Posts

Industry Embraces Big Data

/by

We blogged about Big Data six years ago. Back then, pushing industrial data to the cloud in real time was a novel idea. Collecting industrial data within the plant for on-site use had been going on for decades, but few companies were integrating that data with enterprise IT or analytical systems.

Today, all that is changing. IoT and Industrie 4.0 are ideal for connecting industrial processes to Big Data. Progressive companies routinely use digital transformation to feed analytical systems to improve performance across the enterprise. Others are taking notice, trying to catch up. A recent research project by Automation World points to the growing rate of acceptance and adoption of Big Data among system integrators and end users, and how they leverage it.

Half of the system integrators in the study report that most or all of their clients collect production data to run improvement analysis. A quarter of the end-users surveyed say that they collect data from over 76% of their systems and devices.

While most of the data being collected is for in-plant improvements in equipment and maintenance operations, somewhere between 40% and 54% is also being used for Industry 4.0, smart manufacturing, or digital transformation initiatives. Pulling Big Data from the shop floor has become that important in just a few years time.

Data collection technologies

Despite the move towards Big Data, the most widely-used approaches to collecting data are still hand-written notes entered into a spreadsheet, as well as on-site data historians, according to the report. So for many users, the technology hasn’t changed significantly since the 1980s. However, cloud and edge technologies are gaining acceptance, being used at some level in about one fourth of the facilities reported on.

The survey didn’t specifically address it, but we see that some technologies originally developed for in-plant use—most notably data historians—are now widely used in edge and cloud scenarios. Some of the most well-known real-time data historians have cloud equivalents, or can be run on cloud servers. As a result, there is no clear line between traditional data collection and IoT-based systems, and there doesn’t need to be.

What is needed is secure, real-time data communication between the plant and the office or cloud. As high-quality data communication is more widely adopted, and as companies implement digital transformation in more areas, we can expect to see a huge growth in Big Data applications to optimize resource use, increase production efficiencies, and bolster the profits of the enterprise.

Turning IIoT Data into Value: The 5D Architecture

/by

What’s in it for me? Sure, the Industrial IoT is getting a lot of press—it’s been riding high on the Gartner Hype Cycle for years. But now that most people have beheld the vision and survived the deluge of glowing predictions, they are starting to ask some down-to-earth questions. In particular, engineers who have to assemble the pieces and managers who need to justify the costs are asking, “What are we going to get out of it?”

The benefit of the IoT, according to Finbar Gallagher, CEO and Founder of Fraysen Systems, is its ability to turn data into value. To explain how that happens, Gallagher has boiled down every IoT implementation into a common “5D architecture.” In his article, The 5D Architecture – A Standard Architecture for IoT, he says, “IoT systems are complex, very large scale and present many pitfalls for the system architect. Thinking about these systems in terms of the problem to be solved: turning data into value…”

The article breaks down the process of turning data into value through the interaction of five core elements, the 5D of the architecture, which can be summarized as follows:

  1. Data collection
  2. Detecting events based on changes in the data, and analysis
  3. Dispatching (decide and plan) an action based on events
  4. Delivering the action
  5. Developing value, which underlies and unites all of the above

Surrounding, connecting, and acting upon these 5D core elements are four services:

  1. Communication
  2. Presenting information
  3. Storing data and information
  4. Managing the 5 core elements.

Although these services are sometimes considered to be core elements, Gallagher separates them, because he says they do not in themselves create value. Each of these services relies on a person to extract value from them. Ultimately, value is not intrinsic to the data, analysis, plans, or actions either, but rather depends on human interaction to derive it. To make his point, Gallagher quotes a production manager who once said to him, “So if I don’t look at the charts this system presents, the system doesn’t deliver any value, does it?”

