5 Things to Consider in the Age of Machine Data
“Welcome my son, welcome to the machine.
Where have you been? It’s alright we know where you’ve been.”
If you attended SPF Americas in Chicago or SPF EMEA in Paris last month, you heard about our Industry 4.0 vision of a connected enterprise, and what it means for EHS, risk, sustainability and operations professionals.
Industry 4.0 accelerates the rise of the “smart factory” where cyber-physical systems communicate with each other and with humans through the Internet of Things (IoT). One immediate consequence is the big increase in machine-generated data.
There will still be data entered by humans through computers and mobile devices. But machine-generated data will constitute a much greater share through sensors, beacons, drones and connected assets all linked through the IoT and a smart platform that processes data to deliver insights and intelligence.
Is your organization ready to face this massive onslaught of machine data? This is a critical question because many of your competitors may have already taken steps to procure a competitive edge through Industry 4.0 innovations.
Here are five things that you should consider to help you prepare for the new reality of the prevalence of machine data.
1) Greater Data Security Risks to Mitigate
This item is obvious and there is nothing new about the need to ensure data security. But it’s important to constantly remember the following: Each connected device creates an additional, potential vulnerability.
The increase in systems and devices connected through the IoT creates more points of entry for hackers into your network and platform. If you were already taking effective steps to reduce risks of cyber attacks, be aware that in the age of machine data, cyber risks will rise exponentially with each new connected asset or equipment. As such, be sure that your controls are updated for the new IT landscape brought by Industry 4.0.
2) Automating Data Quality Checks and Validation
Here’s a law of statistical probability: An increase in data records also increases the probability that at least one is an anomaly.
In the age of machine data, it will be common to have a connected machine or equipment send data about its status or performance multiple times per minute, if not every second. This will greatly increase the quantity of data, thus increasing the risk of data inaccuracy.
To address this, be sure to implement automated data quality checks in your platform to detect and flag any potential anomaly. For example, if a pressure reading on a pipe taken every second usually shows a variation range of -20% to +20%, and a single reading shows a variation of -60% while previous and subsequent readings are within the normal range, that may indicate that the reading was inaccurate, and you may want someone to take a closer look (e.g. was something wrong with the pressure sensor or with the processing of the data?).
3) Creating an Inventory of All Machine Data Sources
Each plant, facility, refinery or factory includes many assets, machinery and equipment. Add to this all the drones and wearables that may be used. In addition, you may have many sites spread around the world. Suddenly you end up with a multitude of devices connected to the IoT.
This creates a risk and an opportunity. The risk is that each device becomes a potential point of vulnerability for cyber attacks. The opportunity is that each device can provide useful data that can help you generate valuable insights.
To manage all the risks and opportunities, create and maintain an inventory of all connected devices that send or receive data. If you have already a robust system in place to keep track of all your assets, machinery and equipment, then you’re in a good position to achieve this.
4) The Need for Better Management of Data Integrations
With Industry 4.0 and the connected enterprise, there will be a smart platform receiving data from all IoT devices, and generating insights to predict and prevent incidents, anticipate and mitigate risks, improve safety, and enable operational excellence.
To make this possible, many different data integrations will have to be defined, established, tested and monitored. This includes the mapping of data from different systems and devices to the platform, and steps to ensure secure connections to prevent cyber attacks.
There must also be great scalability. Imagine if it takes a really long time to establish data connections each time a new sensor, beacon, drone, wearable, or connected equipment was changed or added. Operational efficiency would suffer.
In the age of machine data, data connections and integrations must be better managed to bring greater automation, and safeguard data accuracy and security.
5) Evaluating the Value of the Data
There used to be a time when there was not enough data. Now it feels sometimes like there’s too much data. The takeaway is not that too much data is a bad thing. Rather, you must not collect data just for the sake of collecting data.
There are costs associated with the collection of machine data: Sensors must be checked to be sure they’re working well, drones may need to be repaired, etc. And as discussed above, each new source of machine data may create an additional, potential cyber risk that needs to be managed.
But the most important thing is to evaluate the value of the insights and intelligence generated from the data. Is it useful in helping to prevent incidents, anticipate risks or improve operations? Be sure that there is a purpose behind the collection of data from each IoT device, that you’re not collecting data only because it’s something that’s technically feasible.
Address the five items mentioned in this post to be ready for the age of machine data, and to reap the benefits from the analysis of the data.
Download the Verdantix Green Quadrant Operational Risk Management Software 2019 report and learn more about the 17 most prominent operational risk management software vendors: