Towards evolution of M2M into Internet of Things for analytics

Machine-to-Machine (M2M) refers to technologies that enable networked devices to exchange information and perform actions without human intervention. Advances in Radio Frequency Identification technology (RFID), wireless sensor networks (WSNs), embedded systems, wired and wireless networks and reduced cost of transferring bits have accelerated the growth of M2M systems. However, due to lack of standardization, the M2M market is highly fragmented, proprietary and lacks widespread deployment. On the other hand, the Internet of Things (IoT) has emerged as the pioneering paradigm for ubiquitous computing where billions and trillions of devices would be connected together to sense information and to take actions without human intervention. Therefore, the end goal of M2M and IoT remains identical and it is only appropriate that M2M steadily evolves into IoT. Furthermore, with so many interconnected devices, the data generated would be overwhelming. Having this in mind, we chart an evolutionary path for M2M systems towards IoT from analytics perspective. We discuss challenges in M2M systems for analytics and provide solutions and on the basis of these solutions, propose a gateway-based reference architecture for analytics in M2M to facilitate its evolution towards IoT. We also provide a prototype implementation of the reference architecture. This reference architecture consists of data aggregation, data cleaning and data transmission layer at M2M gateway and data comprehension and data analysis layer at M2M server.