As cloud computing continues to grow – seemingly with no end in sight – a key challenge faces businesses: how do they expand their cloud’s functionality to better enable today’s ever-growing workload? That is, how can they run their cloud in the many places where it is now necessary?
To address that concern, Oracle has expanded its hybrid cloud portfolio with a new offering called Oracle Roving Edge Infrastructure, or RED. At the core of the RED solution are server hardware nodes that bring the cloud everywhere. In other words, this a hardware-software combination that enables advanced edge computing. If, for instance, an enterprise team needs cloud support aboard a plane, in a remote location like a desert or any mobile deployment, a RED hardware device can enable that cloud connectivity.
In a world in which the cloud must be everywhere, and where the border between what’s “cloud computing” and what’s “edge computing,” is dissolving, RED is a natural expansion of a hybrid cloud or multicloud platform into the growing sector of edge computing.
To find out more details, I spoke with Ross Brown, VP, Oracle Cloud Infrastructure. In addition to discussing RED and Oracle’s hybrid cloud offering, we delved into key cloud trends and potential future directions for cloud computing.
Download the podcast:
Watch the video:
Edited highlights from the discussion – all quotes from Ross Brown, VP, Oracle Cloud Infrastructure.
Cloud Market Overview
“So at a very high level, I’ll give you a way to think about the market that we look at, which is you can look at hybrid computing options first in the size and capability of the deployment. Is it a small device or is it a big, giant hyper-scale region? And then you look at it in this other control plane aspect which is, “Is it a self-contained device where the control plane, the data plan are running locally? Or is it a remotely tethered where the device is separated from its control plane, so remotely managed?
And if you think about how that builds out, you end up with hyper-scale cloud regions, single server disconnected Edge devices, and there’s now remotely tethered computing storage appliances and what not.
And our competitors have different offers that come in to the different parts of this map. For OCI, our public region, we’re at 29 regions live now. We’ll do more than nine this year. We’ve publicly announced the locations of the nine expansions, but more will get done this year. And then six of these locations, we have a connection between the OCI cloud tendency and the Azure instances that are located in the same facilities, where we can give customers a sub 2 millisecond connection between those tendencies for being able to do cross-cloud computing.”
Snapshot of the OCI Roving Edge
“So the news is we’re introducing the OCI Roving Edge device, the Oracle Rover. The Roving Edge is a single or it can be clustered in 5-15 node configurations disconnected Edge device. And what I mean by disconnected is, it’s provisioned through the cloud, you load cloud applications into it and when you receive it, you can begin operating it in a location that has no network connectivity, low latency network connectivity, or has full bandwidth connectivity. But it will operate without necessarily needing a connection back to the cloud.
So it’s in a ruggedized case, it’s portable. As I mentioned, it’s scalable from a single device to a minimum cluster of five nodes up to 15. It is low latency, high performance data processing. And it creates a very consistent experience around what people do in our cloud. It’s provisioned through our console, it’s managed through our console. All the VMs and image data sits within your tendency. So think of it as a physically separate but still logically connected aspect of your OCI tendency.
It’s a 40 CPU or 80 virtual CPUs, with the way we do vCPUs, 512 gigabytes of memory. We put it in Nvidia T4 GPU unit for helping with machine learning refinement model applications, 61 terabytes of NVMe storage in it for the block and object. And it’s done as a flat fee of $160 per day of possession per node.”
Marc Andreessen famously said that “software is eating the world,” but it turns out that, lo and behold, hardware still plays a pretty critical idea.
“On software eating the world, I think the innovation in software around what we do with data has been incredible, but that innovation has been rooted on top of an assumption about hardware, and I’ll pause for a second and say, I think AWS and Azure and other cloud providers have done a great job building amazing cloud services, but all of them built it on top of a thing that existed for other purposes. AWS is built on top of their e-commerce engine, and Azure’s built on top of GFS. Those existed to run gmail or to run Xbox Live and Bing or to run e-commerce. Oracle built our cloud with the notion that we can re-innovate in hardware.
So instead of you having to adapt your applications to what is considered cloud, we could engineer a cloud that adapts to your applications. And I’ll argue that Oracle is the first bespoke cloud that was built specifically to be a cloud, as opposed to built for something else and then we put cloud software on top of it.”
Cloud’s Future: What Should Be Companies Know?
“Innovation and hardware matters. The innovations that we’re doing in the underlying infrastructure will give benefits for what you can do in the platform on top of it. I will say that is a broad five-year statement.
Things you can do at the platform services layer are driven by what you do in the infrastructure. And the better your infrastructure is, the more performant, the more scalable, the more capable, the more it opens up opportunities for better performance in those things.
And so we focused an awful lot on our infrastructure as a service since we’ve started, because we think there’s innovation still remaining to be done there, and we want to lead that. That doesn’t mean we’re not going to invest just as aggressively on the platform services, but having this foundation gives us so much more capability for what we can do there.”