CN-122001804-A - High-performance intelligent management method and system for network congestion self-optimization

CN122001804ACN 122001804 ACN122001804 ACN 122001804ACN-122001804-A

Abstract

The invention relates to the technical field of network self-optimization management, in particular to a high-performance intelligent management method and system for network congestion self-optimization, comprising the following steps: obtaining a network topology diagram, identifying a plurality of critical path nodes from the network topology diagram, configuring an acquisition module for the critical path nodes to obtain configured nodes, obtaining node network parameters and node association influence characteristic values, confirming root cause congestion nodes, restored nodes or normal nodes, taking the root cause congestion nodes as path optimization nodes if the root cause congestion equipment type is a core router, carrying out route path switching optimization operation on the path optimization nodes to obtain optimized nodes, taking the root cause congestion nodes as nodes to be balanced when the root cause congestion equipment type is network switching equipment, and carrying out load balancing optimization operation on the nodes to be balanced to obtain the optimized nodes. The invention can comprehensively improve the stability of network operation and the resource utilization efficiency.

Inventors

JIA XIAOJIE
WANG XINZHENG
CHEN XINLEI
CAI YUAN
ZHAO YUBING

Assignees

企商在线(北京)数据技术股份有限公司

Dates

Publication Date: 20260508
Application Date: 20260409

Claims (10)

