Multi-path optimization method, device and system for network routing

Abstract

The present invention relates to the technical field of multi-path information processing, and provides a multi-path optimization method, device and system for network routing, including obtaining network node information and obtaining a number of available paths based thereon, evaluating all available paths, obtaining network bandwidth data and initial weights; adjusting the initial weights, obtaining initial path weights and a path priority table, allocating and identifying data packets of network traffic using the path priority table, obtaining traffic data packets with path identifiers and obtaining traffic load logs and network delay data based thereon, adjusting the initial path weights, generating a traffic scheduling scheme and updating the path priority table. By obtaining network node information and using a bandwidth evaluation algorithm, an initial path evaluation reflecting the network situation is generated, and then through real-time data feedback from a dynamic monitoring module, a traffic scheduling scheme that can be adaptively adjusted at any time is generated, thereby improving the shortcomings of the prior art in terms of dynamic adaptability.

Description

Technical Field

[0001] The present application relates to the technical field of multipath information processing, and in particular to a multipath optimization method, device and system for network routing. Background Art

[0002] With the rapid development of the Internet and the Internet of Things, network traffic has increased exponentially, and network architecture has become increasingly complex, placing higher demands on the efficiency and stability of network routing. In addition, the demands for bandwidth, latency, and reliability of various application services in modern networks are also increasing, further promoting the development of multipath routing optimization technology. Multipath routing optimization can improve the utilization of network resources and meet the quality requirements of different applications by accurately allocating and flexibly adjusting network traffic.

[0003] Among the related technical means, multi-path routing optimization technology usually relies on static evaluation information and preset routing strategies. These methods mainly calculate the cost of multiple available paths and select several paths with the lowest cost for traffic distribution and load balancing. This method can improve the throughput and fault tolerance of the network to a certain extent, and provide a fast recovery method when the network is congested or the path fails, thereby improving the overall performance and service quality of the network.

[0004] For the above technical solutions, although the static multi-path selection strategy can effectively meet the basic network traffic management needs, it has the problem of insufficient flexibility and real-time performance in the face of dynamically changing network environments. This defect will lead to the failure of timely routing adjustment when network traffic increases suddenly or path status changes, thus affecting the stability and reliability of network performance. Summary of the invention

[0005] In order to improve the deficiencies of the prior art in terms of dynamic adaptability, the present application provides a multi-path optimization method, device and system for network routing, which performs multi-path optimization through a dynamic monitoring and adjustment mechanism to better adapt to changes in the network environment and improve the overall network performance.

[0006] The present invention provides a multi-path optimization method for network routing, comprising: obtaining network node information, obtaining a plurality of available paths based on the network node information, analyzing and evaluating all the available paths using a bandwidth evaluation algorithm, and obtaining network bandwidth data and an initial weight of each of the available paths; adjusting the initial weight by a preset path priority to obtain an initial path weight and a path priority table, using the path priority table to allocate and identify data packets of network traffic to obtain traffic data packets with path identifiers; inputting the traffic data packets with path identifiers into a preset dynamic monitoring module to obtain traffic load logs and network delay data; adjusting the initial path weight based on the traffic load logs and the network delay data, generating a corresponding adjustment coefficient based on the adjustment result, and generating a corresponding target path weight according to the adjustment coefficient; allocating data packets of network traffic according to the adjustment coefficient and the corresponding target path weight to generate a traffic scheduling plan; and feeding back the traffic scheduling plan to the path priority table for updating through a feedback mechanism.

[0007] As a preferred solution, the steps of obtaining network node information, obtaining a number of available paths based on the network node information, analyzing and evaluating all the available paths using a bandwidth evaluation algorithm, and obtaining network bandwidth data and an initial weight of each of the available paths include: obtaining network node information and a node topology structure of network nodes through a routing protocol, calculating the path between each pair of network nodes in the network node information based on the node topology structure using a Dijkstra algorithm to obtain a number of available paths; evaluating each of the available paths through a bandwidth evaluation algorithm to obtain an evaluation result, obtaining network bandwidth data for the evaluation result using a mean method, obtaining a loss factor for the evaluation result using a variance method, and generating network status data based on the loss factor; and performing a weight analysis on each of the available paths using the network bandwidth data and the network status data to obtain an initial weight for each path.

[0008] As a preferred solution, the steps of adjusting the initial weights by preset path priorities to obtain initial path weights and a path priority table, allocating and identifying data packets of network traffic using the path priority table, and obtaining traffic data packets with path identifications include: adjusting the initial weights by preset path priorities to obtain an adjusted weight set, weighted averaging the adjusted weight set to obtain an average weight value and a weight deviation value, comparing the average weight value with the initial weight to obtain a weight difference value; calculating a weight correction value for each path according to the weight deviation value and the weight difference value, and correcting the initial weight according to the weight correction value to obtain an initial path weight; sorting the initial path weights to obtain a path sorting sequence, and generating a path priority table based on the path sorting sequence; correlating the path priority table with characteristic data of network traffic to obtain a data packet allocation ratio and a data packet allocation rule; allocating and identifying data packets of network traffic according to the data packet allocation ratio and the data packet allocation rule to obtain traffic data packets with path identifications; wherein the characteristic data of the network traffic includes data packet size, protocol type, and source and destination addresses.

