An intelligent routing method for SD-WAN based on link quality
By introducing intelligent data analysis and link quality monitoring in the SD-WAN system, dynamically optimizing the routing strategy is solved, and the problem of lack of flexibility and accuracy in traditional SD-WAN routing is achieved, and higher quality and lower cost network services are achieved.
Patent Information
- Application Number
- CN202211477676.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-11-23
- Publication Date
- 2025-03-25
- Estimated Expiration
- 2042-11-23
AI Technical Summary
The routing of traditional SD-WAN devices lacks flexibility, accuracy and timeliness, making it difficult for users to configure routing strategies that meet business needs, affecting network quality.
Using an intelligent route selection method based on link quality, through the SD-WAN controller and data management platform, combined with intelligent data analysis algorithms and error correction mechanisms, we dynamically monitor link quality and optimize route selection strategies to ensure that the application runs on links that meet quality requirements.
It simplifies the user's routing policy configuration, improves network quality accuracy, reduces link load and application delay, and effectively reduces network cost.
Smart Images

Figure CN115883441B_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of data communication and relates to a multi-link intelligent routing method based on link quality for a SD-WAN. Background Art
[0002] SD-WAN (Software Defined Wide Area Network) is a new generation of network communication technology. Dynamic path selection is one of its core capabilities. Users configure routing strategies to switch applications between different network links to ensure network quality. Currently, most SD-WAN products mainly distribute application traffic between different links by allocating different network links to different applications. This model is simple to implement, and the business system completely relies on the user's configuration strategy to select links. However, simplicity also means a lack of flexibility and intelligence. Users need to have a good understanding of the link quality of the network and the application's requirements for link quality in order to configure a reasonable routing strategy. Although the configuration is simple, it is still very challenging to configure a routing strategy that meets business needs. Summary of the invention
[0003] In order to solve the problems of lack of flexibility, accuracy and timeliness in the terminal routing of traditional SD-WAN equipment and provide users with high-quality network services, the present invention provides an SD-WAN intelligent routing method based on link quality. Based on the user's basic routing configuration, the present invention introduces an intelligent data analysis algorithm and error correction mechanism to provide a more reasonable multi-link intelligent routing method to effectively ensure the quality of network services.
[0004] The present invention provides an SD-WAN intelligent routing method based on link quality, comprising the following steps:
[0005] Step 1: The user configures information on the SD-WAN controller, including: the cost and bandwidth information of the network link on the SD-WAN device terminal, and the level of network quality required by the application on the device; the SD-WAN controller uploads the configuration information to the SD-WAN data management platform;
[0006] Step 2: The SD-WAN device terminal establishes network links with different network operators. The device terminal establishes a virtual tunnel with another terminal device under each network link, monitors the link quality of the tunnel, and obtains the packet loss rate, delay and jitter. The SD-WAN device terminal also detects the packet loss rate, delay and jitter of different applications in the current link. The SD-WAN device terminal uploads the monitored link quality data to the SD-WAN data management platform through the SD-WAN controller.
[0007] Step 3, the SD-WAN data management platform performs:
[0008] Step 3.1, obtaining the routing strategy configured by the user, and starting the expert strategy library to detect the routing strategy configured by the user; wherein the routing strategy configured by the user selects a matching network link for the application according to the network quality requirement level of the device application, and selects a priority for the application;
[0009] In step 3.2, the SD-WAN data management platform intelligently optimizes the routing strategy configured by the user, including: evaluating the currently acquired link quality data according to the set routing strategy evaluation indicators to determine the current link quality level; intelligently optimizing the routing strategy configured by the user for the application based on the current link cost, quality and application network quality requirement level, and sending the optimized routing strategy to the SD-WAN device terminal via the SD-WAN controller for execution.
[0010] The expert policy library pre-stores the priorities of different applications under different categories and the link quality requirements of the applications, and the link quality requirements include packet loss rate, delay and jitter; the user sets the application priority and network quality requirement level according to the expert policy library, or sets the application priority and network quality requirement level based on the expert policy library according to actual needs.
