
Oracle Service uses Autonomous Database to replace Hadoop and Hbase
Oracle Service uses Autonomous Database to simplify data management so that developers can move to a DevOps model and a microservices architecture.
“I was pleasantly surprised by how easy it was to integrate Autonomous Database into our DevOps-based Kubernetes microservices and meet our goal of processing over 400 million records in less than half the time of the previous implementation.”
Business challenges
Oracle Service gives organizations the ability to predict the need for service, automate processes, and deliver tailored responses. It includes B2B, B2C, and field solutions that allow customers to receive the service they want, when and where they need it.
Within Oracle Service, a component called Action Capture Service (ACS) manages license compliance and measures product usage. ACS was originally built on a data platform based on Hadoop and Hbase and was used to manage all the needed data regarding customers, products, service calls, telemetry, and more. But that environment was hard to manage. The development team wanted to move to a DevOps model, and it needed to simplify the data management model in order to do that.
Why Oracle chose Autonomous Database
The team selected Oracle Autonomous Database for analytics and data warehousing. Automated administration capabilities and simpler provisioning made the database easier for developers to work with. They just wanted to use databases, not provision or manage them, and that required autonomous operations.
Results
Today Oracle ACS runs on a single Autonomous Database instance, with plans to grow that to over 2 petabytes. All new data flows into Autonomous Database and works with microservices on OCI Kubernetes Engine (OKE). Migration of more than a petabyte of existing data to Autonomous Database is well underway.
The new system requires fewer compute resources to operate. For example, replacing a Hadoop/Hbase cluster of 2,000 cores requires less than 200 OCPU resources for the same capability. Not only is the system using fewer resources, but performance is two to three times faster. It used to take 6 to 8 hours to process 400 million records—that time is now less than 3 hours.
Developers recently switched on Autonomous Data Guard to give failover to a new region. The process took a developer less than two days to set up and test the new environment, with no DBA assistance.
As demand grows, the development team needs to increase the capacity of its database instances on several occasions. With Autonomous Database, this takes just a call to the right API, with no specific database expertise and no downtime. Developers also switched on autoscaling. In the event of unplanned, transient demand, a database can automatically scale up to 3X the base provisioned OCPUs. Developers prefer to size the database themselves, but autoscaling does act as an insurance policy when things change rapidly.
The original business challenge was to simplify data management so the team could move to a DevOps model and migrate from a legacy Hadoop environment to the cloud, while adopting a microservices approach. With Autonomous Database, the dev team has completed that move successfully.