Pradeep Vincent is the chief technical architect and senior vice-president of Oracle Cloud Infrastructure (OCI). His role involves internal engineering of all the services in OCI. At Oracle Cloud World in London earlier this month, Vincent talked about the work Oracle has done with Microsoft, first with a cloud interconnect, and most recently, putting the OCI database in Microsoft Azure datacentres.
Speaking about the OCI strategy, Vincent describes the importance to the company of its distributed cloud strategy. “Oracle database in Azure is a key part of it,” he says. “When we started talking to customers about what’s missing in OCI and what was causing friction in terms of cloud adoption, we quickly realised it was their on-premise environment and the datacentres they own.”
Vincent feels that the cloud was originally designed for dot com businesses that did not have their own IT infrastructure. “It wasn’t really designed for the enterprise,” he says. “One of the key things that led to our distributed cloud strategy [was the] need to meet customers where they are – not in terms of people, but in terms of the datacentre footprint.”
This led to the development of Oracle Alloy, a platform designed for service providers and large enterprises, which offers the same 100+ infrastructure and platform services that are available in OCI’s public cloud.
Dedicated regions from OCI
Oracle also offers OCI Dedicated Region and is now giving its enterprise customers the ability to run the Oracle Cloud Infrastructure database within an Azure datacentre. Nomura Research Institute (NRI) is one of Oracle’s reference customers. It has moved its retail brokerage accounting system to OCI Dedicated Region to scale the offering and help ensure maximum availability. The accounting system provides back-office support to more than 70 brokerage firms and banks that manage hundreds of millions of trades a day.
When asked about the difference between Dedicated Region and public cloud versions of OCI, Vincent says: “Public regions are shared among many customers. We sometimes do have dedicated pools of capacity, but the region itself is multi-tenanted.” Dedicated regions, he says, are built in a customer’s datacentre, offering them a sense of control over their workloads. “The customer actually deals with the datacentre. In some cases, the datacentre itself is operated both by Oracle and the customer,” he says.
According to Vincent, Oracle has put a lot of effort into making OCI Dedicated Region operate in the same way as the public version. Oracle offers the same services, same deployment model and same service level agreements that it offers its public region customers. “I think this is important,” he says. “On the surface, it looks very simple. But to achieve it, you have to make sure that every service and every feature we have in the public region is also going into every single dedicated region. It’s a lot of commitment, but that’s essentially our goal. This is not a side show.”
Whether it is to meet requirements on data governance, data privacy, compliance or, in some cases, to address latency issues, Vincent says some Oracle customers want a dedicated region located right next to their existing on-premise environment. Vodafone, for instance, runs six dedicated OCI regions across Europe. This, he says, helps customers migrate workloads without the need to take a big-bang approach.
Unlike other public cloud providers, where scale is a measure of the footprint of the regions they operate in, Oracle has a different scalability challenge. “We also have another dimension which is scale of the number of regions, which drives a different type of complexity across the board,” says Vincent. As such, he says Oracle is investing in engineering to keep this on track.
When asked about Oracle’s approach to keeping costs down, Vincent says its wants to minimise the overhead, which means running the least number of server racks. “We want to shrink our initial footprint for our regions and to be as small as possible. This is a cost both for us and our customers,” he says.
Multicloud operations
Looking at Oracle’s approach to multicloud, Vincent says it has moved beyond the OCI Azure Interconnect to allowing Oracle workloads to work in conjunction with Azure workloads. It now offers the Oracle database running alongside Azure.
“There are customers who, perhaps, are already in the Azure world, but they want to use specific Oracle services in conjunction with their Azure workloads, and in some cases, they want to do this using an Azure commercial relationship,” he says.
“We essentially built a small OCI region inside the Azure datacentre that has a direct networking link into Azure.”
Behind the scenes, he says, Oracle has provided a network cluster with remote direct memory access (RDMA), which offers a way for server nodes to transfer network data without involving the central processor. This technology is used to support its autonomous database servers running in an Azure datacentre.
Looking at the various options Oracle now offers in the cloud market to support organisations that want to run its databases and business applications, there is always the option to simply deploy Oracle software on public cloud infrastructure as a service (IaaS).
For instance, Oracle appears to have put in a lot of effort to create an OCI region in an Azure datacentre. Why not simply install the Oracle database in an Azure virtual machine? For Vincent, the main benefit is that Oracle’s approach offers a fully integrated stack. This is one of the reasons the Oracle autonomous database is only available on OCI, he says.