[0011] The step 3.2 includes:
[0012] (1) Setting routing strategy evaluation indicators, including link cost indicator, link bandwidth indicator, link delay indicator and link packet loss rate indicator; the link cost indicator is divided into three cost ranges: high, medium and low. The dedicated line charged by traffic volume is divided into the high cost range, and the link with monthly traffic package is divided into the low cost range. The link bandwidth indicator, link delay indicator and link packet loss rate indicator are dynamic indicators, and the link quality is divided into three levels: high, medium and low.
[0013] (2) Analyze the link quality data collected by the device terminal and regularly evaluate the link quality level;
[0014] (3) Based on the current link cost, quality and the application's network quality requirement level, intelligent optimization is performed on the basis of the routing strategy configured by the user for the application, including: (31) If the quality of the network link currently allocated to the application deteriorates and does not meet the application's network quality requirement level, switch the application to a network link that meets the level; (32) Allocate network links based on application priority to ensure network link allocation for high-priority applications, and select low-cost links from the allocable network links; (33) When network link congestion occurs, limit the speed of low-priority applications based on the application priority.
[0015] The SD-WAN data management platform also sends the thresholds of link quality requirements for different applications under different categories stored in the expert policy library to the SD-WAN device terminals through the SD-WAN controller. The SD-WAN device terminals will detect whether the network link quality currently allocated to the application is applicable based on the thresholds.
[0016] The advantages and positive effects of the present invention are:
[0017] (1) The method of the present invention can simplify the convenience of users in configuring routing strategies. In traditional routing strategy configuration, users must have a deeper understanding of the current link quality and the application's requirements for link quality in order to configure a reasonable routing strategy. However, link quality is not static, so users must constantly adjust the application routing threshold in business applications. If the routing threshold is not configured reasonably, it will cause unreasonable switching of applications between different links, which will increase the link load, increase the application delay, and cause packet loss. The method of the present invention combines link quality and the application's conventional requirements for link quality to provide users with a reasonable threshold reference, thereby simplifying user configuration and improving network quality.
[0018] (2) The method of the present invention can effectively improve network quality. In the past, the system mainly selected routes based on the routing threshold configured by the user. However, user configuration is subjective. Since network routing requires comprehensive reference to multiple indicators, the routing threshold configured by the user based on his or her own subjective consciousness cannot accurately reflect the actual situation of the network. The method of the present invention collects data from multiple parties and provides an expert strategy library to guide users to configure routing thresholds that are more in line with actual needs. The method of the present invention improves the accuracy of user threshold configuration by 50% and reduces the frequency of application link switching by 60%.
[0019] (3) The method of the present invention uses link charges as an assessment indicator for network routing. Users set the assessment weights of the cost and quality of the application. The terminal device conducts a comprehensive analysis of the link cost and link quality when selecting a route, thereby obtaining an application link that better meets the customer's expectations. The method of the present invention can meet the user's needs to reduce network costs while maximizing network quality. It has been proven in practice in a production environment that it can effectively reduce the user's network costs. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] Figure 1 It is a flow chart for implementing the SD-WAN intelligent routing method based on link quality of the present invention;
[0021] Figure 2 is a schematic diagram of configuring network link costs according to an embodiment of the present invention;
[0022] Figure 3It is a schematic diagram of a routing strategy configured by a user for an application in an embodiment of the present invention;
[0023] Figure 4 It is a schematic diagram of network selection based on link quality according to an embodiment of the present invention. DETAILED DESCRIPTION
[0024] In order to make the purpose, technical solution and advantages of this application more clear, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention, but are not used to limit the present invention.
