Service awareness method based on distributed monitoring and management structure

Abstract

The invention discloses a service awareness method based on a distributed monitoring and management structure. The method comprises the following steps of: 1) judging whether service flow is marked by a monitoring agent; 2) when the monitoring agent monitors unprocessed service flow, registering in a service registration form, and inquiring memory record; 3) preprocessing service information submitted by the monitoring agent, namely, vectoring characteristic information monitored by the monitoring agent; 4) putting the preprocessed information in SVM (support vector machine) modules for training by a semi-supervised learning method; 5) storing a sample obtained through training in step 4) in a service data sheet for use in next identification; and 6) after identifying and monitoring the service flow, marking and then turning to a next routing node. In the invention, the modules cooperatively work and the functions of the modules are ordered and unified, thus improving the effectiveness for service identification and ensuring the end-to-end QoS (quality of service).

Description

technical field [0001] The present invention is aimed at the high degree of autonomy of the cognitive network, and proposes a business perception method based on a distributed monitoring management framework. This method mainly studies how to monitor business data through intelligent agents, and carry out relevant intelligent engines to complete online efficient business identification. And design the relevant intelligent algorithm adopted by the recognition engine. The method of the invention relates to related fields such as data mining, traffic classification, network management, intelligent agent, etc., and belongs to a service perception method. Background technique [0002] With the continuous development of network technology and applications, the heterogeneity and complexity of the network are particularly prominent, and the types of services increase sharply, posing challenges for the QoS guarantee of the entire network. Identifying services on the access network a...

AI Technical Summary

Problems solved by technology

[0004] Technical problem: The purpose of this invention is to solve the problem that the traditional business identification research is limited to the expansion and development of the identification engine, and the identification engine has not been well integrated into the network, resulting in many inventions that cannot be reflected in actual value. A method based on A business-aware approach to distributed monitoring and management architecture

In the selection of the recognition engine, the present invention adopts the SVM intelligent algorithm, which solves the local minimum problem of some current intelligent algorithms, and is not limited by the number of training samples, and solves the problem of the disaster of sample dimensionality very well.

Images