00:00
Hi, my name's Michael Eworthy, and today I'm here to talk about the Enterprise Data cloud. Like, what is it? But before we get there, let's really talk about the state of legacy storage today. So if you think about most legacy storage in most organisations, you have a tonne of different, um, boxes, arrays, storage types.
00:21
You might have some block, some objects, some file. Hey, I might even have some various stuff. Over here in the cloud. And then, you know, as a user or as an application, trying to get to this stuff, I'm like putting requests in all the time.
00:38
These things are talking to, these things are talking to these things and trying to manage that. Is a real pain. Think about consistency, think about the gaps, isolation, you know, being able to access the block environment versus the object environment might require different data policies, it becomes a super mess.
00:57
And really what you want to do is have this almost as a black box, and then as a user, have, you know, this sort of unified data access, be able to sort of get to this, get to this layer of data, super easy. So I can script it, I can put it in my terraform, I can put it in my policies, I can create workflow orchestration to it.
01:19
I don't have to use manual tickets, it just happens. And so to do that, you really need to completely re-architect storage in the way that we're thinking of storage today. You have to get away from all these desperate silos of, of, of different architecture and different protocol types and really think about storage as unified, as virtual, and as one. So let's talk a little bit about that
01:47
and what this means. So, the way you need to start this is you need to start with what we call an evergreen architecture. Now the great thing about an evergreen architecture is, this is a, not just, Um, something that we build, but essentially
02:09
a way that we build our technology. So if you think about our all flash systems, so this is our all flash systems, and you would know these today as FA, you know, Flash array, Flash blade, and sort of, you know, some integrated systems. And then in the public cloud, we have this thing called pure cloud services.
02:32
Now, the, the thing about the evergreen architecture, what it's going to give you is a tonne of capability. It's gonna give you non-disruptive updates. Fantastic, it's always gonna be, it's gonna always be getting better. Right, what do I mean by that? Well, what I mean by that?
02:50
I it will consistently update, it will consistently improve both software and hardware not impactful to your organisation. So the good thing about that is that once you've put one of these systems in, this is gonna probably the last system you're ever gonna buy. We have organised customs out there that have the same,
03:10
you know, they, they, they deployed Flash array 10 years ago, and they're still utilising Flash array. Yes, it's been upgraded and updated along the way, but they still architected once and have built upon that, um, capability. Now, on top of that, You need a common operating environment.
03:28
Now this is incredibly important. Now this does a couple of things. One, this gives us all our block, all our file, and all of our object capability, right? So, you know, you're gonna want that in a common operating environment. You don't want 3 different architectures to be able to sort of do that.
03:47
But it's also gonna give you all the data services that you need. What do I mean by data services? Your access control, your performance, your scalability, your replication, your backup, all the things that you need from an enterprise data capability. Now, in pure storage, we have a common operating environment across our evergreen
04:07
architecture. What does that mean? One operating environment unified across Flash array, flash blade, integrated systems, pure cloud um services. So you, you manage one layer in our common operating environment, so that's incredible. Now you can see here where we start really
04:23
thinking about unification. I'm building a foundation of unification here. And why is that important? Because as I build this foundation, you're gonna see here, I can now do a virtual, Cloud of data. Or cloud of storage, let's call it.
04:43
Alright, so this is my virtual cloud of storage. Now what does that give me? Well that gives me a single, You know, virtual pool of storage. Super powerful, enables data mobility. It enables me to get away from copy data management.
05:03
Think about it. Like, if I don't have to worry about physically where that data is stored, and that data's still gonna be stored in multiple different areas, maybe on some flash array or some flash blades or in pure cloud, in the pure cloud in Amazon and AW. But now I have a virtual layer across that.
05:22
I can then use that to get a, get out of copy data management, to be able to get out of data access issues, be able to sort of get out of this sort of isolated, um, you know, distributed sort of data issues that we have sort of have today in traditional storage, we have sort of these isolated silos of data. So this is the virtual cloud storage is the foundation of what an enterprise data cloud is.
05:45
It's in the name. So if you think about Enterprise Data cloud, it's a cloud operating model. How do clouds work, they virtualize everything, they virtualize the hardware, they virtualize the CPU and the GPU. They virtualize the storage, they virtualize the networking.
06:00
What we're talking about here is a virtual cloud of storage, being able to, to have your physical environments wherever you want them to be, cause that's gonna be important for you, but then have a single layer of virtualized storage associated with that. That brings us to the next thing is control. Over here,
06:20
I wanna have a halo. I want an intelligent control plane. And this is super important because an intelligent control plane gives me a tonne of capability. What does an intelligent control plan enables me to do? Well, one, I can start using AI ops,
06:38
um, Telemetry, incredibly well, because I have visibility of all my workloads. I have visibility of all my data, so I have a tonne of telemetry that I can use. What else do I get? I get policies, we all love policies. And what do policies give us? Policies give us consistency,
06:59
they give us governance, and they give us compliance. With a global policy system, you, you create it once and you apply it once and it just happens. Now, the great thing about that is then I can actually build really good work flows. Now we all love workflows, but the great thing about workflows when you think about an
07:19
intelligent control plane is I can create templates for my applications. This is the way I deploy a data blaze, whether it's, you know, um, Oracle or whether it's SQL. This is the way I deploy a CRM system, this is the way I deploy a system. Support a Kuberneti's environment. This is why I support,
07:37
support an inference AI workload. I have a consistent way to deploy that. So it gets away from all this random acts of kindness storage people have to do. I think storage myself is like, in today is like gardening. I don't know if you like gardening, I hate it.
