By Alan Zeichick | March 2020
When enterprises journey to the cloud, many applications and resources come along for the ride—and those applications might kick off a tug-of-war.
For example: An on-premise ecommerce platform built on the Oracle Exadata Database Machine and Microsoft .NET applications. Or perhaps an ERP system that uses Oracle PeopleSoft applications, Office 365, and Microsoft Workplace Analytics, as well as Oracle Database with Oracle Real Application Clusters. Which public cloud is the right place for those critical enterprise workloads? The best answer might not be either Oracle Cloud Infrastructure or Microsoft Azure. Indeed, the best answer might be “Both.”
Until recently, “Both” was not a realistic option for most organizations. Fast-forward to June 2019: An alliance between Oracle and Microsoft produced a fast, secure, easy-to-implement interconnect between their two clouds. That interconnect allows cross-cloud provisioning and low-latency links among different complex applications—each with a variety of parts—protected by a single identity-management system for both clouds. (See figure 1.) And with the recent announcement of a new interconnect site in Tokyo extending new functionality into Asia, more joint Oracle/Microsoft customers stand to benefit.
Many systems integrators, including multinationals such as Accenture, Capgemini, and Cognizant, are working with enterprises that have significant investments in both Oracle and Microsoft technology. Here are four top reasons why those systems integrators are bullish about the Oracle-Microsoft interconnect and alliance.
Want to link applications in two separate clouds? You could build your own bridge code. Another alternative: License a third-party service. Either of those options adds complexity to the connection, will incur high costs, might increase data latency, and can be brittle if something changes. And if something goes wrong, who are you going to call?
By contrast, the Oracle-Microsoft interconnect is direct without any intermediaries and is supported by both companies. You need to call only one phone number if you have questions or concerns.
“The interconnect solves the inherent problem of the connection,” says Chris Pasternak, global Oracle cloud infrastructure lead at Accenture. “If I wanted to connect two clouds independent of each other, and do it myself, I’d have to go through some sort of third-party connection, such as a cloud broker or cloud exchange. That creates a potential pain point, and another bill to pay.”
According to Pasternak, the Oracle-Microsoft interconnect eliminates that complexity. “Through the magic of automation, I could simply write a script and connect two applications,” Pasternak says, recalling his initial experience with the alliance. “It took less than an hour to connect our first application. We’re talking an hour versus weeks. And now that we have the scripts, it takes minutes.”
1.5 Average time, in milliseconds, for cross-cloud communications, according to Accenture benchmarks
When one part of a complex business application talks to another part, such as to query a database or to refresh website content, most applications expect that communication to happen nearly instantly. That’s what happens inside a data center, where a web application server in one equipment rack talks to a database server that’s only a few feet away. The communication between them is said to be low latency, specifying only a very short wait before data being sent is received and processed. When those business applications are migrated into a single cloud, the connection is also low latency.
But multicloud applications create a challenge. If data in one public cloud must exit that cloud, travel across a bridge that might be hundreds or thousands of miles long, enter the other cloud, and finally be routed to the correct cloud server, that lengthy communications path can introduce significant delays—which are doubled when the second part of the application sends its reply back to the first part. The complexity of the communications path can introduce inconsistent delays, especially if the path crosses multiple service providers. And of course, the greater the physical distance between cloud data centers, the longer it takes the data stream’s electronics or photons to flow across that distance—which, even at the speed of light, can become problematic for some multicloud applications.
Such delays, which don’t occur in conventional data centers, can cause applications to malfunction in several ways. For example, the first application might time out—that is, think, “Hmm, this is taking too long,” and resubmit its query, over and over again. Or the first application might simply give up and register a “connection broken” error.
A benefit of the Oracle-Microsoft interconnect is that the two companies’ cloud data centers are physically close to each other. Another is that the connection is direct between the clouds and has been carefully engineered to provide consistent, low-latency links comparable to what would be found inside a traditional enterprise data center.
How fast is that low-latency connection? According to tests from Accenture, run more than 32,000 times, the cross-cloud latency averaged less than 1.5 milliseconds—fast enough, reliable enough, and predictable enough to give customers confidence in this multicloud connection.
Chris Hollies, CTO of the Oracle practice at Capgemini UK, believes that this low-latency interconnect will become the foundation for a seamless, multicloud-architected set of applications and will open up multicloud enterprise adoption in a big way. “The enterprise used to put Windows and Oracle databases alongside each other in the same rack,” says Hollies. “The Oracle-Microsoft low-latency interconnect allows the enterprise to think that way again.”
Within a data center, organizations can standardize on a single directory system to manage access privileges for users, applications, storage, and services—even if those resources come from different vendors. That approach typically breaks down if those applications are migrated to separate nonconnected clouds. That’s one problem that the Oracle-Microsoft alliance addresses, says Roshan Subudhi, vice president and global head of the Oracle practice at Cognizant, who notes that Microsoft Active Directory is often the preferred enterprise directory platform.
When moving to the cloud using the interconnect, enterprises can leverage Azure Active Directory as a single sign-on for Oracle Applications running in Oracle Cloud, Subudhi explains. “The interconnect allows us to leverage multiple security and single-sign-on products that link both cloud stacks across Oracle Cloud Infrastructure FastConnect or Azure ExpressRoute.”
The cloud alliance provides unified identity and access management, including automated user provisioning, to manage resources across Oracle Cloud and Azure. As Subudhi points out, Oracle Applications can use Azure Active Directory as the identity provider and for conditional access.
“You can leverage these dual architectures and deliver a single set of credentials, as well as an improved logging and security experience over the interconnect,” Subudhi continues. “We are able to provide solutions around the cross-application architecture that resides on two different clouds. This capability will create an explosion of opportunities.” Such opportunities include user-facing applications running on Microsoft Azure talking to ERP services within Oracle Cloud, or applications running within both clouds sharing a single access-control model using Microsoft Active Directory or Oracle Access Manager.
Even in a digital age, the chief operating officer and chief technology officer can be buried under paperwork, and that includes managing software licensing and technical support. The complexity of figuring out the right license terms for applications and services that span multiple clouds can be daunting – as can be the challenges of figuring out which cloud provider to call when there’s a technical question. The Oracle-Microsoft alliance addresses those pain points, applauds Ramanan Ramakrishna, cloud center of excellence lead at Capgemini.
“The enterprise’s procurement function is becoming a very powerful gatekeeper,” Ramakrishna says. “This multicloud alliance takes the licensing conversation out of the equation.” Those license terms, offered by both Oracle and Microsoft to their volume-licensing customers, apply to specific applications and services.
Similarly, Oracle and Microsoft have created a collaborative support model, where a customer can call either company in regards to technical or operational questions for the interconnect. This lets developers and systems administrators leverage their existing Oracle or Microsoft customer support relationships and processes.
Ramakrishna is thrilled that the cross-cloud licensing terms provided by the Oracle-Microsoft interconnect encourages IT leaders to adopt the best of what both parties have to offer. “Nobody has to dig really deep or become concerned about ‘how do I license this software?’ or ‘who do I contract this service with?’” he explains—and that flexibility unlocks the multicloud innovation in the enterprise cloud journey.
Illustration: Wes Rowell; Motion graphic: iHua Design
Alan Zeichick is director of strategic communications at Oracle, and is editor-in-chief of Java Magazine. He was previously the editor-in-chief of Software Development Times. You can follow him on Twitter @zeichick.