A centralized control system and method for switch equipment

By collecting and analyzing the traffic data of switch equipment in real time, building a bottleneck detection model, and adjusting the traffic routing path, the low efficiency and response delay of switch equipment under large-scale data traffic and complex topology are solved, and efficient load balancing of the network is achieved and stability improvement.

CN118827520BActive Publication Date: 2024-11-22SHENZHEN KEWANGTONG TECH DEV
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411297469.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-18
Publication Date
2024-11-22
Estimated Expiration
2044-09-18

AI Technical Summary

Technical Problem

Existing switch equipment is inefficient in handling large-scale data traffic and complex network topology, has large response latency, and lacks dynamic perception of the real-time state of the network, resulting in poor network service quality.

Method used

Through the communication protocol between the controller and the switch equipment, traffic data is collected in real time, preprocessing and high-dimensional projection analysis is performed, network bottleneck detection model is built, bottleneck potential value is calculated, and inbound traffic routing path is adjusted according to this value to optimize the network topology.

Benefits of technology

It improves the load balancing and transmission efficiency of the network, enhances the dynamic adjustment ability of network topology and traffic changes, reduces data packet loss and delay, and improves network stability and resource utilization.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118827520B_ABST
    Figure CN118827520B_ABST
Patent Text Reader

Abstract

The present invention discloses a centralized control system and method for switch equipment, which relates to the technical field of switch equipment control. Through the communication protocol between the controller and the switch equipment, the flow data generated by the switch equipment is collected in real time, and the flow data is preprocessed; high-dimensional projection analysis and adaptive network analysis are performed on the preprocessed flow data to build a network bottleneck detection model, and the bottleneck potential value of the switch equipment is calculated; the corresponding switch equipment is sorted in descending order according to the bottleneck potential value, and the inbound flow of the switch equipment that is greater than the preset bottleneck potential threshold and ranks first is allocated to the switch equipment that ranks last in descending order; the topological structure of the switch equipment network is analyzed, and the routing path of the inbound flow is adjusted. By real-time monitoring and analysis of flow changes in the network, the routing path and bandwidth allocation of the flow are dynamically adjusted, the load balancing and transmission efficiency of the network are improved, and network congestion and resource waste are avoided.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of switch equipment control, and in particular to a switch equipment centralized control system and method. Background Art

[0002] With the rapid development of the Internet and data communication technologies, the scale and complexity of networks are increasing. As the core component of the network, the performance and efficiency of switch devices directly affect the stability and transmission speed of the entire network. In the prior art, traditional switch devices usually adopt a distributed control architecture. Although this architecture improves the flexibility of the network to a certain extent, it often faces problems such as low efficiency, large response delays, and poor load balancing when dealing with large-scale data traffic and complex network topologies.

[0003] In addition, existing traffic management methods often rely on static rule configurations and lack the ability to dynamically perceive the real-time state of the network, resulting in the inability to make timely adjustments when network traffic bursts or topology changes occur, which in turn affects the network quality of service (QoS). For large-scale data centers or operator networks, how to achieve efficient and stable traffic control and network topology management has become an urgent problem to be solved. Summary of the invention

[0004] Based on the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a centralized control system and method for switch equipment to solve the above-mentioned technical problems.

[0005] To achieve the above object, the present invention provides the following technical solution: a centralized control method for switch equipment, comprising:

[0006] Through the communication protocol between the controller and the switch device, the flow data generated by the switch device is collected in real time, and the flow data is pre-processed;

[0007] Through high-dimensional projection analysis and adaptive network analysis of pre-processed traffic data, a network bottleneck detection model is constructed to calculate the bottleneck potential value of switch equipment;

[0008] The corresponding switch devices are sorted in descending order according to the bottleneck potential value, and the inbound traffic of the switch device with a value greater than the preset bottleneck potential threshold and ranked first is allocated to the switch device with the lowest value in descending order;

[0009] Analyze the topology of the switch device network and adjust the routing path of inbound traffic.

