2*p2 + w3*p3 +... + wn*pn.

[0129] The beneficial effects of the above technical solutions are: the score determination subunit optimizes network path selection, the data acquisition block collects network state data of the distributed nodes in real time at the SD-WAN control layer, including multiple performance indicators (such as delay, bandwidth, packet loss rate, etc.), to provide a data basis for path evaluation, the score calculation block dynamically adjusts the weights of the performance indicators according to the specific needs of the user, and calculates the score of each path in combination with the preset scoring algorithm, so that the optimal path can be intelligently selected according to the real-time network status and user needs, the flexibility and stability of the network are improved, the limitations of traditional static routing are avoided, and the network performance and user experience are significantly improved.

[0130] Embodiment 8:

[0131] The embodiment of the application provides an intelligent global hybrid cloud routing system based on an SD-WAN architecture, a path determination module, comprising:

[0132] An estimation determination unit determines an estimated path score of each available path based on real-time network state data of all distributed nodes and a preset algorithm;

[0133] A score determination unit determines a comprehensive path score of each available path based on the path score and the estimated path score of each available path;

[0134] An optimal determination unit determines an optimal path as an available path with the highest comprehensive path score, and determines other available paths except the available path with the highest path score as candidate available paths;

[0135] A redundancy determination unit determines a redundancy path as a candidate available path with a comprehensive path score greater than a preset score threshold.

[0136] In this embodiment, the estimated path score is a preliminary score of each available path based on real-time network state data and a preset algorithm. This score is obtained by analyzing the performance indicators (such as delay, bandwidth, packet loss rate, etc.) of the current network to predict the potential performance of each path. The purpose of the estimated path score is to provide a preliminary evaluation in the early stage of the path selection process to help further decision-making. Assuming that there are three available paths in the network, path A, path B and path C. Through real-time collected network state data, the system evaluates that the delay of path A is low, the bandwidth of path B is large, and the packet loss rate of path C is high. Through a preset algorithm, the system generates an estimated path score for each path. For example, the estimated path score of path A is 85, the estimated path score of path B is 90, and the estimated path score of path C is 60. This score reflects the preliminary quality evaluation of each path.

[0137] In this embodiment, the preset score threshold refers to a minimum score standard set by the system in the path selection process. This threshold is used to determine which candidate paths are good enough in performance and can be used as redundant paths. Paths with scores lower than this threshold will be excluded and not selected as redundant paths. For example, assume that the system's preset score threshold is 70 points. When the calculated path score is lower than 70 points, the path will be considered to be insufficient in performance and cannot be used as a redundant path. For example, the comprehensive path score of path A is 80 points, and the comprehensive path score of path B is 60 points. Since the score of path B is lower than 70 points, path B will not be determined as a redundant path, and path A meets the requirements of a redundant path.

[0138] The beneficial effects of the above technical solutions are: the network path selection is optimized by the path determination module. The system generates an estimated path score for each available path based on real-time network state data and a preset algorithm through the estimation determination unit, and calculates the comprehensive path score of each path in combination with the score determination unit. The optimal determination unit selects the path with the highest score as the optimal path, and other paths as candidate paths. The redundancy determination unit determines the redundant path according to the preset score threshold, ensuring the stability and reliability of network connection. Compared with the traditional static path selection method, the present application can dynamically select the optimal path according to the real-time network status and provide redundant paths for possible network interruption, thereby improving the network performance, fault tolerance and user experience.