[0009] As a preferred solution, the step of inputting the traffic data packet with the path identifier into a preset dynamic monitoring module to obtain a traffic load log and network delay data includes: inputting the traffic data packet with the path identifier into a preset dynamic monitoring module to perform traffic tracking on the traffic data packet with the path identifier, and generating a traffic load log based on the traffic tracking result; performing delay monitoring on the traffic data packet with the path identifier through a delay monitor to obtain network delay data.

[0010] As a preferred scheme, the initial path weight is adjusted based on the traffic load log and the network delay data, and a corresponding adjustment coefficient is generated based on the adjustment result, and the step of generating the corresponding target path weight according to the adjustment coefficient includes: performing time series prediction on the traffic load log by a timing analysis method to obtain traffic fluctuations and delay trends, performing pattern recognition on the traffic fluctuations by fuzzy clustering to obtain a traffic fluctuation pattern, performing spectrum analysis on the delay trend by spectral analysis to obtain a delay spectrum; performing a comprehensive evaluation on the traffic fluctuation pattern, the delay spectrum and the network delay data by a fuzzy comprehensive evaluation method to obtain a path quality index, adjusting the initial path weight according to the path quality index, and obtaining an adjustment result by a linear regression method; performing parameter estimation on the adjustment result according to Monte Carlo simulation to obtain an adjustment coefficient, performing global optimization calculation on the initial path weight according to the adjustment coefficient to obtain a weight change vector; calculating the weight change vector by a genetic algorithm to obtain a target path weight.

[0011] As a preferred scheme, the step of allocating data packets of network traffic according to the adjustment coefficient and the corresponding target path weight to generate a traffic scheduling plan includes: normalizing the adjustment coefficient and the corresponding target path weight, and generating a weight matrix based on the normalized data; inputting the weight matrix into a preset allocation model to allocate data packets of network traffic, and generating a traffic allocation table based on the allocation result; selecting the optimal path according to the traffic allocation table, and constructing a traffic scheduling plan based on the optimal path.

[0012] As a preferred scheme, the step of feeding back the traffic scheduling scheme to the path priority table for updating through a feedback mechanism includes: using a deep learning algorithm to perform traffic identification on the traffic scheduling scheme to obtain path scheduling effect data, comparing and analyzing the path scheduling effect data and the path priority table to obtain difference information; and inputting the difference information into a preset feedback mechanism to update the path priority.

[0013] The present application also provides a multi-path optimization device for network routing, including: an acquisition module, used to acquire network node information, obtain a number of available paths based on the network node information, analyze and evaluate all the available paths using a bandwidth evaluation algorithm, and obtain network bandwidth data and an initial weight of each available path; an allocation module, used to adjust the initial weight by a preset path priority, obtain an initial path weight and a path priority table, and use the path priority table to allocate and identify data packets of network traffic to obtain traffic data packets with path identifiers; a monitoring module, used to input the traffic data packets with path identifiers into a preset dynamic monitoring module to obtain a traffic load log and network delay data; an adjustment module, used to adjust the initial path weight based on the traffic load log and the network delay data, and generate a corresponding adjustment coefficient based on the adjustment result, and generate a corresponding target path weight according to the adjustment coefficient; a generation module, used to allocate data packets of network traffic according to the adjustment coefficient and the corresponding target path weight, and generate a traffic scheduling plan; an update module, used to feed back the traffic scheduling plan to the path priority table for updating through a feedback mechanism.

[0014] The present application also provides a multi-path optimization system for network routing, including: an acquisition unit, used to acquire network node information, obtain a number of available paths based on the network node information, analyze and evaluate all the available paths using a bandwidth evaluation algorithm, and obtain network bandwidth data and an initial weight of each available path; an allocation unit, used to adjust the initial weight by a preset path priority, obtain an initial path weight and a path priority table, and use the path priority table to allocate and identify data packets of network traffic to obtain traffic data packets with path identifiers; a monitoring unit, used to input the traffic data packets with path identifiers into a preset dynamic monitoring module to obtain a traffic load log and network delay data; an adjustment unit, used to adjust the initial path weight based on the traffic load log and the network delay data, and generate a corresponding adjustment coefficient based on the adjustment result, and generate a corresponding target path weight according to the adjustment coefficient; a generation unit, used to allocate data packets of network traffic according to the adjustment coefficient and the corresponding target path weight, and generate a traffic scheduling plan; an updating unit, used to feed back the traffic scheduling plan to the path priority table for updating through a feedback mechanism.

[0015] Compared with the prior art, the present application has the following beneficial effects: strong dynamic adaptability. By acquiring network node information and using a bandwidth evaluation algorithm, an initial path resource evaluation that can truly reflect the actual situation of the network is generated, and through real-time data feedback from a dynamic monitoring module, a traffic scheduling scheme that can be adaptively adjusted at any time is generated. Then, the dynamic adjustment and feedback mechanism of the path priority table is used to achieve flexible and efficient traffic allocation, and the ability of the network to adapt to burst traffic and rapidly changing conditions is improved. Finally, by adjusting the initial weight to the target path weight, combined with traffic logs and real-time data, the scientific nature of dynamic allocation and the rationality of path selection are ensured, and the shortcomings of the prior art in dynamic adaptability are improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. 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 paying creative labor.

