Simulation of a Base Station with Integrated Services

Nowadays new key parameters have to be satisfied in the world of telecommunications, in terms of speed, latency and number of users to serve. In order to satisfy these new parameters it is necessary to move from the 4G technology to the 5G technology. It implements many techniques that are aimed to reach the goals described above, and in particular we are interested in the description of network densification and small cells. As the number of users increases, it is necessary to increase also the number of base stations covering the geographical area, in such a way there are many small cells where the bit rates are reasonably high. It is possible to mix the two technologies by putting a small 5G cell in the area of a macro 4G cell, in particular the small cell should be put in the Hotspot, where do we expect to have more traffic. In this way, most of the traffic will be managed by the small cell at high speed, while the other by the macro cell. In this scenario we have also to consider the mobility of the users, because as long as they move, they can change cell of competence, without completing their service. In addition, in modern communications, we can split the services in two main categories: inelastics and elastics. The first ones are the services in which the service rate is fixed by the rate at which data are generated by the source, and cannot be increased or decreased by the network. Elastic customers, instead, requires the transmission of data at the maximum possible speed; in this case, data is available at the source and its data rate depends on network capabilities, and capabilities of transmitter/receiver to transmit or receive data. In small words, inelastic customers require a fixed data rate, while the elastic ones require a minimum data rate, and on top of that, they equally share the capacity not used by inelastic services. So now, the main scope of this thesis is to understand how a small 5G cell acts when it received different types of traffic; once we know, that it is interesting also to study how it interacts with a macro cell, creating the so called HetNet, which are heterogeneous networks comprising two (or more) layers of BSs overlaid in a same area, possibly using different radio access technologies. It is necessary to create, first of all, a queueing theoretical model for the cell, and then a simulation model, in order to get some results. Our base station will be simulated by means of a queue, in which 2 types are traffic can be received inelastics and elastics. The queue has a finite capacity, and a customer can enter the queue if and only if there is enough room to serve it; in particular, an inelastic customer needs a fixed data rate available in order to be served, instead, the elastic needs just a minimum rate, if more is available then it will use all of it. Once the queue has been designed, the generation and service processes have to be described; at the begin, the generation process will be a Poisson one, described by exponential random variable, and the same for service time, it will be described by an exponential random variable. In this way, it is possible to compute analytically the results by studying the Continuous Time Markov Chain, and then we can compare them with the values given by the simulator, in order to check for some error. Simulation model, instead, has been created by mean of a software called Omnet++. Once the simulation model is ready, then many cases can be simulated.