Open AccessAnalysis of the productivity of wireless networks with an entire length of data packages
Author(s) -
В. И. Кравченко,
V. V. Skrypnik,
O. I. Holubenko
Publication year - 2020
Publication title -
zv'âzok
Language(s) - English
Resource type - Journals
ISSN - 2412-9070
DOI - 10.31673/2412-9070.2020.045760
Subject(s) - wireless network , computer science , wireless , computer network , reliability (semiconductor) , key (lock) , point process , entropy (arrow of time) , productivity , principle of maximum entropy , distributed computing , telecommunications , mathematics , statistics , computer security , power (physics) , physics , quantum mechanics , artificial intelligence , economics , macroeconomics
He analysis of productivity of specialized wireless networks in which the number of operating stations changes by the random law is carried out. In this case, this number of stations can not be reliably controlled in the process of transmitting information. To obtain asymptotic characteristics of the transmission duration, it is proposed to use the information entropy parameters of the model distributions. The review of the basic and additional parameters of efficiency from the point of view of their influence on functioning of a network in the given considered conditions is carried out. An assessment of the cross-correlation of key performance parameters was performed. Dedicated wireless networks use a variety of architectures, technologies and standards, so such networks are heterogeneous by definition. However, the basis of specialized wireless networks are usually IEEE 802 standards networks. Such networks have a decentralized control and management system. In this regard, the centralization of control control is associated with significant time, while ensuring the necessary reliability, mobility and security of such a center, especially in an emergency, it is almost impossible. Heterogeneous self-similar traffic circulates, it is necessary to apply nonparametric methods. As the lower threshold of productivity it is possible to receive any asymptotic comparative estimations, for example, information-entropic measures of the considered probability distributions.