Research On Key Techniques Of User Behaviors Control For Trustworthy And Controllable Network

Trustworthy and Controllable Network (TCN) separates control plane from data plane and constructs an independent control plane to realize controllability and trustworthiness. Ensuring network running smoothly and effective utilization of resources is very important to TCN. However, the huge user traffic has already made a huge impact on network resource management. On one hand, it gets harder to meet users’ QoS demands; on the other hand, the network is faced with problems such as low utilization of resources and unstable operation, etc. Therefore, it is a key problem in the research of TCN to study the mechanism of user behavior control and effective management of network resources.The current work has made some achievements in user behavior detection and control mechanisms etc., but there are still many deficiencies:first, with regard to behavior detection, most of existing researches detect user behavior based on the features of protocols instead of the dynamic traffic, making it impossible to detect dynamic user flow and difficult to support user behavior analysis and control; in addition, as for control and decision-making, on the one hand, with the absence of methods which can quantify the impact brought by user behavior to network status, it’s hard to make reasonable control decisions. On the other hand, most existing control and decision-making methods are aimed at a single target, so decisions cannot be made considering the big picture.This dissertation focuses on the theoretical research of user behaviors detection, evaluation, and control mechanisms to improve the limitations of prior works. Firstly, we propose a user behaviors detection method based on selective collaborative learning, which detects user behaviors based on the flow characteristics. Secondly, we propose an adaptive user behaviors evaluation method, which accurately quantifies the correlation between the users’ behaviors and network status fluctuations. Thirdly, we propose a SDN-based inner-domain user behaviors control method, which schedules users’ flow based on multiple target optimization. Fourthly, we propose a multi-path-based inter-domain user behaviors control method which detects multi-path routes and then schedules users’ flows. Finally, several modules on the trustworthy and controllable network experimental platform were developed to verify the feasibility and effectiveness of the theoretical work.This dissertation studies the user behavior control problem. The work in this dissertation can realize the guarantee of user demand meanwhile improve the efficiency of network resources and maintain smooth network running, which contributes to the research of network resource management.