[0010] The present invention is further configured to collect flow data generated by the switch device in real time, including: inbound flow, outbound flow, packet loss rate, port utilization rate and delay;

[0011] The flow data is preprocessed, including: performing outlier detection, missing value filling and denoising on the flow data, and the preprocessed flow data is recorded as ,in, For the Switch devices in Traffic data at the moment, , respectively Switch devices in Inbound traffic, outbound traffic, packet loss rate, port utilization, and latency at each moment.

[0012] The present invention is further configured to perform high-dimensional projection analysis and adaptive network analysis on pre-processed traffic data, construct a network bottleneck detection model, and calculate the bottleneck potential value of the switch device, including:

[0013] Through high-dimensional projection analysis, the preprocessed traffic data is projected into a low-dimensional space to obtain a low-dimensional feature vector ;

[0014] According to the low-dimensional feature vector Compute node affinity matrix , where the node association matrix Each element in Indicates The switch device and The dependence strength between switch devices;

[0015] According to the node association matrix and traffic data Calculate the bottleneck potential of each switch device .

[0016] The present invention is further configured to project the preprocessed flow data into a low-dimensional space through high-dimensional projection analysis to obtain a low-dimensional feature vector ,include:

[0017] Construct the projection matrix W to transform high-dimensional flow data Mapped to low-dimensional space, the construction of the projection matrix W is realized by optimizing the objective function, which is used to maximize the feature retention of high-dimensional flow data in low-dimensional space while minimizing information loss. The optimization objective function is: ,in, To solve the projection matrix , so that the objective function maximum, is the trace of the matrix, is a matrix containing the traffic data of all switch devices. Row indicates the Traffic data of switch devices ; Representation Matrix The transpose of

[0018] Introducing a penalty term into the optimization objective function , use the gradient descent method to iteratively optimize the projection matrix, where the iterative optimization rule of the projection matrix is: , is the projection matrix after iteration, is the projection matrix before iteration, is the learning rate, which is used to control the step size of each iteration. represents the gradient of the projection matrix W, is the penalty coefficient, which is used to control the impact of the penalty term on the optimization objective function. is the penalty term, including the nuclear norm or Frobenius norm of the matrix;

[0019] After each iteration, the change of the optimization objective function is calculated. When the change is less than the preset threshold When , it represents the projection matrix Converged, stop iteration, and get the optimal dimensionality reduction mapping ;

[0020] According to the optimal dimensionality reduction mapping Project the traffic data into a low-dimensional space to obtain a low-dimensional feature vector , the specific logic is: .

[0021] The present invention is further configured to: according to the low-dimensional feature vector Compute node affinity matrix , the calculation logic is: ,in, , and For the , and Switch devices in The low-dimensional feature vector at the moment, is a low-dimensional feature vector and The Euclidean distance of is a scale parameter used to adjust the relevance weight calculation;

[0022] The present invention is further configured to: according to the node association matrix and traffic data Calculate the bottleneck potential of each switch device , the calculation logic is: ,in, For the The maximum bandwidth of the switch device. For the The switch device and The transmission delay between switch devices, To control the delay sensitivity parameter.

[0023] The present invention is further configured to analyze the topology of the switch device network and adjust the routing path of the inbound traffic, including:

[0024] Analyze the topology of the switch device network and calculate the link weights of the links between the switch devices;

[0025] An optimal path is calculated based on the link weights, and inbound traffic is transmitted along the optimal path.

[0026] The present invention is further configured such that the calculation logic of the link weight of the link between the switch devices is: ,in, For the and The link between the switch devices, For the and The link weight between the switch devices, For Link The maximum bandwidth, For Link The delay, and are adjustment factors, which are used to adjust the weights of traffic load and link delay respectively;

[0027] The optimal path is calculated based on the link weights, and the inbound traffic is transmitted according to the optimal path. The optimization goal of selecting the optimal path is: ,in, is the optimal path, It indicates a feasible path from the switch device with the highest bottleneck potential value to the switch device with the lowest bottleneck potential value. To find a feasible path , so that the objective function Minimum.