Be that as it may, people still need an IIoT system to access their data for extracting value.  And the better it functions, the more value they get. A good IIoT service will provide optimal data collection, event detection, dispatching, and delivery of action through secure and rapid communication, accurate presentation, and fully-integrated storage of data and information. Gallagher suggests some specific criteria, such as:

  • The ability to collect data from a wide range of sources, including legacy PLCs, log files, historians, and devices that may use different protocols.
  • Low latency data communication through direct, real-time connections whenever possible, avoiding high-latency approaches such as having a sender write data to files and requiring the receiver to read them.
  • Consistent event detection: repeatable and verifiable.
  • The ability to provide feedback (with or without human input) so that the system supports the ability to learn and modify action plans.
  • Data communication should be easy to use, resilient, and able to preserve structure. To these we would also add secure by design.
  • Data storage should be flexible, fully integrated, and minimal latency.

Anyone familiar with Skkynet’s approach to Industrial IoT will see that it meets the criteria that Gallagher proposes. On our own, we can’t turn data into value. That depends on you, the user. But we can provide you with easy, quick, and secure access to your data, so that you can make the most of it.

Collecting Big Data in Real Time

/by

It was bound to happen.  The two titans meet.  The gargantuan grasp of Big Data turns its ever-open hands towards the firehose stream of real-time data.  “The next evolution of the big data phenomenon has turned out to be real time streaming of data,” says Big Data pundit Rick Delgado in a recent blog: What Real Time Streaming Means for Big Data.  “Organizations have an increased need to gather and analyze their data at the same time, making real time data streaming a must if big data is going to keep up with demand.”

Will Big Data ever be satisified?  Not as long as the demand for informed action continues to grow.  Will we ever run out of real-time data?  Not as long as stuff keeps happening.  The only thing necessary to complete this marriage is to make the connection, and stream real-time data into the welcoming, capable hands of Big Data.

This is what we are keen on.  With our established track record in real-time industrial data communications, we anticipated this need for real-time analytics years ago, along with other thought leaders.  In a blog back in 2011 we quoted Paul Maritz, President & CEO of VMware at keynote address on the future of cloud computing at VMWorld 2011, “People are going to have to be able to react to information coming in, in real time.” Since then we’ve been putting the vision into action, and it’s great to now see the Big Data people coming on board.

Real-Time Analytics from Big Data

The advantage of live connectivity to Big Data is you can now do your analytics in real time. Delgado sees this clearly.  Real-time inputs to Big Data, he says, can fuel near-real-time outputs.  Rather than a two-stage process of storing the data, and then analyzing it, the analysis can take place on the fly, and your system can function like the mind of an athelete, jazz musician, pilot, or soldier. Insights become more spontaneous, and reactive responses are replaced by pro-active initiatives. The competitive advantage goes to those who can better anticipate and immedately meet customer demands, increasing customer satisfaction and establishing greater loyalty.

Delgado lists a number of areas where real-time streaming to Big Data could have a significant impact. For example, certain types of fraudulent or suspicious patterns of trading in the financial sector that don’t show up in the aggregate could be spotted in real time.  Businesses could monitor customer behavior on websites and social media to provide people with exactly what they need, at the moment they want it.

Additional Benefits – Industrial Sector

Among various application spaces that Delgado mentioned, he left out a significant one: streaming real-time Big Data for industrial users.  Imagine an operator of a machine where an alarm light is flashing.  Looking at his smart phone or tablet, he gets not only the alarm and raw data from the machine, but a real-time analysis of what could be wrong.  And along with that, he may receive suggested action steps based on comparing that data in real time to technical specs, historical records, and even live recommendations from its manufacturer, who is also connected to the machine, and monitoring it in real time.

Companies like GE are investing millions in such systems.  They collect and analyze in real time the Big Data coming from power turbines, jet engines, and other equipment during operation.  As the Industrial IoT gains acceptance, we see other companies, big and small, follow suit.  The value inherent in real-time data for making instantaneous decisions is too great to pass up.  The industrial sector, a large and long-time user of real-time data, stands to benefit significantly by connecting to Big Data.