Research On QoS Based Telecom Networks Massive Data Processing Technology

With the rapid development of mobile Internet applications the mobile network face bigchallenges， traditional quality of service (QoS) analysis are difficult to solve these problems. Thisthesis made statistics analysis based on the massive data and focused on the impact oftelecommunications network services, which given the main factors that affect networkperformance and user experience.Firstly, this thesis make an overview of the concept of quality of service, then discussed issuesand concerns about the models for QoS.We also gives the difference and connections between QoSand quality of experience (QoE). After that we make a classficaton of mobile network servicesbased on flow characteristics, and also make a deep analysis of the key factors for the mobileInternet quality of service, then proposed solutions for the isseues that affecting the network QoS.This thesis make a analysis from resource consumption and KPI factors on the whole network basedon data flows,getting the regular pattern of2G,3G data flow characteristics.Finally, We proposed acongestion avoidance mechanism based on Kmeans algorithm,from this way we can guidancenetwork optimization and improve network quality of service.