[0028] The present invention also provides a centralized control system for switch equipment, the system comprising:

[0029] Data collection module: collects the flow data generated by the switch device in real time through the communication protocol between the controller and the switch device, and pre-processes the flow data;

[0030] Bottleneck calculation module: It performs high-dimensional projection analysis and adaptive network analysis on pre-processed traffic data, builds a network bottleneck detection model, and calculates the bottleneck potential value of switch equipment;

[0031] Descending sorting module: sorts the corresponding switch devices in descending order according to the bottleneck potential value, and allocates the inbound traffic of the switch device with a value greater than the preset bottleneck potential threshold and ranked first to the switch device with the lowest value in descending order;

[0032] Path generation module: Analyzes the topology of the switch device network and adjusts the routing path of inbound traffic.

[0033] The present invention provides a centralized control system and method for switch devices. The method collects flow data generated by switch devices in real time through a communication protocol between a controller and the switch devices, and pre-processes the flow data; performs high-dimensional projection analysis and adaptive network analysis on the pre-processed flow data, builds a network bottleneck detection model, and calculates the bottleneck potential value of the switch devices; sorts the corresponding switch devices in descending order according to the bottleneck potential value, and allocates the inbound flow of the switch device that is greater than a preset bottleneck potential threshold and ranks first to the switch device that ranks last in the descending order; analyzes the topological structure of the switch device network, and adjusts the routing path of the inbound flow. The beneficial effects produced include:

[0034] 1. Improve traffic management efficiency: By using a centralized controller to monitor and analyze traffic changes in the network in real time, the present invention can dynamically adjust the traffic routing path and bandwidth allocation according to the real-time changes in the network status, significantly improving the network's load balancing and transmission efficiency, and avoiding network congestion and resource waste;

[0035] 2. Enhanced dynamic adjustment capability: The present invention can respond to changes in network topology and traffic in real time, and uses high-dimensional projection analysis, low-dimensional feature vector extraction and network topology optimization algorithms to ensure that data streams can be transmitted on the optimal path, reducing data packet loss and delay caused by link failures or traffic bursts, and improving network stability and service quality;

[0036] 3. Optimize network topology: Through dynamic monitoring and analysis of network topology, the present invention can timely adjust and optimize the topology in the network to ensure that data streams can be transmitted through the optimal routing path, thereby improving network resource utilization and reducing overall transmission delay and bandwidth consumption.

[0037] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for describing the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0039] Figure 1 A flow chart of a centralized control method for a switch device is shown as an exemplary embodiment of the present invention;

[0040] Figure 2 The present invention is a schematic diagram of a centralized control system for switch equipment according to an exemplary embodiment of the present invention. DETAILED DESCRIPTION

[0041] The following will describe the embodiments of the present invention with reference to the accompanying drawings and preferred embodiments. Those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be understood that the preferred embodiments are only for illustrating the present invention, not for limiting the scope of protection of the present invention.

[0042] It should be noted that the illustrations provided in the following embodiments are only schematic illustrations of the basic concept of the present invention, and thus the drawings only show components related to the present invention rather than being drawn according to the number, shape and size of components in actual implementation. In actual implementation, the type, quantity and proportion of each component may be changed arbitrarily, and the component layout may also be more complicated.

[0043] In the following description, numerous details are discussed to provide a more thorough explanation of the embodiments of the present invention. However, it is obvious to those skilled in the art that the embodiments of the present invention can be implemented without these specific details. In other embodiments, well-known structures and devices are shown in the form of block diagrams rather than in detail to avoid making the embodiments of the present invention difficult to understand.

