Skip to Content
43:57 Video

DBA’s and IT Admins of the Future: The Swiss Army Knives of the Datacenter

The IT workplace is rapidly evolving. An expert panel of Pure Storage industry professionals discuss the major trends shaping the future of the IT profession.
Click to View Transcript
00:17
Narrator: Meet the new builder in the New Age of rolling up your sleeves and getting your hands dirty there in labs, laws and war rooms, launching scrappy startups and reinventing big business. They see what doesn't exist, and they make it exist. They know the question isn't if. But when they know the freedom
00:40
to fail is the freedom to build. And they don't freak out when they run into a problem. These are the new builders. And we're the cloud that's obsessed, turning their dreams into realities with the most capabilities, the most innovation, the most customers and the most experience. We're
00:59
the cloud that breaks down barriers when others say no, we let builders say yes, because we believe everything gets better when you let nothing stand in the builders' way. Amazon Web Services - build on. David: Ladies and gentlemen, welcome to the Pure panel
01:19
discussion. Please welcome your moderator, VP of worldwide systems engineering. Nathan Hall. Nathan Hall: Thanks, David. And thanks, everyone for joining us today. And I am joined by my two esteemed colleagues, Marsha Pierce, who leads our field solutions architects in the
01:36
Americas, as well as Jon Owings, who leads our cloud solutions architects globally. So we're here today to talk about the changing nature of classically siloed IT roles from DBAs, to storage admins, to sysadmins and IT architects, and also how specifically role changes are being driven by new
01:57
technologies, platforms, automation trends, and where these roles might morph into generalist functions, or where specialist roles are potentially likely to continue by adapting to new texts and new technologies to new methodologies. So there's a lot of, you know, technologies that
02:14
are driving these world transformations from automation and the ubiquity of restful API's to the rise of citizen developers outside of it, leveraging low code, no code dev platforms to bring business ideas into digital production, often with zero traditional it or developer engagement. The
02:32
cloud has also brought in, frankly, an intolerance for IT infrastructure complexity from you know, for example, why? Why should it take a specialist a ton of esoteric command line calls to provision storage, or even more ci commands and active monitoring just to keep the stuff running. So as operational
02:50
complexity gets reduced, and manual processes are increasingly becoming this anomaly, we're seeing these traditional IP silos change or consolidate. inversely, if you look at an explosion of platforms, look, think about the ubiquity that VMware virtualization had a number of
03:06
years ago, and how it's given way to the multiple disparate cloud is platforms, multiple Kubernetes, distros, and the growth of data analytics, like from, you know, a few top relational database and Hadoop offerings. To at this point, I don't know hundreds of analytics offerings. Ultimately, that's
03:25
required more multi role generalists. And businesses need to adapt to new technologies quickly. And to do so they have to take existing staff to make those new platforms successful quickly. All of these trends combined are creating just huge opportunities for existing IP practitioners. But it's often
03:42
hard to find practical advice on you know, where do you take advantage of those opportunities? And where do you invest in your own career development in a way that will be the most productive relative to your current background. So let's dig in. Let's start discussing with Marsha and JO
03:56
here. And I also wanted to mention that Q&A is going to be available in this chat for the whole session, we'll be looking at that live, we'll be ready to respond. So please feel free to jump in anytime on chat. Let's start first with Marsha with looking at how classic IT roles are starting to move away from
04:12
specialization and more of a generalist approach. Marsha, this trend from specialization has been happening for some time now. But what what do you see as accelerating it? Marsha Pierce: I think we're kind of hitting that tipping point now. Right? So we were in a long time ago, classic
04:27
