Big data elevates the game

Data management issues have probably plagued the oil and gas industry since the first wells were dug or pounded out with cable-drop tools. More recently, even with the massive improvements in compute power over the last few years, the computer industry has struggled to keep pace with the energy industry.

Now things are ramping up even faster. New play types bring new challenges, and once again Silicon Valley is huffing and puffing to keep up. But companies like Oracle, keenly aware of the obstacles that face them, are suiting up accordingly.

Really big data

Charles Karren, director, oil and gas industry strategy for Oracle, said that the "big data" concept is nothing new for the industry. "This is the most data-intensive industry one can imagine," he said. "You can only get to the oil by data."

Data volume, velocity, variety, and value will force the energy industry to adopt a new paradigm for data management.

Karren's father worked in the industry from 1951 to 1999, and he said the changes in that almost 50-year career are dwarfed by the changes in the past decade. One of the major changes, he said, is the advent of unconventional plays, including not only shale but deep water and oil sands as well.

"Our view at Oracle is that with these new opportunities, you have to think about information technology differently," he said. "Companies need to deploy a new generation of software tools to help manage the complexity and to capitalize on these plays."

Another major change is the big data component. Karren said big data is characterized by "the four V's": volume, velocity, variety, and value.

"Volume is obvious," he said. "Chevron disclosed that it manages more data than Google and is growing it at 2 terabytes per day."

Velocity is the increasing speed at which data are collected. With real-time downhole sensors and other improved equipment, operators are taking measurements in quarter-second intervals.

Variety has always been a characteristic of oilfield data, but the digital age has ushered in additional variety in the form of accounting data, supply chain data, even emails.

Finally, the data have to have value. "You can take all of this: the volume of the data, the speed of the data, and the different types of data, but it doesn't mean much in terms of value, which is getting the right data to the right people at the right time," Karren said.

Naturally data will be stored to have on hand when government officials require an audit trail. But more importantly, the data have to be available to people within the company for them to make timely decisions. "The big challenge for the oil and gas industry is reducing the lag time between the data capture and the data analysis and then implementing and executing on the decisions once you're able to analyze that data," he said. "Really, it's the ability for an oil and gas operator to become a predictive organization, to be proactive."

Help from the computer industry

To react to this need for proactivity, Oracle has acquired several companies in recent years to beef up its business computing offerings. Karren and his industry colleagues mange the marketing in six oil producing regions – Calgary, Rio de Janeiro, London, Abu Dhabi, Beijing, and Perth. Together, this group has put together a stack of solutions and made it applicable to the oil and gas industry.

"We've created a solution called High-Performance Data Management," he said. "This is to boost the execution capability of the asset teams and the operators by closing the gap of being able to integrate this vast amount of data and then be able to execute on it.

"When we go to our customers, we listen to what they have to say, but our aim is to have them achieve their goals and be a predictive organization."

Oracle is not in the business of competing with existing software platforms that enable reservoir modeling and drilling engineering, he said. "What is out there has worked and is working because you cannot produce this amount of oil and scale up without something already in place," he said. "The question is, as these new plays develop and as the amount of data increases, we'll need to look at different ways of going about the business."

Toward this end, Oracle advocates open standards and is a member of Energistics, PPDM, and PIDEX. Its Complex Event Processing software takes data from WITSML and PRODML and directs it to the correct person at the right time with the right recommendations.

"What makes it different is that you're able to not only have that information stored, but based on data mining attributes, you're able to find something that the asset team may not have been aware of," he said. "As you move from these different plays, you have teams that are cross-fertilizing. The people may move around, but the information remains static. This is designed to say, 'We have these issues in this formation in the Eagle Ford, and these were the solutions we devised.'"

This takes the concept of leveraging best practices from what's in people's heads to what exists in the database, he added. "If the information is in a data warehouse, they're able to mine that data and analyze some of these issues as they come up," he said.

An additional tool provides "Exalytics," the ability to analyze key performance indicators, capital and operational expenditures, etc., and to share this information across the organization.

"This is not to replace software that is already out there," Karren said. "It's to elevate the game, to handle the new plays and the explosion of volume. I think the oil industry really is going to have to have a new paradigm. This has always been done at the corporate level, but now the data needs to be brought to the different groups within the organization."

Real-time management

"People want the information they need and they want it now," Karren said. "That's one of the reasons that Oracle is now a hardware and a software company. We want everything in one box. It makes it easier for operators because it lowers their total cost of ownership."

The company plans to spend US $4 billion to $5 billion a year on R&D, he said, "to make that integrated stack accessible to the end-user."