Enabling Hybrid Cloud Anywhere with HPE CloudOps Software

Brad Parks, um, with, uh, HP's tech acceleration team under Hybrid Cloud, uh, representing cloud ops and the private cloud portfolio. I've got a couple of friends, uh, local and on the line to help keep me outta trouble. Hi guys.

My name is Juden Sapa. I'm a solutions architect with Ops ramp. Hey everybody.

I am, uh, Sean Jabber. I'm the director of Enablement with Morpheus for the last five years, and now with, uh, hp. Perfect.

Thank you guys. What we want to talk about today, I wanna dig into the cloud ops portfolio. So this is the really, the combination of a few different products all aimed at helping customers navigate the complexity of hybrid cloud, right?

Heterogeneous runtimes, tool sprawl, process sprawl. This is the, the difficulty of getting to a cloud operating model. And it's what HPE is delivering with the cloud ops portfolio.

So this is the combination of Morpheus ops ramp and Zerto from a software perspective. So this is the, the control plane that is letting us help customers address that complexity on that previous slide. But I wanna dive in, uh, into the tech a little bit.

So, Morpheus data, uh, been around for the last 10 years or so, friend of cloud field day. Our goal in life, our primary function is to deliver on the self-service component of what makes a cloud a cloud, right? The ability to go into a unified catalog of services, a unified API with 1300 commands in the command set, a unified Terraform provider, a unified plugin for ServiceNow, go to one place and then provision VMs containers application services or multi-tier application stacks in a consistent way into absolutely any downstream environment.

That could be bare metal, be hypervisor, could be a Kubernetes cluster, could be a public cloud, native PAV service, consistency of that provisioning experience, and getting all of the dependencies wired together, right? Think about what it takes to provision on multi-node clustered database. It's not just cloning a VM in, in vCenter, right?

It's needs an IP address, needs a DNS entry, need to clone a vm, run an Ansible script, right? Installing a observability agent, create a backup job, assign cost, all of that end-to-end work and all those dependencies is what Morpheus was born to string together. Second, though, within the Morpheus software stack, we have a built-in cluster management engine.

'cause if we're focused on provisioning workloads, those workloads have to run somewhere. Oftentimes when we come into a a customer, we're attaching to an existing, you know, vCenter cluster, a Nutanix cluster, a Microsoft cluster, or a public cloud account. But a lot of customers said, Hey, provisioning, Kubernetes is kind of a pain.

Can you guys build that into the stack? And so we built a cluster management engine years ago to help us provision things like Kubernetes cluster. So we have our own CNCF certified distro.

We can also provision the public cloud managed E-K-S-A-K-S-G-K-E, or attach to external existing clusters and provide some base cluster management. One of the other core pieces of the stack that works is KVM. We've used KVM forever as the core primitive when we were developing a cloud integration.

Um, our engineers, big fans lib big Linux nerds, right? We would rough out a cloud integration using KVM and then model that and apply it to the different public commercial hypervisors that we, uh, integrated with. Third core piece of the stack is attaching meaning to all of those activities, right?

In terms of cost, right? So we can create t-shirt sizes, assign costs to all these activities, give you a unified chargeback showback model for all of the activities of it across all your on-prem and public cloud environments. We're pretty opinionated on the lifecycle, what it takes to provision an app.

'cause we've done it a few million times over the years, but what we're not opinionated on is the tooling that it takes to get there. 'cause every enterprise, most of the SVPs of it, I end up talking to say, my team's never met a tool they did alike, right? Over and over again.

And so we want you to be able to bring your, you know, tech debt, bring your tools into the platform, and give you a meaningful place to organize and automate all of those handoffs. The other thing we're agnostic about is the downstream endpoint. So I talked about integrations.

We integrate out of the box with dozens of tools, dozens of cloud endpoints, but we also have a plugin framework. So the entire platform is pluggable. So if we have independent vendor who, let's say IAM or DNS or ITSM that we don't integrate with, it's an open spec.

We have example code out there. We even have a builder library to get you started. So we're building a robust developer community as part of being part of HPE, we're building relationships into a much broader network of ISVs to help facilitate integration.

But I mentioned, uh, I mentioned KVM and it's something we're gonna talk about in the next session. But this stack, right? We have had a built-in, you know, KVM stack as part of Morpheus for a long time, but we never monetized it.

We never did the final polish over the years. 'cause we had plenty of, plenty of business just making VMware the private cloud. It should be.