[0139] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present application, and not to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still modify the technical solutions recorded in the foregoing embodiments, or make equivalent replacements for some technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. An intelligent global hybrid cloud routing system based on SD-WAN architecture, characterized in that, Comprise: Data acquisition module: acquire the data stream to be transmitted, compress the data stream to be transmitted based on the preset compression algorithm, and then acquire the compressed data stream; Node determination module: based on the preset network performance monitoring tool, the network performance related data is obtained, and at the same time, the demand analysis of the preset network planning demand is carried out, and based on the demand analysis result, the network performance related data and the compressed data stream, a plurality of dynamic distributed nodes corresponding to the SD-WAN control layer are determined; Node connection module: acquire the node type of each distributed node, and then connect all distributed nodes, and connect all distributed nodes to the SD-WAN control layer; Scoring determination module: based on the SD-WAN control layer, the real-time network state data of all distributed nodes is collected and analyzed, and then combined with the preset user demand, the path score of each available path is determined; Path determination module: based on the path score of each available path and the real-time network state data of all distributed nodes, the optimal path and the redundant path are determined; Path switching module: based on the SD-WAN control layer, the network state data of all distributed nodes after data transmission of the optimal path is obtained, and then the redundant switching condition is determined and whether to switch the redundant path is judged, if it is judged to switch the redundant path, the redundant path switching is carried out; Edge routing cache module: acquire the access characteristics of the data to be transmitted, combine the preset cache strategy, cache the user high-frequency access data in the edge router, acquire the network state data and then dynamically adjust the cache content, and then optimize the data transmission process; Among them, the node determination module comprises: Initial node determination unit: demand analysis is carried out on the preset network planning demand, and a plurality of initial distributed nodes are determined based on the demand analysis result; Demand acquisition unit: compress the compressed data stream to determine the compression ratio, and acquire the transmission demand based on the preset compression ratio-demand database; Node adjustment unit: based on the transmission demand, the initial distributed nodes are adjusted for the first time, the network performance related data is obtained based on the preset network performance monitoring tool, the initial distributed nodes after the first adjustment are adjusted for the second time based on the network performance related data, and then a plurality of dynamic distributed nodes are obtained; Among them, the node adjustment unit comprises: Node evaluation subunit: based on the preset network performance monitoring tool, the network performance related data is obtained, and each initial distributed node is evaluated based on the network performance related data; Node determination subunit: based on the preset routing table, the first topology structure of the area in the preset range where each initial distributed node is located is determined, the first topology structure of the area in the preset range where each initial distributed node is located is searched based on the preset graph search algorithm, and then a plurality of associated nodes of each initial distributed node are determined; Node adjustment subunit: when the evaluation result of each initial distributed node does not meet the preset evaluation requirement, the load of each initial distributed node is dispersed based on the associated nodes of each initial distributed node; Among them, the node adjustment subunit comprises: The number determination block: when each initial distributed node evaluation result does not meet the preset evaluation requirement, the number of associated nodes for dispersing the load of each initial distributed node is determined based on the evaluation result of each initial distributed node; The threshold determination block: the distance between each initial distributed node and each associated node is obtained, and then the distance threshold corresponding to each initial distributed node is determined in combination with the number of associated nodes of the load of each initial distributed node; wherein, is a distance threshold of the i-th initial distributed node, is a number of all associated nodes of the i-th initial distributed node, is a distance between the i-th initial distributed node and the j-th associated node, is a number of associated nodes for dispersing the load of the i-th initial distributed node, is a load of the i-th initial distributed node, is a load variation coefficient of the i-th initial distributed node, wherein, is a distance fluctuation coefficient of the i-th node; The node determination block: the associated nodes with the distance less than the distance threshold between each initial distributed node and each associated node are determined as dispersed nodes; The load dispersion block: the load of each initial distributed node is dispersed based on the dispersed nodes of each initial distributed node. 2.The intelligent global hybrid cloud routing system based on SD-WAN architecture of claim 1, wherein, The initial node determination unit comprises: The demand determination subunit: the demand analysis is performed on the preset network planning demand, and then the number of data centers and the number of business demands corresponding to each data center are determined; The type determination subunit: the number of node types corresponding to each business demand is determined based on the type of each business demand and the preset type-node type database; The demand analysis subunit: the demand analysis is performed on each business demand corresponding to each data center, and then the number of nodes of each node type corresponding to each business demand is determined, and then the node type corresponding to each data center and the number of nodes of each node type are determined; The location determination subunit: the location of the nodes corresponding to each data center is determined based on the location of each data center; The node determination subunit: the number of nodes corresponding to each data center is determined based on the location of the nodes corresponding to each data center, the node type corresponding to each data center, and the number of nodes of each node type, and then the nodes of all data centers are determined as initial distributed nodes. 3.The intelligent global hybrid cloud routing system based on SD-WAN architecture of claim 1, wherein, The score determination module comprises: The path determination unit: the source node and the target node are determined based on the preset user demand, and then the number of available paths between the source node and the target node is determined; The score determination unit: the real-time network state data of all distributed nodes is collected and analyzed by the SD-WAN control layer, and then the path score of each available path is determined in combination with the preset user demand.

4. The intelligent global hybrid cloud routing system based on SD-WAN architecture according to claim 3, characterized in that, The score determination unit comprises: The data acquisition block: the real-time network state data of all distributed nodes is collected by the SD-WAN control layer using the preset network monitoring tool, wherein the real-time network state data contains a number of performance indicators; The score calculation block: the weight of each performance indicator is determined based on the preset user demand, and then the path score of each available path is determined in combination with the preset scoring algorithm.

5. The intelligent global hybrid cloud routing system based on SD-WAN architecture according to claim 3, wherein, The path determination module comprises: The estimation determination unit: the estimated path score of each available path is determined based on the real-time network state data of all distributed nodes and the preset algorithm; The score determination unit: the comprehensive path score of each available path is determined based on the path score and the estimated path score of each available path; The optimal determination unit: the available path with the highest comprehensive path score is determined as the optimal path, and the other available paths except the one with the highest path score are determined as the candidate available paths; The redundancy determining unit determines the candidate available path with the comprehensive path score greater than the preset score threshold as a redundant path. The redundancy determining unit determines the candidate available path with the comprehensive path score greater than the preset score threshold as a redundant path.

Images