[0025] The SD-WAN intelligent routing method based on link quality of the present invention involves an SD-WAN device terminal, an SD-WAN controller, and an SD-WAN data management platform. The three parties conduct comprehensive data analysis based on the link data collected by the SD-WAN device terminal to provide guiding parameters for routing. The SD-WAN device terminal collects network quality data, including latency, jitter, packet loss and other data, network link bandwidth, cost and performance index requirements of business applications for network quality, and reports them to the SD-WAN data management platform through the SD-WAN controller; the SD-WAN data management platform calculates a reasonable routing scheme and link switching threshold based on the collected basic data using an intelligent algorithm, thereby providing users with a suggested routing configuration strategy to guide users to configure the routing strategy. At the same time, the data management platform sends the optimized routing scheme and link switching threshold to the device terminal through the SD-WAN controller. When the device terminal has a network quality problem, intelligent network switching is performed according to the network link analysis results, thereby improving the network service quality.
[0026] An implementation process of the SD-WAN intelligent routing method based on link quality in an embodiment of the present invention is as follows: Figure 1 As shown, the following is divided into three parts to explain.
[0027] In the first part, users configure the cost information of the network link and the application's requirement level for network quality on the SD-WAN controller.
[0028] (1.1) The user configures the cost information of the network link on the SD-WAN controller. After the device is started, the SD-WAN controller obtains the interface information of the CPE device by communicating with the device terminal (CPE), and takes over the control of the CPE WAN port through the SD-WAN controller. This step mainly configures the cost information of the CPE WAN port link. For example, the CPE WAN port eth0 is connected to the dedicated line link, eth1 is connected to the Internet link of operator A, and eth2 is connected to the Internet link of operator B; the billing mode of eth0 is based on traffic flow, and the price is 10 yuan / M, the billing mode of the eth1 link is monthly bandwidth billing, and the bandwidth is 1000M, and the billing mode of the eth2 link is monthly bandwidth billing, and the bandwidth is 500M. Users need to submit the billing and bandwidth information of these links to the SD-WAN management platform. These data will be used as the billing basis for subsequent application policy routing.
[0029] like Figure 2 As shown, it is a schematic diagram for configuring network link costs.
[0030] (1.2) The user configures the application's network quality requirement level on the SD-WAN controller. Different enterprises have different requirements for application network quality, so the SD-WAN data management platform provides the ability to configure application network quality requirements. The data management platform first defines some common application identifiers. For those that are not defined by the platform, users can customize application identifiers based on the application quintuple, or define them based on or in combination with the domain name accessed by the application. With the application identifier, the user selects the corresponding application and defines the application's network quality requirement level.
[0031] Different from the traditional configuration of the corresponding threshold of the application, the method of the present invention only needs to configure the application's requirement level for the network, for example, the network requirement level of application A is: high; the network requirement level of application B is: medium.
[0032] The user configures the above information on the SD-WAN controller in advance so that the SD-WAN data management platform can perform intelligent routing for different applications of terminal devices.
[0033] In the second part, the SD-WAN device terminal performs link detection and uploads monitoring data to the data management platform through the controller.
[0034] The equipment terminal uses NQA (Network Quality Analysis) and IP FPM (IPFlow Performance Measurement) technologies to detect the network quality of the current link. NQA is a real-time network performance detection and statistics technology that can collect statistics on network information such as response time, network jitter, packet loss rate, etc. NQA detects the performance of multiple protocols running on the network, allowing users to collect various network operation indicators in real time, such as the total delay of Http, the delay of TCP (Transmission Control Protocol) connection, the delay of DNS (Domain Name System) resolution, the file transfer rate, the delay of FTP (File Transfer Protocol) connection, the error rate of DNS resolution, etc. IP FPM is a passive detection solution for network performance based on IP. It carries statistical information by designing tags (also known as dyeing) for packets to achieve end-to-end performance statistics of three-layer networks.
[0035] (2.1) First, the link quality data to be collected by the SD-WAN device terminal is explained. Link quality is mainly composed of three indicators: packet loss rate, delay, and jitter. The device terminal (CPE) provides a variety of link detection technologies for these three indicators. The following describes the process of measuring link quality data for IP flow performance.
[0036] (2.1.1) Packet loss detection: Packet loss detection technology includes dyeing all packets in a single cycle, sending KeepAlive messages between the two ends of the tunnel, and actively sending detection messages.