But over the last year and a half, two years, we really have put the work into that builtin KVM stack to do things like host to host motion volume, motion for storage, affinity and anti affinity rules, um, uh, distributed resource scheduling, all of the core capabilities you need to land a production workload on a KVM based hypervisor. So that's something we've built into the stack. It's something we spun out as a version of Morpheus last year called Morpheus VM Essentials.

We're gonna talk about that in depth in the next, uh, the next session. But nothing in our KVM stack is new, right? All of this tech has been around for a long time.

KVM, as you guys know, incredibly widely deployed as the underpinning for a lot of the public clouds. You got Nutanix out there, nothing new on KVM. We're just bringing it to the forefront as part of HPE.

Now, talking about, um, day two and Ops ramp might actually let you, uh, dig into this one and do it justice. Cool. Thank you, Brad.

So, ops ramp and Morpheus is really better together, right? So you would use Morpheus for your day zero day one operations. And after you've, you know, used Morpheus to streamline your deployment, um, of your infrastructure, now you wanna focus on day two plus operations that, you know, talking about the observability and the, um, monitoring of your infrastructure.

Um, and that's where OPS RAMP comes in. So, HPE acquired ops ramp about two and a half years ago, um, but prior to HPE acquiring ops ramp, they were a big user of the platform. They eventually became an investor and they were heavily invested in it.

And so they said, Hey, why not just, you know, sprout, right? Just purchase it, right? So, um, ops RAMP has always been vendor agnostic, protocol agnostic, cloud agnostic, right?

Which is why it was ideal for HPE to acquire. Um, so there's, there's three major components of what the ops Ramp platform does. The first one is its ability to do that hybrid discovery, observability, and, and monitoring, right?

And I'm sure you know, everyone's familiar with monitoring and observability, but just as, as a refresher, it's all about, you know, discovering those, those assets, those compute, network storage, virtualization, containerization, dis discovering all those assets, pulling in all those time series metrics, pulling in any events and alerts. And the option app platform can also do log ingest, trace ingest. And you know, as we ingest all these logs and these traces, our ops ramp provides you a kind of a unified query language you can use to kinda, kinda analyze and inspect your logs and traces, and you can alert on anything you see on those logs and traces, right?

So anything related to observability and monitoring, we wanna just, you know, put it all into the ops platform. So whether it's it's full stack compute, network storage, virtualization databases, applications, um, on-prem in or in in your data center or in the public cloud, we support AWS Azure GCP, we even support Alibaba Cloud and Oracle Cloud. This is Kimberly.

Are you expecting the storage admin, the network admin to use ops ramp for their monitoring and observability or Yep. Are they, so they're not using at all the traditional, um, control plane that they have on the, the Like domain specific monitoring versus Yeah. Yeah.

So, so what we're trying to do is that what we see with a lot of our, um, potential customers is that, you know, everyone has monitoring, everyone has observability, but the thing is, they have a siloed view of it. When you talk to the networking team, they'll have a handful of observability and monitoring tools, but it only sees the network. If you talk to the storage team, they'll have their set of tools.

You talk to the compute team, right? Everyone has a siloed view of the infrastructure. And so what happens is, let's say you have a root cause issue, like in maybe a switch port, a switch port going down on a switch, the network team will see all these alerts, but it's gonna cause cascading failures.

You'll, that network issue is gonna cause an issue on the compute side, on the virtualization side, on a database. And so now you have, um, different people looking at basically the same type of problem, right? The root cause was that network port going down, but you know, the, but it's gonna cause 5, 10, 15, uh, other alerts in the stack.

And so you're gonna have multiple people trying to triage all these alerts that are coming in, and they're not knowing that the root cause was that network issue. So the ops ramp platform will give you kind of a un unified visibility into the infrastructure, what's going on in the compute network and storage. Because as these alerts are coming in, SRA can use our AI ops, our machine learning to kind of correlate all these alerts that we're seeing from the compute network and storage and will determine probable root cause.

So instead of having to triage all those alerts that are coming in from that one root cause issue, we identified the root cause issue. And now you focus your time and attention on just that alert. And it should in theory now fix all the other alerts that it, that it caused the cascading failure.

Does that make sense, guys? Mm-hmm. Right.

Um, so that's the, the ultimate aim of the platform, right? A unified view of everything. And if you go to the next slide here, the other use case that ops ramp can handle is event management.

So if you're using Ops ramp to do your native discovery and monitoring and observability ops ramp will be, you know, generating these alerts. But Ops Srap can also ingest these alerts from existing monitoring platforms. If you're using SolarWinds logic monitor, science logic, whatever you're using, you don't necessarily have to go in and rip and replace your existing monitoring tools.

