Research And Implementation Of The Mobile Collaborative Computing Platform Base On Three-tier Cloudlet Model

Along with the rapid development of the mobile Internet,it is much easier for mobile devices to get access to Internet and get information and resources.In the mean time,mobile applications obtain fast development with increasing number of mobile devices.These applications especially those made by computing-intensive and resource-intensive tasks bring new challenges for mobile device since it’s power,storage and processing capabilities is limited. The development of cloud computing give a way to take full advantage of portability of mobile device but also avoid the limits.With cloud computing, mobile device can take full use of rich resources and computing capabilities of cloud by migrating computing tasks to the cloud.This kind of calculation model is called Cloud-Cloudlet two-layer structure migration model.However, mobile network and wide area network are ofen unstable even inrerrupt caused by reasons such as hardware failer, nature disaster or nerwork attack which would make computing migration unavailable.To solve these problems, Cloud-Cloudlet-Client migration model add cloudlet which is within the same LAN with mobile device between cloud and mobile.Cloud and cloudlet have the same computing capabilities for mobile devices. So mobile device can achieve computing migration even when the connection with cloud is down.The advantages and disadvantages of migration model based on Cloud-Client structure and Cloud-Cloudlet-Client structure are analyzed in this paper.A computing migration platform base on Cloud-Cloudlet-Client migration model for the military battlefield and natural disasters intermittent which network is unsstable or down is presented and deployed. The communication between cloud cloudlet and mobile device are based on REST Web Service and HTTP communication protocol in this platform.Cloudlet discovery,computing migration and data synchronization are being developed and deployed. Image recognition applications, wireless sensor network application and message push services are also deployed in this platform. In the end,each module of the platform is being tested and the result shows that the platform is doing well in the military battlefield and natural disasters intermittent environment.