[0044] Embodiment 1

[0045] A centralized control method for switch equipment, such as Figure 1 As shown, including:

[0046] Through the communication protocol between the controller and the switch device, the flow data generated by the switch device is collected in real time, and the flow data is pre-processed;

[0047] Through high-dimensional projection analysis and adaptive network analysis of pre-processed traffic data, a network bottleneck detection model is constructed to calculate the bottleneck potential value of switch equipment;

[0048] The corresponding switch devices are sorted in descending order according to the bottleneck potential value, and the inbound traffic of the switch device with a value greater than the preset bottleneck potential threshold and ranked first is allocated to the switch device with the lowest value in descending order;

[0049] Analyze the topology of the switch device network and adjust the routing path of inbound traffic.

[0050] Specifically, the communication protocols between the controller and the switch device, including SNMP and NetFlow, are used to collect the flow data generated by the switch device in real time. The present invention is further configured to collect the flow data generated by the switch device in real time, including: inbound flow, outbound flow, packet loss rate, port utilization rate and delay. Specifically, the inbound flow is the data flow entering the switch device, the outbound flow is the data flow sent out from the switch device, the packet loss rate is the proportion of data packets that are not successfully transmitted due to various reasons, the port utilization is the use of port resources, and the delay represents the transmission time of data through the switch device;

[0051] The flow data is preprocessed, including: performing outlier detection, missing value filling and denoising on the flow data, and the preprocessed flow data is recorded as ,in, For the Switch devices in Traffic data at the moment, , respectively Switch devices in Inbound traffic, outbound traffic, packet loss rate, port utilization and latency at each moment. Specifically, outlier detection is the process of identifying and processing abnormal data points in a data set that deviate from the normal pattern. Outliers may be caused by collection equipment failure, temporary network problems, input errors, etc., which are obviously inconsistent with the overall trend of the data set. If outliers are not processed, they may introduce deviations in subsequent data analysis and decision-making, leading to misleading conclusions. By detecting and processing outliers, the errors in the data can be significantly reduced and the overall quality of the data can be improved. Specific implementation methods include the Z-score method, the Boxplot method, and the One-Class method in the Support Vector Machine (SVM). SVM, which is not restricted here; missing value filling refers to the process of filling in some missing data in a data set due to various reasons (including incomplete data collection, sensor failure, etc.). If missing values ​​are not processed, it will affect the integrity of the data, resulting in deviations in analysis results, and even make some algorithms unable to work properly. Filling missing values ​​can maintain the integrity of the data set, so that subsequent data processing and analysis are not affected by missing values. Specific implementation methods include mean / median / mode filling and constant filling as well as interpolation, which are not restricted here; denoising refers to identifying and removing noise components introduced by various external factors in a data set. Noise can be random interference data or systematic errors. These noises will mask the real patterns and trends of the data and reduce the quality of the data. Through denoising, the signal-to-noise ratio of the data set can be improved, making the effective information of the data more prominent and reducing unnecessary interference. Specific implementation methods include moving average filtering, Kalman filtering, wavelet transform and Fourier transform, which are not restricted here.

[0052] The present invention is further configured to perform high-dimensional projection analysis and adaptive network analysis on pre-processed traffic data, construct a network bottleneck detection model, and calculate the bottleneck potential value of the switch device, including:

[0053] Through high-dimensional projection analysis, the preprocessed traffic data is projected into a low-dimensional space to obtain a low-dimensional feature vector ;

[0054] High-dimensional projection analysis is the process of mapping high-dimensional data to low-dimensional space in order to more effectively analyze and process data. By reducing the dimension of data, the computational complexity can be reduced and patterns and trends in the data can be captured more easily. The present invention is further configured to project the pre-processed flow data to low-dimensional space through high-dimensional projection analysis to obtain a low-dimensional feature vector. ,include:

