25:02 Video
How Infrastructure-led Innovation Can Change Your Life - For The Better
Discover how changing your technology adoption decision-making process can make a big impact with Julia Palmer (VP Analyst, Gartner).
Click to View Transcript
Unknown: Welcome to our Pure Accelerate digital live session with Julia Palmer VP of research at Gartner. So let's get this straight right from beginning. We are broadcasting live today from our secret bunker in California. And Julia is home in Arizona. Not fake live, not pseudo live, we are live live.
And that means in this second half of the session, Julia is going to take your questions. So fire up your keyboards type your questions into the chat screen, and they'll come magically to me, and I'll put them straight to Julia. So with that, first of all, welcome, Julia. Thank you super excited to be here.
Awesome. Well, let me do just a little bit better job introducing Julia, I think you'll really enjoy this. You know, Julia is one of the most insightful and in my voice practical analysts out there. Because just a few years ago, she was actually in your shoes, Julia built and ran
infrastructure for web hosting giant GoDaddy for over a decade, and is kind of a student and deep thinker on the infrastructure revolution. We invited Julia here today because she has recently published some research on the future of infrastructure. And it's kind of a modern manifesto on the
evolution of modern infrastructure teams. So Julia, why don't we just start right at the top, give us kind of a high level overview of your vision for modern infrastructure teams? Yes, thank you for that. I've been thinking about future of infrastructure and storage professionals role in this for a
while. But what really inspired this research is the data that we saw that CIOs replacing IT leaders, with people with no IT experience, and I was trying to find out why is it happening? And what we learned from surveys conducted by Gartner is that CIOs focusing on change of culture, looking for people with
business acumen and most important thing, they wanted infrastructure that can support change. While it they're still focusing on reliability, quality, performance, but it's really the big gap between expectations from CIO to IT leaders. So right now, they don't want you to find an
infrastructure that will run forever, they want you to focus on the infrastructure that can support change. Well, you know, that's a great point, and honestly, a very different way of thinking. Now, just let that sink in for a bit. You know, for like 20 years, we've all been taught that the
number one thing is reliability, that's how to not get fired. But now there's a shift to agility because just the flexibility and speed at which the business can change and evolve and operate is crucial. And so it's more important to build infrastructure that anticipates that change and is ready for it
while being of course reliable. But you know, what do you think that means practically? How can people build in this agility? Well, first of all, we need to figure out what change means to you. So I encourage you to dig a little deeper on this, what this future holds. Is your infrastructure going to be on
moon, like we just heard from NASA? Or will you have a shortage of storage experts will, you're going to be facing exponential data growth, like some of my customers now collecting more data in a week than they collected over 10 years. And maybe you have a new workloads coming your way
containers, you know, ai ml processing. So first order of business is to figure out what change means to you. And in more practical terms, you know, we're always trying to keep it down to earth. What I recommend you doing is to really forget about storage infrastructure for a little bit and focus on
infrastructure services. So how can you do this? Well build a service catalog, data service catalog and focus on what services you provide. For example, I provide services for virtualized machines, I provide services for HPC enterprise computing, I provide services for data image archiving. And
once you figure out how many data services, you have an enterprise, then you can collect the data requirements and all kinds of other details about particular service. And only then in there, we can focus on architecture, storage technologies and vendors. So I've been doing this with a lot
of Gartner customers right now building data service catalog, elevating the message from provider of infrastructure to provider of data services. Yeah, this evolution to service thinking is just fundamental to how people think about their teams. And, you know, it just
cuts throughout the whole organization. I met a customer recently where the person who used to be the storage architect a few years ago now introduced himself as, you know, basically the storage service owner and on their team, there was a block service owner and a file service owner. So I think when people
embraced this service mentality, it just can cut across the whole way that you think, you know, one of the things that I also see customers doing is really thinking about that whole value chain around how they procure. And there's a whole litany now of as a service buying programs that are one way to add to this
agility, you know, at least from the procurement side, you know, it seems like these buying and consumption programs are coming from many different vendors, and they're a nice way to align the service delivery cost, you know, with this model, are these just different ways to buy? Or is your estimation something much
bigger going on here? You know, the first version of such offerings are definitely more focusing on OPEX versus capex and more focused on on how do you prefer, you know, to acquire your infrastructure. However, now, it's a little bit different story. It's more about
agility of those services, like, for example, I talked to service provider the need to deliver their services in Asia. storage as a service is the only way for them to go forward. Because they need to do it with matter of, you know, couple of months and don't have time to expand. So it provides them agility, other
people really thinking about moving into a mode of operation of provider of services themselves. And they're not interested in less than infrastructure knowledge. So they need this to provide this agility and speed of moving business forward, and invest in what most important things. So