waterfall projects. And then we've moved to kind of agile development, and then DevOps. And as you do that, you accelerate your development cycles, and you become more productive and those companies became more successful. And as they became more successful, the technology started to kind of
04:47
come in and meet them with software applications containerization and data to kind of meet their needs. And then they became, you know, the leaders and their markets for what they were doing. And so I think their competition has Realize that they have to kind of step up their game and move
05:03
along in order just to meet the business needs. I think, if you read some of the books, two of my favorites, when I've talked about this, people kept asking me if I'd read them in, and I finally read them, where the Phoenix project and the Unicorn project that really talked about how accelerating your
05:19
development cycles, and accelerating your routes to market really do impact the bottom line and success of a business. So I think that is the need, right. And as we're starting to do this as well, when I ran these groups, right, talk about those external teams that were developing stuff, we
05:37
inherited those. And so then all of a sudden, my team that may be covered six or seven different database platforms, now had to cover 30 different database platforms. So even if you're specialized as a data professional, because there's so many more out there, you need to know all these other different
05:55
database platforms, you can't just be a sequel DBA you've got to be a DBA, who specializes in many platforms, and know how to integrate and bring that data together. So you don't have silos? Nathan Hall: Great insights marshawn. Lynch is a great book
06:08
recommendations as well, I've read them fantastic, fantastic books. JO jail, in your opinion, is becoming more of an IT generalist. A good step, given the current technology landscape? Jon Owings: Yeah, I would say that, at least from leadership
06:25
of our customers that I talked to, right is is like we want a a team that can I cannot do every, you know, not everything, but a lot of things right? When it comes to infrastructure is like the storage team is kind of absorbed or and VMware teams have kind of merged together, right. And now with things like
06:45
what I do with all day long with, with containers, it's like, while the developers may have embraced that first now they're now the like, VM teams are getting told Whoa, you get to manage containers. And so now there's another layer in there that they get to do. And so I don't know if that really
07:03
answered the question that I'm seeing it and it, I think it's a good thing, because it kind of broadens the horizons of everyone. And as things become automated and become more, you know, kind of abstracted you. You as like an admin, you know, sysadmin, you can move along with the trends. Rather than be
07:22
like, I'm the one person who knows how to script for this one platform, you know, and if that gets abstracted, like, What's going to happen to you, right? So if you can move along like that, that's what's best for you. As as like the consumer of it,
07:38
Nathan Hall: I could definitely see that. I mean, it's just, we saw that Amazon that AWS video, and last I checked, AWS had over 200 services and can't possibly have, you know, 200 specialists in IT organization, there's some you got to take a number of those together. It looks like we do have a few comments here.
07:56
questions in chat. One is around, what are the developing platforms that are the standard now and a follow up question is around DC default language that is prevailing? No Marsha jF, you want to take those? Jon Owings: When I think about platforms, I mean, I think
08:13
about, you know, the way to deploy that, right. And, and I would say 95% of new apps are getting developed on containers. Now the reason why that's so good is because every developer wants to use their own special tool, right? You know, you know, I hear rust or go or, you know, C and Java still exist, right?
08:34
Those, they're never going to go away. And so like, I don't know, the answer both questions is I don't see like one language taking over, like everyone's riding and dotnet, like back when, in 2005, when I was on the customer side, and the sysadmin is like, everything's switching to dotnet do it, you know, and
08:51
guess what Java still exists? And, and I think with containers, that kind of it's opened up that world where like, yeah, I can have this component written in Java, it can talk to this dotnet platform, and Kubernetes or containers just lets them all kind of play together, hopefully nicely.