[0055] Construct the projection matrix W to transform high-dimensional flow data Mapped to low-dimensional space, the construction of the projection matrix W is realized by optimizing the objective function, which is used to maximize the feature retention of high-dimensional flow data in low-dimensional space while minimizing information loss. The optimization objective function is: ,in, To solve the projection matrix , so that the objective function maximum, is the trace of the matrix, that is, the sum of the elements on the diagonal of a square matrix. The trace of the matrix can be used to measure the amount of information contained in the matrix. The goal is to retain the maximum variance information in the original data by maximizing the trace. is a matrix containing the traffic data of all switch devices. Row indicates the Traffic data of switch devices ; Representation Matrix The transpose of; Specifically, the projection matrix W is used to transform the high-dimensional flow data Mapping to low-dimensional space, the construction of the projection matrix is ​​used to retain the key information in the high-dimensional data while reducing the dimension of the data so that subsequent analysis is more efficient;

[0056] Introducing a penalty term into the optimization objective function , use the gradient descent method to iteratively optimize the projection matrix, where the iterative optimization rule of the projection matrix is: , is the projection matrix after iteration, is the projection matrix before iteration, is the learning rate, which is used to control the step size of each iteration. represents the gradient of the projection matrix W, is the penalty coefficient, which is used to control the impact of the penalty term on the optimization objective function. is a penalty term, including the nuclear norm or Frobenius norm of the matrix; specifically, the Frobenius norm is a matrix norm, which is calculated by the square root of the sum of the squares of all elements of the matrix. The Frobenius norm is often used as a regularization term to control the complexity of the matrix and prevent overfitting; the gradient descent method is an optimization algorithm used to find the optimal value of the objective function. By calculating the derivative (gradient) of the objective function with respect to the variable, the variable value is updated along the gradient direction, and the optimal solution is gradually approached.

[0057] After each iteration, the change of the optimization objective function is calculated. When the change is less than the preset threshold When , it represents the projection matrix Converged, stop iteration, and get the optimal dimensionality reduction mapping ;

[0058] According to the optimal dimensionality reduction mapping Project the traffic data into a low-dimensional space to obtain a low-dimensional feature vector , the specific logic is: .

[0059] According to the low-dimensional feature vector Compute node affinity matrix , where the node association matrix Each element in Indicates The switch device and The present invention is further configured to: according to the low-dimensional feature vector Compute node affinity matrix , the calculation logic is: ,in, , and For the , and Switch devices in The low-dimensional feature vector at the moment, is a low-dimensional feature vector and The Euclidean distance of is a scale parameter used to adjust the association weight calculation. Value makes it only with Close nodes have larger weights; larger The value makes more nodes obtain relatively higher weights; specifically, the exponential function part reflects the similarity between switch devices. The higher the similarity, the larger the index value, indicating a stronger correlation. By calculating the node correlation matrix , can accurately quantify the interdependence between nodes in the network, help identify key nodes and potential bottlenecks in the network, and provide a basis for subsequent network optimization.

[0060] According to the node association matrix and traffic data Calculate the bottleneck potential of each switch device The present invention is further configured to: according to the node association matrix and traffic data Calculate the bottleneck potential of each switch device , the calculation logic is: ,in, For the The maximum bandwidth of the switch device. For the The switch device and The transmission delay between switch devices, To control the delay sensitivity parameter, the higher A higher value means that the device is more sensitive to latency, and latency has a greater impact on the bottleneck potential value. Specifically, the calculation logic of the bottleneck potential value is to calculate the bottleneck potential value of the device through weighted summation, which reflects the possibility of the device becoming a bottleneck under the current network load conditions. The size of the weight is determined by the correlation between the devices, the proportion of traffic load to processing capacity, and the latency between the devices. The bottleneck potential value is used to measure the load pressure of the switch device in the switch network and the possibility of becoming a bottleneck. By calculating the bottleneck potential value, the switch device that may become a bottleneck can be identified in advance, helping network administrators take preventive measures to avoid network performance degradation.

