The technological domain of the internet of things (IoT) embraces several developments, as disjointed as they are numerous. As the definition itself is still under heavy discussion, it is quite difficult, even tricky, to set boundaries, in order to determine clearly which technologies are within its range. Considering, for the sake of brevity, that IoT is built by "interconnected smart objects", we can orientate our interest more towards communication technologies, developing the way this connection is established, or else consider the "smart object" perspective, in which for instance, developments related to energy harvesting and conservation, as well as the miniaturisation of printed circuits, and inclusion of transistors into commonly used materials such as plastic, wood or metals are of central importance.
-
Zero-Entropy systems (energy harvesting, energy conservation, energy usage): Energy will be a major technological challenge in the next five to 10 years, and research must be conducted in order to develop systems that are able to harvest energy from the environment and not waste any under operation.
-
Scalability: IoT will be composed of trillions of devices. While it is unlikely that all devices will be connected in a mesh, but rather organized in hierarchical sub domains, the number of interconnected object will outnumber by several orders of magnitude the current internet.
-
Security and privacy: The issue of having sufficient security on devices with limited capabilities has to be addressed and solved convincingly. As well, technological architectures preserving the respect of privacy have to be developed and used as a basis for any future development.
-
Communication mechanisms: The situation in today's IoT domain resembles to the one at the very beginning of the Internet: several communication mechanisms were used, and only the convergence on a particular reference model allowed the development of the web.
-
Integration of smart components into non-standard substrates: The use of non-silicon substrates for developing smart components will reduce the dependency to silicon with all related problems, like packaging and recycling.
The internet as we know today is based on a few, very simple and very meaningful principles. One of those is the "end-to-end" principle: keeping the technologies in the network very simple and dealing with complexity at the end points only, allowed the Internet architecture to be very scalable (Carpenter, 1996). With regards to the IoT domain, there might be a different point of view. First of all, it has to be considered up to what extent IP technology will be used. While many technologists believe that IP will finally be on each and every smart device (Ipso, 2011), there are two particular cases which show the likeliness that different solutions are necessary. Firstly, real-time devices, such as braking systems in cars, which cannot be based on best-effort, connectionless, unreliable protocol (as the IP is, by definition). Secondly, tiny, extremely cheap devices, (such as passive RFID tags) which may be stateless and therefore cannot use complex protocols such as IP. Moreover, it is questionable if the end-to-end principle can (and will) be used in the IoT domain. As the end points of IoT can be extremely simple (as a temperature sensor), even if they will be able to use the IP protocol it is unlikely that they will be able to deal with complexity. Moreover, smart devices do not necessarily need to speak the same language: a medical device such as a nanorobot used to fight cancer cells in the human body has totally different needs than those of a smart fabric needing to communicate its characteristics to a washing machine. Therefore, it is likely that, at some layer, there will be bridges between systems; and these bridges (or gateways) might be considered the end-to-end points between communicating entities. In other words, between two different objects communicating, the communication path may be broken into different sections (object-to-gateway, gateway-to-gateway, gateway-to-object). As this is considered a "curse" in today's internet, and is likely to be a highly controversial topic, there is a strong need to further investigate this matter, and to come up with a commonly accepted set of founding principles.