Rightfully, the practical question of integrating a myriad of (small, low-power) Internet of Things (IoT) devices with (large, scalable, but remote) cloud computing facilities is being addressed by many researchers.
In this publication, the reader will encounter a good but not complete selection of useful but highly superficial vignettes of almost exclusively open-source software packages trying to tackle this problem. The authors have grouped the 27 investigated packages into four groups: IoT middleware technologies (6 entries); cloud platforms (11 entries, almost all HADOOP-based); cloud infrastructure (3 entries); and IoT integration (10 entries). Convenient tabular comparisons of the packages per group are complemented with three (artificial) case studies (e-health, smart home, and smart city), which will help the reader choose the right packages for her project.
At a detailed level, though, the paper is imprecise. Microsoft Azure is not a software as a service (SaaS); one can’t put and compare the constrained application protocol (CoAP) with a platform like LinkSmart in the same category (IoT middleware); and the cloud and IoT integration “architecture” depicted in figure 1 is everything else but an architecture. The paper also tends to overgeneralize without sufficient support: it tries “to provide an overview of existing software,” but practically all commercial software platforms are unaccounted for; and the ability to reconfigure IoT devices through an IoT platform is really not a “novel paradigm.” Additionally, at places the English is plainly wrong, leading to incomprehensible or nonsensical phrases.
If you discount these shortcomings and also disregard the unnecessary (too short in any case) “Challenges and open research issues” and “Conclusion” sections, the compilation and comparison part of the paper may provide a starting point for anyone looking for open-source software for her IoT-cum-cloud project.