[0017] The structures, proportions, sizes, etc. illustrated in the drawings of this specification are only used to match the contents disclosed in the specification so as to facilitate understanding and reading by persons familiar with this technology. They are not used to limit the conditions under which the present invention can be implemented, and therefore have no substantive technical significance. Any structural modification, change in proportion or adjustment of size, without affecting the effects and purposes that can be achieved by the present invention, should still fall within the scope of the technical contents disclosed by the present invention.

[0018] Figure 1 It is a flowchart of a multi-path optimization method for network routing provided by an embodiment of the present invention;

[0019] Figure 2 is a schematic block diagram of the structure of a multi-path optimization device for network routing provided by an embodiment of the present invention;

[0020] Figure 3 It is a schematic block diagram of the structure of a multi-path optimization system for network routing provided by an embodiment of the present invention.

[0021] Description of reference numerals:

[0022] 10. Multi-path optimization device for network routing; 11. Acquisition module; 12. Allocation module; 13. Monitoring module; 14. Adjustment module; 15. Generation module; 16. Update module; 20. Multi-path optimization system for network routing; 21. Acquisition unit; 22. Allocation unit; 23. Monitoring unit; 24. Adjustment unit; 25. Generation unit; 26. Update unit. DETAILED DESCRIPTION

[0023] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.

[0025] It should also be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.

[0026] It should be further understood that the term “and / or” used in the specification and appended claims refers to any combination and all possible combinations of one or more of the associated listed items, and includes these combinations.

[0027] The technical solution of the present invention is further described below with reference to the accompanying drawings and through specific implementation methods.

[0028] Embodiment 1:

[0029] like Figure 1 As shown, the multi-path optimization method for network routing provided in the embodiment of the present application includes steps S100 to S600.

[0030] Step S100: Acquire network node information, obtain several available paths based on the network node information, analyze and evaluate all available paths using a bandwidth evaluation algorithm, and obtain network bandwidth data and an initial weight of each available path.

[0031] In this step, the preset network node information collection module collects detailed information about the topology, connection status, and available resources of each node in the network, and generates a list of potential connection paths between nodes. Specifically, the bandwidth evaluation algorithm calculates the maximum bandwidth and corresponding initial weight value of each available path based on parameters such as path transmission capacity, current load, and historical data, to measure the priority of the path in carrying traffic in the initial state.

[0032] For example, if a node's information shows that it is directly connected to three nodes, multiple available paths will be generated, and the bandwidth evaluation algorithm will record the bandwidth limit of each path and the load balance status of the initial state.

[0033] Step S200, adjusting the initial weight by the preset path priority to obtain the initial path weight and path priority table, using the path priority table to allocate and identify the data packets of the network traffic, and obtaining the traffic data packets with the path identification.

[0034] In this step, the path priority settings are read and the initial weights are adjusted comprehensively according to different network service quality requirements and strategies to ensure the best overall efficiency of path utilization. Specifically, the path priority table records the actual efficiency of the path under different types and priority traffic, and then accurately allocates and identifies the path for each data packet in the network traffic.

[0035] For example, video streaming is preferentially allocated to low-latency and high-bandwidth paths, while general file transfers are allocated to high-load-tolerant paths.

[0036] Step S300: input the traffic data packet with the path identifier into the preset dynamic monitoring module to obtain the traffic load log and network delay data.

[0037] In this step, the flow of the identified data packets is tracked in real time based on the preset dynamic monitoring module, and key dynamic performance indicators such as the traffic status and residence time of each path are recorded. Specifically, according to the changes in network traffic, a detailed traffic load log is generated, the real-time bandwidth utilization of each path is counted, and the network delay of all paths is detected.

[0038] For example, when a path accumulates too many real-time communication requests, the monitoring results will show abnormally high traffic peaks and subsequent signs of increased latency in the log.

[0039] Step S400: Adjust the initial path weight based on the traffic load log and the network delay data, generate a corresponding adjustment coefficient based on the adjustment result, and generate a corresponding target path weight according to the adjustment coefficient.

[0040] In this step, the information fed back by the dynamic monitoring module is analyzed, and the initial path weight is re-optimized in combination with the load balancing requirements to adapt to the real-time changing network status. Specifically, the weight penalty coefficient is increased for overloaded paths, while the penalty is reduced for paths with low latency and light load, so that the target path weight measures the optimal state of the current utilization efficiency of each path.

[0041] For example, during the adjustment process, a path with less authorized traffic and lower latency is given a higher weight, causing more lower-priority traffic to be shared.

[0042] Step S500: Allocate data packets of network traffic according to the adjustment coefficient and the corresponding target path weight to generate a traffic scheduling plan.

[0043] In this step, the traffic allocation strategy is reviewed in combination with the adjusted target path weights to ensure efficient traffic scheduling in a fluctuating network environment. Specifically, based on the target path weights, the packet routing selection is dynamically adjusted to achieve an optimized traffic scheduling solution.

[0044] For example, the updated scheduling scheme directs most non-critical traffic to take low-load paths, alleviating the traffic pressure experienced by high-priority services.

