Oracle Unity Customer Data Platform migrated from Hadoop to Autonomous Database, improving performance up to 30X and providing complete data isolation.
“We migrated to Oracle Autonomous Database from Hadoop, significantly improving performance and security and reducing costs. We manage several hundred terabytes of customer data with no dedicated administrators, so we can focus on delivering sophisticated data pipelines driving advanced analytics and intelligence.”
In our increasingly connected world, customers are asking for a consistent experience across online and offline channels with holistic recommendations and curated experiences for their users. Oracle Unity Customer Data Platform is built to manage this connected ecosystem and effectively sift through vast troves of first-party behavioral, personal, interaction, and transactional data to drive connected experiences across Oracle’s portfolio of Advertising and Customer Experience applications.
Using customer data to understand what matters most is critical, especially when customers can reach a competitor in a single click. Oracle Unity gathers customer signals and applies them to the profile in real time so users always have the most up-to-date information about their customers.
Behind the scenes, Oracle Unity has to blend data from many different sources, build enriched profiles, and conduct analysis with machine learning and predictive models. That used to be done on a Hadoop cluster but Oracle Unity needed something that was simpler to manage, more secure, and provided much better data isolation.
Why Oracle chose Autonomous Database
Over the last year the team embarked on an ambitious journey to move the initial version of Unity based on a legacy Hadoop architecture to a new streaming data architecture using the latest offerings across Oracle Cloud Infrastructure (OCI) and Oracle Autonomous Database. The next-generation data architecture offers new capabilities in streaming data processing, which enables customers to bring large volumes of batch and streaming data to build the most accurate customer profile to power connected applications.
The team migrated from Hadoop to Autonomous Database for analytics and warehousing. It allows for better reliability and management costs and simplifies setup and testing of high availability and disaster recovery configurations. Using Autonomous Database also enables full tenant isolation—a key selling point while talking to customers.
Today Oracle Unity has completed the migration from Hadoop and has been running on Autonomous Database for about a year. All customers have been migrated to the new platform. The current deployment includes over 60 instances of Autonomous Database totaling several hundred terabytes of storage. Each customer runs on a dedicated database instance, providing complete isolation.
All instances of Oracle Autonomous Database are protected using Autonomous Data Guard. This allows Oracle Unity to detect and recover from any failures 50 times faster than was possible with Hadoop both in-region and across regions.
Query performance and job execution has improved from 2X to 30X, depending on the use case. This allows Oracle Unity to serve even more complex use cases while maintaining response times.
Due to autonomous operations, no dedicated administration or DBA resources are required to manage the database instances. This simplified administration and maintenance and significantly reduces operational costs for the platform.