On the system-level statistical characterization of link unreliability in cellular networks

In this paper, statistical characterization of link unreliability in cellular networks at system level is addressed. Traditionally, system-level performance evaluation of mobile wireless communication networks has been performed by considering only resource insufficiency, while the effect of unreliable wireless channel has been largely ignored. Recently, a new system level analytical approach to capture the effect of both link unreliability and radio resource insufficiency in mobile wireless networks was proposed. In that model, the effect of unreliable wireless channel is captured through the appropriate statistical characterization of the unencumbered call interruption time (UCIT). The present paper shows how the probability distribution function of unencumbered call interruption time can be derived from either few system level statistics that can be easily obtained at base stations or link level statistics. Numerical results clearly show that UCIT should be modeled as a general distributed random variable to adequately capture the effect of link unreliability.