[0037] Dyeing all packets in a single cycle: Dyeing is achieved by marking (i.e. dyeing) the reserved bits in the IP packet header. Usually dyeing is done at one end of the tunnel and un-dyeing is done at the other end.
[0038] The two ends of the tunnel send KeepAlive messages to each other: the KeepAlive message sent by the device terminal CPE-A to the device terminal CPE-B carries the message count SndA(T1) sent within the service cycle T1 and the sending time timestamp TsA(T1); after receiving the KeepAlive message, CPE-B sends the message count RcvB(T1) received within the service cycle T1 and the reception time TrB(T1) of the KeepAlive message of CPE-A. After the statistical information from CPE-A to CPE-B is synchronized, both ends can calculate the packet loss rate of the message sent by CPE-A. The calculation formula for the single-item packet loss rate LossA is as follows:
[0039] LossA=(SndA–RcvB) / SndA*100%
[0040] (2.1.2) Bidirectional delay detection: Bidirectional delay detection does not require sending additional messages, because the KeepAlive message of packet loss detection will carry a timestamp (sending time Ts, receiving time Tr), which can be used directly.
[0041] If the time difference between CPE-A sending a data packet and CPE-B receiving a data packet is M, the calculation formula for the single two-way transmission delay Dalay is as follows:
[0042] Dalay=TrB-(TsA+M)+(TrA+M)-TsB=(TrB-TsA)+(TrA-TsB)
[0043] The time when CPE-A sends the message is TsA, the time when CPE-B sends the message is TsB, the time when CPE-B receives the KeepAlive message from CPE-A is TrB, and the time when CPE-A receives the KeepAlive message from CPE-B is TrA.
[0044] (2.1.3) Jitter detection: Jitter detection does not require sending additional messages. This is because jitter is a change in the experiment. Therefore, it is only necessary to calculate the delay obtained from multiple measurements to obtain the jitter, as shown below:
[0045]
[0046] Where pdv is the jitter value to be calculated; N is the number of times the delay is counted within a period; M is the delay at which statistics are started; Delay[i] is the i-th statistical delay; and avrDelay{N,M} is the average value of the N delay statistics.
[0047] (2.2) The SD-WAN device terminal collects data and obtains link quality data.
[0048] In the embodiment of the present invention, the SD-WAN device terminal collects link quality data every 5 minutes. The device terminal will analyze the virtual tunnel established by the device and monitor the delay, jitter, and packet loss indicators of each tunnel. The device terminal will establish different network links with different network operators. As described in step 2.1 above, it is assumed that the current device terminal A and another device terminal B establish virtual tunnels through different network links, and monitor the delay, jitter, and packet loss rate of different virtual tunnels.
[0049] At the same time, since different applications have different requirements for link quality, the measures to ensure the experience are also different. The device terminal will detect the delay, jitter, and packet loss data of various applications in the current link.
[0050] The device terminal in the present invention mainly identifies the application in the following ways: (1) based on the message five-tuple identification; (2) based on the message flow characteristics identification; (3) based on the message load identification; (4) based on the DNS identification, etc. The application identifier can be set according to these identification methods, but other identifiers that can identify different applications can also be used.
[0051] Finally, the device terminal obtains the following data: (1) the delay, jitter, and packet loss rate of each tunnel on the current device; (2) the delay, jitter, and packet loss rate of various applications on the selected tunnel.
[0052] In the third part, the SD-WAN data management platform conducts intelligent analysis on the collected network link quality data and performs network link selection and correction.
[0053] (3.1) The SD-WAN data management platform obtains data, including configuration information about routing strategies and link quality data collected by device terminals from the SD-WAN controller. The SD-WAN data management platform cleans the collected link quality data, eliminates interference data, and obtains the valuable data required for routing, mainly including: (1) the average latency, jitter, and packet loss rate of each tunnel of the device; (2) the actual latency, jitter, and packet loss rate data of each application on the tunnel.
[0054] (3.2) The SD-WAN data management platform obtains the routing strategy configured by the user, starts the expert strategy library to detect its rationality, and provides feedback on the strategy configuration recommendations.