[0045] Step S600: Feedback the traffic scheduling solution to the path priority table through a feedback mechanism for updating.

[0046] In this step, the feedback mechanism ensures that the dynamic adjustment process of the cyclic update is well-founded, thereby optimizing the suboptimality and durability of the path selection decision. Specifically, based on the feedback data, the path priority table will perform incremental learning with reference to the scheduling effect of the past round, gradually improving the network's predictive ability and decision-making efficiency.

[0047] For example, by setting routes that have performed well in multiple feedbacks as long-term priority options, it is possible to adapt more quickly to future traffic changes.

[0048] In this embodiment, network node information is obtained, and several available paths are obtained based on this information. Then, the bandwidth evaluation algorithm is used to analyze and evaluate these available paths, and the network bandwidth data and initial weight of each path are calculated. The initial weight is adjusted by the preset path priority, and the initial path weight and path priority table are formed to allocate and identify the data packets of the network traffic. After that, the traffic data packets with path identification are input into the preset dynamic monitoring module to obtain the traffic load log and network delay data. According to these data, the initial path weight is adjusted in real time, the corresponding adjustment coefficient is generated, and the target path weight is generated accordingly. Finally, according to the adjustment coefficient and the target path weight, the data packets of the network traffic are redistributed, the traffic scheduling plan is generated, and the path priority table is updated through the feedback mechanism to realize the intelligent optimization of multi-path routing, so that the network can adaptively adjust the traffic distribution in a dynamic environment. The network bandwidth utilization is improved, the network delay is reduced, and the robustness and flexibility of the network when the traffic increases suddenly and the path state changes are improved, and the shortcomings of the existing technology in dynamic adaptability are improved.

[0049] Embodiment 2:

[0050] In step S100, network node information and node topology of network nodes are obtained through routing protocols, and the path between each pair of network nodes in the network node information is calculated using Dijkstra algorithm based on the node topology to obtain several available paths.

[0051] Through the communication protocol between routers and switching nodes, detailed information and connection relationships of each node in the entire network are obtained in real time to ensure that the node topology is complete and correct. Specifically, the Dijkstra algorithm is applied to gradually calculate the shortest path or multiple available path combinations between each pair of source and destination nodes to ensure that the extracted path has potential and efficient carrying capacity.

[0052] For example, in a simple topology consisting of five routing nodes, the Dijkstra algorithm can quickly identify two pairs of paths, facilitating subsequent quality assessment and weight analysis.

[0053] Each available path is evaluated by a bandwidth evaluation algorithm to obtain an evaluation result. The mean method is used to obtain network bandwidth data for the evaluation result. The variance method is used to obtain the loss factor for the evaluation result and network status data is generated based on the loss factor.

[0054] By continuously monitoring the real-time data flow transmission capacity of the path, and combining historical data with the mean method and variance method, we can accurately analyze the bandwidth level and potential transmission loss that each path can achieve. Specifically, the mean method is used to grasp the average transmission efficiency of the path, and the variance method is used to identify the sensitivity of network load changes, thereby comprehensively describing the current network status of the path.

[0055] For example, when one of the three paths exhibits a high bandwidth mean and a low variance, this suggests that the path is relatively stable under the current and expected load conditions.

[0056] The network bandwidth data and network status data are used to perform a weight analysis on each available path to obtain the initial weight of each path.

[0057] By performing multi-level analysis on the acquired bandwidth data and status data, a comprehensive evaluation index reflecting the path capacity and robustness is generated. Specifically, the initial weight is set high or low depending on the weighted calculation results of multiple factors such as transmission rate, loss impact, and load adaptation.

[0058] For example, if a path exhibits the highest bandwidth and the lowest state loss, its initial weight is relatively large, so it is preferred in the allocation in the initial stage.

[0059] In step S200, the initial weights are adjusted using the preset path priorities to obtain an adjusted weight set, the adjusted weight set is weighted averaged to obtain an average weight value and a weight deviation value, and the average weight value is compared with the initial weight to obtain a weight difference value.

[0060] Through the pre-defined path priority criteria, the initial weight of each path is slightly modified to better meet the changing characteristics of business needs. Specifically, the overall attractiveness value of each path is obtained through weighted average, while the weight deviation value and difference value reflect the dynamic deviation degree of path performance.

[0061] For example, when a path shows a gradually strengthening trend in multiple business changes, its weight difference value is positive, which strengthens its participation in future traffic distribution.

[0062] The weight correction value of each path is calculated according to the weight deviation value and the weight difference value, and the initial weight is corrected according to the weight correction value to obtain the initial path weight.

[0063] By calculating and applying the specific deviation data of each path, more accurate weight adjustment is performed to make the weight closer to the optimal state in daily operation practice. Specifically, the weight correction value will function like a filter to remove the interference of random traffic fluctuations so that the path performance will no longer lag.

[0064] For example, since a path has no obvious usage change during a period, its correction value is close to zero, thus stabilizing its role in subsequent scheduling.

[0065] The initial path weights are sorted to obtain a path sorting sequence, and a path priority table is generated based on the path sorting sequence.

