Oracle uses Autonomous Database, Oracle Analytics Cloud, and Apache Spark to operate a 5 terabyte data mart in a lakehouse architecture.
“Our internal reporting platform, CP360, was built to help democratize data for over 250 diverse users in a secure, self-service environment. Our cost to operate the system is 30% to 40% lower as a result of using Oracle Autonomous Database. ”
Companies like Generali, FedEx, and Skanska use Oracle Analytics Cloud to understand their businesses and run them more effectively. At Oracle, the internal analytics team uses the analytics platform to run the Oracle Analytics, Fusion Analytics, and NetSuite Analytics business.
In total, there are about 250 different internal Oracle users with very diverse needs. Product managers need event analytics to understand what features are used and to make consumption predictions; support engineers need to proactively detect what causes service ticket issues and understand time to resolution; and executive decision-makers need insights on user growth, product engagement, service health, and customer health scores of products. All this requires pulling data from over 20 data sources with different types of connections, including consumption and billing data, usage logs, and service requests.
Two years ago Oracle had no unified data platform to integrate the various stream and batch datasets to meet the needs of all those users. What it needed was a holistic view across all areas with common tooling and reporting.
Why Oracle chose Autonomous Database
To address that business challenge, the analytics team built a self-service insight data platform called Cloud Product Analytics (CP360) using a lakehouse architecture built on a stack containing both open source and Oracle services. This also helps with rapid development and helps Oracle shorten the time from ideation to user adoption.
Data from those 20-plus different sources is initially moved into an object storage-based data lake, where the data is processed using Spark applications on Oracle Data Flow. The transformed, enriched, and curated data is then loaded into Oracle Autonomous Data Warehouse and made accessible to the users in Oracle Analytics.
The CP360 team and environment were built from the ground up in just 8 months. Data is currently loaded daily, though it has the capability to load data in near real time when the use case dictates that. And it’s in widespread use across the business. Development teams, for example, start their daily meetings with a live dashboard so they know exactly the status of their systems and customers.
CP360 is managed by only a few people and, thanks to Oracle Autonomous Database, they need no DBAs or data modelers. The team estimates that autonomous operations has saved 30% to 40% in overall costs.
One other strength of Oracle Autonomous Database is the ability to easily add capacity. As users and use cases multiply, the system can scale up several times. Resizing a system based on Autonomous Database is as simple as clicking on a menu to change current OCPU count or storage capacity: no downtime, no special expertise required. Today it stands at 10 OCPUs and 5 TB of storage, with growth planned for the future (10 TB to 20 TB). The team is also investigating use of autoscaling/ autotuning to improve performance when large, transient workloads occur.
CP360 helps Oracle’s analytics business to improve engineering agility and customer experience. Autonomous Database helps accomplish that by integrating smoothly with the rest of the open source and Oracle development stack, and reducing operating costs by 30% to 40%.