09:08
Marsha Pierce: Yeah, I agree. There isn't one that is like the predominance people are containerizing. They're virtualizing. You know, those remaining workloads that they hadn't virtualized, are starting to go virtual. And then they're containerizing, the rest, and it is hodgepodge of everything out
09:25
there. Nathan Hall: is interesting, especially in some of the comments around, I think he said, Jay 95% of new apps are being written to containers. You combine that with another stat. I believe, the average enterprise app lifecycle is
09:41
about four years. So that implies just this massive transformation. we're on the cusp of right where, you know, lots of apps, traditional apps are going to retire get replaced by microservices architectures. And the world is changing at the same time. As you mentioned, Java has been around for a while
09:57
and and there's a lot of fundamental languages and scripting languages that are going to endure and stay the same. So Marsha, you know, a great example of this is, you know, I think of how I started my career as a solera sysadmin. And Oracle DBA. kind of one of those are still left, I guess
10:14
there are a few swear sysadmin still out there, but not not quite as many. You were a DBA for for many years, what are the traditional tasks that DBAs have had to worry about? Marsha Pierce: Well, you know, our first task is always to be able to backup the database, and then to make sure that it's
10:30
ready and available, open for transactions and responding. But then the tasks that we don't talk about that sucks up a lot of our time, is all of the moving of data around, I used to call it my DBA hamsters in a wheel, because everybody thought it was automatic, and that it just happened. And it was not it
10:47
was my DBAs, picking up backups, copying data moving around for test Dev, QA, right for your reporting system for your data warehouse. Even when we went to the cloud, right, my team had to set up a whole other process, they did it to get rid of our data lake. But the funny thing was, is we had to, we had to
11:07
actually set up a second data lake that just the cloud pulled from. So we've spent a lot of time moving data around. And I think that's one of the ways where we come into play is that we kind of help you out with some of those tasks that are sucking up your team's time. And I think taking away from
11:25
valuable things that they could doing, right? Those are not fun tasks. And there's much more important things you want your team to be focusing on. Nathan Hall: Yeah, I remember my storage admin days, just wondering like, What is it? What is it with DBAs? They certainly
11:38
love their copies. Is there anybody that loves copies more than DBAs? And then I got introduced to SAP basis admins? And the answer was yes. So So Jao, it evolved mainly from this, obviously starting off, and then bare metal. Moving to virtualized. Now, it's cloud as a service containers are taking
12:00
over? How does that transition change admin roles for for storage and sis admins? Jon Owings: Well, it kind of it changes what you what you're looking at, right? So they're still letting you know, like, I like to joke, if serverless still has servers, right? This
12:22
infrastructure, you're, you're an admin, you're still going to be managing storage, you're still going to be managing infrastructure, but you're going to be doing it through an API, you'll be doing it through some kind of set like that, where now I can define it as code. And so a lot more of what are kind of
12:38
at least mine. And my team's transformation is like, now we're doing this with Git. Like, if I can hit Git pull, and it, you know, some system automatically grabs that code and changes the infrastructure to what we want, like that is, you know, more reliable, less prone to me, you know, clicking
12:56
on the wrong thing, like, even though, you know, the Pure GUI, I've been, you know, playing with flash arrays for over seven years now. It's amazingly simple, but you can still, you know, maybe type in something wrong or mount a volume to the wrong place. Like, if it's all in code and codify like,
13:11
hopefully error free at that point. So it makes it makes life a lot easier. Nathan Hall: Yeah, certainly, certainly that if you think about the most common source of downtime isn't the platform, it's the human error. And it kind of goes to this next
13:25
question, which is, you know, we certainly hear a lot about automation more than ever now. Marshall, do you think automation, how much of that is a opportunity for DBAs and other it rolls Marsha Pierce: up? I think it's a great opportunity. And I think
13:41
a lot of times, it's one of the, in my career before I joined cara, I specialize a lot in automating and consolidating legacy shots, right? Their workflows. And when you come in to do that, a lot of people thought as a threat to their current job, that you were going to take away their little widget
14:00
that they're moving from here to here. That's what they do. But when we started working with our teams, and showing them, like, what do you want to do in your career, where do you want to be? And they would talk about all these aspirations that they had. And I'm like, What if I could give you time to learn that and
14:17
we started automating the repetitive tasks and building on prem cloud. I think the satisfaction rate and the the career trajectory you saw for those employees went way up when they got out of doing the Monday things and got chances to do all those like dream projects that they had seen for a long time