07:54
But you know, I, I live on this sort of half acre property and I have probably 50, 60, 70 different types of plants, and every one of them requires something different, you know, different watering schedules, pruning schedules, feeding schedules. It is a pain in the butt to be able to keep up with that. And then, you know what happens? My wife comes home with a new plant and now
08:14
I've got something else, and that's exactly what happens to storage today. Right, every new request is a new plant. We want to get away from that, we want to get into farming, where I create 5 crops and you consume from the five crops. How do I do that? Through policies and through workflows. Right, this is the way you deploy this application, and this application could be
08:37
deployed into these zones, right? And then day 01 and 2, workflows can happen. How many times have we as storage admins deployed, you know, a workload, and then we have to come back and think about replication, think about backup, think about security, think about encryption,
08:52
all these operations that we have to continually sort of think about, and what happens if they change? With an intelligent control plane, I now have all these orchestrated workflows that when I deploy that application, all those day 1 or 2 operations could be deployed on day 0, so when that database is deployed, it's automatically replicated.
09:12
It's automatically backed up. It's automatically secured to our security policy. This is sort of fantastic. And that gives us, you know, the, the great thing is then, is I can also hook a tonne of APIs into this, right? So if I've got terraform or a compliance system or a,
09:30
or an external backup system, or a governance system or an AI workflow system or a DevOp system, I can hook storage straight into those. I can get away from opening a manual storage ticket. Who wants to do that? That's like planting a brand new plant in my garden. I don't want to do that.
09:47
Right, I can completely get away from this and integrate through this intelligent control plane API and integrate all storage into my my workflows. So it means my DevOps can self-service storage through the terraform, for example, which I think is absolutely fantastic. Now, if you think about them, is I want to deliver all this.
10:11
As a service. Now this is super important because when you think about the cloud, what's the cloud delivered as? It's delivered as a service, right? And what do I get from the cloud? I get two amazing things. I get guaranteed SLAs.
10:28
I, this needs to be super important, right? Because when I think about guaranteed SLAs, um, I want to ensure that whoever I'm consuming that service, service, um, from, I can actually put a contract to, so if they don't meet my SLAs, there's some form of impact, right? So when we think about SLAs that we can deliver,
10:49
we can deliver things like Easy scaling and usage-based billing and power and um, Uh, you know, um, Watts, there's a tonne of different SLAs that are built into our, um, service delivery aspect as well. The good thing about also this is it can deliver storage as a service, exactly what you think from a cloud perspective, right? So when you go into a public cloud, you are
11:13
consuming storage as a service and networking as a service and uh lots of other things as a service. This can bring you storage as a service as well. Not just self-service storage. You know, where uh an individual can request storage automatically through the system, but again, because now we have this integrated API and workflows,
11:32
any system can then go and request storage, and I can make make sure through my security policies and my government policies, I'm giving them the right storage or access to the right storage associated with as well. Now, if you think about here, what we've just drawn is quite amazing, right? Now, forgive all my messiness here, but let's
11:50
go walk through it again, right? So I have a consistent evergreen architecture, it's the way that I build things, you have to build a consistent architecture. Or none of this doesn't work, like, right? So our Flash ray, flash blade, integrated systems, our pure cloud services have a consistent way that we architecture architect them.
12:11
We think about the distribution of hardware and software. And how we, how we, you know, create the gap between those, we put a tonne of our effort into the software as well as into the hardware. We provide the best all flash systems on premises, but also we provide a tonne of capability in our software so we can have the same consistent experience through our common
12:32
operating environment across on premises and the public cloud. And then that gives us the ability, cause now we have a common operating environment with a common block file and object that are native, a global pool of storage, um, and integrated data services. I can now create that virtual cloud of storage, um, that just brings all my unified access into one, helps me with copper data management,
12:58
helps me with AI ops, helps me with policies, helps me with its workflows. And now I can build a consistent control plane on top of that to be able to orchestrate work workflows. You know, um, and then be able to sort of move from sort of storage infrastructure into data management and then deliver this all as a service.
13:18
And the good thing about this is this supports all the workloads that your, um, organisation could require, whether they're traditional workloads, like, you know, databases, you know, virt, all the things that run your workload. Today, or whether they're modern um environments, you know, things like AI or um machine learning or you know, Kernetes.
13:42
The great thing about this is you can build an entire, your entire workload infrastructure on this single, EDC architecture. So if you think about like the bottom line here, the Enterprise data cloud is just not about storage. It's really about how control and agility and outcomes. I mean it changes how IT operates and how
14:06
businesses move. I mean it's the foundation of the future of private and hybrid cloud. And you know what? It's available for you today with our pure storage platform. So with that, I say learn more about it on our enterprise data cloud at uh pureStorage.com.
14:26
Um, get into the pure storage platform which is the engine behind this, and really take yourself from managing storage to really managing your storage and your data and enabling a greater sense of control in your organisation. Thank you.