[0061] The present invention is further configured to analyze the topology of the switch device network and adjust the routing path of the inbound traffic, including:

[0062] Analyze the topology of the switch device network and calculate the link weights of the links between the switch devices. Specifically, analyze the topology of the switch device network, identify the switch devices (nodes) in the network and the links between them, and calculate the link weights of the links between the switch devices in order to measure the importance of each link in the network. , the present invention is further configured such that the calculation logic of the link weight of the link between the switch devices is: ,in, For the and The link between the switch devices, For the and The link weight between the switch devices, For Link The maximum bandwidth, For Link The delay, and are adjustment factors, which are used to adjust the weights of traffic load and link delay respectively; specifically, the above calculation logic balances the relationship between network load pressure and transmission efficiency by combining the ratio of link traffic load to link capacity and link delay, avoiding network bottlenecks caused by excessive use of certain links, and also avoiding the decrease in transmission efficiency caused by high-latency links.

[0063] The optimal path is calculated based on the link weight, and the inbound traffic is transmitted according to the optimal path; the present invention is further configured to calculate the optimal path based on the link weight, and the inbound traffic is transmitted according to the optimal path, and the optimization goal of the selection of the optimal path is: ,in, is the optimal path, It indicates a feasible path from the switch device with the highest bottleneck potential value to the switch device with the lowest bottleneck potential value. To find a feasible path , so that the objective function Minimum; optimal path It refers to the path with the smallest total weight among all possible paths. This path can minimize the total cost of traffic transmission and is the optimal traffic transmission path. By calculating and selecting the optimal path, it can ensure that traffic transmission in the network is carried out at the lowest cost and highest efficiency, avoiding the waste of network resources.

[0064] Embodiment 2

[0065] See also Figure 2 , the exemplary centralized control system of a switch device includes:

[0066] Data collection module: collects the flow data generated by the switch device in real time through the communication protocol between the controller and the switch device, and pre-processes the flow data;

[0067] Bottleneck calculation module: It performs high-dimensional projection analysis and adaptive network analysis on pre-processed traffic data, builds a network bottleneck detection model, and calculates the bottleneck potential value of switch equipment;

[0068] Descending sorting module: sorts the corresponding switch devices in descending order according to the bottleneck potential value, and allocates the inbound traffic of the switch device with a value greater than the preset bottleneck potential threshold and ranked first to the switch device with the lowest value in descending order;

[0069] Path generation module: Analyzes the topology of the switch device network and adjusts the routing path of inbound traffic.

[0070] It should be noted that the centralized control system of a switch device provided in the above embodiment and the centralized control method of a switch device provided in the above embodiment belong to the same concept, wherein the specific manner in which each module and unit performs operations has been described in detail in the method embodiment, and will not be repeated here. In actual application, the centralized control system of a switch device provided in the above embodiment can allocate the above functions to different functional modules as needed, that is, divide the internal structure of the system into different functional modules to complete all or part of the functions described above, and this is not limited here.

[0071] The above embodiments can be implemented in whole or in part by software, hardware, firmware or any other combination. When implemented by software, the above embodiments can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from one website site, computer, server or data center to another website site, computer, server or data center by wired (e.g., infrared, wireless, microwave, etc.). The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium can be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium can be a solid-state hard disk.

[0072] It should be understood that the term "and / or" in this article is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. A and B can be singular or plural. In addition, the character " / " in this article generally indicates that the associated objects before and after are in an "or" relationship, but it may also indicate an "and / or" relationship. Please refer to the context for specific understanding.

[0073] In this application, "at least one" means one or more, and "more than one" means two or more. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can be represented by: a, b, c, ab, ac, bc, or abc, where a, b, c can be single or multiple.

[0074] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0075] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0076] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0077] In the several embodiments provided in the present application, it should be understood that the disclosed system can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0078] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0079] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0080] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium and includes several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage media include: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROM), random access memories (RAM), magnetic disks or optical disks.

