On March 29, LNS Research and MESA International hosted a webinar based on the results of the most recent edition of the biennial Metrics that Matter research study: "Metrics that Matter: Digging Through the Data in an IoT World,"
The presentation focused on the development of the Industrial Internet of Things (IIoT) platform, including Big Data, Cloud, and Analytics capabilities, and how manufacturers today are beginning to implement these into operations, as well as how it is affecting the way they approach data and key performance indicators (KPIs) across plants and the enterprise. Indeed, as the buckets of the IIoT platform continue to be developed, manufacturers are developing new ways of connecting and sharing information from within the plant to higher levels of the organizations, and using analytics to come up with ever more sophisticated levels of intelligence.
During the webinar we received a number of questions during the webinar that we unfortunately unable to address within time constraints, which I'll now revisit below. In case you were unable to attend the live event, the free on-demand recording can be viewed here.
Q. Can you explain what periodic licensing is and how it differs from other licensing models?
A. Periodic licensing is simply renting instead of buying software. You pay an annual fee (often the contract is renewed once every two or three years) for the software you need. Periodic licensing models are often combined with user based fees so, for example you might rent 8 seats of some design software and if your usage increases during the rental period, you can simply add seats onto your rental agreement. It gives flexibility to users while makes a steadier income stream for vendors.
Q. What percentage of participants would you say feel they have a deficiency in data aggregation/historian/monitoring systems vs. those that are actively collecting data? For those that have data, do they know how to best use it, or is there a perceived lack of value in the data?
A. It is hard to answer the first part of the question because most manufacturers who collect data feel they use it well. I do not think there is a perceived lack of value, but rather that there is much more potential value available than people realize. For example, SPC is often cited as a valuable set of tools to improve quality and performance of manufacturing processes. Many companies use out of the box SPC charts from their software vendors; the opportunity to delve much deeper is there, but you need to employ real SPC expertise to dig deeper and get insights that can drive value specific to your process. The same issue will arise as we move to Big Data Analytics.
Q. Do you think the ISA Level 3 functionality will split into cloud and site domain?
A. Yes. In fact I think it will split more into functional domains that can then be run in Cloud or on premise. As manufacturers roll out IoT platforms, the integration between many of the level 3 functions will take place via the IoT platform infrastructure. This should (in a perfect world) lead you to be able to choose whether a particular application runs on premise or in the Cloud but for the foreseeable future the decision will be made by vendors solution architectures
Q. What are the different skills needed from an analytic perspective for IIoT vs. MOM or ERP? Do you see confusion regarding the needed skillsets?
A. MOM analytics skills are often focused on trending and SPC
Q. Is there a difference between private and public Clouds as far as data storing is concerned? How does it effect security or analytics capabilities?
A. The main difference between private and public Cloud is the location. A private Cloud will sit inside a company's firewall and run on the company's data center. Therefore the company will be responsible for maintaining the infrastructure and managing data storage and upgrades to the Cloud systems. Expansion as data storage grows will require capital investment.
The public Cloud model means that a company simply pays for the resources required and the Cloud company manages the infrastructure, shared across many clients. One common issue with public Cloud is security. However, one of the goals of the IoT is to be able to combine data from many different sources to allow sophisticated analytics that could use data from suppliers, customers, social media and many other sources. There will be a need to gather data from outside the company and in a private Cloud setup this will involve data connectivity beyond the firewall. Clearly in both public and private Cloud scenarios, security is an issue. At first glance private Cloud does offer more innate security, but public Cloud service providers consider security as their number one technical issue and are constantly working to improve the privacy of their customers' data.
Q. Can you think of any real examples of data analytics used to introduce new KPI/business metrics that contributed to large savings for a company? Something that was outside the box?
A. great question. A couple of examples we have heard about:
- A Big Data solution for optimizing rail operations that saved an operator $1Billion in fuel savings.
- A major trucking logistics provider put in IIoT, Big Data, and reduced maintenance cost by 30%
Although these are dramatic examples, similar claims have been made using traditional solutions like data historians and enterprise manufacturing intelligence. Modern Big Data tools certainly make this type of application easier to implement and may well increase the benefits but these are not out of the box solutions. Ask again later in the year and we hope to have evidence of real game changing examples.