Performance modeling of wireless networks with generally distributed handoff interarrival times

Impact of mobility on call block, call drops and optimal cell size in small cell networks

“…Often, in literature the handover arrivals are also modeled by Poisson arrivals (see for example [1], [2]). Further, it is easy to see that new arrivals into a cell are totally independent of handed over calls from the neighboring cells.…”

“…As discussed earlier, the cell is modeled by an M/G/K/K queue and its service time is given by (2). Using the theory of M/G/K/K queue, the condition for stability of the system is that the load factor,…”

“…Because of the memoryless nature of S, this probability does not depend upon the number of the handover. Probability P ho,D can be calculated by first conditioning on the event that the call is completed in the current cell (call it as C) and then on the event that 2 When a call is not completed in the current cell, it invokes a handover arrival to the next cell, which is modeled as a Poisson arrival. By PASTA, handover fail probability, i.e., the probability that all servers in the next cell are busy at the instance when user reaches the boundary of the current cell, exactly equals PBusy.…”

Performance Analysis of Queuing Priority Schemes in cellular communication

“…When the MS strays beyond a certain range from the BS the RSS drops beyond a required level, there is the need for a transfer of service (channel) from one BS to another. This process of transferring an ongoing call from BS to another is called Call Handoff or Call Handover [2], [3], [4]. When the assigned channels of the receiving BS are occupied the call is forced to terminate.…”

Contribución a la gestión de recursos en redes de acceso celulares. Modelos analíticos y evaluación de prestaciones

“…La conclusión principal es que si el tiempo de residencia en una célula sigue una distribución Gama -suposición que está avalada por diferentes estudios como se verá mas adelante-con una varianza alta, el tráfico de handover no se puede caracterizar mediante un proceso de Poisson. Motivados por los trabajos antes citados, en [DTL03] los autores estudian un sistema en el que el tiempo entre llegadas de los handovers se modela mediante una distribución de Erlang o hiperexponencial. Los parámetros de interés para las prestaciones del sistema se comparan con los obtenidos cuando el tráfico de handover es de Poisson, manteniendo en ambos casos la misma tasa de llegadas.…”