Design And Implementation Of Traffic Generator For Internet Of Things

The Internet of Things(IoT)has a complex structure and different types,and the traffic characteristics of different Internet of Things applications vary widely.In order to analyze the impact of complex IoT network environments and traffic on networks and other applications,it is increasingly important to analyze and predict the characteristics of IoT traffic.It is very difficult to obtain large-scale and complex type of application traffic data under various IoT scenarios.To complete IoT traffic prediction and model evaluation,it is very important to use the traffic simulation tool to generate IoT traffic.However,most of the current mainstream simulation tools have problems such as long simulation time,insufficient simulation scale,inconsistent traffic modeling,and inflexible network configuration.These simulation tools are often not designed for the Internet of Things and cannot meet the forecast of IoT traffic.The core work of this thesis is to design an Internet of Things Traffic Generator(NS-IoT)to solve the above problems.The paper first analyzes the system's functional requirements through background investigation and analysis of project requirements.Then,with the simulation scale and simulation time as the entry point,the requirements analysis,summary design and detailed design of the system are designed to improve the simulation efficiency,and the overall software structure and business process of the system are designed and implemented.NS-IoT uses MapReduce,a distributed computing model in Hadoop,to increase the scale of simulation by multi-device co-simulation to improve system resource utilization.At the same time,this paper proposes a Time-Scaling Data Generation(TSDG)algorithm,which solves the simulation time problem through non-stop operation,continuous packet transmission,pseudo-parallel startup,and transmission process decomposition.In this paper,the TSDG algorithm is applied to the traffic generator,and combined with the MapReduce distributed computing model,by compressing the simulation time and improving the system resource utilization,the simulation scale can be improved while the simulation time can be effectively reduced.At the same time,the system provides a configurable operation interface for the user to customize the configuration and simulate the traffic model data of multiple types of traffic distribution,multi-environment network delay,and multi-scene node configuration in accordance with the actual IoT scenario.Based on NS-3 development,the simulator uses its node code execution capabilities to run real software on the analog nodes,allowing users to generate repeatable simulation experiments in discrete event network simulations.This paper also introduces the implemented operation interface,and tests and evaluates the system in detail according to the test process of functional test,performance test,recovery test and compatibility test.In order to reflect the improvement of simulation efficiency of this system,this paper compares the simulation data of NS-IoT with the native NS-3 system and cloud computing simulation platform CloudSim.