Research Of Behavior Reconfiguration Methods In Large-scale Network Service System

Under the condition of limited system resources and event disturbance,the large-scale network service system user scale is rapidly expanding in a short time,resulting in the rapid decline in service quality and service capability.How to dynamically adapt to the changes of users and environment,that is,how to maintain the load dynamically balance of the adaptive behavior reconstruction,to guide the user behaviors and maintain the stability of the system service,that is a current hot spot in the field of network service system research.Most of the existing researches consider user needs or changes of the individual needs and whether the system meets the needs of users.This thesis mainly considers the system anomaly caused by the short time aggregation of the rigid demand behaviors of large-scale users,and puts forward the model and method of large-scale network service system behavior reconfiguration,the main work is as follows:1.The system behavior reconfiguration is to provide better adaptive behavior reconstruction while the system load is dynamically balanced to guide the group user behaviors and maintain stable operation of the system.A system behavior reconfiguration model and method based on abnormal user behaviors detection is proposed to deal with the problem of the rapid rise of large-scale network service system user amount,which causes the increase of system load and the unstable operation of the system,the behaviors of system users are divided into normal user behaviors and abnormal user behaviors,according to whether they are repeated or not,and the abnormal ones will be controlled.On this basis,this thesis proposes a model and algorithm of the system behavior reconstruction based on the Petri Net with priority and timed color double transition.The model and method can reduce the system load to a certain extent,ensure the timely response of the normal user request behaviors,and improve the stability of the system.2.A system behavior reconfiguration model and method based on user behaviorclassification is proposed to solve the problems of system anomaly and system load overload caused by the rapid expansion of system user scale in a short time.According to the characteristics of user interaction behaviors,the user behaviors are classified,and this thesis proposes the model and algorithm based on the timed stochastic fuzzy Petri Net.By dynamically controlling the time delay of different user groups,the system load is dynamically balanced and system adaptive reconfiguration.The model and method can reconstruct the system model and reduce the system load when the system load exceeds the warning point,and at the same time,it improves system availability and ensures good and stable operation of the system.In short,this thesis proposes a system behavior reconfiguration model based on abnormal user behaviors detection and a system behavior reconfiguration model based on user behavior classification to solve the problems of system anomaly caused by the rapid expansion of system users scale in short time.It provides technical support to solve the problem of system anomaly caused by load overload in large-scale network service system.