Modeling And Performance Analysis Of Fog Network Based On Stochastic Fliud Model

Fog network is an extension of cloud computing to edge computing.It can perform computing hierarchically and regionally according to needs,so as to solve possible network congestion and other phenomena.Fog network has multiple central nodes that process data,called fog nodes.As fog nodes are closer to end users,fog networks are considered to be a key enabling technology for future networks.Fog nodes have buffers with limited capacity.Due to the wide distribution of fog nodes,the energy consumption of fog network will be a big problem.However,because the energy collection process is unstable,and the arriving data is also highly bursty,if the capacity of the node buffer is too small,data overflow or even data loss may occur in the node buffer.If the buffer capacity is too large,may reduce the utilization of buffer devices,so it is crucial to design techniques that can improve the efficiency of these nodes.In this paper,based on the Stochastic Fluid Model(SFM)framework,the fog network nodes are modeled by SFM,and the performance of the node buffer is analyzed.The fog network node is modeled,and the energy harvesting process and the data arrival process are modeled as a multi-state continuous random process.The Matrix-Analysis Method is used to perform a stationary analysis of the fog node performance,and the data occupancy level in the data buffer in the fog node is obtained.and the joint distribution of the energy occupancy level in the energy buffer,as well as the exact algorithm for the probability of data overflow in the joint buffer and buffer idleness.Through numerical analysis,the feasibility of the performance indicators proposed in this paper is verified,and the system is analyzed.The effect of parameters on performance indicators.Since the instantaneous dynamics of the fog node system is also very important to the stability of the system,this paper controls and analyzes the data processing of the fog node buffer based on the multi-layer two-dimensional stochastic fluid model(ML-2DSFM).Different from the traditional SFM and two-dimensional stochastic fluid model(2D-SFM),the reward process in ML-2DSFM depends not only on the background process but also on the level process.Secondly,the Matrix-Analysis Method and the Level Crossing Arguments are used to derive the Laplace-Stieltjes Transform(LST)of the reward function at different time intervals.Based on ML-2DSFM,various time-dependent performance measures are converted into integrals of the reward process,and three First Passage Time(FPT)of the fog node buffer is obtained,using this performance evaluation technology to analyze the performance of the fog node buffer.Finally,the validity of the relevant theory is verified by numerical simulation.The theoretical results in this paper have practical significance for the design and management of fog network nodes.