Abstract
Recently, Internet of Thing (IoT) systems enable an interconnection between systems, humans, and services to create an (autonomous) ecosystem of various computation-intensive things. Software architecture supports an effective modeling, specification, implementation, deployment, and maintenance of software-intensive things to engineer and operationalize IoT systems. In order to conceptualize and optimize the role of software architectures for IoTs, there is a dire need for research efforts to analyse the existing research and solutions to formulate the vision for futuristic research and development. In this research, we propose to empirically analyse and taxonomically classify the impacts of research on designing, architecting, and developing IoT-driven software systems. We have conducted a survey-based study of the existing research - investigating challenges, solutions and required futuristic efforts - on architecting IoT systems. The results of survey highlight that software architecture solutions support various research themes for IoT systems such as (i) cloud-based ecosystems, (ii) reference architectures, (ii) autonomous systems, and (iv) agent-based systems for IoT-based software. The results also indicate that any futuristic vision to architect IoT software should incorporate architectural processes, patterns, models and languages to support reusable, automated, and efficient development of IoTs. The proposed research documents structured and systemised knowledge about software architecture to develop IoT systems. Such knowledge can facilitate the researchers and developers to identify the key areas, understand the existing solution and their limitations to conceptualize and propose innovation solutions for existing and emerging challenges related to the development of IoT software.