Hybrid software defined wide area network device upgrading method with key performance perception
By establishing key equipment indicators and heuristic algorithms in hybrid software-defined wide area networks, optimizing switch upgrades and controller deployments, the problem of insufficient fine-grained evaluation in the existing technology is solved, and network performance improvements and delay reductions are achieved.
Patent Information
- Application Number
- CN202510580563.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-05-07
- Publication Date
- 2025-08-08
AI Technical Summary
The prior art lacks fine-grained flow path programmability evaluation and key link identification in hybrid software-defined wide area networks, resulting in the inability to effectively improve network performance.
By establishing the key device indicator DIRi, using Apollo heuristic algorithm, priority is given to upgrading switches with high programmability in key flow paths, and rationally deploying SDN controllers to form fine-grained upgrade priority evaluation standards to realize the mapping between switches and controllers.
It improves the critical programmability of the network, reduces the propagation delay of control information, and improves the overall performance of the network.
Smart Images
Figure CN120455416A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of computer network technology, and in particular to a key performance-aware hybrid software-defined wide area network device upgrade method. Background Art
[0002] Software-defined networking (SDN) technology has revolutionized modern network architecture by enabling flexible network management and reducing operational costs. Its applications have spread across multiple scenarios, including wide area networks (WANs), data centers, enterprises, and campus networks. In software-defined wide area networks (SD-WANs), since network devices are widely distributed across different locations, they are typically managed collaboratively using multiple distributed SDN controllers. These controllers regularly synchronize and maintain a global network view, forming a physically decentralized yet logically centralized control plane. Based on this architecture, SDN controllers can monitor the entire network status in real time and optimize network performance by dynamically adjusting traffic paths. However, in actual deployments, due to upgrade costs and technical compatibility constraints, traditional network equipment and SDN switches will coexist for a long time. This makes maximizing network benefits through gradual upgrades a key challenge.
[0003] The device upgrade strategy in a hybrid SD-WAN significantly impacts network performance. Traditional WAN architectures lack flexible routing adjustments, making them difficult to adapt to dynamic traffic changes. By upgrading traditional switches to SDN devices, the controller can intelligently schedule flows passing through the SDN switch, a process known as path programmability. This path programmability is a core advantage for improving network performance. Existing solutions primarily employ two optimization paths: one that maximizes the scale of programmable flows, and the other that focuses on geographic deployment to reduce propagation latency between the controller and switches. The most representative solutions attempt to balance these two dimensions, but they still have significant limitations.
[0004] Through in-depth analysis, we found that the performance improvements of existing solutions in actual deployments are limited mainly due to two core issues: First, the current coarse-grained method, which uses the number of programmable flows as an evaluation metric, fails to effectively distinguish the scheduling value of the paths of different flows. In fact, the path programmability of different business flows varies significantly, and highly programmable flows are more important for improving network adaptability. Second, flows on critical links (such as the most congested links) have a decisive impact on global performance. Under the constraints of limited upgrade resources, prioritizing the improvement of the programmability of such critical flows should be the focus of optimization. Therefore, there is an urgent need to establish a fine-grained evaluation system that can quantify the differences in the criticality and programmability of flows. Summary of the Invention
[0005] In view of this, the present invention provides a key performance-aware hybrid software-defined wide area network device upgrade method, which can improve the key programmability of the network and reduce control delay during switch upgrades and controller deployment.
[0006] The present invention provides a key performance-aware hybrid software-defined wide area network device upgrade method, comprising the following steps:
[0007] Step 1: Calculate the device criticality-aware programmability of switches in a set S in the wide area network based on device criticality, and sort S from highest to lowest programmability, where the device criticality is the ratio of the sum of the programmability of key flows in the switch to the sum of the programmability of all flows; establish a set X;
[0008] Step 2: Select switch i from S in sequence. If the sum of the upgrade policies of S is not greater than the threshold, set the upgrade policy of switch i to 1, update its sequence number to set X, and execute step 3; otherwise, execute step 3;
[0009] Step 3: Delete it from S and create set Y. If S is empty, execute step 4; if not, execute step 2.
[0010] Step 4: Select switch i from X. When the sum of feasible deployment strategies of Y is not greater than the threshold, deploy the controller at position i, grant the control of switch i to the controller and update the remaining control capability. Then execute step 6. When the sum is greater than the threshold, establish the set Y. * Save the location and proceed to step 4;
[0011] Step 5. From Y * Select a position, and when the remaining control capability of the corresponding controller is not less than the threshold, assign the control right of switch i to the controller, update the remaining control capability and execute step 6; otherwise, assign it to the controller from Y * Delete, Y * If it is not empty, go to step 5; if it is empty, go to step 6;
[0012] Step 6: Delete switch i from X. If X is not empty, execute step 4. If it is empty, end the process.