[0066] By considering multi-dimensional path performance indicators, a strict priority arrangement is formed to provide clear support for subsequent packet processing. Specifically, in the path priority table, a higher priority mark is set for efficient paths to guide the execution of traffic distribution strategies.

[0067] For example, if a path is strategically located next to a high-bandwidth node and has a small delay confirmation, it is assigned the first priority.

[0068] The path priority table is associated with the characteristic data of network traffic to obtain the data packet allocation ratio and data packet allocation rules.

[0069] The flow type and target indicators given by the characteristic data are reasonably combined with the priority table to ensure that each data packet is delivered accurately and the priority resources are maximized. Specifically, allocation ratios and rules are created based on traffic characteristics to control transmission bandwidth, timeliness and reliability.

[0070] For example, large file transfer programs are mapped to the most stable path, while small data packets such as conversations are preferentially transmitted on low-latency paths.

[0071] The data packets of the network traffic are allocated and identified according to the data packet allocation ratio and the data packet allocation rule to obtain the traffic data packets with path identification; wherein the characteristic data of the network traffic includes the data packet size, protocol type and source and destination addresses.

[0072] By comprehensively evaluating feature data and allocation strategies, routing identifiers are attached to traffic data packets in real time to ensure accurate and intelligent transmission paths. Specifically, for traffic data packets of different sizes and protocols, exclusive transmission lines are designed and the current optimal facilities are identified.

[0073] For example, to ensure stable video calls, specific path identifiers will be maintained as routing priorities during such sessions.

[0074] In step S300, the traffic data packets with path identifiers are input into a preset dynamic monitoring module to perform traffic tracking on the traffic data packets with path identifiers, and a traffic load log is generated based on the traffic tracking result.

[0075] Through the high-precision monitoring module function, the data volume and timeliness fluctuations experienced by traffic with different identifiers in the past links are observed in detail to generate detailed load logs. Specifically, it records parameters including but not limited to the packaging path, stop time on the way, and transient changes caused by traffic fluctuations.

[0076] For example, the traffic load log identifies a surge in the amount of data flowing into a certain path during a certain period of time, and estimates its congestion related to the external system.

[0077] The delay monitor is used to monitor the delay of the traffic data packets with path identifiers to obtain network delay data.

[0078] The full-path delay tracking method ensures that the real-time delay status of each path selection is well documented, laying the foundation for the next optimization analysis. Specifically, a dedicated delay monitor is used to evaluate the time it takes for a data packet to reach the destination station. The data feedback is immediately integrated into the delay database.

[0079] For example, a path may show abnormal latency during peak traffic periods, prompting timely path replacement or load balancing adjustments during the operation phase.

[0080] In step S400, the traffic load log is predicted by time series analysis to obtain traffic fluctuation and delay trend, fuzzy clustering is used to perform pattern recognition on traffic fluctuation to obtain traffic fluctuation pattern, and spectrum analysis is performed on delay trend according to spectrum analysis to obtain delay spectrum.

[0081] Advanced time series analysis algorithms are used to evaluate logs vertically and horizontally, forming a theoretical basis for predicting dynamic traffic and delay trends. Specifically, the traffic pattern is fuzzy clustered to obtain the optimal law that is close to the sample concentration trend, and by scanning the delay spectrum, interactive nodes where delays frequently occur are presented.

[0082] For example, spectrum analysis clearly points out the early large-discrete operation of a certain node, providing a basis for subsequent correction and transmission.

[0083] The fuzzy comprehensive evaluation method is used to comprehensively evaluate the traffic fluctuation pattern, delay spectrum and network delay data to obtain the path quality index. The initial path weight is adjusted according to the path quality index, and the adjustment result is obtained using the linear regression method.

[0084] By integrating multiple available trait information, fuzzy evaluation is used to ensure that the path quality results are highly effective. Specifically, linear regression compares the index with historical reference data to accurately drive the subsequent weight distortion plan.

[0085] For example, regression coefficients are used to adjust the weights of noisy paths to capture certain potential paths that can be reused to generate additional value.

[0086] The adjustment results are estimated according to the Monte Carlo simulation to obtain the adjustment coefficients. The initial path weights are globally optimized according to the adjustment coefficients to obtain the weight change vector.

[0087] By adjusting the data stability through large-scale simulation settings, the accuracy gradually increases and tends to a reasonable range after multiple recursive calculations. Specifically, the adjustment coefficient makes corrections to the path observation efficiency and comprehensively aggregates the data items, thereby obtaining a weight change vector diagram. Adjusting the stability vector in the self-assessment randomness result can enable excellent paths to smoothly adapt to environmental changes.

[0088] For example, the adjustment coefficient significantly reduces the frequency of matches that are not usually used, further guided by the concept of priority paths.

[0089] The weight change vector is calculated by genetic algorithm to obtain the target path weight.

[0090] By introducing an evolutionary genetic algorithm, we explore and find the optimal path of the various solutions that were initially skipped, just like optimizing the branches of a tree. Specifically, through heuristic query and seed point rewriting, a new round of weight setting is formed and assimilation continues.

[0091] For example, experimental setups will compensate for problems caused by unexpected failures, using genetic algorithms to examine multiple generations of samples to successfully reach a certain value.

[0092] In step S500, the adjustment coefficients and the corresponding target path weights are normalized, and a weight matrix is ​​generated based on the normalized data.

