The Management Model and Implementation for Ubiquitous Device of Internet of Things Based on Cloud
|School||Beijing University of Technology|
|Course||Computer Science and Technology|
|Keywords||Plug-and-Play Sharable gateway Internet of Things Cache optimization Scheduling Periodic task model|
In recent years, typical applications for the Internet of Things has been graduallydeveloped with the progress of the Internet of Things technology. However, the issuesof device heterogeneity is becoming more and more prominent. There must be auniform standard for interface and data model. We need to unify technical standardsand form a management mechanism as soon as possible. Emerging cloud systems areexpected to present tremendous, intelligent, flexible and scalable services penetratedinto people’s daily life especially when combined with Internet of Things (IoT) orCyber Physical systems (CPS). The capacity of device is different in the Internet ofThings. We need to take different measures for devices which have differentprocessing capabilities to drive them to work. In response to these problems, Thispaper proposes a gateway architecture by sticking to which all kinds of serviceproviders can share the common data-path to the cloud without providing anyproprietary data-processing protocols/algorithms at the shared gateway. In themeantime, the gateway is Plug-and-Play.For the unified management issues of ubiquitous device, a ubiquitous devicemodel was proposed through traditional computer I/O system management idea.Taking into account a combination with the cloud system, a driven framework underthe environment of Internet of Things is put forward. The driven framework includesdevice discovery mechanism, driver deployment and unified application interface.According to autonomous management and Plug-and-Play problem, there is a wholearchitecture which includes two modules: Gateway Main Manager and Gateway Audit,Gateway Main Manager consists of Driver Manager, Cache Manager, SchedulingManager and Application Manager. In the case of the gateway resource-constrained,the cache and the scheduling are optimized in this paper so as to improve the systemperformance. In order to increase throughput of system, this paper uses the periodicapplication working model, event-driven cache management model and schedulingmanagement model of mixed event-driven waveform and regular waveform. Based onthe driven framework and the whole architecture of gateway, the prototype system isrealized by using simulation method, and design experiment plan for analyzing thecorrelations between the performance of cache and scheduling and its influencingfactors. According to the experimental results, this paper tests the performance ofcache and scheduling through selecting the best value. On the basis of the above research, this paper realizes automated loading driverof physiological data measuring equipment, transparent processing of datainformation and uniform application interface for the service.