14:39
that they were kept from doing because they were doing those mundane tasks. So I think it really is an opportunity for everybody to do the things they love. Nathan Hall: Makes makes sense. Nobody loves typing the same command over and over again, cut and paste etc. And just a
14:56
reminder to everyone out there watching. We do have chat open. We're watching So anything that you want us to cover any questions you have type them in there, and we'll address them live to jail, you focus very heavily around DevOps space for Pure, how would you characterize the relationship between DBAs
15:12
and developers? And how can DBAs work better for support DevOps and vice versa? Jon Owings: DB DB DBAs work better with with DevOps? And the other way around, right? Yeah. Well, my personal experience is always the SIS admins and the DBAs always fought, right, like,
15:30
the deep end and the DBAs. And the developers were always out back smoking cigarettes, right? So like, my, that was my experience back in the old days. But I think now you have this DevOps team, they're a little bit Ops, right. They're a little bit like me, the sad sysadmin. Guy. And, and but they're
15:47
developers, right? And they are developing not only the application, but how to deploy it. So they're, you know, they're hitting the Go button for that to go live. And if there's a DBA, who's in control of the data, right, that is that relationship has to be solid. And if it's something where now,
16:06
you know, I update some code, and it breaks the database, like, how do I roll back? Right, like a lot of those things that were that we talked about with snapshots and stuff in Kubernetes? Like, it doesn't really it doesn't really exist, right? It's not like built in it is there CSI, but you know, and
16:25
you have to have a correct array and things like that to do it. But having that those those fallback points to be able to not upset the DBAs like, that's what I my goal always was like, Don't make DBAs mad and run into my office. And so you know, now that we have these these systems that could Hey, I can hit a
16:45
button and scale to 10,000. You know, and if that topples things like being able to work with the DBA is, is, I think priority one, you know, area DBAs, or data architects, whatever we want to call them today, like, we need to be able to work together so that when I do hit that button, they understand
17:03
what's going to happen or not like, Oh my gosh, you just a crash the whole system because they weren't planning on it. Marsha Pierce: Well, I said, you said do you think that DevOps and DBAs get along, I think DevOps folded DBAs into their process, because they didn't get along, right? It's that classic
17:21
mean that you've seen with a little girl with the house on fire, and like it works fine. In Dev, it's a production problem now. And so DBAs have always been at stop games that are being the naysayers that tell everybody No, because they're worried about production being on fire. But to be fair to deaf
17:40
people, the reason why production is on fire is that we haven't given them a good dev environment. There's that term size of data operations. So my dev team would sit there, and they were required to show us missing indexes and performance results. But they did it for a database with five rows of data.
18:00
So nothing showed up as a performance problem. Because the copying of that data down to those large data sets was problematic. When we one started, including them, like into that DevOps cycle. When we got on call, the person who wrote the bad code also had to get up in the middle of the
18:16
night. That was one fix. But the other fix was stopping it before it ever went to production by giving them good datasets, right, and so that we started using storage, cloning technologies to do that. But what we never got, which is why I love Jao, what you're doing with containers, is getting
18:37
every developer their own copy of data. And I think containers is finally getting right. Instead of just having your dev your QA in your UI at every dev can have their own little area to work in and with our stores our own full copy of data, right?
18:55
Jon Owings: Yeah, and what I was probably not clearly communicating, which is just the way that I am pleased way too many words is that now Dev and Ops and DBAs can get along because they have those tools to do that. And my experience in the old days was I had a developer and a DBA. And who was
19:15
also a developer and they just did that they just they were that they threw it over and said it worked on my laptop and my laptop is not the same as my servers and and they were out on the on the loading dock talking to each other while I was trying to fix everything.
19:32
Marsha Pierce: Complaining out on the loading dock, why you would have a copy of our database. Jon Owings: There's no like why why can't Jon fix that? And so now I have like a magic button. Right? I can do blue green and roll it back and do all kinds of cool stuff. Yeah.
19:44
Nathan Hall: Marshall with your comments there. I had just visions of that Oprah Winfrey meme, like you get a copy and you get a copy and you get a copy. Everybody gets a copy, which I guess is the holy grail Marsha Pierce: is the holy grail, right? And I talked to
19:58
like storage admins all the time, and I I'm like, I haven't been somewhere that I do. But I'm like, what happens when you ask a DBA for more space and like they want 10 terabytes more, like it's never enough like if you give them 10 terabytes, they they're like where's my next tip?