1. A high performance intelligent management method for network congestion self-optimization, the method comprising: receiving a high-performance intelligent management instruction, and acquiring a network topology graph according to the high-performance intelligent management instruction, wherein the network topology graph comprises a plurality of nodes; Identifying a plurality of critical path nodes from the network topology map, performing the following for each critical path node of the plurality of critical path nodes: configuring an acquisition module for the critical path node to obtain a configured node, and acquiring node network parameters and node association influence characteristic values based on the configured node; confirming the root cause congestion node, the restored node or the normal node according to the node network parameters and the node association influence characteristic values; Confirming the type of root cause congestion equipment according to the root cause congestion node, wherein the type of the root cause congestion equipment is a core router or network switching equipment; if the root cause congestion equipment type is a core router, taking the root cause congestion node as a path optimization node, and performing route path switching optimization operation on the path optimization node to obtain an optimized node; If the root cause congestion equipment type is network switching equipment, taking the root cause congestion node as a node to be balanced, and carrying out load balancing optimization operation on the node to be balanced to obtain an optimized node; summarizing the optimized node, the recovered node and the normal node respectively to obtain an optimized node set, a recovered node set and a normal node set; and completing the high-performance intelligent management of the network congestion self-optimization based on the optimized node set, the recovered node set and the normal node set.
2. The method for intelligent management of network congestion self-optimization according to claim 1, wherein the obtaining node network parameters and node association influence feature values based on configured nodes comprises: Acquiring the historical byte number and the interface maximum bandwidth, acquiring data of the configured node by using a preset acquisition interval to obtain the current byte number, and calculating the node bandwidth utilization rate according to the historical byte number, the current byte number and the interface maximum bandwidth; Confirming a sending target and a detection packet, and sending the data packet for multiple times to the sending target by using the configured node and the detection packet to obtain sending times, receiving times and node average delay values; And calculating the node data packet loss rate according to the sending times and the receiving times, calculating the node association influence characteristic value according to the configured node, and confirming the node network parameters based on the node bandwidth utilization rate, the node average delay value and the node data packet loss rate.
3. The high-performance intelligent management method for network congestion self-optimization according to claim 2, wherein the calculation formula of the node bandwidth utilization rate is as follows: Wherein, the Indicating the bandwidth usage rate of the node, The number of bytes present is indicated as a current number of bytes, The number of bytes in the history is represented, The acquisition interval is indicated as such, Indicating the interface maximum bandwidth.
4. A high performance intelligent management method for network congestion self-optimization as recited in claim 3, wherein said calculating node association influence feature values from configured nodes comprises: Confirming an associated target node set in a network topological graph according to the configured nodes, and screening the associated target node set to obtain a downstream direct connection node set; According to the downstream direct connection node set, a downstream direct connection device set is confirmed, downstream direct connection devices are sequentially extracted from the downstream direct connection device set, and network access parameters and device operation parameters are obtained based on the extracted downstream direct connection devices; Performing abnormality judgment on the extracted downstream direct connection equipment according to the network access parameters and the equipment operation parameters to obtain performance judgment results, wherein the performance judgment results are abnormal performance or normal performance; if the performance judging result is abnormal, taking the extracted downstream direct connection equipment as abnormal performance equipment; Summarizing the performance abnormal equipment to obtain a performance abnormal equipment set, and respectively carrying out summation calculation on the performance abnormal equipment set and the downstream direct connection equipment set to obtain the number of the performance abnormal equipment and the number of the downstream direct connection equipment; and calculating a node association influence characteristic value according to the number of the abnormal performance devices and the number of the downstream direct connection devices, wherein the node association influence characteristic value is a numerical value obtained by dividing the number of the abnormal performance devices by the number of the downstream direct connection devices.
5. The method for intelligent management of network congestion self-optimization according to claim 4, wherein said identifying the root cause congestion node, the restored node, or the normal node according to the node network parameters and the node association influence characteristic values comprises: If the node network parameters do not accord with the preset standard node network parameter interval conditions, the configured node is used as a suspected congestion node, a first proportional threshold is obtained, and whether a node association influence characteristic value of the suspected congestion node is larger than the first proportional threshold is judged; If the node association influence characteristic value of the suspected congestion node is larger than a first proportional threshold, taking the suspected congestion node as a root cause congestion node; if the node association influence characteristic value of the suspected congestion node is not greater than a first proportional threshold, taking the suspected congestion node as an affected node, and performing self-optimization operation on the affected node to obtain a restored node; And if the node network parameters meet the standard node network parameter interval conditions, taking the configured node as a normal node.
6. The method for intelligent management of network congestion self-optimization of claim 5, wherein said obtaining a first proportional threshold comprises: acquiring a history alarm root cause device set, wherein the history alarm root cause device set comprises a plurality of history alarm root cause devices, and each history alarm root cause device corresponds to an alarm trigger time stamp; Sequentially extracting historical alarm root cause equipment from the historical alarm root cause equipment set, acquiring a time window based on an alarm trigger time stamp corresponding to the extracted historical alarm root cause equipment, confirming a direct connection equipment set according to the time window, and calculating the abnormal proportion of historical downstream equipment according to the direct connection equipment set; Summarizing the abnormal proportions of the historical downstream equipment to obtain a historical downstream equipment abnormal proportion set, extracting the median of the historical downstream equipment abnormal proportion set to obtain a median abnormal proportion, and taking the median abnormal proportion as a first proportion threshold.
7. The method for intelligent management of network congestion self-optimization according to claim 6, wherein performing route path switching optimization operation on the route optimization node to obtain an optimized node comprises: Retrieving a standby route path set from a pre-constructed route information base according to the route optimization node, sequentially extracting standby route paths from the standby route path set, and confirming a route node set based on the extracted standby route paths, wherein the route node set comprises one or more route nodes; Calculating the path node bandwidth utilization rate and the path node delay value of each path node in the path node set to obtain a path node bandwidth utilization rate set and a path node delay value set; Performing maximum value extraction operation on the path node bandwidth utilization rate set to obtain maximum path node bandwidth utilization rate, and performing summation operation on the path node delay value set to obtain a total path delay value; Calculating a standby path evaluation value based on the maximum path node bandwidth utilization and the total path delay value; Summarizing the standby path evaluation values to obtain a standby path evaluation value set, and taking the standby route corresponding to the standby path evaluation value with the largest standby path evaluation value in the standby path evaluation value set as the optimal standby path; and switching the routing path of the path optimization node according to the optimal standby path to obtain the optimized node.
8. The method for intelligent high-performance management of network congestion self-optimization according to claim 7, wherein the performing load balancing optimization operation on the node to be load-balanced to obtain an optimized node comprises: The method comprises the steps that high-load source equipment is confirmed according to a node to be balanced, a plurality of terminal equipment and a plurality of similar equipment are obtained according to the high-load source equipment, minimum load rate retrieval is conducted on the similar equipment, and candidate target equipment is obtained; Identifying a non-key terminal equipment set from a plurality of terminal equipment, and counting the number of the non-key terminal equipment in the non-key terminal equipment set; Obtaining a high load rate and a target device load rate according to the high load source device and the candidate target device, and calculating the theoretical migration device number based on the non-key terminal device number, the high load rate and the target device load rate; obtaining the physical maximum connection quantity and the current connection quantity of candidate target equipment, and calculating the residual connection quantity according to the physical maximum connection quantity and the current connection quantity; Correcting the number of theoretical migration equipment according to the number of the residual connection and the preset migration safety proportion to obtain the number of final migration equipment; And matching the equipment set to be migrated from the non-key terminal equipment set according to the number of the final migration equipment, and migrating the equipment set to be migrated to the candidate target equipment to obtain the optimized node.
9. The high-performance intelligent management method for network congestion self-optimization according to claim 8, wherein the calculation formula of the theoretical migration equipment number is as follows: Wherein, the Representing the number of theoretical migration devices, Indicating a high load-carrying capacity, the load-carrying capacity, Indicating that the load rate of the target device is high, Indicating the number of non-critical terminal devices, Representing rounding up symbols.
10. A high performance intelligent management system for network congestion self-optimization, the system comprising: the system comprises a critical path node identification module, a network topology graph and a network control module, wherein the critical path node identification module is used for receiving a high-performance intelligent management instruction, acquiring a network topology graph according to the high-performance intelligent management instruction, wherein the network topology graph comprises a plurality of nodes, identifying a plurality of critical path nodes from the network topology graph, and executing the following operation on each critical path node in the plurality of critical path nodes; The node state confirmation module is used for configuring the acquisition module of the critical path node to obtain a configured node, acquiring a node network parameter and a node association influence characteristic value based on the configured node, and confirming the root cause congestion node, the restored node or the normal node according to the node network parameter and the node association influence characteristic value; the root node optimizing module is used for confirming the type of root congestion equipment according to the root congestion node, wherein the type of root congestion equipment comprises a core router or network switching equipment, if the type of root congestion equipment is the core router, the root congestion node is used as a path optimizing node, the path optimizing node is subjected to route path switching optimizing operation to obtain an optimized node, if the type of root congestion equipment is the network switching equipment, the root congestion node is used as a node to be balanced, and the node to be balanced is subjected to load balancing optimizing operation to obtain the optimized node; the network self-optimization completion module is used for summarizing the optimized node, the recovered node and the normal node respectively to obtain an optimized node set, a recovered node set and a normal node set, and high-performance intelligent management of network congestion self-optimization is completed based on the optimized node set, the recovered node set and the normal node set.

