"Everything to the Cloud." Really?

I get the opportunity to discuss Cloud strategies with many Oil & Gas operators and services companies. When I ask them to tell me about their corporate mandate for Cloud implementation, I often hear “Our CIO is telling us everything has to go to the Cloud”. Many of the people that I interact with, aren’t the C-Suites of these organizations who have the bug for the Cloud. Instead, they are the IT strategists who are tasked with making this happen. My questions back to them: “How is that strategy working for you?”

First of all, I am a huge Cloud proponent! For several years now I have seen firsthand many successful implementations that are providing real business values to the implementers of solutions in the Cloud: 

• The ability to rapidly spin up and prototype test environments. 
• Data security now surpassing what almost any individual company can achieve on their own. 
• Machine learning and AI techniques designed to extract insight from vast arrays of data. 
• The ability to scale volumes of data to analyze, and the processing compute capacity to extract those business insights in a timely manner. 

You don’t have to search hard to find these success stories. But, as our creativity expands into new ways of extracting such business insight gems, and volumes of data from ever increasing numbers of sources grows, so does the cost of doing business in the Cloud. My challenge is: can the Cloud alone be the answer to our business analytics quest for operational greatness?

At the same time as CIOs are shouting from the upper reaches of their corporate headquarters, “I want to see all our data and business systems fully operating in the Cloud.”, I am keenly aware of other trends in data growth that appear to be continually outpacing the reality of moving 100% of everything to the Cloud. Enter the age of the Industrial Internet of Things (IIoT). According to a recent Gartner report, 8.4 Billion connected “Things” will be in use in 2017, up 31 percent from 2016. My belief is that the technologies that are driving down costs of sensors, proving out AI, cost of storage and computing, are outpacing the growth of the world’s capacity in the Cloud, and any cost-efficient means in which to get our data to the Cloud.

The major players in Cloud infrastructure: Microsoft Azure, Amazon Web Services, IBM Bluemix and Google are some of the largest technology companies in the world. But at the same time they are expanding and innovating, trying to pave the way for us all to migrate to the Cloud, let’s not forget about some of the other noteworthy tech-giants: GE, Siemens, ABB, Schneider Electric, Honeywell, Emerson, etc. This list of industrial heavyweights who are the drivers of IIoT is not insignificant and their impact should not be under-estimated. Where our Cloud infrastructure providers are trying to enable your digitalization transformation, my contention is that these major industrial players-turned-software juggernauts, are the ones driving today’s digitalization transformation. Who is outpacing whom? And where is your company at this time?

Take a step back in your drive for the Cloud. With data sources and data volumes outstripping your practical abilities to “move everything to the Cloud”, have that conversation with not only your CIO, but also your business leaders. Ask them “What business problems are we trying to solve for?”. “What data is truly needed to address these needs?”. Take the time to develop a Cloud strategy.

Much of the growth in data volumes, and the need for business analytics, we are seeing is happening at the Edge: sensors on industrial systems to determine non-optimal operating conditions or failures, monitoring systems for worker health and safety, field surveillance in the form of fixed or drone based cameras, and so on. The workflows and benefits being advocated by the Industrial players address operational business needs like machine & equipment health, worker and plant safety, maintenance and equipment inventory optimization. But are they advocating that all your data must be in the Cloud?

Let’s look at the practicalities of having all this IIoT data in the Cloud: The cost of transmitting and storing all this data in the Cloud, and the time cycle it will take in order to pass all that sensor data to the Cloud, analyze it, and affect some action back in the field.

Today you likely have several pilot studies ongoing whereby you have time series sensor data streaming from business-critical equipment. In most cases I’ve seen, 50 to 500 devices are targeted for monitoring in a pilot study. Even for per-second data feeds, with 5-10 sensors on each device, this volume of data is manageable. Telecommunications bandwidth and associated costs, along with the data IO costs from your Cloud provider are manageable. A Cloud only strategy seems feasible. But, now your business unit announces success with the pilot, and commands “Let’s scale this up!”. 

Let’s take a typical Oil & Gas industrial site: A refinery or LNG export terminal. Don’t be surprised to uncover functional devices, processing and control systems amassing over 200,000 sensors all with associated live data feeds. Now let’s cycle back to examine the go-forward operating costs of an all Cloud strategy. There’s the problem! Any business efficiency gains and profits to be had at this one site will be quickly gobbled up by your new ongoing telecommunications and Cloud costs. And what of your entire enterprise? How many pieces of industrial equipment and sensors are out there? Millions?

But let’s say, for your organization, “Money is no object”. I personally don’t know of such an Oil & Gas company, however, bear with me. I’m sure that your trusted Industrial provider of all those sensors on all that critical equipment they sold you, is expressing the time sensitivity of the information being collected. Meaning, if the data you are collecting and analyzing is telling you “There is a workman down”, or that “A mission critical device has just failed”, you don’t want to find out about this too long after the occurrence. Implementing on-site converged IT infrastructure (storage & compute) is the only viable way to ensure real-time alarms and alerts are received and acted upon. “Real-time” is not going to occur if this time sensitive data is pushed to the Cloud, analyzed, instructions returned, and then acted upon.

Again, I’m a huge fan of the Cloud. It has its place in all our digitalization initiatives. My belief is that every Oil & Gas operator and services company must embark on a Cloud implementation journey. But “Move everything to the Cloud” is just not practical, feasible, or warranted. Take the time to facilitate the business needs discussion within your organization. Discuss the needs within the context of what the Cloud is excellent for, and what it is not. My belief is that you will end up with a more cost effective, more functional, Hybrid Cloud solution in the end.