Chris Rogers VMware Explore 2022
Transcript
>> Welcome to VMware Explorer 2022 with the CTO advisor studio. Come on in and consume some content. >> Welcome back to our coverage of VMware Explorer 2022. " Day two of VMware Explorer, I can honestly say this feels like VMware Explorer. It doesn't feel like VMworld anymore, VMworld as we know it, ended in 2019 VMware Explorer, still a big conference- >> Chris: A hundred percent. >> Still intense, still huge investments, but the pre poll going it is a different event.
And I'm okay with that because the conversations are deep. We're about to go into one of these deep sponsored conversations with Chris Rogers, Evangelist at Zerto. Chris, welcome to the show. >> Oh, thank you, thank you having me. Yeah, I completely agree with you. I'm loving VMware Explorer, VMware Explorer this year and the conversations have been vastly different to we had at VMworld, right? We're talking about this multi-cloud world and it it's been great. It's been really, really fun so far.
0 or whatever the current version ones. That was the deep conversations. 3 actually mean or Tansu mission control? Or how do I integrate this theme with red head or Google cloud, et cetera? So we're going to go into not the application side of that conversation, but the I've talked to quite a few data protection companies and Zerto does do some data protection, but we're going to talk about workload protection because at the end of the day disaster recovery, ransomware recovery, etcetera is a big part of what this community does.
>> Chris: Yeah, absolutely. >> Zerto's been a part of that. I can't help but notice the commonality of VMware vSphere across all of the major cloud providers, not the way that we thought it would be maybe five years ago very different way. Each cloud provider has a different solution. So let's- off the bat, which cloud providers does Zerto recover doing a disaster recovery process or solution into their VMware cloud compartments? >> Yeah. Great question. So Zerto's really blessed in the fact that we are pretty agnostic in the way we do it.
We approach the data protection and disaster recovery. So Google cloud, VMware engine, Azure, VMware solution, Oracle VMware solution. All of those are supported, native Azure, native AWS as well. And we have 350 plus managed service providers who run every own VMware clouds as well, you know like OVH that are here as well. And those big companies that are doing their own thing as well as the public cloud providers. So we support all those platforms to give customers and users the ultimate flexibility to adopt that hybrid multi-cloud approach which is what this conference is all about right?
Like trying to intertwine your vSphere and your hybrid cloud and your Tansu and your Kubernetes all of those concepts into one kind of holistic view, right? >> Yeah, so we're not going to get into the details of minutia of how Zerto does orchestration. And DR, that's kind of a told story. We spent some time, a couple of years ago, doing some pretty deep research comparing the public cloud providers. >> Chris: Yeah. >> What I am interested in is this veneer of vSphere across the multiple cloud providers.
Once I'm into the console, once I'm into vCenter, vSphere is vSphere. Whether I'm in vCenter, that's being hosted in AWS or vSphere that's been hosted in Azure. It's vCenter. >> Chris: Yeah. >> Where I start to experience differences is issues that I think are very relevant for disaster recovery. If I have a VM that needs a public IP address, it's a different process to do that in each one of the clouds. And if I'm protecting let's say my DMZ in the private data center and I want to fell over to Oracle cloud infrastructure.
There's a temptation for me to say, "You know what? " >> Chris: Yeah, yeah. >> What am I missing? >> I think just making sure that you understand that each platform has those unique commonal-, unique commonalities but those unique differences throughout the platform. So SRM will work when vSphere to vSphere, right? But what happens if you want to start adopting native cloud technologies as well, or you want to start adopting Kubernetes on those environments as well, right? With SRM you're it kind of stuck with that platform right?
It's VMs only, it's two vSphere to vSphere, does an okay job of doing that right?. But if you start looking at the wider community and the wider world you want to get to, it's very unlikely you're just going to be choosing one platform or one a way of running applications right? Your new applications coming out next year. You're probably not going to develop that on VMs. You're probably going to use Tansu or OpenShift or somewhere like that right?. So if you go down the path of an SRM, for instance you're going to be tied into another solution then.
So going out and purchasing another solution a more complexity in that hybrid cloud world. Whereas if you chose a solution like Zerto, for instance then we kind of do the data protection in a different way. We try and make it simple and easy throughout all your hybrid cloud estate. So that includes Kubernetes applications as well. So we can do Vms, Kubernetes, in a very consistent manner. So people are understanding what the tool's doing and how it's working across those VMs across whether it's on Oracle cloud or native Azure or vSphere or any of the other platforms we mentioned.
So it's just bringing that simplicity back down cause running hybrid estates, isn't easy, right? Its complex, but if we can take the hassle out of the DR piece of that and make it nice and easy for people to understand then I think that's where the real value proposition comes from. >> So I guess I do have to penetrate this, this, layer of technology, weed, and talk through that a little bit. Cause when I'm thinking about protecting a Kubernetes workload versus protecting a VM, a monolithic application those are two very different things.
Like, getting the data, replicating the data for a persistent volume in Kubernetes and replicating a VMD K pretty much same thing, it's just, that just needs to get over to a degree, it's the same thing. But the orchestration of recovering that is different. So where's that- where does data replication kind of stop and Zerto start? I know Zerto does both, but what does that, yeah. >> Yeah. So I think it's adding all those layers around it. So yeah, as you said, moving the data from A to B isn't really the challenge that we solve right?
There's many tools out there, you know can manually yourself and copy and run a script to copy your persistent data from A to B wherever that may be, what the complexity comes is how are you going to orchestrate spinning that up? How are you going to keep multiple thousands of different checkpoints in check when they've all got just UID's? How are you going to understand which one of those is the one you need to recover to? How are you going to build analytics around that?
How are you going to understand what your RPOs and RTOs really are? How are you going to test all of those things? So that's where Zerto brings that whole kind of automation orchestration. And the simplicity back down to it is that we do the data replication for you as well. So using our CDP engine that we've actually put inside of you know, the ESX I host, we've now put that inside of Kubernetes as well. So we're using the same type of technology inside the Kubernetes stack to replicate their data every five to 10 seconds.
So what we're used to in our customers are used to seeing in their VMs with that super quick RPOs they're going to get in Kubernetes as well. But then we are capturing every single piece of the application that makes up that Kubernetes app at the same time as well. So it's not just about the persistent data. It's about secrets, it's about services. It's about you know all those other pieces that make up applications. And the way that we've designed our Kubernetes data protection is actually with developers in mind.
So instead of it being breakout into UI and then protect everything afterwards, we're doing data protectionist code. So inside of those Kubernetes applications where you're actually configuring Kubernetes, tell you how many pods and how many things you want to restart and all those types of things, that's where you're actually annotating your application to say protect my applications Zerto. So soon as Kubernetes spins up those pods and those containers they're born protected with Zerto. So it's never an afterthought and we are trying to make sure we make it as simple as possible for those developers and the DevOps guys to kind of adopt data protection.
Cause we've seen, you know, containers have taken this massive up curve in the last five years or so when they came from the really upstart quick start businesses right? But now as we start moving into enterprise we are seeing compliance come into it and having these data protection strategies. " I'm a developer, I want, I want to develop applications. So if we can give them the tools to say it's one line of code to protect your whole application, just use that across your whole application group it together.
We capture everything that you want inside of that protection group. And then we recover the whole thing as one, not just persistent data. >> So let's, let's penetrate that protect- protection group a little bit. You know, we have, have my state information from a database perspective. So my application information, I have my services my control plane, my Kubernetes control plane. What people don't think about a lot is that, yes, I can have auto scaling, but the application control plane is always constantly running.
It's rare when I have to restart my entire application cluster, that cluster infrastructure is running. So the dependencies, is the database working? Is the persistent volumes mounted? That sequence of recovery. >> Chris: Yeah. Is Zerto orchestrating all that in the DR location? So recovering my control plane, and my data plane? >> Chris: So- >> In a way that I can actually recover the state, the app in a stateful manner? Because the microservice may be one UID and then the database, another UID.
>> Yeah. So we're capturing. So every time we do a checkpoint, we're capturing the state of everything that you've tagged inside of that protection group from that point in time. So every time we create those five or ten second intervals we're capturing the state at that point in time. So we are then recovering the whole Kubernetes application to that point in time throughout the whole application. So we're recovering from five seconds 10 seconds previous to that whole application. We need a Kubernetes somewhere to stand it up right?.
