Research And Implementation Of Service Support And Management Technology In Internet Of Things(IoT)

In recent years, with the further development of sensing equipment and network technology,the Internet of Things shows the enormous potential and space. Many applications have begun toprovide convenient and efficient service in smart home, smart medical, smart agriculture, and so on.The Internet of Things applications have just started, they need more open service platforms tosupport the development, fusion or unified management of the applications. The lack of platformsrestricts the popularization and development of the Internet of Things applications.To explore and solve the above problem, a platform named USPIOT has been developed in thelab. The main work of this paper is to study the service support and management technology ofUSPIOT.The paper proposes a service support and management mechanism for USPIOT. Contraposingthe large visits of the platform, the mechanism adopts the two-stage buffer pool in the platformcushioning module. Meanwhile, the mechanism adopts the service library to achieve the quickloading of the new service in the platform. The running result shows that the mechanism caneffectively support the service of the platform and also can quickly load, discover and manage thenew service.The paper carries the development and function demonstration of USPIOT in the three aspectsof the basic functions, the loading of applications and the typical applications and also proposes averification system to load and test the new service. The test result demonstrates the effectiveness ofthe service support and management mechanism and the experimental system provided in this paperfor new service development, load and test.The paper designs the task allocation and scheduling module carefully, and proposes a taskallocation algorithm named TABGA (Task Allocation Based Genetic Algorithms) for thedecision-making module of task allocation. Comparing with the random allocation, the experimentresult shows that with the computing nodes increasing from2to3to5, the task execution timereduces7%,13%and17%averagely.