[0093] By linearly transforming all values ​​of adjustment coefficients and target path weights, their range is limited to a standardized interval, which facilitates stable operation and consistent trade-offs in subsequent calculations. Specifically, the weights of all paths are numerically scaled to ensure that the importance ratio of the entire path set remains clear and accurate.

[0094] For example, assuming that the preliminary path weights are in the range of 0 to 100, they are converted to 0 to 1 through normalization, so that the weight matrix can reflect the comparative differences within the standardized range.

[0095] The weight matrix is ​​input into the preset allocation model to allocate the data packets of the network traffic, and a traffic ratio table is generated based on the allocation results.

[0096] Through the advanced allocation model, under the guidance of the weighted matrix, each package data flow is automatically identified and matched with the appropriate path to form an effective flow flow ratio structure. Specifically, the allocation model will determine which path is the best choice for different types of data flows according to the matrix value to alleviate bandwidth bottlenecks and efficiency losses.

[0097] For example, in multimedia data transmission scenarios, stable and low-latency paths are preferred to ensure smooth transmission.

[0098] Select the optimal path according to the traffic ratio table, and build a traffic scheduling plan based on the optimal path.

[0099] By analyzing the generated traffic allocation table, the path with the highest utility is selected and the transmission strategy is adjusted to formulate a comprehensive and refined traffic scheduling plan. Specifically, the allocation strategy is changed based on real-time demand. For example, during light load periods, some temporarily idle paths are activated to transmit traffic between different servers.

[0100] For example, when a traffic burst occurs, the candidates for temporarily open paths are automatically balanced to evenly share the traffic.

[0101] In step S600, a deep learning algorithm is used to perform traffic identification on the traffic scheduling scheme to obtain path scheduling effect data, and the path scheduling effect data and the path priority table are compared and analyzed to obtain difference information.

[0102] The adaptive neural network model is used to judge the environmental adaptability of the traffic scheduling scheme in terms of multi-dimensional indicators, and the real-time decision-making accuracy of the output is continuously optimized. Specifically, the path management smart contract proposes a path update strategy for differences by learning and analyzing the scheduling data set.

[0103] For example, if the experimental scheduling of an existing path proves to be frequently below the effectiveness threshold, the deep model can be used to retrieve the details of the plan for further adjustment.

[0104] The difference information is input into a preset feedback mechanism to update the path priority.

[0105] Potential efficiency gaps are identified and corrected by regularly comparing the current path with the expected data generated by the recommendation mechanism. Specifically, the feedback system integrates historical usage trends and standard model results to improve and gradually evolve the policy standards.

[0106] For example, if feedback shows that a path in the priority list is chronically inefficient, system feedback will trigger resource reallocation, changing its order in the list to regain operational efficiency.

[0107] In this embodiment, by combining data analysis technology and intelligent scheduling models, the core goal is to optimize the overall utilization efficiency of network resources. During the data packet transmission process, a multi-parameter adjustment optimization strategy is adopted to ensure that each path selection is based on the current network environment, and the rationality of traffic distribution is dynamically adjusted to reduce the risk of delay and blocking. Normalization processing and advanced allocation strategies provide intelligent selection for complex and diverse network traffic, ensuring real-time and efficient data packet transmission. In addition, introducing deep learning into the feedback loop strengthens the scientificity and foresight of scheduling decision updates, promotes the gradual and coagulation of network strategies, and effectively improves network stability and user experience.

[0108] Embodiment 3:

[0109] like Figure 2 As shown, the present application also provides a multi-path optimization device 10 for network routing, including an acquisition module 11, an allocation module 12, a monitoring module 13, an adjustment module 14, a generation module 15 and an update module 16.

[0110] The acquisition module 11 is mainly used to acquire network node information, obtain several available paths based on the network node information, analyze and evaluate all available paths using a bandwidth evaluation algorithm, and obtain network bandwidth data and an initial weight of each available path.

[0111] The allocation module 12 is mainly used to adjust the initial weight through the preset path priority to obtain the initial path weight and path priority table, and use the path priority table to allocate and identify the data packets of the network traffic to obtain the traffic data packets with path identification.

[0112] The monitoring module 13 is mainly used to input the traffic data packets with path identifiers into the preset dynamic monitoring module 13 to obtain the traffic load log and network delay data.

[0113] The adjustment module 14 is mainly used to adjust the initial path weight based on the traffic load log and the network delay data, and generate a corresponding adjustment coefficient based on the adjustment result, and generate a corresponding target path weight according to the adjustment coefficient.

[0114] The generation module 15 is mainly used to distribute data packets of network traffic according to the adjustment coefficient and the corresponding target path weight, and generate a traffic scheduling plan.

[0115] The updating module 16 is mainly used to feed back the traffic scheduling solution to the path priority table for updating through a feedback mechanism.

