Analytics

The IOT Continuum – How do you see it?

28th Oct `16, 05:03 PM in Analytics

Over the past months I have read a great deal about the Internet of Things (IoT) and how…

mark_torr
Mark Torr Contributor
Follow

Over the past months I have read a great deal about the Internet of Things (IoT) and how companies are moving forwards with it.

Looking across all the stories on the web it is clear that there is an IoT Continuum when you look at the different types of projects out there which many vendors are talking about.

This post shares some phases which I have summarized. I am interested to know if you think there are intermediate phases or even phases before or after these. I would also love to hear if you would alter the names and the focus!

Introducing the IOT Continuum

There are three things I see happening today with a fourth one maybe being the future.  Certain projects will contain multiple phases of the continuum over time with each building on the previous one and each phase might be a milestone in a much bigger goal.

IOT continuum

You might have ideas for different names for these and I would be very open to hearing your thoughts on that. The main question I want you to think about, after reading the phase descriptions below, is if you see different phases I have missed altogether? If so please share your thoughts and pointers to the stories that support them as I am eager to learn.

The IOT Continuum – Monitor

The monitor phase of the continuum seems to be the start point for companies who are getting going with IoT. Examples could be companies who putting sensors on existing “things” to gain insight they did not have before or it could be people getting started with a more ambitious project deploying new “things” with sensors. Nonetheless the aim of monitor is for organizations to:

  1. Get sensors onto  “things”, or deploy “things” with sensors, such that they have data to capture. Sometimes it could just be as simple as recognizing that they already have sensors they were not tapping into.
  2. Get an infrastructure into place to allow them to capture and manage that data. This could be a real-time streaming and “hot” infrastructure, it could also be a data capture and “colder” storage infrastructure or it could be a mixture.
  3. Get a reporting front end into place that can expose that data in real-time, if required, or in a way that enables actionable and timely insights.

In this phase of the IoT Continuum the focus is really on surfacing insight so someone can see what is happening. For example – how is oil flow through a pipe, where are specific vehicles in the fleet etc. This is an important first step as it can highlight areas to focus on.

The value here is providing awareness so people can start to make better decisions based on a data view they did not previously have access to.

From a personal perspective this is where most consumer wearables such as GPS tracking watches are today. I use a Garmin running watch and it captures huge amounts of data which I can display in a dashboard to help me make informed decisions about how my training is going and if I need to train more or less. Unfortunately they do not do much else yet but I see signs some vendors are moving towards the Advise phase.

The IoT Continuum -Advise

In the Monitor phase of the IoT Continuum the focus is very much on capturing data and displaying the results. The Advise phase takes it a little bit further by starting to alert and suggest based on the data which is being captured. I put alert in here as it is a trigger (or advice) to suggest someone should take a closer look but it could equally be in the monitor phase too.   The bigger part of this is when you start to apply analytics to proactively advise on upcoming issues based on models built over time. So the aim of the Advise phase is for organizations to:

  1. Get an alerting infrastructure put into place that can monitor for threshold anomalies and then issue alerts. Alerts should be things that can be pushed in real-time to a dashboard, email, SMS or other places humans can receive them and then take action.
  2. Get an analytics infrastructure into place that can be deployed against either real-time “hot” streaming data or against stored “colder” data. This sounds simple but models degrade over time so that infrastructure needs to be constantly testing the quality of the models and pitting new models against the incumbents. Also the target of the analytics will need to continuously shift.

This for me is where I see applications of analytics to things such as streams of telemetry data, in conjunction with stored colder data, to be able to do things like identify potential brake failures in cars before they will happen or identify assets that need replacing on oil rigs before they will fail so they can be replaced at a time when cost implication can be kept low.  Such technology can also be deployed to suggest changes to routes of vehicles to optimize how they get from A to B or to predict demand so that assets can be moved or sourced appropriately.

The value here is that organizations can not only see what is happening right now (as they can in the monitor phase) but they start to look out into the future, using the hot data combined with the colder data, and can use analytics to provide advice helping decision makers get ahead of the curve as the systems can advise them of upcoming issues.

The IOT Continuum — Automate

Thus far the IoT Continuum has dealt mostly with the delivery of information to a human consumer. The challenge is that a human cannot possibly be involved in every decision organizations need to make. This is where the Automate phase comes in.

In the automate phase organizations need to:

  • Map out their business processes and identify places where there is a potential for automation. This means they need to map all places where decisions are taken today by humans or via some other method. Normally this will be targeted at a very specific part of the business so as not to try map the whole company.
  • Get an infrastructure into place that can manage, store and use business rules which can be integrated within those processes to automatically take decisions based on the data.  Of course there will need to be a way to track the decisions that were automated and to push edge cases to people as required.

Automate seems simple but it is not. In this phase we will be relying on the rules we have in place to determine how things execute. For example if a car suddenly appears in a totally new location it could be a sign it has been stolen or it could just be a new destination for the driver.

When you combine location with the driving style telemetry it is likely you can tell if it is the regular driver or someone else.  All of this could be combined together to help automate if the police should be informed directly or if an operator should call you to check all is ok.

The value here is that we can speed processes and possibly make them more repeatable and intelligent. The outcomes of that depend very much on the process involved.

The IOT Continuum — Adapt

Adapt is the holy grail. In this phase processes will adjust based on what is happening with the data and self learning algorithms will adjust. Here the intelligence of the machines will help constantly refine the business rules, analytical models and the processes to deliver the best outcomes based on the constraints the system has been provided. This could be as simple as one machine automatically adjusting another. An example could be that if load on the electricity network is high then a machine could instruct all connected air conditioning systems to cool less to enable the network to not have outages.

It could also be much more complex though. In the previous simplistic example the assumption is that we were at max power on the grid already. What if the grid could see everything consuming electricity coming online, realize there was a capacity issue coming, automatically request more output from the power providers, ramp up power output to meet demand or to the max, dial down the air conditioning in peoples comes during the day (or businesses at night) automatically to coincide with when a max load is predicted to be reached, see things normalizing as things go back offline, lower the power output from the power providers, see capacity returning and then re-return the air conditioners to the original state.

This would require standards in many layers, and a lot more, but these are the sorts of things that might be possible going forwards.

Extend this out and it is easy to see how IoT in the Adapt phase could be a significant game changer not only for businesses but also for consumers and the planet.

Conclusion

This is my first attempt to put together the IoT Continuum and stretch it out to a possible future.

I think most success stories out there today focus on Monitor, Advise and Automate as it pertains to the IoT Continuum as I have defined it.  I would love to hear of someone who thinks they have an example of Adapt in action. I am sure there are some examples even if they are on a smaller scale to the grand idea above.

Originally appeared on Mark Torr.

MORE FROM BIG DATA MADE SIMPLE