20:12
Nathan Hall: Alright, so we've we've devolved a little bit into meme discussions we're not a Reddit thread with us is real webinar, how if we turn our attention to practical ways that IT professionals can invest in expanding their responsibilities and what they can plan for today. So when we start with
20:31
just specifically like the types of projects and what skills you would advise DBAs to invest in today, Marsha to make them more valuable for the future? Marsha Pierce: Well, I think no SQL and data science is really kind of a really good thing to kind of like be studying and
20:49
looking at containerization of data. And think no SQL and SQL actually really have a lot in common. So learning that Structured Query Language set, understanding how large data sets work, that's the hurdle that I see most people really kind of fighting with today. Because whether it's structured
21:09
or unstructured, most of the issues that people are running or running into are size of data operations, how does your data function when it becomes very large, and that used to be 510 years ago, only really large enterprise, you know, those tier one companies suffered from that issue. And as data sets are
21:29
growing, and we're getting more Internet of Things, data, everybody is suffering from those issues. So they really need to learn how to make things work at large scale sizes. Nathan Hall: is such a challenge. I mean, you look at what all the hyper scalars have gone through and virtually all
21:46
of them go through these repetitive re architecture and exercises as they hit a certain scale limits. I remember just seeing, you know, example being a picture of if you want to go faster, probably a bigger sale for your sailboat only has a certain amount of you know, headroom. And at that point, you
22:02
have to switch architectures. And we've certainly seen Google do that. Amazon do that Google, you know, example, the Google file system they came up with in 2003. Totally threw it away in 2012, and created something differently, because it didn't, didn't scale enough for the the payload of data that they had.
22:18
So, Jon, on the cloud and container side, how would you? How do you think that traditional practitioners should look to get started with building some of those cloud native skills and learning containers, technologies, Jon Owings: they probably all have some VMware somewhere,
22:38
right? And VMware is as embraced Kubernetes. You know, whether it's on prem or in the cloud some way like, figure this, there's a net like, looking at what VMware has done that with tanzu, weather, you know, we'll talk about like, how many versions they have, but it's, you know, there's a lot of
22:56
opportunity to learn there. And it's natural, it's like a natural next step. Like maybe in from there, once you learn that, right, you kind of get the basics of Kubernetes, you have a system that you can run in production and those types of things. And now I can branch out into, you know, some of the
23:15
other projects don't, there's what, here's what I did, that was a huge mistake, when I first started looking at containers, I went to the CN CF website, and hit the little ecosystem infographic. And there's like, 10,000 logos on there of ecosystem partners. And and you know, you go, you go on there,
23:32
and you're just overwhelmed. You're like, there's no way I can learn all of this, right? But what's nice is like, there's some of these distros and you can go find any of them, like what Amazon does with Eks. You know, there's, they're all out there, where they've, they've simplified it for you. Right?
23:49
You don't need, you know, make sure you can go on there and run a database, right, you're providing you know, you know how to do storage, you know, how you already know how to do VMs. Now I want to be able to run a MySQL database in a container, right? That's a great goal. Don't get into like, services and load
24:06
balancers. Like there's lots of ways to make you crazy and hard, you know, get do it step by step. And that's something that will provide value to like people like, like Marsha, who are like, I want to learn how to do Jupiter analytics and all these things. And how do I get that and right, and then we're
24:22
meeting right? So then now we're working together with the DBAs and the developers and start just absorbing as much as you can from them when it comes to code, and how to apply that to what you do. Right? When it comes to, you know, being just like infrastructure, IT sysadmin right. So I think it's a great
24:46
step. I mean, there's lots of things that you could after you do that, there's 1000 other things to learn, but it's a good little start. Marsha Pierce: I was gonna add in if you're a data professional, right and you database professional, and you
24:59
you're wanting to get into thi space, Microsoft has got righ on their, you know, books onlin area, if you go into thei containerization, they hav courses that you can ste through there are free that yo can run on your laptop, you ca run up and get a small thin open Azure, that you can ru
25:18
that course and learn how to d development and databases and d those kinds of deployments in containerized environment. An so you know, those ar simplistic containerize environments, the really coo thing is like, you know, in th real world, I won't have t manually deploy my containers
25:34
and I get to work with some coo kind of automation system. Tha does it for me, but it gives yo a good understanding of how al that works. And then wha architectural considerations yo need to really think abou because that is the funny thing because it is so easy. I don' think sometimes think peopl
25:51
think with these ne technologies, what does m storage strategy going to b right? Because so much i stateless. They don't thin about what they're going to d with all that stateful data tha they're going to star containerizing
26:03
Jon Owings: Yeah, yes, if you start with knowing the data, you're going to provide value because a lot of people don't understand how the data works in containers. So then you're already setting yourself apart from everyone else, to provide great value to your organization.
