Leveraging Big Sensed Data over Ubiquitous Networks

点击翻译按钮获取中文摘要

Our everyday lives are growing ever more dependent on the assimilation of information. In the advent of our digital world, a monumental amount of data is collected to aid our activities, prevent and mitigate dangers, and synchronize complex systems (e.g., transportation). However, the increasing adoption of sensing systems has created unprecedented amounts of data that traverse an already strained communication network. As we plan for better connectivity, abundance in data collection and scalable sensing, a growing dependence on the Internet to provide the communication backbone is unrealistic. More importantly, the sheer volume of data and the heterogeneity of its sources with varying quality, are pushing us into the era of Big Sensed Data (BSD). As such, building services on top of such data impedes decision making and real-time processing of information. I will build on my research in dynamic resource management to build a platform for BSD in the Internet of Things (IoT) era. Specifically, I will devise novel data collection, cleaning, pruning and management schemes over heterogeneous IoT systems, to feed high-quality data to IoT services. Our paradigm will address information services from three dimensions. First, data collected over a multiplicity of heterogeneous sensing systems will be pruned at the source, and evaluated under Quality of Resource metrics to dictate their viability relative to all other sources. Thus, a uniform metric will dictate the Quality of Data (QoD) before traversing (and loading) the network infrastructure. Second, I will devise Adaptive feedback protocols to enable load-balanced management of sensed data sources, reduce the data production rate of superfluous/inferior sensors, and enable faster access to high-quality information. Third, I will adopt localized data management by introducing Edge processing that enables real-time fusion of QoD-measured data, yielding our Edge Fusion BSD (EF-BSD) framework. High quality data will then be pushed onto a promising candidate for the next-generation Internet, namely Information Centric Networks (ICN), which inherently handle content dissemination. Training of highly qualified personnel in this program will include experience in advanced Big Data management and fusion techniques over next generation IoT technologies. They will be trained in real-time data fusion, dynamic resource management, ICN protocol design, managing heterogeneous networks, and caching techniques. I expect that two PhD, two Master's and two undergraduate students will receive training in this research program. There is a strong demand for HQP in the information technology and telecommunications sectors and their future employment will accelerate disseminating next-generation communications technology to Canadian industry, and present a competitive edge for Canada in global IoT initiatives for services and systems.