[0013] Furthermore, the equipment key indicator DIR i The calculation formula is as follows:
[0014]
[0015] in, Indicates switch s i For the flow f l Programmability, if the flow f l Flow through switches i And si There are at least two paths to reach flow f l The destination point is β i l =1, otherwise Represents the flow f l On switches i The number of feasible paths at the switch; i is the number of the switch, the minimum value is 1 and the maximum value is N; l is the number of the flow, the minimum value is 1 and the maximum value is L; x i Indicates switch s i Upgrade strategy, if switch s i Can be upgraded to an SDN switch i =1, otherwise x i =0.
[0016] Furthermore, the programmability includes key device indicators of the switch, programmability of the flow, number of feasible paths of the flow, and upgrade strategy.
[0017] Furthermore, the flow through switch s i The set of flows is F = {f 1 ,f 2 ,…,f l ,…,f L}, the set F is divided into two subsets, where {f 1 ,…,f K} contains all key flows, {f K+1 ,…,f L} includes all non-critical flows, and K is the maximum value of the critical flow number.
[0018] Furthermore, the method of selecting switch i from X in step 4 is: selecting according to the order of programmability from large to small in the set S.
[0019] Furthermore, the set Y * The locations stored in are sorted in ascending order according to their physical distance from switch i.
[0020] Furthermore, in step 5, * The method of selecting the position is: according to the set Y * The sort order is selected.
[0021] Furthermore, the remaining control capability of the controller is updated by the difference between the original remaining control capability and the total number of flows required to be controlled by the control switch i.
[0022] Furthermore, the key flows are identified by constructing a traffic matrix based on network status information extracted from the wide area network, and identifying the key flows from the traffic matrix.
[0023] Beneficial effects:
[0024] By establishing device criticality indicators, the present invention quantifies the value of programmability of key flows in switch upgrades, forms a fine-grained upgrade priority evaluation standard, and proposes a heuristic algorithm called Apollo to determine the switches to be upgraded, the location of deploying SDN controllers, and the mapping between the upgraded SDN switches and controllers. This can effectively improve the programmability of device criticality perception in real topology environments and effectively reduce the propagation delay of control information. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 A schematic diagram of the processing flow of a key performance-aware hybrid software-defined wide area network device upgrade method provided by the present invention. DETAILED DESCRIPTION
[0026] The present invention is described in detail below with reference to the accompanying drawings and with reference to the embodiments.
[0027] The present invention provides a key performance-aware hybrid software-defined wide area network device upgrade method. The core idea is: by establishing device key indicators and proposing a heuristic algorithm called Apollo, it is possible to improve path programmability in software-defined networks and restore the path programmability of offline flows in the network.
[0028] The present invention provides a key performance-aware hybrid software-defined wide area network device upgrade method, the processing flow is as follows Figure 1 As shown, the specific steps include:
[0029] Step 1: Build a traffic matrix based on the network status information extracted from the WAN, identify the key flows on each switch from the traffic matrix, and define the device criticality index of the switch as switch s i The ratio of the total programmability of the key flows to the total programmability of all flows in the switch, expressed as DIR i ; The programmability of the device criticality perception of the switch is calculated based on the device criticality index and is expressed as in, The value of the flow f l Flow through switches i And s i The arrival flow f exists l The number of paths to the destination point is determined by the switch s i For the flow f l Programmability, if the flow f l Flow through switches i And s iThere are at least two paths to reach flow f l The destination point otherwise Represents the flow f l On switches i The number of feasible paths at the switch; i is the number of the switch, the minimum value is 1 and the maximum value is N; l is the number of the flow, the minimum value is 1 and the maximum value is L; x i Indicates switch s i Upgrade strategy, if switch s i Can be upgraded to an SDN switch i =1, otherwise x i =0.
[0030] Among them, the key indicator DIR of the switch i , as shown below:
[0031]
[0032] Among them, the path programmability is the flow f l On switches i The path programmability at , expressed as When 1≤l≤K, the corresponding flow is the key flow, and K is the maximum value of the key flow number.
[0033] Typically, a WAN consists of N traditional switches located at N locations, represented by S = {s1,…,s i ,…,s N The set of flows passing through switch S is F = {f 1 ,f 2 ,…,f l ,…,f L}. The set F can be divided into two subsets, {f 1 ,…,f K} contains all key flows, and {f K+1 ,…,f L} contains all non-critical flows. If flow f l Flow through switches i , then it is expressed as otherwise x i Used to indicate the upgrade strategy of the switch. If the switch s i If it is upgraded to an SDN switch, it is represented by x i =1, otherwise x i =0.y j Used to represent the deployment strategy of the controller. If the controller is deployed at position j, it is represented as y j =1, otherwise yj =0. z ij Used to indicate the mapping policy between switch domain controllers. i is controlled by the controller at position j, which is represented by z ij =1, otherwise z ij =0.
