Article Preview
Top1. Introduction
In contemporary times, the Internet of Things (IoT) has appeared as a new opportunity. Consequently, all devices such as smartphones, public services, transportation facilities and home appliances are considered as data creator devices. The term Internet of Things is largely used, but there seems to be no common definition or perception of what the IoT encompasses (Wortmann & Flüchter, 2015). We can attribute the background of this term to Auto-ID Labs at the Massachusetts Institute of Technology (MIT) on networked radio-frequency identification (RFID) infrastructures (Iera & Morabito, 2010). In (Nguyen, Astaloš, & Hluchý, 2016) authors refer IoT as the domain of devices linked to the Internet, by means of which the enormous amount of data is endlessly acquired, assembled and controlled. Therefore, these continuous communications between large amounts of heterogeneous objects characterize IoT as a disruptive technology that permits ubiquitous and pervasive computing platforms (Van Kranenburg, 2008). In (Tabrizi & Ibrahim, 2016) authors have perceived the IoT, as smart system model, using smart objects with perception, computation and communication abilities where diverse information is accumulated from the physical world and connect the physical objects identifying each other, the key components being Radio-Frequency Identification (RFID), sensor networks and Machine to Machine (M2M). Methods such as big data and data mining can be used to improve the efficiency of IoT and storage challenges of a large data volume and the transmission, analysis, and processing of the data volume on the IoT. A cloud-based framework represents the combination of components like development tools, database services and middleware, required for cloud computing, which helps in developing, deploying and managing cloud-based applications actively, accordingly making it an efficient paradigm for enormous scaling of dynamically allocated resources and their complex computing. Big Data Analytics (BDA) delivers data management solutions in the cloud architecture for storing, analyzing and processing a huge volume of data.
In (Cai, Xu, Jiang, and Vasilakos, 2017) authors have provided functional framework for cloud centric IoT applications and highlighted how voluminous heterogeneous data generated by massive amounts of distributed sensors in IoT can be acquired, integrated, stored and processed in cloud platforms. The different areas like Data Storage, Data Management, Data Disposing, Data Mining and Application Optimization for cloud centric IoT applications have been explored.
In (Song, Wang and Chen, 2013) also authors have proposed an integrated framework for handling voluminous heterogeneous sensor data on cloud platforms. In (Jiang, Shen, Chen, Li, and Jeong, 2015) a secure and scalable IoT storage system founded on revised secret sharing scheme supported by scalability, flexibility and reliability at both data and system levels is proposed. In (Kakanakov and Shopov, 2017) authors have presented an example sensor-cloud architecture in IoT and cloud technologies with special focus on data security. The architecture is based on multi-layer client-server model. The physical and virtual instances of sensors, gateways, application servers and data storage are separated. The introduction of virtualized sensor nodes as a requirement for increasing security, privacy, reliability and data protection is proposed. Authors in (Nguyen, Astaloš, & Hluchý, 2016) have discussed IoT as the world of devices connected to the Internet, by means of which enormous volume of data is endlessly collected, assembled and managed. Other processes like information retrieval, database systems, web monitoring etc. also produce raw data. Data Mining in such data resources of analysis to acquire practical results and/or knowledge is worth. Authors have paid attention towards large-scale data, data processing and data mining using machine learning techniques through technological experiences in the direction of high-performance computing (HPC), Apache Spark and GPU. Authors in (Dissanayake and Jayasena, 2018) have discussed a proper framework that can evaluate the big data in the internet of things in a more efficient way. In (Mahdavinejad, Rezvan, Barekatain, Adibi, Barnaghi, and Sheth, 2018) authors have provided the assessment of several machine learning approaches used to deal with the challenges offered by IoT data.