Towards a RISC Framework for Efficient Contextualization in IoT
Dimitrios Georgakopoulos, Ali Yavari, Prem Prakash Jayaraman and Rajiv Ranjan
Swinburne University, RMIT University, Swinburne University, Newcastle University

The Internet of Things (IoT) is a new internet evolution that involves connecting billions of internet-connected devices we refer to as IoT things. These devices can communicate directly and intelligently over the Internet, and generate a massive amount of data that needs to be by a variety of IoT applications. This paper focuses on the automatic contextualisation of IoT data, which also involves distilling information and knowledge from IoT aiming to simplify answering the following fundamental questions that often arise in IoT applications: Which data collected by IoT are relevant to myself and the IoT Things I care for? Related work around context management and contextualisation ranges from database techniques that involve query re-writing, to semantic web and rule-based context management approaches, to machine learning and data science-based solutions in mobile and ambient computing. All such existing approaches have two main aspects in common: They are highly incompatible and horribly inefficient from a scalability and performance perspective. In this paper, we discuss a new RISC Contextualisation Framework (RCF) we have developed, implemented key aspect off, and assesses its scalability. RCF provides fundamental contextualisation concepts that can be mapped to all existing contextualisation approaches for IoT data (and in this sense, it provides a common denominator that unifies the contextualisation space). RCF can be easily implemented as a cloud-based service, and provides better scalability and performance that any of the existing content management and contextualisation approach in the IoT space.