Five ways to avoid IoT integration nightmares
18 November 2016 | 0
We all love shiny new technology and there’s an eagerness to get onboard with the latest “big thing” rather than risk getting left behind. Right now, the Internet of Things (IoT) is the hot topic — the idea of connecting devices with an IP address and not just people. But before you jump on the bandwagon, you need to make sure that you have the infrastructure in place to get a return on your investment.
These five key IoT integration considerations are drawn from hard lessons learned by the earliest adopters of IoT integration. They are a sound foundation on which you can start your own IoT Integration journey.
1. Choose your backend system carefully
If you want a temperature sensor for your IT room or a movement tracking system for customers browsing in your shop, you need to know where the data is going to be stored. Cloud-based solutions currently on offer are Microsoft’s Azure IoT Hub, which allows users to spin a free 30-day evaluation copy to integrate their devices. All architecture requirements are provided, from edge devices to dashboard reports.
Another supplier is IBM’s Watson IoT Platform which allows users to build IoT applications and analyse data from any source.
When picking a provider, you should choose a solution that will be around for the life cycle of your IoT device so big names like Microsoft and IBM may be the best place to start.
“IoT device weaknesses are increasingly being discovered in IT security audits and throwing up all sorts of issues around data protection and the privacy of customer information”
2. Standardise where possible
Getting data from the IoT device into backend systems typically involves new technology, which is mostly open source. No matter what the vendor says, based on time in the market alone, it is unlikely to be “enterprise grade”, so ensure that it aligns as much as you can with your current core infrastructure skills.
If you are a Microsoft house and using .Net apps, then the Azure IoT hub is a natural choice. If you choose some technology based on Node.Js and JSON and don’t have the internal skills, it’s going to be seat-of-the-pants stuff. Where possible, standardise your approach with the vendor’s.
3. Mitigate the security risks
Are you sitting in a cool room? Is the air con keeping you nice and comfortable? Today devices are often hooked into central monitoring systems to save on costs and may well be sharing data with the Internet through your local network. That’s how the biggest, mass data breach to date was carried out, with hackers tapping into an air con unit of Target and accessing its corporate LAN. From there they accessed the EPoS and stole credit card details for months on end.
If you don’t know the IoT technology you are putting in place, in terms of what data it transfers, to whom and what systems, you need to completely lock it right down from the get go. IoT device weaknesses are increasingly being discovered in IT security audits and throwing up all sorts of issues around data protection and the privacy of customer information.
4. Control the flow of data
Data centres of the world are the recipients of petabytes of data streaming from IoT devices. People are storing everything from Twitter feeds to customer transactions, and data processing will be the largest single segment in the data centres workload in the future, with PWC predicting a total of $162 billion (€152 billion) in 2019 in associated revenue. Imagine your office room temperature again. I’d guess it is fairly constant. So why would you stream temperature data every minute of every day back to a cloud server? Choose edge devices and gateways which use threshold alerts that only fire when temperatures fall outside acceptable ranges. This results in only a fraction of the data being sent to a cloud solution.
5. Extract value from the data
With all this data flowing around, how do I make it pay for itself? HVAC systems cool buildings more when prices of electricity are cheaper. That makes 100% sense but in many offices, this is probably not feasible. Data on office hours and room usage levels, for example, should be monitored to inform the heating levels. So the total cost of the devices and cloud platform would need to be lower than the savings on heating costs to make the payback. Without the data, you cannot make the business case for change. The more complex the data, the more you will need data skills beyond the Excel-jockey level. You will need to understand the data you gather and have the right statistical and predictive analytics skillsets available to make sense of it all.
IoT demands a complex balancing act. There are clear benefits to be gained from IoT integration such as reduced operating costs, but you should also consider the security risks and any potential internal skills gaps. An IoT project will also need cross functional considerations from finance, IT, compliance and management.
Only when you take all of this into consideration will you be able to make an informed decision and know whether the call of shiny new IoT technology is a blessing or a curse to your organisation. Follow the above key points and hard lessons learned closely and you will be able to build out your IoT strategy on a fit-for-purpose foundation.
Diarmaid Flynn is client service director with Ergo
For more information, see ergogroup.ie/blog