So we are not doing DR for the actual Kubernetes runtime or anything else like that. So you'd use EKS, AKS, Tansu, OpenShift, whatever it is and replicate. And again, being agnostic, we can go between those as well. We don't have to choose one or one or the other. So we're doing data protection a different way instead of being tied to a platform or a storage place and using snapshots, trying to take it, you know, to the next level. Like we did a bit with like hypervisor replication.
" We're doing the same thing with Kubernetes and actually saying our customers needed that for VMs. Kubernetes is a completely different monster right? The containers change all the time and sometimes VMs last for decades right? Sometimes containers last for seconds right?. So how do we make sure we're protecting the data that those are containers are creating? Well, CDP is the answer to that right? So making sure we can keep all those checkpoints in sync recovering, the whole app is one cohesive unit.
>> So let's zoom back out and let's talk about that portability across public cloud providers. So today I'm replicating my VMware, vSphere, VMs. Let's zoom, let's even zoom out further than that. I'm replicating my application >> Chris: Yeah, yeah. >> From one data center or cloud provider to another data center cloud provider using Zerto. >> Chris: Yeah. >> We talk about those nuance and differences. How portable, either from a vSphere or Kubernetes perspective, is that blueprint? Let's use the term blueprint, that I use.
So I've set up my application replication, and data protection policy and orchestration for AWS. >> Chris: Yeah. >> And I want to move that process now from AWS to Oracle cloud, how difficult of a migration is that? >> So with Zerto, very easy. Moving the workloads we have a, I think it might be four clicks fail over to be able to move that application into there. And then essentially once you've moved that data into the new cloud environment, using Zerto it's going to look and feel identically the same right?
We it's the same buttons, the same workflows. And then we know which platform your VMs are running on or your Kubernetes contains are running on. So we can then just subtly change what we're asking you and what questions right?. So if we're moving inside of Azure, for instance we're going to be asking you about V nets. If we're going into EC 2, we're going to be asking you about availability zones and those. So we're going to subtly change the questions we ask you in the workflow, but it's going to feel the same to a user.
They're going to be able to answer those questions without really thinking about it. But just depending on where you are landing as we're going to subtly change the workflow, to make sure we capture the information we need to recover that data into where you're going to. But for Kubernetes, we don't even see what platform you're really running on because we're in the K8 stack itself. So actually if you're moving from EKS to AKS to Tansu, doesn't matter to us, we don't really see the underlying platform, we see inside the Kubernetes stack.
So it'll be the same annotation, the same tag inside the app inside of the Kubernetes application itself which can then re-protect that application across, so. >> So, last question, we've talked a lot about the experience. >> Chris: Yeah. >> What we haven't talked about is how is this packaged? Like, what is this? Is this a, is this a box? Is this a piece of software? Is it a SAS offering? How how's it packaged? >> Chris: It, it's completely software only right?
So I think the multi-cloud world is becoming becoming harder to kind of justify hardware, right? If we're, we want to be agile we want take the data where we want to, and kind of spin the- we don't want to exit data centers. It's going to take 24, 46 months. So being software only is a great advantage to us. And so software only, licensed on a per VM basis or in K8s on the amount of CPU or work nodes use. So again, we we've had that licensing model that allows users to scale the way they want.
So instead of saying per worker node, which usually means people have big worker nodes but fewer of them, right? We're saying, well, we are going to do it on a per CPU which means you can scale your work nodes, how you want. You can have hundred with two CPUs or you can have one with 200 CPUs. It's going to cost you the same either way, right? So it allows you to adopt Kubernetes or how you want to architect, not based on licensing but based on your best practice how your company wants to do it rather than what's going to save you the most amount of money.
>> Well, Chris, I really appreciate you taking out the time to come and talk to me about Zerto. I've asked you some pretty tough questions. >> Chris: (laughter) You always do Keith, you always do, that's what we're here for. >> If you don't think I've asked Chris tough enough questions. And frankly, I find that hard to believe, you can DM me on Twitter @ctoadvisor you want to learn more about the CTO advisor. com Make sure to tune in for more VMware Explorer content.
We've just scratched the surface. We went deep unexpectedly into Kubernetes and data protection and workload protection. Who knows what we're going to go when we talk to many more people from VMware themselves. Talk to you next episode of CTO advisor studio at VMware Explorer, 2022.