right now, I encourage everybody to have another look of as a service offering and see how can they deliver on agility and not just like an acquisition methodology, but also, and not all of them created equally, let's be honest, if you embrace as a service offering, underlying infrastructure needs
to be built for it, it needs to be API driven, it needs to be completely seamless, you know, upgrade should be happening without any disruption. So it's, it's something that, like, they say, devil is in detail. So I encourage you to look at the details of underlying infrastructure is it built to
become as a service? Yeah, that's a great point. I mean, certainly just slapping a fancy leasing program on top of the same old storage that was clunky, and not very agile yesterday, doesn't change much. You know, I know from purist point of view, as we've done
this, with our periods of service program, we really thought about this as a product strategy, not just a way that customers can buy storage. And so we have a whole bu focused on it. We have engineers, they're building technology, it's going kind of core to how we build our products. And so it's a huge
embrace in the company right now. You know, another big change that I hear a lot from customers and infrastructure teams, is learning how to kind of evolve their infrastructure to run next gen cloud, native apps, containers, microservices, scale out architectures, all these ingredients are and modern
applications are changing infrastructure. So what advice do you have for folks who need to think about how to evolve their infrastructure for this next generation of applications? So first of all, biggest mistake is to think you don't have to involve evolve your your strategy. It's not just another
workload, you know, it's very unique. workload was very unique requirements. So when you think about cloud native applications, you need to think about cloud native principles. If your cloud native applications will be deployed on top of containers and microservices, the infrastructure needs to adhere
to those principles, for example, to be modular, to be programmable, to be intelligent, you know, those are those things and capabilities you need to look for in underlying infrastructure. Otherwise, you are pretty much taking legacy infrastructure in put modern applications on top of them,
which will not deliver on any agility going forward. So what specifically are we talking about? We're talking about being software defined, we're talking about being able to be deployed in the public cloud and the edge, we're talking about having Rarity and be driven by API's only. And we're talking about
being able to be managed by developers. Yeah, I think that's spot on. And, you know, I've seen, just in our experience with this, one of the first things that really just blows customers away is the change rate. You know, they're used to a traditional application world where we build
infrastructure and try to never touch it, because we want it to be reliable again. And so you know, you might have a storage array that gets a new volume or new file system, you know, a couple times a day. It's not uncommon in containerized environments to have 10,000 of these events a day. And so it's
just simply not possible to use the old processes to spin up storage for 10,000 containers that might only be around for a minute or two at a time. You know, I think another area that's kind of interesting in this is the rise of object storage, because many of these application environments need
object. Do you think that's kind of a key ingredient for this world? Oh, yeah, absolutely, you know, modern applications being designed on top of seamless storage. And object storage becomes standard since arise of the cloud native infrastructure.
That's definitely something that people looking for when they build the next generation applications. You know, why would you use anything else, because we can have object storage that can be geo distributed, it can have rich metadata, which your developers just love. And it can be faster
as well. So what's not to love? Yeah, that that rise of kind of fast object storage has been a more fun area of our business where folks, we're just kind of used to this land of object being the kind of the slow, you know, dirt tier of storage, if you will. And now, it's actually
one of the most exciting areas of next gen applications. You know, another area that I think is exciting in this world is the next gen apps around big data and analytics and AI. You know, there's a lot of hype around big data and AI over the past decade. But it feels like it's settling down now. And that's
getting real. And so I'm curious how infrastructure teams, from your point of view, can help drive next gen analytics projects and support data science teams. You know, first of all, is we need to focus on data, when you talk to those guys. And I encourage everybody to go and
learn about data for AI, where is it coming from, where it's going to be stored where it's going to be processed, does it need to be archived, all of the data flow is very unique to AI ml process, because it's not a single workload. It's many different workloads. So if you're not careful, you might
end up with a sprawl of infrastructure everywhere, and you will be in the world of hurt. So you need to first learn about data. And then really, if you talk to those people, they don't talk about storage, they talk about data operations, that's what they focused on. Because for them, you know, data
is the fuel, it flows from one thing to another, and they need your help to manage it for them. So to really be part of the IML workflow, you need to be in data operations, no longer storage operations. But in data operations, of course, it requires very specific storage that can accommodate, you know,
low latency and high throughput workloads, which is pretty demanding. But it's even more scary to end up with a system that cannot handle the giant data flow. Yeah, I think this notion of just knowing your customer, in this case, the data scientist is key. And, you know, part of the
role of data scientists, they just tend to be, you know, agnostic infrastructure, they're even disinterested, I would say in infrastructure. And so part of your goal is to provide an as a service experience to them, where they don't have to think about it, they just come in and do their data science work, and
they can leave the infrastructure to you, you know, that's your job. You know, we talked a lot about the people there, all this change come always comes down to the people and inspiring the team, you've got to the rise the occasion, attracting new talent, the organization, you know, what
