Oracle Generation 2 Public Cloud aims to beat AWS in price and performance

In a keynote address at Oracle OpenWorld in San Francisco, Larry Ellison introduced Oracle’s second generation cloud architecture featuring “Star Wars cyber defenses – an impenetrable force field with autonomous robots for finding threats and killing them.”

Current generation public clouds, use commodity servers built with standard silicon. A common base of cloud control code is used to manage multiple clients on same bare metal machine. The shared CPU, memory and storage introduce security risks for enterprises sharing these resources. In contrast, Ellison said the Oracle Gen 2 Cloud will not have cloud control software on each server. Instead, Oracle will use a separate network of dedicated cloud control computers to provide full isolation of customer data and separation of its control plane.

Design goals for the Oracle Cloud 2 also focus on automation for the easy migration of workloads from on-prem resources, including via Kubernetes.

Ellison said that unlike AWS, Oracle will not significantly mark up interconnect bandwidth costs it obtains from telecom carriers for either ingress or egress from its public cloud. On pricing and performance, Ellison affirmed that Oracle will offer a better deal than AWS. In terms of performance, Oracle will see gains from using RDMA cluster networking.

Oracle Gen 2 public cloud infrastructure is now ready for service, including its government cloud. In 2019, Oracle will also offer the option of having Oracle Gen 2 Cloud infrastructure deployed directly in the data centers of its largest enterprise customers.

Ellison also stated that the most important component of this new cloud infrastructure is the Oracle Autonomous Database, which was introduced at last year’s event.