Oracle blends analytics with MySQL database service

Oracle blends analytics with MySQL database service
Nipun Agarwal, vice president (research and advanced development), Oracle Labs
9 Dec, 2020

Legend has it that the name ‘Redshift’ for Amazon’s data warehouse product, released in 2012, was christened thus to signal a shift away from Oracle. Red alludes to Oracle, the database software corporation often referred to as ‘Big Red’.

On 2 December, Oracle’s chief corporate architect Edwin Screven stole a 15-minute head-start at Oracle Live, over AWS Databases vice-president Shawn Bice’s presentation at AWS: Reinvent 2020, scheduled for 9.15 AM (Pacific Time), to announce Oracle’s new data warehouse product.

Screven’s pitch entails a shift back to Oracle’s Gen 2 Cloud. In his half-hour address, Screven announced how Oracle’s latest cloud-based MySQL service integrates the open-source MySQL database for transaction processing with an analytics engine. This means enterprises will not require two separate databases (or a database each for transaction processing and analytics) anymore.

The Oracle offering faces off against a combination of AWS Aurora (a relational MySQL database service) and AWS Redshift (for data set storage and analysis). 

“Our MySQL product is 2.7 times faster than Redshift and at one-third the cost,” Nipun Agarwal, vice president (research and advanced development) of Oracle Labs, told TechCircle in an exclusive interview.

“It is a single database service to process OLTP (online transaction processing) that has introduced support for OLAP (online analytics processing),” Agarwal explains. “There is no need to move data to a separate database—all existing tools and applications function as is without needing a separate database.”

Oracle’s new data warehouse product has been possible because of its Sun Microsystems buyout in April 2009. This brought the MySQL database team to Oracle. By then, while Oracle had a stronghold in the enterprise market, MySQL had become the preferred open-source database for startups (then) such as Facebook, Twitter, Netflix and even Flipkart in India.

Amazon Redshift has 9.7% market share in data warehousing, compared to Oracle Autonomous Data Warehouse Cloud’s 7.3%, according to Datanyze, a firm that captures over 9,000 technologies used or implemented by 35 million companies globally. SAP Business Warehouse leads with 15.3% of the data warehousing market, which is estimated to touch $35 billion by 2025.

Two worlds

Oracle’s move now is to draw and converge enterprise-buyers and startups from the general-purpose database and analytics processing segments. Currently, the analytics database market is served by Amazon Redshift, Google’s BigQuery and Azure Synapse, among others.

“When it came to analytics processing, people would have to run the data out of the MySQL database to some other database for running analytics,” Agarwal says. “When you move data from one application to another, it has security repercussions, is expensive and takes time. Managing two different databases began to be a growing pain.” 

The new Oracle MySQL service has enhanced two components, Agarwal says. First, the query optimizer in the MySQL database: when a query comes, it determines whether the query should be executed by the OLTP engine, or the analytics engine.

Second, the analytics engine is an in-memory engine, which is why it is fast. “The data is populated from the MySQL database to the memory of the analytics engine. The second integration is with the storage layer,” Agarwal says. “We are able to support real-time analytics for subsequent queries. It is extremely efficient.”

The new MySQL service supports in-memory hybrid columnar processing, which means data can be represented in a columnar format, as well as accessed in a row format. Oracle has also developed new state-of-the-art algorithms for many of the distributed analytics operations, Agarwal says.

Oracle has been a late mover to the cloud, and the MySQL service is a new bid for enterprises and technology startups to harness its Gen 2 cloud infrastructure.

Data generated by organisations worldwide is increasing constantly, driving the growth of data warehousing. “Telecommunications and banking, financial services, insurance industries need to store billions of records on a daily basis,” according to a Transparency Market Research report tracking trends from 2016 onward. “Proliferation of cloud technology has been helping the data warehousing market to grow in recent years,” it states. 

For enterprises, budget constraints and the time required to build in-house software are the main factors for adoption of cloud-based data warehousing solutions. “Enterprises went to the cloud for elasticity, agility, scalability and security, but also to save money,” Edwin DeSouza, vice president-MySQL Product Management, told TechCircle. Cloud entails translating the capital expenditure of infrastructure into operating expenditure. “What ended up happening is month after month, the bills keep getting bigger and bigger. You increase consumption and utility—and you pay more,” he adds. 

With the MySQL service, Oracle is trying to also play up the price benefit of combining transaction processing with analytics. It is compatible with on-premises MySQL, making it easy for customers to move existing MySQL applications and workloads to Oracle Cloud.