[0116] In this embodiment, an efficient network routing multi-path optimization device is constructed by integrating the collaborative work of the acquisition module 11, the allocation module 12, the monitoring module 13, the adjustment module 14, the generation module 15 and the update module 16. The acquisition module 11 uses the network node information to carry out multi-path calculation and evaluation, and provides accurate basic data for subsequent traffic distribution. The allocation module 12 effectively distributes data packets by analyzing the initial weight and path priority to ensure that the traffic is evenly and efficiently transmitted in the network. The monitoring module 13 tracks the dynamic changes of the identified traffic path in real time, generates accurate traffic load logs and delay data, and provides real-time data support for the adjustment module 14. The adjustment module 14 optimizes the path weight based on detailed monitoring feedback, so that the network adapts to environmental changes and improves the flexibility and stability of data transmission. The generation module 15 implements a formal traffic scheduling plan based on the optimized path weight to further improve the efficiency of network resource utilization. The update module 16 acts on the decision feedback formed during the adjustment process, so that the path priority table is updated in a timely manner, so that the system has self-learning and adaptability, and forms a benign optimization closed loop. Overall, the design of this device can significantly improve the refined management capabilities of network routing and is suitable for complex and changing network environments.

[0117] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and conciseness of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned multi-path optimization method embodiment of network routing, and will not be repeated here.

[0118] Embodiment 4:

[0119] like Figure 3 As shown, the present application also provides a multi-path optimization system 20 for network routing, including an acquisition unit 21, an allocation unit 22, a monitoring unit 23, an adjustment unit 24, a generation unit 25 and an update unit 26.

[0120] The acquisition unit 21 is mainly used to acquire network node information, obtain a number of available paths based on the network node information, analyze and evaluate all available paths using a bandwidth evaluation algorithm, and obtain network bandwidth data and an initial weight of each available path.

[0121] The allocation unit 22 is mainly used to adjust the initial weight through the preset path priority to obtain the initial path weight and path priority table, and use the path priority table to allocate and identify the data packets of the network traffic to obtain the traffic data packets with path identification.

[0122] The monitoring unit 23 is mainly used to input the traffic data packets with path identifiers into the preset dynamic monitoring module 13 to obtain the traffic load log and network delay data.

[0123] The adjustment unit 24 is mainly used to adjust the initial path weight based on the traffic load log and the network delay data, and generate a corresponding adjustment coefficient based on the adjustment result, and generate a corresponding target path weight according to the adjustment coefficient.

[0124] The generating unit 25 is mainly used to distribute the data packets of the network traffic according to the adjustment coefficient and the corresponding target path weight, and generate a traffic scheduling solution.

[0125] The updating unit 26 is mainly used to feed back the traffic scheduling solution to the path priority table for updating through a feedback mechanism.

[0126] In this embodiment, through the efficient integration of the acquisition unit 21, the allocation unit 22, the monitoring unit 23, the adjustment unit 24, the generation unit 25 and the update unit 26, the high-precision execution of the network routing multi-path optimization system is achieved. The acquisition unit 21 makes full use of the network node information, accurately constructs several available paths, and generates important bandwidth data and path initial weights through the bandwidth evaluation open evaluation method. The allocation unit 22 reasonably distributes the traffic according to the path priority table, and marks the transmission path of the traffic to ensure that traffic of different priorities is properly processed. The monitoring unit 23 is responsible for dynamically observing the actual operation of the traffic packet in the network, recording the traffic load and network delay, and providing a data basis for the decision of the adjustment unit 24. The adjustment unit 24 adjusts the path weight in real time through these monitoring data, and continuously optimizes the target path weight to adapt to the changing network needs. The generation unit 25 generates a traffic scheduling plan based on the optimized path weight to ensure that the data packet flows along the optimal path. The update unit 26 adjusts the path priority table through a feedback mechanism to ensure the continuous optimization and improvement of the system. Overall, this system effectively promotes load balancing between network paths, improves the efficiency of network resource allocation, and is suitable for deployment in network environments that require high stability and high efficiency.

[0127] It should be noted that technicians in the relevant technical field can clearly understand that for the convenience and conciseness of description, the specific working process of the above-described device and each module can refer to the corresponding process in the aforementioned multi-path optimization method embodiment of network routing, and will not be repeated here.

[0128] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A multi-path optimization method for network routing, characterized in that: include: Acquire network node information, obtain a number of available paths based on the network node information, analyze and evaluate all the available paths using a bandwidth evaluation algorithm, and obtain network bandwidth data and an initial weight of each available path; The initial weight is adjusted by a preset path priority to obtain an initial path weight and a path priority table, and the data packets of the network traffic are allocated and identified by using the path priority table to obtain a traffic data packet with a path identification; Input the traffic data packet with the path identifier into a preset dynamic monitoring module to obtain traffic load log and network delay data; Adjusting the initial path weight based on the traffic load log and the network delay data, generating a corresponding adjustment coefficient based on the adjustment result, and generating a corresponding target path weight according to the adjustment coefficient; Allocate data packets of network traffic according to the adjustment coefficient and the corresponding target path weight to generate a traffic scheduling plan; The traffic scheduling solution is fed back to the path priority table through a feedback mechanism for updating.

