Mobility Prediction for Traffic Offloading in Cloud Cooperated mmWave 5G Networks

Future cellular networks are predicted to witness an extraordinary increase in mobile related traffic load in the next 10 years. This is the catalyst for the creation of the 5th generation (5G) cellular networks that could potentially accommodate much higher data rates by a factor of 1,000. Currently, there have been quite a few different proposed architectures that promise to support such an overwhelming demand. The utilization of the ultrawideband aspect of the mmWave bands is considered at the moment one of the most promising approaches, since it makes use of very high frequencies and therefore it offers a much higher theoretical channel capacity for data transfer. Under the umbrella of mmWave bands to be used for the implementation of 5G networks, many studies have proposed the incorporation of the currently dominant 4G/LTE technology to function alongside 5G and to be solely responsible for signaling and control data transfers (C-Plane), so as user data (U-Plane) will be given priority over higher 5G data rates whenever and wherever available. This heterogeneous network that could operate in a range of different frequencies over the same area and at the same time, may be enhanced even further with the use of a cloud infrastructure for radio access network (C-RAN) that would be responsible for overseeing the entire network topology's optimized functionality. Such a complex architecture is certain to bring to the surface some very challenging problems. The switching between 4G and 5G, whenever a User Equipment (UE) exits a pico cell or enters a new pico cell, is not as simple as normal handovers between cells that operate under the same technology. Service break ups and disruption of service are only two of the devastating results in user experience when dealing with sudden handovers between technologies and not just cells. In this paper, a mobility prediction scheme is proposed that makes use of C-RAN, titled Cloud Cooperated Mobility Prediction (CCMP) and instructs UEs under a certain probability whether or not they are predicted to exit a pico cell in the near future. If there is a positive chance for this to happen, the UE will take all the necessary actions to offload its data traffic from the U-Plane to the C-Plane in a much smoother and more efficient way.