26:18
Nathan Hall: It's a really exciting time, you look at the what appears to be inevitable in terms of containerization adoption. But combined with the fact that it's not like there's a ton of people out there that have years of experience in this right, so you have this opportunity to become an expert,
26:36
at least relative to the rest of the population relatively quickly. You know, that doesn't necessarily mean that you're not going to see every once in a while those job advertisements asking for 10 years of experience for a product that's, what about six, seven years, but
26:53
Jon Owings: it happens every day. Nathan Hall: So it looks like we got another question here in the chat. I'm not certain if we can, if we have time to address this as a pretty pretty technical specific question. But But Marsha, I'll have you look at that, if that's something that
27:09
you think we can address here live, or if we just reach out to that person offline, with the recommended configuration that they're looking for. Marsha Pierce: Okay, so they have read out the question. So the question is, we have a stretched data center that load
27:26
balances VMs, a complete data center can fail, and then the VMs will spin up in the other? What is the recommended ha for the SQL Server in the VMware environment, always on availability groups? Or is this overkill? So high availability, I always ask the question, what is your uptime requirement? A
27:46
G's are really perfect for this kind of scenario, if you're going between data centers, you're going to probably have it depending on how far apart your data centers are, you're going to want to have one asynchronous or synchronous. So if you need it to be synchronous ad is probably not what you're going
28:04
to do. And you're not gonna be able to failover to that data center. But probably ag is asynchronous, with like a one minute Blip. That's usually what people do. Nathan Hall: Great. Thanks, Marsha. And, and not certain who asked that. But if you have other questions, we're happy to
28:18
kind of dig into that a little deeper offline as well. So I look behind me here, it looks like I've got a few Pure logos. So I'd be remiss if I didn't talk about how Pure is helping to empower IT leaders and dmins, specifically around some f the features that can help ddress the things we're talking
28:36
bout today, how they can ddress how to make it ltimately easier for these ypes of roles to to do their obs to automate more, to think ess about that kind of manual rovisioning and operation. So tarting with the database pace, and particularly, you now, like we just saw an
28:55
xample of here, one of the reas where we help customers he most around SQL Server. So ow does Pure build integrations o make it easier to operate and onnect to cloud and Azure? Marsha Pierce: I think the first thing is that we're all API driven, right? So I'm being API driven means that anything we
29:15
do, can be coded. And so you can then automate that function. The other thing is, is we're building integrations right into the tools that these people are using. So you know, snapshots have been around for a long time. But we recently just built our SQL server management tool, and we built it right into SQL
29:36
Server Management Studio. So that if people want to start automating these things they can and to the point that the customer was just asking babies, right, one of the pain points with a GS is seeding that AG because it has to be done to where the database is in a restoring state. So
29:53
traditionally, that's a backup and restore or this little kind of drip process that you can run with her own SQL server to see That, and that can take days long time you got a multi terabyte database, you could be waiting over 24 hours to see that secondary ag. And an ability or scenario, you always
30:12
have to receive that secondary, if they've lost connectivity, leveraging snapshots and tools and integration, we're able to do that in a few seconds and help you get your data back. So we're talking about augmenting your current processes, right? Not replacing with our tools, but augmenting them and speeding
30:30
up that recovery time for those data movement type things. Nathan Hall: If you have any just anecdotal examples of some of the customers you've worked with, as to how this is impacted them in terms of kind of accelerating deployment or lowering operation?
30:45
Marsha Pierce: Oh, yeah, I mean, well, one of my favorite stories was one of our early adopters of containers that Jao worked with, was actually a company I used to work for. And we were an agile development shop. But we never got to the point where we were able to give every developer a copy of their data. And so they
31:05
went with a pure service orchestrator and containerization of data, and we met with them, and they were kind of deploying that out rapidly. And then I've got like some managed service providers that are using this, to rapidly clone out and restore, right, they may have on a given
31:24
instance 1000 databases. And one of those databases for our customer needs to be restored. And they are required to keep I want to say years worth of backups. And so they were able to leverage some of our rapid resort technology and snapshots in order to act economically and meet those RPO RTO times of
31:45
rapidly restoring that data. And, and recovering from ransomware events. That's the you know, big thing that every worried about your data set has to be protected. Because if you lose that, unlike your web servers or your apps, or if you can just redeploy those, those are pretty quick, especially as
32:04
we go to services. But that data set has to be restored no matter what, or it's going to cost your company a lot of money, or be kind of a catastrophic event that can put you out of business. And so being able to get those datasets back up and running, we've helped out quite a bit in that space as well.
