Analysis On The Queueing Of Mobile Instant Messaging In Cellular Networks

With the development of information communication technology,mobile instant messaging(MIM)traffic should not be underestimated due to the increasing prosperity of mobile devices and social software applications,such as WeChat.In this thesis,we study the mechanism of MIM,and investigate the application of queueing theory to MIM in cellular networks.Researchers have found that the arrival interval of MIM received by the base station obeys log-normal distribution,indicating that the commonly used M/M/1 queueing model is not suitable for this scenario.Hence,we think of the G/M/1 queueing model for MIM in cellular networks.First,we focus on the single-user scenario with accompany of keep-alive(KA)mechanism,and deduce the relationship between KA duration and average delay.Both theoretical derivation and simulation results show that the average delay will reach its minimum if the KA duration is chosen appropriately.Furthermore,we also simulate the multi-user MIM case and obtain the relationship between the average delay and the number of users and the KA duration.Due to the roaming of cellular network users,we analyze the queueing network to model the home-routed roaming mobility.With the nature of queueing networks and Little’s Law,we obtain the expression of average delay in this model.The simulation results show that the average delay of roaming is related to the ratio of roaming users and the ratio of E-UTRAN and EPC service rates.Moreover,supposing that the computing resource of EPC is allocable,we investigate the allocation strategy which minimizes the average delay,and give the expression of the optimal allocati’on ratio.Finally,we examine the variation of optimal service rate ratio under different arrival rate ratio.These research results are basic but vital features of MIM and should be fully considered in the deployment of future cellular networks.