Figuring out challenges in IoT at the IoTSC2014
The opening keynote at the #iotsc2014 today was presented by Prof. Carsten Bormann from the University of Bremen, Germany and IETF. It was kinda lengthy, however interesting – so interesting, that we never noticed how the time flew. He presented challenges in IoT nowadays.
The greatest challenge in IoT nowadays is developing products which will operate in really constrained networks. The situation, since ever communications of any kind exist, changed a lot. In the beginning, it was all about connecting places. Every city had a telegraph. As the technology advanced, it was all about connecting people (Nokia, you alive, huh?) and the present mission communications face is to connect things.
We agreed to divide things in three separate classes:
- Class 0 – devices too powerless to run securely on the Internet.
- Class 1 – devices processing between 10kb of data and 100 kb of code.
- Class 2 – devices processing between 50kb of data and 250 kb of code.
As we all know, you can achieve a lot with devices that fell under Class 0, however this is considered as heroic act. And we cannot build the Internet of Things based on heroic acts. They are too sporadic, too clumsy and there is no strict formula of doing things that way.
These classes are not clear-cut but we used them to structure the discussion and help avoiding talking at cross purposes.
The importance of IP (Integration Protocol, not the everyday thing) was firmly stressed out. More and more things are getting connected to each other and the Internet. The approach to integration matters. At the present moment, however, two major situations exist:
IP is too expensive for my microcontroller app (my hand-knitted protocol is better) vs. IP already works well as it is, I just go ahead and use it. This gives us a good glimpse on the urgent need for developing more efficient integration and cross-platform protocols. Integration must kick in as soon as possible and it must be unified.
Another challenge we discussed is the code complexity. In the era of servers, we forgot how to measure code’s efficiency. We need to recalibrate our code complexity measurement system. In the world of IoT, code is expensive stuff.
Class 1 devices are fine with processing sketches of 100kb each and class 2 are doing fine with 250kb. This, however will not remain forever. On the other side, packets our devices exchange are really expensive. Listening, in power constrained environment is the most expensive thing you can get nowadays. Multicast, well.. it does not work at all.
Professor Bormann has also shown us reports from Gartner (widely spread over the Internet in the past couple of months) that measure IoT to be in its very hype phase. We tried to distinguish between what is it now (IoT hype) and what will it be when things get real (what we called real IoT). It is based on the assumption that when a technology is in its hype phase, expectations regarding it are high and often false. Let us see what will happen when the IoT hype folds down and people start developing in real-world conditions:
IPv4 will be replaced by IPv6. I read a great paper on that subject on the train couple of days ago. This is happening already, probably in your town too.
Connected devices now communicate mostly through cloud platforms. When IPv6 kicks in, they will communicate directly, ruling out part of the gateways and cloud providers surrounding us.
Security is really a fragile factor of IoT these days, primarily because it is not yet too popular. When it is massively adopted, applications will be designed with security in mind. Currently, security layers are being added, usually, in the last phase of development a device, which makes it hard to implement and maintain it further, thus exposing its users to various threats.
Most Internet of Things things cost above $40 at the present. Technology is aimed to bringing costs down to below $5 (wow!). I want to witness this. Hopefully I will live by that moment.
Power consumption of Internet of Things devices is currently measured in Watts, because.. well, it is not that optimized. Days ahead will bring us devices that consume as less energy as it is measured in microwatts.
If you are interested taking a look at the entire lecture. Here go the slides.
Currently, it is the Internet Engineering Task Force (IETF) that is behind challenging all these factors, which will surely help end consumers adopt the IoT idea faster. They have proven to do well and they have excellent engineers there – all we need to give them is time and support. The last is a must – support researches in this area, preach it to your friends and family, make the world know that future is coming!