Service Data Management In Intergrated Sensing Network

Combining organically modern communications and sensor networking, the Integrated Sensing Network (ISN) provides a feasible scheme for the development and application of Internet and next generation network. It is necessary to establish a platform with efficient data access and processing, so that can realize the scientific management of data for each tier in ISN, owing to the features of decentralization, diversification, and multi-speed for the data of ISN. This paper focuses mainly on the ISN service data management system in management layout, modeling analyses and demonstration.Firstly, the function of the service data management framework of ISN is introduced and the system is designed, including data processing, data storage and user query. Then the performance index for service data management system is proposed. Successively the model of single node service data management system is established, and the multi-level multi-priority queuing algorithm based on the M/GI/1 is put forward. Results show that with the increase of arrival rate, the delay of all priority and the ratio of the arrival rate and throughout increases, and with the priority increasing, the delay decreases. Next the model of multi-node service data management system is established and the dynamic feedback load balancing algorithm of ISN is proposed based on the consistent hash algorithm, including static initialization and dynamic data allocation method. Then the load balancing rate is analyzed respectively according to the ISN load balancing algorithm and consistent hash algorithm. Results show that the average load rate variance of the former is 5.37×10-6 and the latter is8.67×10-5. The performance of system load balance is improved significantly with the result of the former decreasing an order of magnitude. Finally, the distributed ISN service data management system with ten nodes based on Internet is set up to realize the function of service data real-time monitoring, real-time or stationary synchronization and real-time query. Successively the load balancing rate is tested. Results show that the dynamic feedback load balancing algorithm ensures the database load balancing rate is all below 0.10 which improves the performance of ISN's load balancing.