[0034] In the SDN architecture, the path programmability of a flow on an SDN switch is reflected in the ability to dynamically adjust the flow path through the switch. It should be noted that the calculation of path programmability is only meaningful when there are two or more feasible paths for a flow; if the feasible path is unique, the SDN controller will not be able to modify its forwarding path. Specifically, if the flow f l Flow through switches i , and s i There are at least two paths to f l The destination point is expressed as otherwise variable Represents the flow f l On switches i The number of feasible paths at the destination. Since the switch can only adjust the next hop routing of the flow, the number of feasible paths here refers to the total number of valid routes starting from the next hop of the switch and finally reaching the destination of the flow. l On switches i The path programmability at can be expressed as
[0035] The number and scale of key flows passing through different switches vary significantly, so the benefits of upgrading traditional switches to SDN switches are also very different. Based on the analysis of real-world traffic matrices, existing work proposes a reinforcement learning-based method to select key flows in the network. This method uses historical traffic matrices to train the selection model. These traffic matrices are usually measured by SDN switches and collected periodically by SDN controllers. The selection of key flows is represented as a neural network that connects the original observation data (such as traffic matrices) with the combination of key flows. Neural networks provide an efficient and scalable solution for integrating various traffic matrices into key flow selection strategies. The training of the neural network involves a customized reinforcement learning algorithm. Once the training is completed, a small number of key flows are periodically selected based on the real-time traffic matrix as input. These selected flows usually exhibit high traffic loads and are more likely to cause network congestion. Based on this, we propose a device criticality indicator to quantify the extent to which the upgrade value of each traditional switch improves the programmability of the entire network. The device criticality indicator is defined as: switch s iThe ratio of the total programmability of the key flows to the total programmability of all flows in the switch, expressed as DIR i :
[0036]
[0037] Therefore, the programmability of device criticality awareness can be expressed as
[0038] Step 2: Sort the switches in the switch set S from largest to smallest according to the programmability of device criticality perception calculated in step 1, establish a set X containing the serial numbers of all upgradeable switches, and initialize the set X to an empty set.
[0039] Step 3: Select switch s from the switch set S i .
[0040] The best selection method is to select switches according to the sorting order of the switches in the switch set S.
[0041] Step 4: If the current switch set S satisfies Then x i is a feasible upgrade strategy, let x i =1, switch s i According to the strategy upgrade to the SDN switch, update the sequence number i to the set X, and s i Delete from the set S and execute step 5; otherwise, i Delete it from the set S and go to step 5. Where H is the maximum number of switches that can support upgrades.
[0042] Step 5. If the set S is not empty, execute step 3; otherwise, exit the loop and execute step 6.
[0043] Step 6: Create a set Y containing all controller positions and initialize the set Y to an empty set.
[0044] Controller deployment and switch-controller mapping. Ideally, the more SDN controllers deployed, the smaller the propagation delay between switches and controllers. However, these physically dispersed controllers need to periodically synchronize with each other to maintain a global network view, which can incur high synchronization overhead. Therefore, a maximum of M SDN controllers are deployed to control these SDN switches. These controllers are capable of handling flow requests from the SDN switches without incurring high synchronization overhead.
[0045] Step 7: Select the switch with sequence number i from the set X and set j = i.
[0046] The best way to select the switch number is to select it according to the sorting order in the set S.
[0047] Step 8: If the set Y satisfies Then deploying the controller at location j is a feasible deployment strategy, let y j = 1, deploy the controller to location j according to the strategy, and update location j to set Y. At the same time, switch s i The control authority of is given to the controller at position j, and z ij = 1, and update the residual control capability of the controller at position j Execute step 11; otherwise, establish set Y * , deploy the controllers in set Y in the same position as switch s i The physical distances between them are sorted from small to large and saved in set Y * , go to step 9.
[0048] in, is the residual control capacity of the controller, which represents the total number of flows that it can control without introducing additional delay (e.g., queuing delay); i To control switches i The total number of flows that need to be controlled is expressed as In the present invention, the residual control capability of the controller is required to remain non-negative.
[0049] Step 9. From set Y * Select the controller position j.
[0050] Among them, the best way to select the controller position is: according to the set Y * The sort order selection in .
[0051] Step 10: If satisfy Then switch s i The control authority of is given to the controller at position j, and z ij = 1, and update the residual control capability of the controller at position j Execute step 11; otherwise, remove j from the set Y * Delete, if Y * If Y is not empty, execute step 9. * If it is empty, exit the loop and execute step 11.
[0052] Step 11: Delete i from the set X. If X is not empty, execute step 7; otherwise, end this process.
