Traffic-aware overload control scheme in 5G ultra-dense M2M networks

Due to a huge number of M2M devices expected to access the network simultaneously, congestion and overload of networks are the core challenges to the 5G ultra-dense M2M networks design. In this paper, an optimal traffic-load control scheme is proposed to overcome the challenges. Since the base station lacks accurate information about the traffic-load of network resulting in an unachievable solution to the optimal control scheme in practice, we propose two different traffic-load estimation schemes corresponding to the two different access protocols. For the first protocol that the collision devices are not barred by an access probability, we propose a markov-chain based estimation scheme according to the collision state of network. For the second protocol that the collision devices are barred by an access probability, we estimate the traffic-load using the number of success devices and the average number of new active devices in each slot. Both estimation schemes ensure that the practical control schemes are very close to the theoretical optimal control scheme. In addition, we prove that the optimal performances of the two protocols are exactly the same in theory. Finally, the simulation results exhibit our conclusion and show the superior performances of our proposed scheme.