The Next Great Frontier: Automating Data and Application Deployments
DevOps, DataOps, AI, and containers all lead to one important innovation for enterprises seeking to be more data-driven—and that is greater automation. Data-driven enterprises cannot function if data resources and applications are in any way being manually administered, deployed, remediated, or upgraded.
The ability to move fast, make decisions in real time, and respond quickly to events requires automated processes for ingesting and managing data. Organizations that fail to effectively leverage and deploy their data assets will find themselves falling behind. Data managers are turning to automation and autonomous databases and platforms, a recent survey of 217 data managers by Unisphere Research, a division of Information Today, Inc., found. According to the research, three in four DBAs feel that applications can be deployed faster with increased database management automation, and seven in 10 expect increased database automation to boost the impact of their roles (“2019 IOUG Autonomous Database Adoption Survey”).
Already, database functions such as backup and recovery are highly automated, and plans are underway to automate such day-to-day functions as monitoring, provisioning, and maintenance. Data managers welcome the advance of automation of these tasks and see greater roles for themselves in higher-level business decision making.
Moving to a high-functioning, automated enterprise is not without its challenges—both at organizational and technological levels. But corporate culture may be the most significant obstacle, industry observers agree. “Many people and teams who do not come from an automation mentality struggle to appreciate the value,” said Ben Klang, vice president of business technology at Power Home Remodeling. “This is
fundamentally different than manual processes,” Klang said. Enterprise teams have been deployed for a long time, and processes often come from a previous generation. Many safeguards and sign-offs are in place to enforce the wisdom of lessons learned the hard way and to help prevent future mistakes, he noted. “It’s easy to believe that those steps are still necessary, though they come into conflict with the goals of automation.”
IT departments themselves have historically grown up in cultures emphasizing risk avoidance, added Miha Kralj, managing director of cloud strategy, architecture, and delivery at Accenture. “Data centers, hardware, and software licenses—the foundation of IT until the cloud—were all expensive. To avoid any risk of an investment failing, companies instituted rigorous processes to analyze and justify these investments before making them.”
Ben Newton, director of operations analytics at Sumo Logic, agreed that culture change is the greatest inhibitor to enterprise automation, suggesting that many organizations may have difficulty adopting newer technology trends such as microservices and DevOps, which are all “built with small-team, agile structures in mind. Applying those ideas to older, siloed organizations is a recipe for failure and disappointment. Enterprises that want to compete with modern technology giants entering their markets must be willing to rearchitect their organizations, not just their tech stack.”
Organizational and data structures both play a role in slowing the drive to automation. “It’s challenging for companies to deliver real-time insights and analytics if their infrastructure can’t accommodate the data being captured,” said Rajan Kohli, president of Wipro Digital. “Data-hungry applications require a growing number of pipelines, which have caused data silos to become disconnected from other pipelines, datasets, and producers.”
Anders Wallgren, vice president of technology strategy at CloudBees, went further, saying databases themselves “are one of the last holdouts in the digital transformation.” He noted that he frequently sees “organizations that are quite broadly automated and reap the benefits of that automation but still grind to a halt if a push involves schema or database changes, which are still handled manually by DBAs.”
The sheer amount of data flowing into and through enterprises will also create challenges for the move to automation. “It’s no longer just active transactional data,” said Dheeraj Remella, chief product officer of VoltDB. “In the quest to understand their customers, internal processes, supply chain, employees, and systems, telemetry collection points are being deployed. Organizations need to act on the incoming data almost immediately. How do you incorporate machine learning into event stream processing in a continuous manner? Where do you make these decisions—in the cloud or near the edge?” Organizations cannot fall into the trap of assembling too many technologies in the interest of a loosely coupled architecture and give up on the primary business objectives of low latency, accurate decision making driving actions and responses to events, he noted.
Discover Past Posts