We can use a web hook to ingest the alerts that those tools are generating. And again, whether it's a native alert that ops shop is generating or third party, we can correlate those alerts across the, across the silos. So that's what we can do for event management.

Okay. Now, for incident management, we have an integration with whatever ITSM platform your customers may be using, whether it's ServiceNow, sharewell, JIRA, right? So as these alerts are coming in ops ramp could potentially auto ticket on these alerts, we'll create the ticket on the ops ramp platform.

And with our integration, we can push it out into your IT TSM. Okay? It's a bi-directional integration.

So if the tickets originate on your platform, it'll get pushed into ServiceNow. And regardless of where you make the update, whether it's in ServiceNow or within sra, it's, it's bi-directional, so it'll sync up. Okay?

Um, and so that's our story for event and incident management. And the third piece here where we can tie in is intelligent automation. As we see these alerts, we can potentially run automation to try to remediate the issue, right?

Take the human element out of the picture. That way you can reduce your mean time to resolution. Okay?

This is actually one of the areas we've, uh, we've looked at from a, from a loose coupling on the current integration stack, right? So Morpheus today, right? We're in the control plane, we're talking upstream till we get repo mounting all the task task types, Ansible, chef, salt, puppet bash, whatever.

Customers have to do that remediation or to do that care and feeding of the infrastructure level, we can actually host those as operational workflows in Morpheus, right? OP ramp can detect an incident and can actually call those operational workflows and we can then execute and remediate. So that's one of the places where these are already starting to come together.

Similarly, this observability data, um, I don't think I have it in the demo environment I'm gonna get into today, but you have a UI tab inside Morpheus that actually shows a subset of that observability data in context for the database analyst or the application developer who's using Morpheus to provision their workload. So these are starting to come together across these three areas. Question?

Yes. Um, on the observability side, have you guys, are you guys looking at building any, um, user experience monitoring capabilities into the offramp platform? Words that already there, um, Talking about like, uh, like real user monitoring rum?

Yes. That, that has been, um, uh, it's been on our roadmap for a while, so yeah, it is definitely on our roadmap and we'll hope to roll it out in the very near future. Yep.

And I think that's, um, okay, so this next slide here just kind of shows are, um, you know, 2,500 plus integrations. So we use these integrations to, you know, pull in those time series metrics, pull in those events and alerts, and obviously you can see we're, we're vendor agnostic, protocol agnostic, cloud agnostic. Yeah, we can hop in.

I think this other was more of an architectural diagram, But, and this is the last slide I'll show. Um, so this kind of shows the architectural diagram. How does ops ramp gather, you know, that data.

So on the left side of the screen here, for infrastructure running on-prem, typically we'll deploy what's called the ops ramp gateway. This is a virtual appliance. You can consider this like a collector.

So using this ops ramp gateway will discover your networking devices, your router switches, firewalls, VPNs, um, we'll discover your storage infrastructure, virtual infrastructure, and we can also do synthetic monitoring using this ops ramp gateway, right? That's gonna be like agentless monitoring. On the top left there you see servers.

For servers we can monitor in two fashions. Again, if we leverage and use our ops ramp gateway, we can monitor your Windows or Linux servers via the gateway. That's gonna be agentless monitoring for Windows.

We'll connect via WMI for the various Linux distributions. We'll connect via SSH. Okay?

Now, optionally, you do have the option to download an Ops Ramp agent and install it on your Windows or Linux distributions With our agent, it's a lot quicker and more secure to do monitoring. And plus you can leverage additional functionality like automation. So, for example, you know, if you're monitoring, let's say, a SQL service running on Windows, if that, you know, SQL Server goes down, you can try to restart it by, you know, executing a PowerShell script.

So as automation using PowerShell Shell or even Python scripts, you can, you can utilize, um, with our agent, because of our automation capabilities, we can also automate patch management. Okay? So, so D**e for, so the agent, the agent give you sort of an active role versus no agent is more past collection.

Is that fair? Or are you, or is when you started talking about the agent, you talked about being more, uh, quicker Or something Yeah, Speedier. But is that speedier action or is that, um, more real time data collection?

Um, the, the, the type of data, it's, it's still like, you know, we're monitoring at a like one minute granularity, right? So whether you're doing agent or agent list, it's still about one minute granularity. The agent just provides you like additional insight into the operating system, right?

Because we have a little agent running on there. So like on the Windows platform, you'll see like the services that are running and you could potentially start or stop these services. Now It gives you a wider band Yep.

But then also gives you active Role. Yep. Yep.

And so our platform has role-based access control. So obviously you wanna restrict who has access to do what. And when you install our agent, you can even say, look, I wanna disable that automation capability.