[0053] Finally, it should be noted that it is apparent that those skilled in the art may make various modifications and variations to the present invention without departing from the spirit and scope of the present invention. Thus, to the extent such modifications and variations fall within the scope of the claims and their equivalents, the present invention is intended to encompass such modifications and variations.
[0054] The above is only an embodiment of the present invention, but it is not intended to limit the scope of the present invention. Any structural changes made according to the present invention, as long as they do not lose the essence of the present invention, should be considered to fall within the scope of protection of the present invention and be subject to restrictions. Those skilled in the art will clearly understand that for the convenience and brevity of description, the specific working process and related instructions of the method described above can refer to the corresponding process in the aforementioned method embodiment, and will not be repeated here.
[0055] The term "comprise," "comprising," or any other similar term is intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus / method that comprises a list of elements includes not only those elements but also other elements not expressly listed or inherent to such process, method, article, or apparatus / method.
[0056] Thus far, further embodiments have been listed to describe the technical solutions of the present invention. However, it is readily understood by those skilled in the art that the scope of protection of the present invention is obviously not limited to these specific embodiments. Without departing from the principles of the present invention, those skilled in the art may make equivalent changes or substitutions to the relevant technical features, and the technical solutions after such changes or substitutions will fall within the scope of protection of the present invention.
[0057] In summary, the above are only preferred embodiments of the present invention and are not intended to limit the scope of protection of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A key performance-aware hybrid software-defined wide area network device upgrade method, characterized in that: The following steps are involved: Step 1: Calculate the device criticality-aware programmability of switches in a set S in the wide area network based on device criticality, and sort S from highest to lowest programmability, where the device criticality is the ratio of the sum of the programmability of key flows in the switch to the sum of the programmability of all flows; establish a set X; Step 2: Select switch i from S in sequence. If the sum of the upgrade policies of S is not greater than the threshold, set the upgrade policy of switch i to 1, update its sequence number to set X, and execute step 3; otherwise, execute step 3; Step 3: Delete it from S and create set Y. If S is empty, execute step 4; if not, execute step 2. Step 4: Select switch i from X. When the sum of feasible deployment strategies of Y is not greater than the threshold, deploy the controller at position i, grant the control of switch i to the controller and update the remaining control capability. Then execute step 6. When the sum is greater than the threshold, establish the set Y. * Save the location and proceed to step 4; Step 5. From Y * Select a position, and when the remaining control capability of the corresponding controller is not less than the threshold, assign the control right of switch i to the controller, update the remaining control capability and execute step 6; otherwise, assign it to the controller from Y * Delete, Y * If it is not empty, go to step 5; if it is empty, go to step 6; Step 6: Delete switch i from X. If X is not empty, execute step 4. If it is empty, end the process.
2. The method for upgrading a hybrid software-defined wide area network device according to claim 1, wherein: The key indicators of the equipment DIR i The calculation formula is as follows: in, Indicates switch s i For the flow f l Programmability, if the flow f l Flow through switches i And s i There are at least two paths to reach flow f l The destination point otherwise Represents the flow f l On switches i The number of feasible paths at the switch; i is the switch number, the minimum value is 1 and the maximum value is N; l is the flow number, the minimum value is 1 and the maximum value is L; x i Indicates switch s i Upgrade strategy, if switch s i Can be upgraded to an SDN switch i =1, otherwise x i =0.
3. The method for upgrading a hybrid software-defined wide area network device according to claim 1, wherein: The programmability includes the key device indicators of the switch, the programmability of the flow, the number of feasible paths of the flow, and the upgrade strategy.
4. The method for upgrading a hybrid software-defined wide area network device according to claim 2, wherein: Flow through switches i The set of flows is F = {f 1 ,f 2 ,…,f l ,…,f L }, the set F is divided into two subsets, where {f 1 ,…,f K } contains all key flows, {f K+1 ,…,f L } includes all non-critical flows, and K is the maximum value of the critical flow number.
5. The method for upgrading a hybrid software-defined wide area network device according to claim 1, wherein: The method of selecting switch i from X in step 4 is: selecting according to the order of programmability from large to small in the set S.
6. The method for upgrading a hybrid software-defined wide area network device according to claim 1, wherein: The set Y * The locations stored in are sorted in ascending order according to their physical distance from switch i.
7. The method for upgrading a hybrid software-defined wide area network device according to claim 6, wherein: Step 5 is as follows: * The method of selecting the position is: according to the set Y * The sort order is selected.
8. The method for upgrading a hybrid software-defined wide area network device according to claim 1, wherein: The remaining control capability of the controller is updated by the difference between the original remaining control capability and the total number of flows required to be controlled by the control switch i.
9. The method for upgrading a hybrid software-defined wide area network device according to claim 1, wherein: The key flow is identified by constructing a traffic matrix based on network status information extracted from the wide area network, and identifying the key flow from the traffic matrix.