Description

High-performance intelligent management method and system for network congestion self-optimization Technical Field The invention relates to the technical field of network self-optimization management, in particular to a high-performance intelligent management method and system for network congestion self-optimization. Background The network congestion self-optimization is an automatic network management process, which can automatically detect the congestion condition in the network, automatically adjust the distribution of network resources and the path selection through a series of intelligent algorithms and strategies, so as to relieve the congestion and improve the transmission efficiency and the performance of the network. The high-performance intelligent management is an advanced network management mode, and utilizes high-performance computing power and intelligent algorithm models to carry out omnibearing and automatic management on the network. The traditional network congestion management generally adopts a mode of combining manual monitoring with manual intervention, an administrator needs to manually collect whole network flow data by means of a monitoring tool, analyze flow characteristics and locate congestion nodes, and manually adjust core configurations such as routing strategies, bandwidth allocation rules and the like. The method has the technical limitations that on one hand, the manual operation link is long, the decision period is long, the response timeliness is insufficient, the instantaneous congestion caused by the burst flow is difficult to deal with, the data transmission delay is easy to increase, the packet loss rate is easy to increase, and on the other hand, the accurate identification and classification management mechanism of the terminal equipment is lacking, and the differential scheduling scheme cannot be formulated for different service attribute equipment such as an industrial control terminal, an office terminal and the like, so that the resource allocation is unbalanced. Therefore, how to comprehensively improve the stability of network operation and the resource utilization efficiency is a technical problem to be solved urgently. Disclosure of Invention The invention provides a high-performance intelligent management method and system for network congestion self-optimization, which mainly aim to comprehensively improve the stability of network operation and the resource utilization efficiency. In order to achieve the above object, the present invention provides a high-performance intelligent management method for network congestion self-optimization, including: receiving a high-performance intelligent management instruction, and acquiring a network topology graph according to the high-performance intelligent management instruction, wherein the network topology graph comprises a plurality of nodes; Identifying a plurality of critical path nodes from the network topology map, performing the following for each critical path node of the plurality of critical path nodes: configuring an acquisition module for the critical path node to obtain a configured node, and acquiring node network parameters and node association influence characteristic values based on the configured node; confirming the root cause congestion node, the restored node or the normal node according to the node network parameters and the node association influence characteristic values; Confirming the type of root cause congestion equipment according to the root cause congestion node, wherein the type of the root cause congestion equipment is a core router or network switching equipment; if the root cause congestion equipment type is a core router, taking the root cause congestion node as a path optimization node, and performing route path switching optimization operation on the path optimization node to obtain an optimized node; If the root cause congestion equipment type is network switching equipment, taking the root cause congestion node as a node to be balanced, and carrying out load balancing optimization operation on the node to be balanced to obtain an optimized node; summarizing the optimized node, the recovered node and the normal node respectively to obtain an optimized node set, a recovered node set and a normal node set; and completing the high-performance intelligent management of the network congestion self-optimization based on the optimized node set, the recovered node set and the normal node set. Optionally, the acquiring the node network parameter and the node association influence characteristic value based on the configured node includes: Acquiring the historical byte number and the interface maximum bandwidth, acquiring data of the configured node by using a preset acquisition interval to obtain the current byte number, and calculating the node bandwidth utilization rate according to the historical byte number, the current byte number and the interface maximum bandwidth; Confirming a