Oracle Cloud and 3rd Gen Intel® Xeon® Scalable processors (Ice Lake)

14:26 · Watch on YouTube ↗

Transcript 2,222 words · about 15 min to read

Auto-generated captions from YouTube, not hand-corrected, so names and technical terms may be imperfect. The video is authoritative.

(Introduction music) >> Keith: Hey, it's Keith Townsend, principal of the CTO advisor. And welcome to the sponsored episode of the CTO Dose. We have Clive D'Souza, who is a returning guest. Clive, welcome back to the program. >> Clive: Good to be back Keith. >> Keith: So it's been about a year since we did our VMware cloud bake off. I don't know if bake-off is a fair description cause we didn't pick a winner. We kind of educated the market around the overall landscape of the various VMware cloud solutions and Oracle sponsored that piece.

You manage that solution. You're the general manager and senior director of the VMware cloud solution or VMware solution for Oracle cloud VMware solution. Tell me, what's the update since we ran our report. 7 support for the VM. What we have now done and in fact we were the first cloud to actually support the Vphase Additionally, what we have done is we've got an optimized operational mechanisms. When we launched a solution in August of last year, Customers needed about roughly about a little under hours to put a initial software defined data center with three nodes.

With the optimization we have shrunk that now to close to two hours and we're well on our path to further shrink that by another 50%. We have also taken our time taken to add in new node to an existing data center, a software defined data center and we have shrunk that for 25%. So, from an overall user experience perspective for our customer to add a new note or expand sheet a new software defined data center, we brought the time down 33% and 25% respectively.

We've also gone ahead and now supported STS enterprise for customers who want to bulk migrations. As STS standard comes as part of our VMware cloud foundation. We now support STS and outbox literally a checkbox and NCX gets extanstiated for the customers in there and the tendencies. We have also done our preview transfer and we support that in preview mode. And we, one of the launch partners with me and, and and the thing is, you know, as we look at the roadmap we have more and more architectures being implemented.

You know, we talked about isolate. You're also looking at offering different radiation in combination. So a lot more is going happen in this space. I was pretty excited about what's coming down the road here. >> Keith: So Clive, what else is new since the past year? >> Clive: As you rightfully said it's been a year and it's been very eventful. We have gone mainstream. We obviously we jade the solution and gave a very healthy customer update of your large enterprises, such as Ramon logistics deployed their entire data center migrations onto OCI.

We also have the largest telephony providers and mobile in Brazil, who are shutting down the data centers and moving all their on firm marginalized work roles to VMware solutions So it definitely seeing the large enterprises adopt and are gravitating towards, you know, the ASEC and nutrition. So it's been good. So we've seen a very positive interest and we have tens of customers in multiple thousands right now to be honest, we have workloads you know which are few thousand shipping deployed globally, all over the world.

So we are using the company call right now. >> Keith: So then big customer focus. And when we talked last, you know we talked about 56 cores in a single platform, big Intel footprint for each individual bare metal workload. Have customers reacted to this? You know, let's give them these super powerful, super big workloads compared to some of the other solutions on the market. >> Clive: You know the customers are actually reacting really well to that. What we're finding in most of the on-prem workloads, especially when you move from an on-prem data center there is a level of consolidation.

Right sizing applicant downs of the actual workloads moving to the cloud. The other inherent, native benefit is because we use an X7 which is an Intel's Xeon processor From our native fleet, it's not a custom shape just for the VMware solution. Like all the other clouds do. It comes from a native fleet, the fleet management becomes super easy for them. And we are also finding that the number of nodes being a customer needs to managing the data center, it gets simplified because you're seeing consolidation from an on-prem where they have a hodgepodge collection of different mode, configuration, different sizes.

They all get consolidated and the overall management, the whole orchestration it's becoming a lot simpler for the customers. So from that perspective, selecting the, you know the 52 core Xeon CPU with the NVMA attached to it in many ways, you know, it's, it's shaping up with the right decision and you seeing customers actually warm up to that. >> Keith: So thanks for correcting me 52 cores versus 56, 4 extra cores matter. Well, that topic of Intel big CPUs. Talk to me about this relationship.

You, you, you guys specifically seem to have doubled down on the Intel relationship. Tell me about the importance of the Oracle cloud Intel relationship especially as it pertains to VMware customers. >> Clive: So if you look at pretty much any larger enterprise, and if you look at the technologies, you know, VMware, Oracle, seem to be in common. So in many instances what are finding is that large enterprise have optimized the VM rates translations, in addition to the other deployments, are very unique in Dell offerings.

The way we offer our orchestra, it's a bare metal solution. So we literally give access from the east side operating system hose. Anything negative. the Intel allows, we let them do. So that level of granularity, not many cloud vendors can do. Second is, because we give them the complete access to the environmental, customers have full control of the level of performance and scale they can get. Keep in mind, they're not over provisioning. We don't give them a slice of the CPU, or a slice of a VM.