[0081] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A centralized control method for switch equipment, characterized in that: include: Through the communication protocol between the controller and the switch device, the flow data generated by the switch device is collected in real time, and the flow data is pre-processed; Collect the traffic data generated by the switch device in real time, including: inbound traffic, outbound traffic, packet loss rate, port utilization and delay; pre-process the traffic data, including: abnormal value detection, missing value filling and denoising of the traffic data, and record the pre-processed traffic data as ,in, is the traffic data of the i-th switch device at time t, , are the inbound traffic, outbound traffic, packet loss rate, port utilization and delay of the i-th switch device at time t; Through high-dimensional projection analysis and adaptive network analysis of pre-processed traffic data, a network bottleneck detection model is constructed to calculate the bottleneck potential value of the switch equipment, including: through high-dimensional projection analysis, the pre-processed traffic data is projected into a low-dimensional space to obtain a low-dimensional feature vector ; According to the low-dimensional feature vector Compute node affinity matrix , where the node association matrix Each element in represents the dependency strength between the i-th switch device and the j-th switch device; according to the node association matrix and traffic data Calculate the bottleneck potential of each switch device ; Among them, through high-dimensional projection analysis, the pre-processed traffic data is projected into a low-dimensional space to obtain a low-dimensional feature vector , including: constructing a projection matrix W to transform high-dimensional flow data Mapped to low-dimensional space, the construction of the projection matrix W is realized by optimizing the objective function, which is used to maximize the feature retention of high-dimensional flow data in low-dimensional space while minimizing information loss. The optimization objective function is: ,in, To solve the projection matrix W, the objective function maximum, is the trace of the matrix, X is the matrix containing the traffic data of all switch devices, and the i-th row represents the traffic data of the i-th switch device ; represents the transpose of matrix X; a penalty term is introduced in the optimization objective function , use the gradient descent method to iteratively optimize the projection matrix, where the iterative optimization rule of the projection matrix is: , is the projection matrix after iteration, is the projection matrix before iteration, is the learning rate, which is used to control the step size of each iteration. represents the gradient of the projection matrix W, is the penalty coefficient, which is used to control the impact of the penalty term on the optimization objective function. is a penalty term, including the nuclear norm or Frobenius norm of the matrix; after each iteration, the change of the optimization objective function is calculated, and when the change is less than the preset threshold When , it means that the projection matrix W has converged, stop the iteration, and get the optimal dimensionality reduction mapping ; According to the optimal dimensionality reduction mapping Project the traffic data into a low-dimensional space to obtain a low-dimensional feature vector , the specific logic is: ; According to the low-dimensional feature vector Compute node affinity matrix , the calculation logic is: ,in, , and is the low-dimensional feature vector of the i-th, j-th and k-th switch devices at time t, is a low-dimensional feature vector and The Euclidean distance of is a scale parameter used to adjust the calculation of the association weight; according to the node association matrix and traffic data Calculate the bottleneck potential of each switch device , the calculation logic is: ,in, is the maximum bandwidth of the i-th switch device, is the transmission delay between the i-th switch device and the j-th switch device, To control the delay sensitivity parameters; The corresponding switch devices are sorted in descending order according to the bottleneck potential value, and the inbound traffic of the switch device with a value greater than the preset bottleneck potential threshold and ranked first is allocated to the switch device with the lowest value in descending order; Analyze the topology of the switch device network and adjust the routing path of inbound traffic.

2. A centralized control method for switch equipment according to claim 1, characterized in that: Analyze the topology of the switch device network and adjust the routing path of inbound traffic, including: Analyze the topology of the switch device network and calculate the link weights of the links between the switch devices; An optimal path is calculated based on the link weights, and inbound traffic is transmitted along the optimal path.

