As hybrid environments become the norm within organisations, considering what you can run where is a key component of system architecture and IT Asset Management. In this episode, Rich talks to Dean Bolton from LicenseFortress and Tim Gormon from Bridgeton Data about the dos and don’ts of running Oracle software in the Microsoft Azure cloud – from both a licensing and a technical perspective.
If you have Oracle software and Microsoft Azure in your environment, this episode will help you define what can (and can’t) be done as they come together.
Table of contents
Introduction
Rich Gibbons:
Welcome to the ITAM Review podcast, news, reviews, and resources for ITAM, SAM, and software licensing professionals. Hi everyone. Welcome to this latest ITAM Review podcast, and joining me today I have got Dean Bolton from LicenseFortress. Hi Dean.
Dean Bolton:
Hello all. Thank you, Rich.
Rich Gibbons:
And I have got Tim Gorman from BRIDGETON Data. Hi Tim.
Tim Gorman:
How are you?
Rich Gibbons:
I’m well. Good, thank you. It’s great to have you both with me.
Oracle in the Cloud Overview
Rich Gibbons:
So today we’re going to be talking about Oracle in the cloud, specifically Oracle in Microsoft Azure. Many people have Oracle to some degree, many are using Azure, and the question arises: Can we put Oracle on Azure? Should we put Oracle on Azure? What about Oracle OCI? These are questions that we hear often, and I think that hopefully between us, we will come to some kind of agreement by the end of this podcast.
Can You Put Oracle on Azure?
Dean Bolton:
Sure. The short answer is yes, you can, but there are a lot of nuances in there, and I think there are still a lot of misconceptions around running Oracle workloads in Azure. Hopefully, we can address these from a licensing perspective, a support perspective, and a technical perspective during the course of this podcast.
From a licensing perspective, at a very high level, there is nothing in your Oracle contract that would prevent you from moving your workloads wherever you see fit, including Microsoft Azure. Azure is a fantastic platform with many advantages as a public cloud, especially in terms of scalability and elasticity. Your Oracle workloads can fit right into that.
Tim Gorman:
Agreed. As Dean was mentioning, there are many different aspects to consider—licensing, technical, and otherwise. Oracle’s licensing policy is relatively flexible. Many people mistakenly think that Azure is a Windows-only environment because it’s from Microsoft, but in reality, it’s 65% Linux and growing. There’s really no technical limitation preventing Oracle workloads from running on Azure.
Azure-Oracle Partnership
Tim Gorman:
The partnership between Microsoft and Oracle is relatively new and is currently being rolled out to different regions globally. However, neither Oracle nor Microsoft is particularly enthusiastic about running Oracle on Azure. Oracle prefers Oracle Cloud (OCI), and Microsoft would rather you use Azure-native services like SQL or Postgres. However, customers are pushing to run Oracle on Azure because of the central role Oracle plays in their IT estates.
Dean Bolton:
That’s a great point, Tim. I think the key takeaway is that customers now have options. Exadata is an interesting solution, but it’s high-end, like a Ferrari. For many use cases, you don’t need that level of power. Azure offers a variety of compute and storage options that can suit many Oracle workloads without requiring Exadata.
Cross-Cloud Interconnect and Exadata
Tim Gorman:
In 2019, Microsoft and Oracle introduced the Cross-Cloud Interconnect, which allowed Azure and OCI datacenters to connect directly. The idea was that you could run, for example, a front-end application in Azure and connect it to an Oracle database in OCI. But that concept didn’t work out as planned due to latency issues. The Exadata at Azure solution that was introduced more recently is a response to that, providing a more integrated option for running Oracle workloads natively in Azure.
Dean Bolton:
The Exadata partnership at Azure allows customers to avoid the latency issues that came with the Cross-Cloud Interconnect. It’s an interesting option, but you need to determine if you really need Exadata-level performance for your workloads. Many organizations don’t, and Azure’s standard compute resources can be more than enough.
Migration and Licensing Considerations
Dean Bolton:
There’s often confusion when organizations are migrating to the cloud. Many think they need to silo their workloads—keeping Oracle on OCI and everything else on Azure—but that’s not the case. You can move your entire Oracle stack to Azure, just like your other workloads. What’s important is to right-size your environment and ensure you’re only paying for what you need.
Tim Gorman:
One licensing concern comes from Oracle’s processor core factor table, which was introduced in the 1990s. It gave a 50% discount for Intel processors, but that discount does not apply to Oracle workloads in AWS or Azure. That’s led to the myth that Oracle is twice as expensive on Azure or AWS, but really it’s about right-sizing your VMs and being mindful of licensing.
Dean Bolton:
Customers also need to think about migration time and license optimization. Most organizations aren’t going to pick everything up and drop it into Azure overnight. It takes time, and you need to manage your licenses accordingly. In some cases, you can realize cost savings immediately, but in other cases, you may need to wait until your next renewal.
Oracle RAC and Alternatives
Dean Bolton:
Oracle Real Application Clusters (RAC) is not officially supported on Azure, but there are technical workarounds if you really need to run RAC in the cloud. That said, many organizations are moving away from RAC in favor of alternatives like Oracle Data Guard, which offers high availability without the overhead that comes with RAC.
Tim Gorman:
That’s right. RAC was originally developed for scalability, but today, most organizations are using it for high availability. However, Data Guard is a more cost-effective and technically simpler solution. For the vast majority of workloads, you can achieve the same level of availability with Data Guard as you would with RAC, but with less complexity.
Conclusion
Rich Gibbons:
This has been a fascinating conversation, and I think for some people listening, it’s perhaps going to raise additional questions. Feel free to get in touch with any of us via email or LinkedIn. If there are more questions, we can cover them in a future session or address them directly.
Dean Bolton:
Thank you, Rich, and thanks to all the listeners. We appreciate you taking the time to listen.
Rich Gibbons:
Thank you to both of you, Dean and Tim, and thanks to our listeners. Hopefully, this made your day a little more interesting. Catch you all on the next one.