[0055] (3.2.1) The routing strategy configured by the user is to match the corresponding network link for the application according to the network quality requirement level of the device application and select the priority for the application. Figure 3 , which is a schematic diagram of a user configuring a routing strategy for an application. The routing strategy configured by the user will be sent to the data management platform through the controller.
[0056] Normally, the user-defined routing strategy will specify that a certain type of application is forwarded through a certain tunnel link, but this cannot guarantee the session quality of the application, because the network quality of the link is not static. Therefore, in the SD-WAN intelligent routing method based on link quality of the present invention, the user can define the link based on the application session quality target. For example: the user defines video applications as high-priority applications, and at the same time defines that the video application can select two links A and B, where link A is a high-priority link and link B is a low-priority link. Then, at the device terminal, traffic distribution will be performed according to the user-defined routing rules, and the video application will be forwarded to link A. When quality problems occur in link A and cannot meet the network quality requirements of the video application, the device terminal will switch the network traffic of the video application to link B according to the user-defined routing strategy, thereby ensuring the network quality of the video application.
[0057] (3.2.2) The SD-WAN data management platform starts the expert policy library to test the routing policy configured by the user and determine whether the routing policy configured by the user is reasonable. If it is unreasonable, it will provide user configuration suggestions.
[0058] The expert strategy library is pre-set to guide users to configure routing strategies. The following detection steps will be performed:
[0059] a. Determine whether the application routing policy configured by the user has an escape link configured. If not, provide suggestions for configuring an escape link.
[0060] b. If the user configures a load balancing link, it is recommended to configure two links with the same network quality and billing level;
[0061] c. Specify the priority of the link. You can specify multiple links to use the same priority or different priorities.
[0062] d. If the load balancing link mode is used, a load sharing algorithm based on link bandwidth weight and link billing weight will be run on network links of the same priority, so that links with high bandwidth and low cost can share more traffic;
[0063] e. Configure application priority: Detect application priority. When the link is congested, low-priority applications can give way to high-priority applications, thereby ensuring the network quality of high-priority applications.
[0064] Users check and configure the application routing policy according to the policy configuration guide.
[0065] In the embodiment of the present invention, the priorities and link quality requirements of different types of applications are pre-set in the expert policy library as shown in Table 1. The user can directly set the application link quality requirement level and priority according to the expert policy library, or set the application link quality requirement level and priority based on the expert policy library according to actual needs.
[0066] Table 1 Link quality requirements and priorities for different applications
[0067]
[0068]
[0069] As shown in Table 1, for collaborative applications, VoIP has the highest priority, but the bandwidth it occupies is lower than that of video conferencing, and the requirements for packet loss rate, delay, and jitter are also lower than those of video conferencing, that is, the requirements for network link quality are lower than those of video conferencing. As shown in Table 1, the link quality requirements of each application are the set link selection thresholds. When the threshold is not met, it will switch to a link that meets the threshold.
[0070] (3.3) The data management platform performs intelligent optimization of the routing strategy. The intelligent routing of the present invention provides intelligent optimization capabilities of the application routing strategy through real-time monitoring of link quality based on the user routing strategy and intelligent algorithm. Intelligent routing mainly includes the following steps:
[0071] (3.3.1) The data management platform configures routing strategy evaluation indicators, which mainly include link cost indicators, link bandwidth indicators, link delay indicators, link packet loss rate indicators, etc. The link cost indicators divide the link billing mode and price into three intervals: high cost, medium cost, and low cost. Among them, dedicated lines charged by traffic belong to the high cost interval, monthly packages belong to the low cost interval, and others belong to the medium cost interval. The link bandwidth indicator, link delay indicator, and link packet loss rate indicator will be reasonably divided and continuously adjusted and optimized based on the user's actual business bandwidth, delay, packet loss rate and other data, and statistical historical data. The above three indicators are dynamic change indicators. The network link quality can be divided according to these three indicators, and the network link quality can be divided into three levels: high-quality link, medium-quality link, and low-quality link. Dynamic change indicators are also sent to device terminals so that terminals can perform network quality classification.