I wanna disable certain features for added security, right? So there's multiple levels of security on our platform. That's probably one of the areas we're looking at from an integration perspective.

And I'm gonna steal a seat here and drive a little demo is, you know, can we bring, we also have an op, uh, optional agent on the Morpheus side. So we've talked about bringing obviously those together. So we're doing some of the same things in the same place, but also, uh, from a catalog perspective, if I flip into, into, we can make sure by policy machines that need that tramp agent installed, you know, it just happens outta the gate.

That's true with any, you know, security agent, other things. So bringing these together just to make sure it's one, uh, contiguous workflow. So I've got about, I think five minutes or so if I'm keeping, uh, keeping timing right in my head.

So it's just gonna be a quick, quick show Intel and, uh, Sean jump in if I, uh, I miss anything key. But this is the Morpheus ui. Um, so we are not actually SaaS hosted, that's a, you know, a core part of the platform, and that's intentional.

We, uh, do extremely well in highly regulated and disconnected environments. Government, finance, healthcare, where people want their control plane for all of that data pretty close to their data. We can install up in the public cloud, uh, reach down into the, uh, into an on-prem environment.

So it's really deploy it how you need it. We also have a ton of service providers who do provide a hosted managed version of Morpheus. And over time, that's something we'll be looking at, taking some of that business logic and actually putting it up into the, the Greenlight Cloud platform for things like updates.

So that's, uh, engineering work that's happening real time. Um, we talk a lot about provisioning, so it's probably the, uh, the one of the best places to start from a, uh, from a demo perspective, if we think about that, you know, death march across different handoffs for different tools that can take, you know, hours or days in a big enterprise. You know, provisioning something like, uh, you know, MySQL database is probably a good one to start with.

So Morpheus very focused on the application, less about the underpinning technology. So we represent application services first. I need my database.

This could be a container based, image backed, uh, reference. Could be a VM backed, could be DynamoDB up in the public cloud. So I need my thing like, click next.

I'm a member of a group. So we talked about RAC. We have an incredibly granular multi-tenanted, um, RAC models.

So I, as a user, I'm a member of a group that informs which clouds I can provision into. And a cloud for us is anything that we can touch and turn into a cloud. So I'm not talking only AWS or only GCP could be a bare metal provider, could be your hypervisor.

We are creating that cloud experience. Um, I've got dozens of policies I can apply. That's just an example of a naming policy.

So, uh, lab guys don't want me naming this after my, uh, my cap click next we can conversion software. We can maintain updates, we can integrate into things like your build pipelines. Um, mentioned we've got either a Docker or a VM back.

I get a t-shirt size a plan around pricing. And, uh, the platform teams that we work with can fully create all their own price sets. We can also, if I were provisioning into the public cloud, we're gonna download and synchronize all the public cloud data on a nightly basis, including your pricing plan.

If you've got an EA agreement with one of the hyperscalers, we're gonna knit all that together. So you get a, a single place to go assign cost to this activity. Um, and we separate cost from price.

So as an IT team, I might have a certain price for an EC2 workload in the public cloud that I've negotiated. Let's say I wanna add software markup or services markup for my own time. So at the end of the day, I can do chargeback or value assignment.

Those are concepts that we bring into the platform, uh, having to do with that service provider backend need part of a resource pool. I'm gonna select a network, and this is actually gonna go out and grab an IP address from an IPM tool. So starting to bring those integrations together in context, I can mount my user config in here, um, hitting next workflows.

So I, uh, I am managing the state of this database, right? So I can assign tasks to trigger every time this machine is provisioned, started, stopped, torn down, or reconfigured all that manual work that usually is some poor IT guys punch list for the day or the week. We just build into the life cycle of this application.

So it really is a force multiplier for a lot of teams and jump in and ask questions as we go. Um, you know, if you have any. Um, but here's where some of those integrations start to, to come to bear.

I can actually deploy a database schema or a website. If this were a WordPress instance, I can set up a backup job integrations with things like Veeam, Commvault, Rubrik, others. I can set up a replication job.

Here's where that Zerto integration I mentioned already existed. So at provision time, if I need to make sure this is replicated, 'cause it's a, let's say it's going into prod and then lifecycle management. So things like power schedules or expiration or approval policy all built into the workflow.

And next, and I'm done that really click, click, click, done. That, that is the experience we're trying to bring together to help simplify and automate a lot of those handoffs. We can also even go into the UI and we have, uh, kinda a catalog view.

