The Big Picture of Big Data

Digital Realty VP of information delivery and analytics Paul Balas eyes the future of data-driven business.

As the top information expert at the largest data center real estate investment trust operator in the world, Paul Balas has a front-row seat to the technology struggles and triumphs that all companies face. And, in the modern era, most of those battles revolve around the need for, or the use of, data. Today, companies that harness data to enable operations, close gaps, find efficiencies, and relieve pain points will find an advantage over their competitors. Data scarcity is no longer the problem; most companies can fill hard drives with mined data. The real trick comes in improving how data gets transformed into actionable information. Together with chief information officer Michael Henry and other Digital Realty leaders, Balas has found a solution.

The road has been a difficult one, though. The challenge that Digital Realty faces is familiar to most organizations. With more than 150 data centers spread across the globe, operating in thirty-three markets, Digital Realty supports the IT infrastructure strategies of some of the industry’s largest players, from financial services, cloud, and technology services to manufacturing, energy, life sciences, and gaming.

To do so effectively, the company has to answer some fundamental questions. How do leaders use the information that they have to make quality decisions? How do they get people to agree on what information is important? How do they unite information from numerous business systems and deliver it where and when it’s needed? And, perhaps most importantly, how does a company achieve these goals with the right investment for the maximum return? Balas answers all of those questions with one word: “leverage.”

“Use the people who know the data,” Balas says. “Digital Realty is embracing the expertise of our subject matter experts and enabling them to do their own ETL and dashboards with more efficient tools.”

By harnessing those tools to share accurate information, the company delivers dependable data center platforms and world-class ecosystems. This methodology requires modern tools and a new approach. “Traditional ETL is dead. The largest problems in delivering trusted information aren’t technical,” he says. “We have to know what the data is and how it will be used to discover or manage a business issue.”

Recently, he looked on as a new class of data integration vendors disrupt ETL platform sellers by building tools for subject matter experts such as financial analysts, marketing managers, and others. These experts who know their data are accustomed to using Excel—the world’s most used analytical and data integration tool. What was done for end-user dashboard development and exploratory analysis is now allowing these SMEs to work without the expense and support of large IT staffs.

This trend is also taking companies from IT-centric to IT-enabling. A company’s IT professionals have to understand what their business counterparts want to accomplish. The process takes time and involves many people. These companies have to hire, train, and groom people to deliver anything of value. Agile methods may help communication, but alone fail to change the speed of delivering solutions, and leaves frustrated business leaders tired of doing cumbersome data analysis themselves looking for new tools.

These new tools are coming from data wrangling vendors. Data wrangling, or data munging, is the practice of applying automated tools to convert data into visual and consumable sets. These vendors (such as Digital Realty’s vendor of choice, Paxata) are using big data (NoSQL) platforms like Hadoop and SPARQ to create powerful, flexible, and simple to use products.

The new breed of vendors are allowing businesspeople to collaborate faster as they transform and move their data. A financial leader who needs to blend invoice with a sales pipeline data, for example, no longer needs to wait months for IT to deliver a reporting solution or integrated system.

As Digital Realty offers these new capabilities, its leaders are watching other emerging trends that impact data-driven companies. “There is more to the puzzle than simply giving our business users access to their own ETL platform for a cohesive information strategy. We don’t want to make them work any harder than we need to,” he says. His team is implementing a data virtualization solution via Denodo that allows clients to source their own data more simply by integrating it across business systems. Denodo brings rapid delivery of business views and allows Digital Realty to deliver an abstraction layer between transactional and analytical systems.

Balas and his peers are taking these steps because they know that, for all companies, time is money. “By leveraging these tools and enabling our business users, we’re letting data analysts do a much more effective and efficient job at preparing their own data,” he explains.

The new approach is paying dividends. In 2016, Balas’s team completed a successful prototype and later sent that model into full production. In the test, they built a unified data view and new dashboards for a trusted business user. Previously, he would have asked for twelve weeks and five developers to complete the work. By applying data munging methods combined with cloud-based virtualization, data discovery, and dashboard delivery, Balas completed the project in a little more than a month with fewer developers.

Balas joined Digital Realty in late 2015. After establishing some early wins, he’s excited to see what’s next as users delve into memory computing, deep learning, and embedded AI. As they do, he will use those tools to create predictive models, better infrastructures, and more robust data.