Call Admission Control under a Probabilistic Bandwidth Reservation Policy and Handover Queueing in Mobile Hotspots

Performance evaluation of a mobile hotspot supporting quasi‐random traffic

“…In Figures 8 and 9, we present the total CBP of new calls B new (Figure 8) and the CBP of handover calls $${B}_{s,h}^{\ast }$$ (Figure 9) as a function of the values of a new (for a new = 20, 24, 28,…,80) for the proposed model using three different values of p s ( p s = 0, 0.5, 1) for three cases of number of sources: (i) N h = N new = 50, (ii) N h = N new = 100 and (iii) for comparison reasons, the case of random call arrivals (Poisson arrivals) [14].…”

“…The opposite happens when p s tends to 1.0, (iii) as in the case of B new , higher initial number of call sources lead to higher $${B}_{s,h}^{\ast }$$ and vice versa. Again, as the number of new call sources increases the achieved $${B}_{s,h}^{\ast }$$ values are upper bounded by those of the corresponding random call arrivals case (Poisson arrivals) [14].…”

“…[13] and for loss/ queueing systems in Ref. [14] which serve random traffic (traffic generated via an infinite number of users). To the best of our knowledge, this is the first work that studies quasi-random traffic in loss/queueing models applied in a mobile hotspot.…”

“…The motivation of studying the case of quasi-random traffic relies on the fact that a finite number of users (and not an infinite number as in the case of Poisson [random] traffic) do create traffic in a mobile hotspot. Based on the above, the contribution of our paper is the following: (i) we propose the case of a mobile hotspot that serves quasi-random traffic under the probabilistic BR policy applied to new calls during the stop phase while handover calls have the option to enter a queue of finite size, (ii) we present an accurate analytical method for the determination of time congestion probabilities, CBP, system's utilisation and mean waiting time of handover calls in the queue and (iii) we compare the proposed model with the models of Refs [11,14] in order to show its necessity.…”

“…In Section 3, we present the proposed quasi-random probabilistic BR loss/queueing model, we describe the analytical model (Section 3.1), we present the corresponding global balance equations for the steady state probabilities (Section 3.2) and present an analytical methodology for the determination of the steady state probabilities and the various performance measures (Section 3.3). In Section 4, we compare the proposed model with the models of Refs [11,14] by presenting analytical results for the various performance measures. We conclude in Section 5.…”