Spiking neural networks (SNN) are expected to enable several use-cases in future communication networks (beyond 5G and 6G), as edge AI and battery-constrained systems can leverage the fast computation and high-power efficiency offered by SNNs. In this work we consider a Distributed Wireless SNN (DW-SNN) system and we analyze its performance in terms of inference accuracy and total neural activity when radio losses are applied to spikes transferred during the inference phase. Our aim is to understand how radio losses impact performance when considering different SNN spike communication types, i.e., input, excitatory, and inhibitory spikes. Then we evaluate the impact of different traffic prioritization approaches among SNN spikes when considering a shared channel capacity being available for SNN activity. From these analyses, we derive some key insights and features that can be considered when applying a DW-SNN and handling its traffic over wireless communication systems. Finally, we report a prototype implementation of DW-SNN using custom-built IoT components, which we use to further investigate different coverage scenarios.

We consider the connection between the packet loss ratio (PLR) in a switch with a finite buffer of size L and the tail distribution of the corresponding infinite buffer queue Q. In the literature the PLR is often approximated with the tail probability P(Q > L), and in practice the latter is often a good conservative estimate on the PLR. Therefore, efforts have mainly focused on finding bounds and asymptotic expressions concerning the tail probabilities of the infinite queue. However, our first result shows that the ratio PLR/P(Q > L) can be arbitrary, in particular the PLR can be larger than the tail probability. We also determine an upper bound on this ratio yielding an upper bound on the PLR using the tail distribution of the infinite queue. The bound is fairly tight for certain traffic patterns. In many situations it clearly improves the estimation with the tail probability, and it is rarely significantly larger than the estimate P(Q > L), while it is an upper bound. On the other hand, if the PLR is much smaller than P(Q > L), then our bound is usually loose. For this case a practically good approximation on their ratio is proposed.

It is widely accepted that the video frame size process is long range dependent (LRD), which has been indicated by a number of statistical tests. This approach suggests the use of LRD traffic models for traffic engineering, which in many cases leads to subexponential queue tails in the network buffers. Basically, there are two concepts that are considered as the origin of LRD in video modeling: self-similarity and heavy tailed level durations. In this paper we argue for a new approach, where the frame size sequence is described by three processes representing three explicit time scales, all of them having a physical meaning. We demonstrate that this concept serves as an explanation for the outcome of the LRD tests. The properties that are attributed to long range dependence in fact seem to be the result of non-stationarity of the large time scale process. We also show that the high frequency component is short range dependent, thus questioning the presence of long range dependence in video traffic.