A lot of consumers, things like higher ed, it's a big, uh, big industry. If they've got a grad student on the weekend who just needs a database, he may not need all of those choice points. You might just want to very simply say, what password do you want on this box?

And we have taken all the choice away so we can, uh, give customers this persona based view and they can dial in what of the sausage making do they want to expose to different communities inside, uh, inside their, uh, their team stack based on time. We could demo Morpheus for about four hours straight and not get into every dook and cranny. But any questions from, uh, what we've talked about today, What's the experience gonna be like for customers who are using Morpheus and Ops ramp and they want to maybe build in automated day two operations?

Meaning, okay, it's set up now, it needs to be monitored by this group for these services, and these alerts need to be generated, right? Like, I didn't see that in the demo is available planned. Yeah, the, I mentioned, uh, earlier, we actually already have, uh, one of our engineers built a a UI tab plugin so that some of the tabs for day two management, I can bring in things like storage network, uh, remote console access.

So by bringing the agents together, I think we're gonna get common data aggregation, being able to actually show and have an ops ramp tab inside my environment so that I can see basic telemetry, but in the context of the, the user as opposed to maybe that day two, you know, I Tom knock guy who's using the main ops ramp dashboard. So that, that's probably one of the first places that'll come together. And, and then furthermore, um, what about within either platform or maybe in development, some kind of self-documenting type of infrastructure so that, hey, who deployed this?

Why'd they deploy it, what they're using it for? Who needs to be monitored, alerted, and all those kinds of things so that we're not left with a bunch of applications that we don't know when they were deployed, if they're still in use, who's using them? Uh, those kinds of things.

I think as far as, uh, a lot of things we already do, uh, with a lot of customers, things like tag policy, you know, the, the data model. I mean, there's several hundred tables of normalized data across all the cloud endpoints. So being able to scrape that data in a much easier way.

And that's where some of those ENT interfaces are coming into with GreenLake intelligence that I mentioned. All of that populating, you know, in a common data model up in, uh, the GreenLake cloud. One of the other things, just usability, we've got a wiki built in.

Um, that's something a lot of customers use, actually better than a sticky note on somebody's door, right? For each app. So you can have key data and, and collaborate across teams for sure.

Yep. And I can add to that on the op app side, um, we have a concept called process automation. So as infrastructure is deployed and, you know, discovered, we can automatically monitor it using whatever policies you define.

Um, and also, you know, the platform has a full audit log, so you'll know exactly who created the vm, who did what to it, right? A full audit log. And we also have, uh, on ops ramp, a knowledge base article, so you can use a knowledge base, um, to attach information like, you know, to, to, to databases, to infrastructure.

So there's a lot, there's multiple ways you can, you know, structure that automation either through Morpheus or through ops ramp natively. Sean, anything, uh, we may have missed you wanna add in? Otherwise I may, uh, give a little time to our, our next session.

No, no, I think that was pretty robust just on, you know, getting that information in. You can add things in that would script out filling out all of this information. And we've been looking at the API integration between Ops Ramp and Morpheus.

When we provision something, we can actually push some of that information into ops ramp as the agent comes live too, so. Perfect. Uh, last thing, uh, a little shout out.

I don't have the, the link here, but we do have a community edition of, uh, of Morpheus, right? And, uh, I know a lot of home lab in the, in the blogosphere. So hit me up after this.

Uh, if you guys want to get into that community edition, and we've got some, even some, uh, training assets that Sean's team helped pull together to get you, get you bootstrapped you. Just more quick question over there. Is It, is there a way to drive Morpheus from an API or an infrastructure as code perspective, or is it strictly Click ops?

Oh, no, for, uh, like I said, the, uh, APIs got 1300 commands in it. Okay. So, you know, this thing was built by a DevOps team for their own use.

That's where a lot of the classic tools kind of fell short, right? They were built by ops four Ops. Mm-hmm.

And they said, Hey, development team, guess what? We built a thing and they'd turn it over and, you know, development teams be like, yeah, no, uh, that is not, not the way we want to do it. So yeah, we've got a Jenkins plugin.

The API sets very robust. And one of the things I'm excited about as part of the acquisition, right? We, we've had a Terraform provider, but it was, uh, those remember Martes Cloud Field Day, uh, forever, right?

It was like his fifth job was, was building our Terraform provider, uh, pre-acquisition. We now have an entire team in Galway who's built a unified Terraform providers so that you go to us, we handle all the downstream. So instead of having to have, you know, 10 different TF spec files for each cloud you're doing and doing this mountain of, of just digging a hole, you write to us, we'll handle all the downstream complexity.

So that's a core use case.