The Φ-stack for smart web of things

With the emergence of smart IoT, a key problem has becoming increasingly acute: the "Paradox of Classis Insecta", which in industry is called the fragmentation problem. We propose an architecture called Smart Web of Things (SWoT) to address this problem. The SWoT architecture extends the REST architectural style that has been validated by PC Web and mobile Web to the smart IoT arena. However, previous researches have shown that such extension is nontrivial. According to Vint Cerf, intelligent Web is heavy, difficult to be realized in resource constrained smart IoT devices. We propose a novel co-designed hardware-software stack, called Φ-Stack, to natively support Web and intelligence (e.g., pattern recognition and deep learning), with minimized cost, footprint, response time, and energy consumption. This hardware-software stack must be an open architecture, to encourage innovations and accommodate diversity inherent in the smart IoT field. Preliminary experiment results show that our approach is promising in extending the Web technology to resource constrained smart IoT environments and helpful in partially unifying APIs and interactions among smart IoT devices.