In addition, “the traditional demarcation lines we understood for various data architectures are increasingly blurring together,” said Lewis Carr, senior director of data management product marketing at Actian. Many SQL databases are now able to support NoSQL as well; there are on-premise platforms that also run in the cloud; transactional databases that can also handle analytics processing; and a range of document, relational, time series, and other fit-for-purpose architectures “are increasingly trying to do more than what they were initially designed to do,” Carr noted. “Hadoop was going to be the answer to how organizations handle unstructured and semi-structured data, but the luster wore off and the demand for Hadoop as a solution shifted to the big three public cloud data stores.”
Becoming stream-capable and cloud-ready are two essential elements of today’s architectural change, said Shriram Natarajan, director of digital and cognitive solutions with ISG. “The value of unstructured data used in conjunction with structured data is undeniable. With enterprise systems ready to handle both, the next stage of evolution is adding the ability to process data streams and event streams along with batch and synchronous data. Cloud readiness and cloud adoption have been introduced into enterprise architectures in greater scale over the last decade. With these improvements, the capability of the data stack to provide understanding has gone up substantially.”
THE MAKINGS OF GREAT DATA ARCHITECTURES
Four key traits that enterprises “need to look for in business-responsive architecture are availability, security, understandability, and trustworthiness,” said Tyler Warden, senior vice president of product and engineering at Syniti. “A common question I hear from many business decision makers is: ‘Don’t we have this data somewhere?’ In response to this question, a good data architecture makes the right data available to the right people at the right time.”
This doesn’t just mean changes are limited to the way data is managed and optimized—the 2020s data architecture extends well into other parts of the digital technology stack. Expect more innovation around passing the results of data analysis in forms that minimize bandwidth usage and costs, said Eric Schabell, portfolio architect director for Red Hat. “Look for innovations in how applications are architected and distributed around the world to minimize their cost profiles while maximizing your customers’ experience.” The cloud is making things easier—at least from a planning standpoint. Expect a bigger focus on managed data service offerings that are attempting to provide the tools and services to achieve these innovations, Schabell added, noting, “These are going to head in the direction of push-button cloud catalog offerings you can deploy as required to serve your application needs and then decommission when no longer needed.”
With data architecture 2020s-style, delivering great digital experiences is more important than ever. “Business success lies in offering agile digital services with high customer satisfaction,” Abeysinghe said. “The role of enterprise architecture is shifting from a centralized governing framework to a value stream generator—where it connects business and IT.”
It’s essential, then, that data architectures be as responsive to business requirements as possible—and that’s the new thrust of priorities in the 2020s digital era. The key is transparency and flexibility, said Gitenstein. “If you’re unable to see and analyze where and how you’re using data, it’s impossible to determine when to scale up to avoid workflow disruption. Our world is constantly changing, and data is a huge contributing factor to this. It’s imperative to have data architecture that is flexible enough to adapt to changing data, whether that’s by growing by petabytes or streamlining accessibility.”