3. A centralized control method for switch equipment according to claim 2, characterized in that: The calculation logic of the link weight of the link between switch devices is: ,in, is the link between the i-th and j-th switch devices, is the link weight between the i-th and j-th switch devices, For Link The maximum bandwidth, For Link The delay, and are adjustment factors, which are used to adjust the weights of traffic load and link delay respectively; The optimal path is calculated based on the link weights, and the inbound traffic is transmitted according to the optimal path. The optimization goal of selecting the optimal path is: ,in, is the optimal path, It indicates a feasible path from the switch device with the highest bottleneck potential value to the switch device with the lowest bottleneck potential value. To find a feasible path , so that the objective function Minimum.

4. A centralized control system for switch equipment, characterized in that: include: Data collection module: collects the flow data generated by the switch device in real time through the communication protocol between the controller and the switch device, and pre-processes the flow data; Collect the traffic data generated by the switch device in real time, including: inbound traffic, outbound traffic, packet loss rate, port utilization and delay; pre-process the traffic data, including: abnormal value detection, missing value filling and denoising of the traffic data, and record the pre-processed traffic data as ,in, is the traffic data of the i-th switch device at time t, , are the inbound traffic, outbound traffic, packet loss rate, port utilization and delay of the i-th switch device at time t; Bottleneck calculation module: Through high-dimensional projection analysis and adaptive network analysis of pre-processed traffic data, a network bottleneck detection model is constructed to calculate the bottleneck potential value of the switch equipment, including: through high-dimensional projection analysis, the pre-processed traffic data is projected into a low-dimensional space to obtain a low-dimensional feature vector ; According to the low-dimensional feature vector Compute node affinity matrix , where the node association matrix Each element in represents the dependency strength between the i-th switch device and the j-th switch device; according to the node association matrix and traffic data Calculate the bottleneck potential of each switch device ; Among them, through high-dimensional projection analysis, the pre-processed traffic data is projected into a low-dimensional space to obtain a low-dimensional feature vector , including: constructing a projection matrix W to transform high-dimensional flow data Mapped to low-dimensional space, the construction of the projection matrix W is realized by optimizing the objective function, which is used to maximize the feature retention of high-dimensional flow data in low-dimensional space while minimizing information loss. The optimization objective function is: ,in, To solve the projection matrix W, the objective function maximum, is the trace of the matrix, X is the matrix containing the traffic data of all switch devices, and the i-th row represents the traffic data of the i-th switch device ; represents the transpose of matrix X; a penalty term is introduced in the optimization objective function , use the gradient descent method to iteratively optimize the projection matrix, where the iterative optimization rule of the projection matrix is: , is the projection matrix after iteration, is the projection matrix before iteration, is the learning rate, which is used to control the step size of each iteration. represents the gradient of the projection matrix W, is the penalty coefficient, which is used to control the impact of the penalty term on the optimization objective function. is a penalty term, including the nuclear norm or Frobenius norm of the matrix; after each iteration, the change of the optimization objective function is calculated, and when the change is less than the preset threshold When , it means that the projection matrix W has converged, stop the iteration, and get the optimal dimensionality reduction mapping ; According to the optimal dimensionality reduction mapping Project the traffic data into a low-dimensional space to obtain a low-dimensional feature vector , the specific logic is: ; According to the low-dimensional feature vector Compute node affinity matrix , the calculation logic is: ,in, , and is the low-dimensional feature vector of the i-th, j-th and k-th switch devices at time t, is a low-dimensional feature vector and The Euclidean distance of is a scale parameter used to adjust the calculation of the association weight; according to the node association matrix and traffic data Calculate the bottleneck potential of each switch device , the calculation logic is: ,in, is the maximum bandwidth of the i-th switch device, is the transmission delay between the i-th switch device and the j-th switch device, To control the delay sensitivity parameters; Descending sorting module: sorts the corresponding switch devices in descending order according to the bottleneck potential value, and allocates the inbound traffic of the switch device with a value greater than the preset bottleneck potential threshold and ranked first to the switch device with the lowest value in descending order; Path generation module: Analyzes the topology of the switch device network and adjusts the routing path of inbound traffic.