面向物联网传感器信息的数据分配策略

A Data Allocation Strategy for Sensor Information of Internet of Things

物联网作为国内外新兴的热门技术,正在深刻地影响着人们的生产生活,它在带来诸多好处的同时也给信息存储领域带来挑战.物联网信息存储中心需要根据其数据特性结合分布式实时数据库信息存储管理的优点,设计与之相适应的数据存储方案,而数据分配策略作为数据存储方案的关键技术是研究的重点.根据物联网传感器信息的海量性、时空相关性、访问失衡性和连续变化性,需要一种基于时域的数据分配模型与之相适应,以此设计出基于自适应时域负载反馈的动态数据分配策略(adaptive time domain data allocation,ATDA).策略根据数据特征,将静态数据分配问题归约成简单线性规划问题,同时采用自适应时域对负载信息进行反馈,最后设置动态负载门限函数实现数据的动态分配.实验表明,该策略与同类Random、Bubba算法相比,在系统短时域负载均衡(LBST)、系统数据迁移量(DM)方面具有更好的性能.

The Internet of Things(IoT)is one of the most popular research issues,and it is changing the way of people's work and life.IoT brings a lot of benefits as well as challenges to information storage community.In order to design a suitable data storage scheme,IoT storage center needs to combine the characteristics of its information with the advantages of distributed real-time database, and data distribution strategy as a key technology for data storage scheme is the focus of the study. The features of sensor information of IoT include magnanimity of data,correlation of temporal and spatial,imbalance between the update and the inquiry and continuous changes of the update. According to these features,a data allocation model based on time-domain is proposed,and a data allocation strategy based on adaptive time-domain load feedback is designed.In data allocation strategy,static data allocation problem is reduced to a simple linear programming problem,and load information is feedback dynamic in adaptive time-domain.To achieve dynamic data allocation, dynamic loading threshold function is set.Experimental results show that,compared with algorithm of Random and Bubba,our strategy performs better in load balanced in short time period(LBST)and the number of data migration in overall time(DM).