[0072] (3.3.2) Analyze the network link quality data collected from the device terminal, perform regular calculation and evaluation of the link quality based on bandwidth utilization, latency, and packet loss rate, and determine the current network link level. In this embodiment of the present invention, the link quality data is collected from the device terminal every 5 minutes and the link quality is evaluated once.
[0073] (3.3.3) Based on the current link cost and quality and the application's network quality requirements, intelligent optimization is performed on the basis of the routing strategy configured by the user for the application to achieve intelligent selection of the links required by the application.
[0074] After obtaining the link billing ladder data, link quality data, and the latency and packet loss rate of each application on the link, the current network link of the application is adjusted in combination with the user's demand for application quality. The adjustment principles are: (1) First, detect whether to switch the network link based on the application's network quality requirement level. If the quality of the network link currently allocated to the application is not good, that is, the quality of the current network link deteriorates and does not meet the application's network quality requirement level, it will switch to the network link of the current corresponding level; (2) Secondly, allocate network links based on the priority of the application, first ensure the network link allocation of high-priority applications, and select low-cost links on the allocateable network links. (3) When network link congestion occurs, limit the speed of low-priority applications based on the priority of the application.
[0075] like Figure 4 As shown, when an application of two device terminals communicates, if the current link is interrupted or seriously delayed, the link will be reallocated and a link with suitable network quality that meets the link quality requirement threshold and the network quality requirement level set by the user will be selected.
[0076] Ensure high-priority applications, and try to allocate low-cost links to high-traffic applications. For example: video conversation applications are high-priority applications. In order to ensure the quality of the conversation, they can be first allocated to the Internet link. The cost of using the Internet link is relatively low. If the link quality can be guaranteed under low cost, the Internet link will be used. Otherwise, the application will be upgraded to a high-cost dedicated line link. The method of the present invention comprehensively utilizes the user-defined routing strategy and intelligent analysis algorithm to perform comprehensive scheduling. For example: Link A is a dedicated line link, and Link B is an Internet link. Under normal circumstances, dedicated line links can often provide higher quality network services, but the price is also relatively expensive. Internet links may have higher latency, but the cost is relatively low. In the intelligent algorithm of the present invention, cost-benefit will be taken into account for a comprehensive evaluation to provide users with the highest possible network service quality while ensuring low cost.
[0077] Intelligent routing is to further monitor and analyze the link quality of the application based on the user application routing strategy, and adjust and optimize the link. Although the user has configured the routing strategy, there may still be link congestion caused by imperfect considerations. At this time, intelligent routing optimizes the link. For example: when the link is congested and multiple links cannot guarantee the link quality, the low-priority link will be speed-limited. Speed limits generally include average speed limits and discard speed limits. Average speed limits refer to speed limits for multiple low-priority applications at the same time. Discard speed limits refer to ranking the bandwidth occupancy of low-priority applications and giving priority to limiting low-priority applications with larger bandwidth occupancy.
[0078] The data management platform sends the dynamically changing indicators and optimized routing strategies to the device terminals through the SD-WAN controller. The device terminals monitor the network quality of the applications on the link in real time. When some applications experience severe packet loss or delay, the overall link quality is analyzed. At the same time, according to the importance and priority of the application, it is determined whether to switch the link of the application or perform multi-link load sharing and link optimization. The SD-WAN data management platform collects link quality monitoring results from the device terminals for big data analysis, and guides and corrects the routing strategies configured by users through intelligent algorithms. At the same time, through big data analysis, the expert strategy library is continuously optimized to improve the accuracy of the expert strategy library.
