Oracle today made available its converged cloud database, Oracle Database 21c, in the Autonomous Database Free Tier Service.

This means anybody can spin up 21c as an autonomous database, which is essentially the Oracle database engine plus some additional cloud operational functions running on the vendor's Exadata infrastructure.

Included with the release are 200 new updates such as immutable blockchain tables, In-Database JavaScript, native JSON binary data type, AutoML for in-database machine learning (ML), and persistent memory store, as well as enhancements for in-memory, graph processing performance, sharding, multitenant, and security.

“Our goal is to provide the best converged database,” said Jenny Tsai-Smith, Oracle VP for overall database product management. “And by that what we mean is being able to support any data type, any data model within the same database engine, as well as any workload.” According to Tsai-Smith, the objective is to provide transparent scale up and scale out. 

The launch follows the vendor's introduction of Oracle Database 21c in the Oracle Cloud Database Service Virtual Machine (for RAC and single instance) and Bare Metal Service (single instance) late last year. General availability of Oracle Database 21c for on-premises platforms, including Exadata, Linux, and Windows, will follow in the first half of 2021.

Oracle says its single, multi-model converged database hides complexity of management and maintenance, which makes the enterprise database accessible to small businesses or individual developers. 

“The 200 new built-in innovations, including immutable blockchain tables and AutoML for in-database machine learning, elevate Oracle Database 21c to a new level of functionality, eliminating the need for specialized, isolated cloud services and tools to do those jobs,” said Carl Olofson, research VP for data management software at IDC, in a prepared statement.

The single-engine database does not require users to integrate different silos of services and absolves the compounding of costs and operational complexity that comes with each additional cloud service that organizations ordinarily use. 

Steve Zivanic, global VP of database and autonomous services at Oracle, likened Oracle 21c to buying a completed car as opposed to buying a steering wheel, a chassis, seats, and then eventually assembling an entire vehicle to get from point A to point Z.

“In this way, Oracle is effectively slicing away at this disjointed set of services with a simplified, more technically elegant, and integrated approach that is far better suited for the enterprise needs of 2021,” Olofson added.

The latest generation of Oracle's converged database introduces Blockchain Tables to give users some much needed security to build applications that can support a distributed ledger. A cryptographic hash creates an insert only table to ensure that no row can be changed at a later date. Additionally, users are prevented from truncating data, dropping partitions or dropping Blockchain Tables within certain time limits. 

"It's really about protecting your data in the database and creating this immutable ledger that is easy to implement and is scalable," Tsai-Smith explained.

It also offers support for all data types including relational, JSON, XML, spatial, graph, OLAP and others in addition to scalability, availability, and security for operational, analytical and other mixed workloads. It also provides Intel Optane Persistent Memory support, which Tsai-Smith said makes it “cheaper than DRAM, but much faster than flash cache,” for better performance at a lower cost.  

Sharding – which involves breaking up data into two or more smaller chunks – is part of Oracle’s strategy to support any data type, any data model, as well as the workload. It allows a database to be sliced up into different shards and spread across a geographically dispersed database. Each shard is independent in terms of software and hardware to apply the same parallel and optimization across any of sharded workloads.

According to Zivanic: “Most databases are technically using sharding to overcome issues like scalability, availability, and reliability, we're only using sharding when the application actually requires it, and not to cover up holes in the product. So for us it's a feature, for other databases it's more of a workaround. I think that's an important distinction.”

Despite being a database vendor with decades of experience, Oracle is still an underclassman in a cloud market of seniors. 

Amazon Web Services (AWS) is kind of King of the Hill as far as the cloud platform, said Tsai-Smith, “we're [Oracle] trying to make sure that people know we have a solution that's even better than all those different specialized databases services on AWS, that we actually have this offering available in a single database engine.”

Amazon – the oldest and largest cloud service provider in terms of its infrastructure footprint – by contrast, does not offer a similar multi-modal database. Instead, it offers 15 specialized managed database services. And while it does offer comprehensive database migration services, a fragmented database portfolio makes cross-silo data analytics more complicated as capabilities vary between database engines.

“Ease of use is not something that Oracle has been known for," Tsai-Smith added. "But over the last two years we've really been focused on taking features that are already in the database, making them simpler to use, and adding in the autonomous capability so they [customers] don’t have to make all of the decisions."