Network equipment health degree evaluation method based on dynamic and comprehensive weights

Abstract

The invention discloses a network equipment health degree evaluation method based on dynamic and comprehensive weights. The method comprises the following steps: S1, determining a plurality of performance indicators for network equipment health degree evaluation, constructing a performance indicator system, and determining subjective weights of the performance indicators; S2, acquiring running data of network equipment to be evaluated, calculating values of the performance indicators, and establishing an evaluation matrix for the network equipment health degree evaluation; S3, calculating objective weights of the performance indicators; S4, calculating the dynamic and comprehensive weights of the performance indicators according to the corresponding subjective weights and the corresponding objective weights; and S5, according to the established evaluation matrix and the dynamic and comprehensive weights of the performance indicators, calculating a health degree of the network equipment to be evaluated, and carrying out the evaluation according to a health degree calculation result. The method provided by the invention has the advantages of high simplicity in implementation, high health degree evaluation precision and reliability and the like, and can achieve rapid and automatic evaluation.

Description

technical field [0001] The invention relates to the technical field of data center network performance evaluation, in particular to a method for evaluating the health of network equipment based on dynamic comprehensive weights. Background technique [0002] The data center concentrates various resources and business systems, is the center of data computing and storage, and realizes automatic computing and management based on standard building blocks, and provides services such as shared infrastructure, information processing, and application environments in a supply chain manner. Unlike traditional networks, data center networks have large-scale data processing and storage, and play an important role in applications such as the Internet of Things and cloud computing. Modern data center networks mainly have the following characteristics: [0003] ① The scale is getting larger and larger, and the network is heterogeneous and complex. [0004] Modern data centers usually have...

AI Technical Summary

Problems solved by technology

[0003] ① The scale is getting bigger and bigger, and the network is heterogeneous and complex

[0005] ②The traffic behavior is complex and the congestion is intensified

This type of method is to obtain the deviation degree of network delay, throughput rate, and normal connection rate by combining historical data, which can reflect the overall operation of the network. However, Tcpflags, which is used as an evaluation parameter, will cause tcp retransmission to increase to normal after the Tcp connection times out. The ratio of tcp connections may cause a large evaluation value, which cannot reflect the real network situation

[0016] (2) The network performance evaluation method combined with the cloud model and classification algorithm, through the classification training of normal samples and fault samples, such as using the SVM classification algorithm, combined with the cloud model to evaluate the network performance, can improve the objectivity of the evaluation results, However, in the process of sample selection and sample normalization, if the sample gap is relatively large, or the normalized parameter data deviates far from the normal operating value of the parameter, the evaluation result will be inaccurate

However, because this type of method only completes the evaluation of the network equipment’s operating state performance based on the calculated objective weight of the collected data, if the initially collected performance index data differs greatly, the weight values ​​of some indexes obtained will vary greatly, thus It makes it difficult for the final calculated comprehensive evaluation value to truly reflect the current operating conditions of equipment and networks

[0018] At present, the overall network performance is usually evaluated from the network link level. For a large network, the number of network links is far more than the number of devices in the network. In order to maintain the robustness of the network, the general large network , redundant backup links are usually established. Therefore, for the data center network, if the overall network performance is evaluated from the network link level, it will not only increase the difficulty of calculation, but also cause a large amount of calculation. In the main path, there will be a large difference between the final performance evaluation model and the actual performance value

Images