A whole note is dedicated to that. So for customers, when they move, they're moving their tool set they're moving their policies which they have hardened on Intel on prime or hard drive, whatever. And they convey as it is with very, very minor tweaks but literally they can take an on-prem recommendation point it to the cloud, use V motion and the whole environment boards over. And literally the focus then becomes I move the workload. That is literally the least amount of my accounts.

I'm going to focus now on innovating, you know, customers we see customers come in, use the VMware implementations. They, you know, the stabilize, the migration and to start drawing high-performance compute nodes they're starting to deploy autonomous databases they're starting to deploy more cloud native service sites. We'll be seeing more of that happening. So in many ways that migration is predicated based on their familiarity with the Intel architecture which they had on track. And now they know that they don't have to go and literally go down the stack and tweak more things because something changed.

So we give them that ability that nothing changed. It's just VMware, like we said, it's just, VMware which was on-prem. Its normal to the cloud. So it's resonating really well. >> Keith: So I get two things out of that, that I really want to dive into. The first is we had Rebecca Weekly who is in charge of the hyper-scaler relationships. So you know her well. And when I challenged her on how do we expose or as customers, how do we leverage a third generation Xeon scalable processors in the public file.

She specifically pointed to a ISV such as VMware and the cloud providers as a key component to leveraging isolate the, the cold name for the third generation Xeon, scalable processors what are customers and not just customers. What is your team excited about when it comes to isolate and OCVS Oracle cloud VMware solution? >> Clive: Well, strictly speaking from what isolate brings, right. We've seen the performance update, you're seeing how they've done the hyper scaling and hyper threading on and can give already have we have isolated the labs.

We are paying with it to watch the space very closely, very soon we will give you an update on maybe. So coming back to, when you look at the way the cloud workloads are designed and the way Isolate architecture has been designed you know, they have dedicated core pinning is not becoming possible. And they can actually dedicate very specific compute cycles based on 1:00 PM new workloads. When we talk to our customers what you're hearing more and more often is this whole construct of right-sizing the work.

When they're moving from a on-prem where they have a combination. And sometimes they even have mixed architecture where they want to consolidate. Isolate in combination with OCBs will give them the plan. We can actually reduce total cost of ownership over. They have a disparate set, of course, keep in mind VMware has a licensing mechanism based on number, of course, we can consolidate the codes. We can have the density. And because the way we designed the architecture you can take multiple disparate architecture multiple disparate workloads you have, in different data centers different implementation, and a hybrid cloud and consolidate all of them into one, you know, fully customer control and customer managed software defined data center.

So that level of being able to consolidate the same and get the best performance for the price is very appealing to customers. Especially when they're trying to consolidate their data. That's one aspect of it. The second aspect that I mentioned is because it's all standardized and Intel. Their mechanisms, all the governance enforcement policies everything can be used in space. So them having to re-architect their migration journey to the cloud is very minimal. They focus more on doing other things.

So that is where we are seeing the customer getting excited about and most of the performances, right? Especially, when we reduce the number of nodes the level of performance is very visible and people see it. Customers are seeing it. It can actually drive high level of throughput. We actually see then we actually giving more in terms of proof. or current Oracle cloud VMware solution customers perspective, right? That's still a mouthful to say, if I'm a customer >> Clive: OCVS >> Keith: OCVS If I'm an OCVS customer and I want to migrate from the current platform to new workloads with the Z, with these third generation Xeon scalable processors, it sounds like this is just you know, I can pick my solution that can stand up but another close cluster and just do a V motion is it's VMware vSphere underneath.

>> Clive: That's right. Literally is, and you can mix and match them. You want to get the broadband B part or you want to get a nice steak depending upon what your workload you want to do for that specific mode, or you want the workloads which are very unique to a specific instantiation of Intel CPUs, you can mix and match them. They're going to allow it. In your software defined data center. Your nodes can have different configuration. So we are heading down that group again.

But the key point, Keith here is what we heard from customers is there is a level of flexibility they have. They are not looking for a complete homogeneity and you know it needs to be only a broad real part or only a high stand part. You know the total cost of ownership So we realize that. So we said, you know what? As long as you are able to get in node into the software defined data center, we will have that. We allow you to merge different portions of VMware at the same time, change up all the underlying intel CPUs We allow customers to do that.

>> Keith: Clive, Thanks a lot for joining the program. If you want to find out more about the Oracle and Intel relationship click the link associated with this video, you want to find out more about the CTO advisor and download the report that we did with Oracle or sponsored by Oracle. com and clicking on a project button. You'll find that Oracle important list of various proof of concepts. We've partnered with various enterprise IT vendors to do in the CTO visor hybrid infrastructure until the next CTO dose, You can follow me on the web @CTOadvisor on Twitter, Dms are open.

Challenge anything that Clive said or anything that I've said. I would love to have a conversation with you about your hybrid IT infrastructure needs. Talk to you next CTO Dose.