buchachon - Fotolia
Oracle Autonomous Data Warehouse aiding analytics developers
Early adopters of Oracle's new AI-based, partly self-managing data warehouse say it can free up time to develop value-adding enterprise applications, workflows and analytics.
As enterprises strain to keep up with the Ubers and Amazons of their respective industries, they often struggle migrating on-premises databases and ERP applications to the cloud. The cloud holds the allure of rapid innovation using new approaches like microservices and short development cycles. But the databases that sit behind key enterprise applications, such as ERP, have not been so nimble.
Oracle is one enterprise software vendor attempting to bring a measure of agility and automation to the databases powering such applications via its Autonomous brand, including the Oracle Autonomous Data Warehouse, which was released in fall 2017 at the Oracle OpenWorld conference.
"Autonomous is based on machine learning and decades of innovation by Oracle," said Andy Mendelsohn, executive vice president of database server technology at Oracle. "We want our users to be able to write the applications, give us their data, and we drive the database for them."
The vision includes an optimized infrastructure for databases that handle everything from memory management to patching themselves seamlessly in the background. Oracle claims this automation could help companies save 80% of the cost of running their applications.
By the product's one-year anniversary at the 2018 OpenWorld, it was clear that the Oracle Autonomous Database, including the data warehouse version, was now central to Oracle's cloud efforts. Early adopters of Oracle Autonomous Database interviewed at the conference largely confirmed that moving to the automated data warehouse has freed them up to develop new applications by minimizing database administrator (DBA) tasks.
Shifting DBAs to new roles in development and analytics
At an Oracle media event this past spring, senior IT executives shared their experiences with the Oracle Autonomous Data Warehouse cloud service.
Accenture is starting to use the Oracle Autonomous Data Warehouse internally as part of a tool to allocate about 400,000 employees across upcoming engagements. Patrick Sullivan, a managing director at Accenture, said the consulting firm is seeing a tenfold improvement in performance in the cloud at one-tenth the cost.
Sullivan said Accenture is finding less need for core Oracle DBA skills and is already looking at how to reskill their DBAs for other tasks, like building better analytics and application development workflows.
"We are looking at figuring out how to bring them to focus on what is in the data ... [and how] we understand the data and make it useful for our clients."
Better automation will make it easier to spin up and manage databases, according to Sullivan. Accenture has found that employees with modest technical skills can launch a new database in under five minutes. But this improved automation is also likely to highlight other bottlenecks in data quality and the ability to weave data into new applications.
For example, 97% of respondents to an Accenture survey were using data to make more informed decisions, but 90% said some of this data was wrong, Sullivan said. "There is a problem in finding a truth in the data."
Greater agility in spinning up new databases won't directly solve the data quality problem, but it could make it easier to experiment with and improve processes for moving data between ERP, HR and analytics systems, he said.
Improving maintenance across the enterprise
Aker BP, a petroleum producer, adopted the Oracle Autonomous Data Warehouse for a variety of applications. For example, it consolidated an archive of alarms that contained almost a billion rows of data. But the data was only needed sporadically. Oracle's autonomous offering allowed Aker BP to spin the database up quickly, file a report and spin it back down when it wasn't needed.
"I don't need to be a DBA anymore," said Erik Dvergsnes, a database architect at Aker BP. "I used to be a DBA, and it was quite a relief to move data to the cloud."
The cloud makes it easier for Aker BP to scale up applications for specific processes, such as scheduling maintenance. It used to take 15-20 minutes for maintenance teams to run reports, which delayed actual work for large teams.
"We can scale up the CPU so they can be out on the job much quicker," Dvergsnes said.
One major attraction of this transition of database technology to the cloud is that it dramatically reduces the anxiety around things like routine database maintenance. Oracle's autonomous technology automatically manages patch updates with guaranteed service-level agreements. In the past, Dvergsnes would hold off on patches because it was hard to find a window when some team wouldn't need the system.
"With patching, you are kind of scared of doing it because it will affect many databases if you do it wrong."
Experimenting with a better process
Quality Metrics Partners (QMP), a medical lab analytics service, has started experimenting with Oracle's new autonomous capabilities to improve the workflow of its core business process. One of the big challenges in the medical industry lies in analyzing the complex data sets captured from blood tests. Many labs can take up to two weeks to return results, but QMP worked with Oracle to reduce this to one or two days.
Michael Morales, CEO of QMP, said the business was started with the intention of helping five labs to be more competitive with larger enterprises. He attributes part of QMP's growth to using Oracle Autonomous Data Warehouse cloud to make it easier to experiment with different processes for analyzing data. For example, it has helped find new ways to add value in the summaries delivered to doctors.
As larger competitors started inquiring about their process, the company saw an opportunity to shift its focus from owning labs to being a back-end lab management service.
"We are seeing the lab management side grow faster than our core business," Morales said. Oracle's autonomous features make it easier to spin up new instances and fine-tune the workflow to deliver more valuable results faster. "Autonomous will change the future," he said.
