A Holistic Approach to a Context aware IoT Ecosystem with Adaptive Ubiquitous Middleware

Abstract

Present-day devices are becoming increasingly smarter than their predecessors. From a newlinesimple passive light switch to an intelligent wristwatch, great strides have been made in newlinenetworking smart devices, creating an autonomous ecosystem, the so-called Internet of newlineThings (IoT). In an increasingly information-driven world, context-awareness supports newlinethe intended applications and their constituent devices, making them conscious of and newlineadaptive to the specific scenario in real-time. Moreover, heterogeneous devices in the newlineIoT ecosystem peruse disparate data formats and semantics, leading to interoperability newlineand information sharing challenges. Based on applications, devices, and individuals newlineinvolved in an ecosystem, the nature of the network formed varies. Therefore it is necessary newlineto create an intelligent ecosystem where clients, data sources, smart objects, and newlineservices can coexist and interact among themselves. We present AUM-IoT, Adaptive newlineUbiquitous Middleware for context-aware IoT ecosystems, that considers the situational newlinecontext of applications, devices, or people and the contexts of the IoT network formed newlineand accordingly adapts the behavior of the ecosystem. To develop and implement such an effective and robust context-aware IoT ecosystem, we need to identify the fundamental building blocks required. In reviewing the literature, we found that a context-aware ecosystem should, at a minimum, require the following components: context modeling, context organization, and context-aware middleware. Context modeling is a core feature that facilitates interoperability and information sharing among applications. Although generic context models exist, they do not consider pertinent dimensions of context to provide a generic vocabulary, and therefore, the generalization of situations commonly encountered in the IoT environment is not possible in these models. Furthermore, most existing context models lack considering newlinethe state transition of the context and an endpoint criterion for the event to address IoT ecosystem ...