Claims
1. An SD-WAN intelligent routing method based on link quality, characterized in that: The steps include: Step 1: The user configures information on the SD-WAN controller, including: configuring the cost and bandwidth information of the network link on the SD-WAN device terminal, and the level of network quality required by the application on the device; the SD-WAN controller uploads the configuration information to the SD-WAN data management platform; Step 2: The SD-WAN device terminal establishes network links with different network operators. The device terminal establishes a virtual tunnel with another terminal device under each network link, monitors the link quality of the tunnel, and obtains the packet loss rate, delay and jitter. The SD-WAN device terminal also detects the packet loss rate, delay and jitter of different applications in the current link. The SD-WAN device terminal uploads the monitored link quality data to the SD-WAN data management platform through the SD-WAN controller. Step 3, the SD-WAN data management platform performs: Step 3.1, obtaining the routing strategy configured by the user, and starting the expert strategy library to detect the routing strategy configured by the user; wherein the routing strategy configured by the user selects a matching network link for the application according to the network quality requirement level of the device application, and selects a priority for the application; Step 3.2, the SD-WAN data management platform intelligently optimizes the routing strategy configured by the user, including: evaluating the currently acquired link quality data according to the set routing strategy evaluation indicators to determine the current link quality level; intelligently optimizing the routing strategy configured by the user for the application according to the current link cost, quality and the application's network quality requirement level, and sending the optimized routing strategy to the SD-WAN device terminal for execution via the SD-WAN controller; The step 3.2 includes: (1) Set up routing strategy evaluation indicators, including link cost indicator, link bandwidth indicator, link delay indicator and link packet loss rate indicator. The link cost indicator is divided into three cost ranges: high, medium and low. The dedicated line charged by traffic volume is divided into the high cost range, and the link with monthly traffic package is divided into the low cost range. The link bandwidth indicator, link delay indicator and link packet loss rate indicator are dynamic indicators, and the link quality is divided into three levels: high, medium and low. (2) Analyze the link quality data collected by the device terminal and regularly evaluate the link quality level; (3) Based on the current link cost, quality and the application's network quality requirement level, intelligent optimization is performed on the basis of the routing strategy configured by the user for the application, including: (31) If the quality of the network link currently allocated to the application deteriorates and does not meet the application's network quality requirement level, switch the application to a network link that meets the level; (32) Allocate network links based on application priority to ensure network link allocation for high-priority applications, and select low-cost links from the allocable network links; (33) When network link congestion occurs, limit the speed of low-priority applications based on the application priority.
2. The method according to claim 1, characterized in that In the step 1, different users have different network quality requirements for each application, and the user configures the application identifier and the application's requirement level for network quality on the SD-WAN controller; wherein the application identifier is used to identify different applications; and the application's requirement level for network quality is set to three levels: high, medium, or low.
3. The method according to claim 1, characterized in that: The step 3.1, detecting the routing strategy configured by the user and feeding back configuration suggestions, includes: Determine whether the routing policy applied in the user configuration policy has an escape link. If not, provide feedback on how to configure an escape link. If the user configures a load balancing link, feedback is given to suggest configuring two links with the same network quality and billing level; Specify the priority of network links, and specify that different links use the same priority or different priorities; If the user configures a load balancing link, a load sharing algorithm based on link bandwidth weight and link billing weight is run on network links with the same priority, so that network links with high bandwidth and low cost share more traffic; Detect application priority settings and make low-priority applications give way to high-priority applications when network links are congested.
4. The method according to claim 1, characterized in that: The expert strategy library stores the priorities of different applications under different categories and the thresholds of link quality requirements of the applications, where the link quality requirements include packet loss rate, delay and jitter; The user sets the application priority and the network quality requirement level according to the expert policy library, or sets the application priority and the network quality requirement level based on the expert policy library according to actual needs.
5. The method according to claim 1 or 4, characterized in that: In the above step 3, the SD-WAN data management platform also sends the thresholds of link quality requirements for different applications under different categories stored in the expert policy library to the SD-WAN device terminal through the SD-WAN controller. The SD-WAN device terminal will detect whether the network link quality currently allocated to the application is applicable based on the threshold.
Citation Information
Patent Citations
SD-WAN network intelligent link selection method based on cloud computing
CN111835639A
Service quality charging method of internet business based on service grade protcol
CN1815965A