2. The multi-path optimization method for network routing according to claim 1, characterized in that: The step of acquiring network node information, obtaining a plurality of available paths based on the network node information, analyzing and evaluating all the available paths using a bandwidth evaluation algorithm, and obtaining network bandwidth data and an initial weight of each of the available paths includes: Obtaining network node information and a node topology structure of network nodes through a routing protocol, and calculating a path between each pair of network nodes in the network node information using a Dijkstra algorithm based on the node topology structure to obtain a plurality of available paths; Evaluate each of the available paths using a bandwidth evaluation algorithm to obtain an evaluation result, obtain network bandwidth data using a mean method for the evaluation result, obtain a loss factor using a variance method for the evaluation result, and generate network status data based on the loss factor; The network bandwidth data and the network status data are used to perform a weight analysis on each of the available paths to obtain an initial weight of each path.

3. The multi-path optimization method for network routing according to claim 1, characterized in that: The step of adjusting the initial weight by a preset path priority to obtain an initial path weight and a path priority table, and using the path priority table to allocate and identify data packets of network traffic to obtain traffic data packets with path identifications includes: The initial weights are adjusted using a preset path priority to obtain an adjusted weight set, weighted average is performed on the adjusted weight set to obtain an average weight value and a weight deviation value, and the average weight value is compared with the initial weight to obtain a weight difference value; Calculating a weight correction value of each path according to the weight deviation value and the weight difference value, and correcting the initial weight according to the weight correction value to obtain an initial path weight; Sorting the initial path weights to obtain a path sorting sequence, and generating a path priority table based on the path sorting sequence; Correlation analysis is performed on the path priority table and characteristic data of network traffic to obtain a data packet allocation ratio and a data packet allocation rule; The data packets of the network traffic are allocated and identified according to the data packet allocation ratio and the data packet allocation rule to obtain traffic data packets with path identification; wherein the characteristic data of the network traffic includes data packet size, protocol type, and source and destination addresses.

4. The multi-path optimization method for network routing according to claim 1, characterized in that: The step of inputting the traffic data packet with the path identifier into a preset dynamic monitoring module to obtain the traffic load log and network delay data includes: Inputting the traffic data packet with the path identifier into a preset dynamic monitoring module to perform traffic tracking on the traffic data packet with the path identifier, and generating a traffic load log based on the traffic tracking result; The delay monitor is used to monitor the delay of the traffic data packet with the path identifier to obtain network delay data.

5. The multi-path optimization method for network routing according to claim 1, characterized in that: The step of adjusting the initial path weight based on the traffic load log and the network delay data, generating a corresponding adjustment coefficient based on the adjustment result, and generating a corresponding target path weight according to the adjustment coefficient includes: Perform time series prediction on the traffic load log by using a time series analysis method to obtain traffic fluctuation and delay trend, perform pattern recognition on the traffic fluctuation by using fuzzy clustering to obtain traffic fluctuation pattern, perform spectrum analysis on the delay trend by using spectrum analysis to obtain delay spectrum; A fuzzy comprehensive evaluation method is used to comprehensively evaluate the traffic fluctuation pattern, the delay spectrum and the network delay data to obtain a path quality index, the initial path weight is adjusted according to the path quality index, and a linear regression method is used to obtain an adjustment result; Perform parameter estimation on the adjustment result according to Monte Carlo simulation to obtain an adjustment coefficient, and perform global optimization calculation on the initial path weight according to the adjustment coefficient to obtain a weight change vector; The weight change vector is calculated by a genetic algorithm to obtain the target path weight.

6. The multi-path optimization method for network routing according to claim 1, characterized in that: The step of allocating data packets of network traffic according to the adjustment coefficient and the corresponding target path weight to generate a traffic scheduling solution includes: Normalizing the adjustment coefficient and the corresponding target path weight, and generating a weight matrix based on the normalized data; Inputting the weight matrix into a preset allocation model to allocate data packets of network traffic, and generating a traffic allocation table based on the allocation result; An optimal path is selected according to the flow ratio table, and a flow scheduling solution is constructed based on the optimal path.

7. The multi-path optimization method for network routing according to claim 1, characterized in that: The step of feeding back the traffic scheduling solution to the path priority table for updating through a feedback mechanism includes: Using a deep learning algorithm to perform traffic identification on the traffic scheduling scheme to obtain path scheduling effect data, and comparing and analyzing the path scheduling effect data with the path priority table to obtain difference information; The difference information is input into a preset feedback mechanism to update the path priority.

8. A multi-path optimization device for network routing, characterized in that: include: An acquisition module is used to acquire network node information, obtain a number of available paths based on the network node information, analyze and evaluate all the available paths using a bandwidth evaluation algorithm, and obtain network bandwidth data and an initial weight of each available path; An allocation module, used to adjust the initial weights by a preset path priority to obtain an initial path weight and a path priority table, and to allocate and identify data packets of network traffic using the path priority table to obtain traffic data packets with path identifications; A monitoring module, used for inputting the traffic data packet with the path identifier into a preset dynamic monitoring module to obtain a traffic load log and network delay data; An adjustment module, configured to adjust the initial path weight based on the traffic load log and the network delay data, generate a corresponding adjustment coefficient based on the adjustment result, and generate a corresponding target path weight according to the adjustment coefficient; A generation module, used to distribute data packets of network traffic according to the adjustment coefficient and the corresponding target path weight, and generate a traffic scheduling plan; An updating module is used to feed back the traffic scheduling solution to the path priority table for updating through a feedback mechanism.

Images