Oracle Corp. today announced that the Globally Distributed Exadata Database on Exascale Infrastructure offering it announced early this year is now generally available.

The service is aimed at artificial intelligence-driven applications and those requiring high levels of uptime and reliability while running across cloud regions. It combines global data distribution with Exascale’s serverless architecture to support a wide range of distributed workloads while maintaining data locality and minimizing operational complexity.

Exadata Exascale uses a multitenant resource pool, hardware-based remote direct memory access and predictive preprocessing to deliver query performance. Oracle says it’s more than 50 times faster than comparable services from other cloud vendors.

The service automatically synchronizes data across Oracle Cloud Infrastructure regions to provide continuous availability, even during regional outages. Its Active/Active/Active architecture provides high availability by running the same application or database across three or more data centers or nodes using Raft replication, a consensus algorithm that maintains consistency across replicated data. The combination ensures that even if some nodes fail, the system maintains a single, consistent view of data to enable failover without data loss. Each node is capable of handling live traffic at any time.

The distributed architecture also helps organizations meet data residency and sovereignty requirements by automating data placement based on regulatory or performance needs.

Bid for simplicity

Oracle said the service reduces the cost and complexity typically associated with operating large-scale distributed databases. “We are providing a mission-critical distributed database to the masses,” Wei Hu, Oracle’s senior vice president of high availability technologies, said in a statement.

The serverless architecture allows for dynamic scaling without requiring manual infrastructure management. That’s particularly useful in running fluctuating AI workloads, which tend to spike based on user demand.

Rob Strechay, managing director at theCUBE Research, SiliconANGLE’s sister research firm, praised the architecture’s scalability. “AI and agents will be distributed, so your data must also be,” he said. He described the offering as “a game changer” for organizations building next-generation applications. “Given that this can scale from very small to very large clusters, organizations will be able to build applications that can start as a single agent and grow into a more complex collection of agents.”

The database supports traditional transactional workloads alongside newer AI requirements, such as real-time image recognition and vector search. Oracle said it can support petabyte-scale capabilities for analytics and online transaction processing  of millions of transactions per second.

The service supports full Oracle Database and SQL functionality across regions, which reduces the need for application rewrites.

Image: SiliconANGLE/DALL-E