advice have you got for influential leaders to manage through just the human side of all this change? You know, this is probably the most important change that we need to embrace. And it is very uncomfortable. I'll be honest with you, as it practitioner myself. I like technology, I
like to deal with new vendors, new products and trying things on. But we need to get out of our comfort zone, we need to go and really become embedded in our business not aligned, it's no longer good enough to be aligned, you need to be embedded, we need to get out, go talk to the developers, people
that designing applications, talk to database owners, and talk to the end users of your services to really figure out what are the problems they facing? And what are the business key performance indicators, they have to tie in into your infrastructure technology, performance
indicator. So I encourage people to be heroes, to be superheroes to be data champions, to learn everything that is to their to learn about data their team is holding and expand their knowledge to become versatile is in fact, Gartner predicts already that in five years, more than 70% of storage
professionals, you guys will expand your job responsibilities and scope of work and knowledge beyond managing storage infrastructure. So Tommy, his time is now to embrace it and expand your role. It can be data operations, it can be infrastructure versus a list. It can be data Ops, with an AI ml
angle, because we know those workloads are coming your way. But it's really time is now to get out of your comfort zone. Yeah, that's awesome advice and you know Ultimately very inspirational, right? Like what's better in life than to embrace something new, and grow as an organization. And when I
look at this whole journey around modern infrastructure, that's what people are doing. That's what companies are taking advantage of is this is a real opportunity to uplevel their infrastructure teams to deliver the as a service model. And so you know, there you have it, modern infrastructure and modern
processes to block to evolve what we do as infrastructure teams. So with that, let's transition now into q&a. So as I said earlier, please go ahead and type questions into the chat window. And we are going to go right now. So a question from Jeremy, what are some of the storage technologies that we
should be evaluating now, that will have an impact in the next 12 to 24? months? Wow, it's such a good question. You know, I always say, before you related every any technologies, and you know, I love those, you know, I cover emerging technologies and
Gartner I also need to pull myself back and say, okay, but what is your objective? So the first order of business when I work with people who asked this question is, let's focus on what are you trying to achieve? Is it simplicity, is it better performance is agility, adaptability, maybe hybrid cloud
integration. So based on your direction, and you cannot do them all at once. So pick one and two to go forward with that, how you going to pick what technologies you're going to be evaluating. For example, if simplicity is something that you need to deliver to the business, I would really evaluate AI
driven operations. All of the vendors have different offerings, they also have different capabilities. But you need to figure out how will AI driven operations will impact your business. So that's number one, for simplicity, for for a number two, for simplicity, I would say, as a service
offering, if this is something that you want to build your business on, you don't want to be you know, in a world where you maintaining your own infrastructure, and doing upgrades, then as a service offering is a way to go. So kind of a long list I can go on and on. But in terms of things that
topple the mind, a lot of people I talk to, is obviously NVMe and NVMe over Fabrics to achieve performance. We'll also talk to a lot of people that designing infrastructure for persistent storage for containers. It's a big project for the most people. So definitely, we'll have to have a look at this. And also
distributed system and object storage. Great on your on your point there around containers, we got another question around that. Just asking for kind of practical advice. This company is just getting going on the container journey. They were used to kind of all the change
that they had to go through with virtualization over the last, you know, 15 plus years. And so what should they be thinking around design for containers? And maybe how is this different than then what they went through for the design for virtualization?
Yes, gigs, it's a good question. Because, you know, if you're designing the big platform for the scale out operations, for containers, it's very different from when you just getting started. So when you just get started, I always encourage people to leverage infrastructure they already
have, because there is a lot of unknowns in infrastructure for containers, from a customer requirements to performance to scale and capacity. So just leverage what you have right now, modern storage infrastructure has CSI container storage interface plugins. So you can plug in with your flavor
of Kubernetes and get you going pretty fast. Another option is to use freemium solutions. A lot of infrastructure providers now have a freemium solution to get you started, and it's usually a software defined storage offering. So I would say either you leverage your existing infrastructure or try some
premium. Yeah, I think that's great. And, you know, one of the nice things about the software approaches to container storage, is that they do allow you to get going really easily, you know, it's just software, you can install it on whatever storage you've got. And
it just gets you up and going as fast. And you know, the other thing I would say is, if I look at kind of the containerized world, many customers just have many different projects within their companies at different stages of evolution. And so you might have a dev team building something, you might have a
production application, and everyone's making different choices. And so having a storage infrastructure that's kind of flexible to support all of these is really critical to the design. So we got another question that kind of goes back again to the discussion of object storage. And this
particular person is asking, what are the use cases for fast object, you know, I'm used to object storage being essentially a really, really big book tier at the lowest cost per gig. And so, you know, how should I use this fast object that's coming on the market now?
