Internet of Business pulled up a smart-city stool to chew the data-enriched fat with Basho CTO Dave McCrory over issues encompassing the Internet of Things (IoT), data gravity the use of ‘time series’ data in business.
What does ‘data gravity’ mean?
McCrory: Imagine data in its entirety, as though it were a planet or an object with similar mass. When data accumulates and grows larger, there is a greater likelihood that additional services and applications will be drawn to it. This is the same concept as a planet’s gravitational pull, the larger the planet/body of mass, the greater the gravitational pull, which will attract other objects.
As applications and services get closer/rely more on a source of data (as they often do in the Internet of Things), they are drawn towards it in order to take advantage of the speed and performance advantages of being closer. Ultimately, wherever there’s data with high volumes of interaction there is a likelihood that you will see the effects of data gravity.
If you’d like to learn more about the theory you can check out my blog on the subject.
Why does IoT time series data matter?
McCrory: Time series data is integral to IoT applications and functions. It’s data that literally has a time components built into it, usually transmitted in regular intervals but not always. Time series data can be analysed and queried as part of a specific time range, or from start to finish to give often powerful insights into an organization’s data.
I think most have come around to the idea that IoT is really about data generated by the devices and the feedback to the devices because of that data. Countless tiny sensors constantly transmitting time series data to gateways and datacentres where it can be aggregated, analysed and normalized. The real value of IoT comes out of the insights that can only be extracted through analysis of data. A sensor on its own adds little value to a business.
How is time series data used?
McCrory: In terms of broad characteristics, time series is ideally suited for monitoring, real time analytics, infrastructure management applications, anomaly/fault detection and countless other uses.
To get more specific, time series would be a perfect fit if you’re a water company such as Temetra, wanting to keep a constant eye on the status and wellbeing of your infrastructure. Sensors could transmit pressure data constantly so that you know in real time if there’s a fault that needs repairing. Alternatively if you’re a farmer and all your animals wear GPS sensors, you’ll know immediately if an animal strays off property and you need to take action.
Will time series data fuel the IoT boom?
McCrory: Definitely. For me the fascinating aspect of IoT is that we’re seeing new and emerging business models flourish out of the possibilities the technology offers. Businesses are now only limited by their ability to innovate, it isn’t dependent on capital. I like to think that IoT has shattered a few glass ceilings, and in the next five years we’ll see more and more disruptors and (most likely) subsequent acquisitions of said disruptors. Time series data and the insights we can glean from it is certainly playing its part here.
It also goes without saying, but more broadly speaking IoT is all about scale. If there’s tens of billions of devices transmitting time series data, not only are you going to need a database solution that can process time series data, but it’s also going to need to be hugely scalable.
We’re finding that the aforementioned innovative IoT applications and business models are hungry for a NoSQL solution, as opposed to a traditional relational database. We’ve talked a lot about the insights we can distil from time series data, however if your back end can’t handle the sheer volume of data to begin with, you’re setting yourself up for failure.
What’s you’re takeaway message?
McCrory: Well now… I’d like to issue just a small challenge if I may.
You can actually test your IoT/time series knowledge with a quirky quiz we’ve put together here.