Internet of Things II) System architecture considerations.

In my last article I explained, why we would need a regulated base to make the IoT to a real business and/or society success. Within this second chapter I want to go deeper into the technology – first the actually available hardware and second to show a vision about what may be the future in terms of hardware.

Theoretically all available hardware worldwide like mobile phones, industrial control equipment, car-integrated systems, personal computers … would qualify for the internet of things. In fact, all of these devices interact today through many sensors and actors with their surrounding world. In the developed countries the density of sensor information is enormous. Although I must say this is purely “virtual” as the hardware on which all this is running differs from device to device. It can be based on Intel, Motorola, Arm and other CPU architectures with different bus structures, and different operation systems like Unix (Android, Debian, Ubuntu…) Microsoft or Apple or other derivates and proprietary systems – unlimited numbers of possible variations. This variety is of course a positive structure once we talk about stability and safety. It is not really easy for hackers to target such a variation of systems. However, this leads us towards problems of standardization.

Next, the Network side of the structure. Networking is pretty standardized today. The limitations of addressing through IP4 are planned to be overcome by the introduction of IPV6. Which would allow that every computing device on the network can have its own telephone number. As of today we are still running IPV4 networks in parallel to IPV6 especially on the endpoint routing side most network connections are not able to run V6. On the other side it would make no sense to put future IoT standardization on V4. I consider this being in development and more or less solved within the next years.

The IoT is a combination of sensing and acting devices on the network. Considering the different hard and software – why not implementing a new hardware that controls the IoT information, ensures the safety layers and standardizes the flow on the network side. This would mean, that between the Network interface of today we would integrate a piece of hardware to standardize IoT applications. The functionalities would be encryption, Authority layer, standardized presentation of sensor and actor information, activation of streams and so on. With older devices we could just plug inbetween the network line of the appliance. However, for each hardware an individual driver is necessary to convert the own data presentation to the new hardware. With future developments I consider just that this hardware would be included inside the device. Because of the high scale of application, the price for this kind of hardware would become insignificant.

At this point of vision, the question about data presentation comes up. All the possible information like temperature, speed, acceleration, moisture, media stream, pressure, frequency, strain etc. need to be presented with all background information like geotags, data precision, timestamps, refreshing intervals in standardized formats. There might even be data sets not present today, but we need to make sure that the future data can be made compatible, too. An answer might be dynamic objects programming. Advantage is that there is a basic growing set of well-known objects directly useable in the application. But enough openness to key in own sets – that may get standardized in the future once the feasibility is proven.

My personal conclusion is the feasibility of such a future system. Possible businesses are not countable. Services for agriculture, travel, infrastructure, civil engineering, and more…. Worth to have a closer look into during the next years.

The next chapter will be devoted to the future options. 

Bernhard Rau, Radolfzell 20.2.2017