Well, we Have a lot of third party software vendors that certifying storage, object storage news with any of the applications. So list is growing, we already have a Splunk. And we have Apache Spark that you can run with object storage. But I think the number one thing is actually custom
applications. If you think about it, you know, in the public cloud, any applications running on top of object storage, and it just started to happen on prem? Because perception was, like you said earlier gigs that object storage cannot perform. But now when we have, you know, flash optimized object storage system,
that really any applications and design is object storage in mind can live there. Yeah, I mean, I think one of the the shocking statements I've heard in the space is imagine now we're hiring software developers graduate at a college, who may have never
built an application that didn't use object storage, like they may not really ever have used the file system. And so how do we provide to them what they're used to from the cloud, and what's comfortable in that way? I got another question from Fidel. So with all the news, we are seeing around ransomware,
what is your recommendation for how to have a strategy to protect against ransomware? And how to recover if you do get hit? Well, that's a very topical question this week, right? I think, as a storage professionals, we our last way
of defense, if they coming to us to recover from ransomware, it means a lot of things have failed them already. That's why I spent sibility my lioness. So there's a different approaches for ransomware detection. I think we have a lot of emerging technologies coming in right now, that really will be
focusing on how to detect attack while it's happening. But at this point, right now, we mostly will be relying on recovery. At this point, we have technologies, that story, you know, read only snapshots that are not corrupted that you can restore from them as well. different vendors
implementations have different angles, but it's definitely something you need to look at. And when you're relating new storage offerings, is how far back can you go? How can you detect which snapshots were last known good methods? Do you have air gapped copy of your data elsewhere? That's definitely
something we'll be looking into writing very, very soon about. Great. Yeah. And it's been, you know, from my point of view, one of the more surprising parts of our business over the last few years, when we had when we started Pure, twelve odd years ago, and had the vision of an all flash data center, you know,
kind of assumed that mission critical databases and virtual machines and apps would go to flash, but I never really expected backup would go to flash. And that's exactly what we've seen over the last couple years, because of this need for fast recovery. And ransomware is one of these things is driving
the need to do potentially a mass recovery event and make sure you're prepared for that. So another question that comes in is related to your concept of a service catalog. And this person's asking, I get the idea of basically creating standardized service tiers. But how many services? should I
have? You know, do I just need a few? Do I need, you know, 10s 20s? How do I think about th right number of services for m organization I would say, in terms of service catalog, you can have as many services as you want. In fact, I would encourage you to have very
unique services providers. But on the back end, you can consolidate them on a very few offerings. So right now, we really try to work on consolidation of infrastructure services, you know, modern technologies now can provide performance and scale at the same time. So sometimes, there
is no need to have different tiers, for example of block storage, because the way both very efficient and not very expensive. So I say service, because a log establishment is number one goal, but behind the scenes, you should be looking at consolidating of the consolidation of those services,
as much as you can. So I would not focus on file block and object, I would focus on just application delivery. And on the back end, you get to choose what type of storage they're going to use. And as little as possible. That's my goal, you know, infrastructure sprawl, is what killing innovation and it's
really impeding agility. That's an awesome quote to end on. Well, all there you have it. Modern apps need modern infrastructure, both modernization of the data infrastructure itself, but maybe more importantly, modernization of the delivery processes and of
the user experience. Julia, it's always a pleasure and I can't wait for us to see what you write next.
For the last twenty years, reliability has been the watchword for data infrastructures. But in the modern data revolution, agility is fast becoming a top concern. What does this change mean to your organization? Listen to Julia Palmer (VP Analyst, Gartner) discuss how organizations should identify and catalog their own data services in order to support the evolution to services-based thinking.
In an as-a-service world, your underlying data and storage infrastructure needs to be purpose-built: seamless, API-driven and updated without disruption. Businesses that wish to remain agile in this environment should adapt their strategy to handle the emergence of cloud-native applications, which require a modular, programmable, intelligent, data infrastructure.
Meet with an Expert
Let’s talk. Book a 1:1 meeting with one of our experts to discuss your specific needs.
Questions, Comments?
Have a question or comment about Pure products or certifications? We’re here to help.
Schedule a Demo
Schedule a live demo and see for yourself how Pure can help transform your data into powerful outcomes.
Call Sales: 800-976-6494
Media: pr@purestorage.com
Pure Storage, Inc.
2555 Augustine Dr.
Santa Clara, CA 95054
800-379-7873 (general info)