Architecture Design Of Wireless Sensor Network Cloud Simulation Platform

With the increasing research interest on sensor network, the simulation technology, as an essential study, has also been developed rapidly. Wireless sensor network simulation tools widely used in recent years can be divided into two categories, one is the extension to traditional simulation software, and the other is dedicated to wireless sensor network protocol simulation. Whether the traditional or the new simulation software special for sensor networks simulation environment has certain limitations. As the focus and stages of development of all these software are quite different from each other, and realization of the diversity of their own, resulting in learning to use and deploy a simulation platform to be a hard task.So, the performance, scalability, compatibility and ease of deployment and use of simulation software has become a key issue in the field of wireless sensor study.In this thesis, a wireless sensor network cloud simulation platform has been proposed based on a detailed analysis of the advantages and disadvantages for the existing simulation software. The aim is to fundamentally solve the problems in usability, compatibility and easy manageability on the deployment of existing software. User interaction, simulation and data storage, are three kinds of sub-clouds within the cloud simulation platform. They are independent with each other and work together through the loose coupling. For users of the system, they just use the interactive interface to submit and query simulation tasks. The installation and implementation, management, data exchange among sub-clouds, simulation task scheduling and all the other details have been abstracted out, and are transparent to the user. Thus it greatly reduces the difficulty and the time for the user to carry out a simulation, and enables them to focus on the research protocol or application itself. Within the system, virtualization technology is used to integrate different simulation software, that is to construct virtual machines corresponding to different simulation environments and performance requirements. By monitoring task queue and resource usage, simulation tasks and resources are automatically managed. Thereby it will increase the system availability, scalability and compatibility, and full use of limited resources, improving performance while reducing power consumption. Finally, this thesis provides the function to query, download and dynamic demonstrate the simulation results. The storage of custom protocol and simulation results ill help the further statistical analysis on them, and rich the protocol resource libraries and enhance the supporting tools also.