Thought Leadership

Autobahn in the Cloud

How Gen 2 Oracle Cloud Infrastructure delivers for your business

By Alan Zeichick

July/August 2019

Fast transactions mean happy customers. Fast transactions mean happy employees. And fast transactions mean more business agility and scalability. Oracle has worked hard to ensure that every part of its second-generation Oracle Cloud Infrastructure is geared toward performance. Part of that effort means using the latest technology within the servers, memory, storage, and network fabric.

Equally important is the networking used to connect resources within the Oracle Gen 2 cloud and to provide the fastest-possible highways into and out of the cloud as well. That means high-speed links from on-premises data centers to applications running in Oracle Cloud—and recently to services running in Microsoft’s Azure cloud.

Together, these advanced technologies provide rapid ingress and egress of data, drive applications and transactions, and help deliver success for enterprise customers and independent software vendors (ISVs) that use Oracle’s business applications or deploy custom applications on the platform.

Fast Interconnects Pave the Superhighway

The high-performance servers and internal networks within the Gen 2 Oracle cloud ensure that cloud applications run quickly, delivering user experiences, analytics, AI, and other benefits directly to businesses.

Many Oracle Cloud users, however, run their applications across clouds from different vendors and in a hybrid environment, where part of the workload resides in Oracle Cloud Infrastructure and other parts are in the customer’s own on-premises data center. In such scenarios, it’s essential to provide a fast, reliable interconnection between Oracle Cloud Infrastructure and the customer’s data center. Reliable, in this case, means more than ensuring that the link doesn’t go down. It also means that the performance of the interconnect is predictable.

You can choose how to run an application in the cloud and how to connect it with your on-premises apps—or not. We provide that flexibility and isolation between what you do in the cloud and what you do on premises.”—Charlie Baker, Senior Director, Oracle

The interconnect is critical to performance, because some queries and transactions may require dozens or even hundreds of information transfers, some in parallel but others sequential, one after another. A slow interconnect, or one with unpredictable delays or throughput, might bog down one-off or routine tasks and cause a terrible user experience. In some cases, job processing might even time out and fail, requiring the process to start over.

“If I run in a hybrid cloud environment and I’m connecting the cloud to my data centers, what kind of latency can my application tolerate to be split across those things?” asks Charlie Baker, Oracle’s senior director of product management for network services in Oracle Cloud Infrastructure. “Can I have my database in one place and my application in another?”

The answer, says Baker, is yes, and the reason is Oracle Cloud Infrastructure FastConnect, which offers fast, predictable connectivity across on-premises and cloud networks. Using Cloud Infrastructure FastConnect is an alternative to using the unpredictable public internet, where data routing is based on how the internet is doing at that moment. Instead, Cloud Infrastructure FastConnect uses dedicated, pre-established paths to and from Oracle Cloud.

“It’s like having your data center extended with a wire to the cloud,” says Baker, adding that Cloud Infrastructure FastConnect was built specifically to take advantage of the network model, scalability, and compute capacity of Oracle Cloud Infrastructure. For example, customers can use Cloud Infrastructure FastConnect for disaster recovery and database backup and to bridge their data centers’ applications and on-premises databases to Oracle Autonomous Data Warehouse, Oracle Autonomous Transaction Processing, and Oracle Autonomous Database Dedicated services running on Gen 2 Oracle Cloud Infrastructure—with better performance, improved reliability, and easier administration.

“Oracle Cloud Infrastructure is built with the requirements of the Oracle Autonomous Database services at the forefront,” adds Baker.

Embrace Your Network Differences

Business and cloud networks are not all the same, and Cloud Infrastructure FastConnect enables administrators and network architects to choose how to deploy their virtual networks, including how to connect, route, and scale, says Baker. “You can choose how to run an application in the cloud and how to connect it with your on-premises apps—or not,” he says. “We provide that flexibility and isolation between what you do in the cloud and what you do on premises.”

If I run in a hybrid cloud environment and I’m connecting the cloud to my data centers, what kind of latency can my application tolerate to be split across those things?”—Charlie Baker, Senior Director, Oracle

How does Oracle pull this off? It’s a team effort: The Cloud Infrastructure FastConnect service is supported by dozens of top-shelf connectivity partners around the world.

What about customers that use multiple clouds, such as Oracle Cloud Infrastructure and Microsoft Azure? The two companies recently announced a direct connection using a fast interconnect. This connection is initially between Oracle’s Ashburn Gen 2 cloud region and Microsoft’s Azure US East cloud facility, both physically located in Ashburn, Virginia. The companies have plans to expand to additional regions.

Fast Gen 2 Servers Have Plenty of Horsepower

Cloud Infrastructure FastConnect is far from being Oracle Cloud Infrastructure’s only performance driver—the entire architecture is built for performance.

Oracle Cloud Infrastructure completely separates the servers running customer workloads—code, data, and resources—from the servers Oracle uses to administer those servers. This approach has two key benefits:

  • Security. Customers deploy their workloads on compute and storage hosts on which no Oracle software of any kind runs. Oracle’s administrative software can’t see the customer’s code—and customer code can’t see Oracle’s administrative software. This isolation is an important key to effective security.
  • Reduction of overhead. Essentially, 100% of the customer’s cloud server’s processor, memory, storage, storage bandwidth, and network bandwidth is available for the customer’s applications.

Workload isolation also helps to drive service reliability. Oracle’s service-level agreement (SLA) specifies between 99% and 99.99% uptime for Oracle Cloud Infrastructure, depending on the services, so everyone can depend on Oracle Cloud Infrastructure for business-critical applications.

Next Steps

LEARN more about
Oracle Cloud networking.
Oracle Cloud Infrastructure FastConnect.

READ about the Microsoft and Oracle cloud interoperability partnership.

Illustration by Wes Rowell