Databases designed for embedded world

Article By : Max Maxfield

Here's a relational database targeted at embedded systems and intelligent mobile devices.

About a month or so ago, I had a very interesting discussion with Sasan Montaseri, the founder of ITTIA, which develops, manufactures, licenses, and supports data management software products and provides related services.

We started off by considering the Internet of Things (IoT) and bouncing ideas back and forth as to what actually comprises the "Things" part of the IoT. My knee-jerk reaction is to think of small sensor and actuator modules at the very edge of the system, but Sasan introduced me to a new perspective by asking me to consider an autonomous car of the future.

The example he chose was that of a self-driving, battery-powered taxi. In addition to being charged while at home base, it also makes sense that such a vehicle would have solar panels on the roof to help maintain as high a charge as possible.

The next step was to consider what sort of data might prove useful to the owner of such a vehicle. The sort of things one might wish to monitor would certainly include the charge and discharge rates associated with the battery and the current draw versus the electric motor's RPM. There would doubtless also be numerous other sensors monitoring things like temperature and tire pressure — possibly also sounds and vibrations.

[EETI 2016JUN27 blog1](Source: ITTIA)

Obviously, this is going to generate a lot of data over time. We tend to think of data as being immediately handed off into the cloud, but you can't always guarantee a good wireless connection from a moving vehicle, so you will need the ability to store at least some amount of data locally and then sync up to the cloud whenever it's possible to do so.

Actually, even when your data is backed up to the cloud, it makes sense to keep as much data as possible locally so as to facilitate quick and easy access, especially if that access ends up being required in a remote, non-connected location (see also this video).

And it's not sufficient to simply store the data in its raw form — in order to reap the maximum benefit from having such data, it needs to be stored in a relational database so that a technician, for example, can quickly and easily access the battery charge/discharge data and compare it to the engine's RPM and temperature over the course of past days, weeks, months, and even years. It may be, for example, that the charge and discharge rates vary with the seasons due to environmental conditions like temperature, pressure, and humidity (Ooh, there's some more data we need to sense and store).

The reason we ended up talking about databases is that I am a little "fluffy around the edges" in this area. If you had asked me to name as many databases as I could prior to my conversation with Sasan, I'm afraid my response would have been limited to Microsoft Access and Oracle. Having said this, in my own defense, I did learn quite a lot of useful information in Nine Algorithms That Changed the Future by John MacCormick (see my review).

You can only imagine my surprise to discover that ITTIA has its own database — the ITTIA DB SQL. This little scamp is a relational database targeted at embedded systems and intelligent mobile devices. In addition to running on modern operating systems such as Windows, Linux, INTEGRITY, QNX, ThreadX, Nucleus, μcOS-II, μcOS-III, and VxWorks, ITTIA DB SQL can even work with custom file systems and without an operating system.

I recently attended the Embedded Vision Summit and if there's one thing I learned there it's that "data is king." I don't think that most of us realize just how important it is to gather as much data as we can, even if we're not exactly sure what we're going to use it for.

In the case of the data gathered by autonomous cars as described above, it may be that performing data mining using deep neural networks (DNNs) could reveal unexpected relationships and facilitate predictive modelling of potential problems. It may also be that unusual data patterns observed in one vehicle can be used to predict problems in another.

The older and more knowledgeable I get, the more I realize how little I know as to what the future holds, but I think it's going to be a very interesting journey (part of which I expect to take in an autonomous car LOL).

Leave a comment