Loss probability estimation and control for OFDM/TDMA wireless systems considering multifractal traffic characteristics

Admissible number of users in OFDM/TDMA based wireless systems through loss probability estimation considering multifractal traffic modeling

“…The scaling function plots for a self-similar traffic trace and for two wireless traffic traces are shown in Fig. 2 [12], where we can observe that the self-similar trace presents a linear relationship between () q  and q , which is fully consistent to the monofractal scaling behaviour. In contrast, the () q  function versus q for the wireless traffic traces, such as the USCtrace01, collected from the USC_06spring_trace packet from USC (University of Southern California), available at [14] and the ISF_wifidog traffic trace available at [16], show evidences of a nonlinear behavior, suggesting a multifractal structure.…”

“…The workload process () Wt is the total amount of work stored in the buffer in the time interval [0, ) t [13], i.e., ( ) ( ) W t A t ct  (12) where () At is the accumulated amount of work that arrives to the queue model and the service rate is c .…”

“…Another approach, similar to MSQ, is called CDTSQ (Critical Dyadic Time-Scale Queue) which takes into account only the queue length distributions at dyadic time scales [3]. Now, we can enunciate the following proposition regarding data loss as defined in (14) for multifractal traffic [12]. Proposition 1.…”

“…In this work, we argue that multifractal analysis can enhance the performance of OFDM/TDMA based wireless networks. Thus, by assuming a multifractal traffic model, we propose an expression for calculating the byte loss probability of wireless traffic flows that can be applied to evaluate the queueing performance in terms of byte loss rate of an OFDM/TDMA based wireless system, which was previously presented, by us, in [12]. Afterwards, based on the loss probability expression, we present, as a novelty of this paper, an admission control scheme that decides whether or not to allow new users to the OFDM/TDMA based wireless system.…”

Adaptive Rate Control Based on Loss Probability Estimation Considering a Cascade Based Multifractal Model for the Network Traffic

“…11 Among them, we can mention models based on Markov chains, autoregressive models, self-similar and multifractal models. 2,33 These traffic models are intended to provide ways of characterizing the network traffic behavior. Once the traffic is modeled, one can predict the network performance through analytical techniques or by simulation.…”