32:23
Nathan Hall: Sure, and I remember the beginning of your what you're talking about, you said we're 100% API driven for those out there that want to kind of see examples of that code, Doctor storage, comm has essentially our entire, you know, RESTful API kit, but also a bunch of different modules
32:39
there with examples of putting that into action integrations with terraform. I think one of my favorites is there's a, there's a few API calls for changing the color or the frequency that the bezel lights go off. So that's a really, really important one,
32:53
Jon Owings: very important. Marsha Pierce: Here's a cool thing. That's like we tried to do but the eight, there was no common API call. And now I've talked to about customers that some of our customers do, they have a monitoring system, they get a space alert. So what
33:07
happens when you get that space alert, usually what happens is by the time the on call, person gets up in the middle of night to go and dress that space alert. They've already had a downtime, right? Because it's a log file was growing, and it rapidly build up, but it causes an issue. But if you're API
33:23
driven, you can create an automated response when you get a space alert to go to go automatically expand that drive, and give somebody time to go address it while you're logging in. So I think those are some of the cool things you can do once your API driven, right? If you can think about it, you can do
33:40
it. Nathan Hall: Yeah, that's that's interesting. You could that you could definitely do that. You could you could also throw on for example, do an API call if it's if you're getting that to go apply QoS and start throttling that, that workload
33:51
down. So you're also not filling it up as fast. So there's a bunch of things you can do. Again, if you think about it, you can do it, you can automate it. So JO, how about on the cloud native side here, there's, you know, we have the the acquisition of power Portworx, that's added a ton of value on
34:07
top of some of the other integrations we already have been doing with Kubernetes and stateless storage Portworx being foreign above the leader, according to you know, the analysts out there in terms of the Kubernetes data services space. So what what is the combination of Portworx now with
34:23
Pure Storage enable customers to do now? Jon Owings: Definitely. So obviously, I think they announced in the first week of accelerate, but there's a there's a new version of Portworx that will actually use API Yay, API's? Well, there's a theme today to actually automate
34:40
the storage from a FlashArray or a FlashBlade, right? So Portworx, you know, is a virtualized layer of software. And it needs physical storage from somewhere, right. It can be anything, but we actually now you just give it an API key, and it can go and it can go and get new storage from the FlashArray,
34:59
for example, and like the example Marsha was just giving to us is that it'll actually monitor, it'll actually monitor that and expand the pool. So if I start filling it full of containers, and it reaches a threshold that I set 70%, it triggers an action that will expand the pool, what that means
35:19
is either one, it goes to the array and says, give me give me more FlashArray volumes boom in the pool gets bigger, super, super easy. And something where, you know, now I'm not, you know, having a database crash in the middle of the night, because the drive filled up, right. And so that's always a bad thing. And I
35:37
don't like getting called in the middle of the night, so none of us do. So that's one thing with that. The other is, is that now Portworx. Because of that layer, it provides this data services that just aren't, they're not there natively in Kubernetes. And a lot of people don't expect that they expect they expect to
35:56
be like vSphere. And so when they deploy it, they go and the project or when they're about to go into prod someone asked a question like, what is our dr plan? What how do we back it up, right? And Portworx provides those hooks to do that, right? It has a way to do Dr. It's either synchronous or
36:14
asynchronous, right? It has a way to do backup, right? With PEX backup things that like we that people aren't considering when you're when you're just deploying Kubernetes and you're like, hey, look at me, I'm deploying code. And it's awesome and great, but then someone is gonna come along, and eventually
36:29
when the, when that code that you wrote is so important, they're gonna say, hey, how does how does that fail over? How does that data get moved to the cloud? If If this whole data center disappears, you know, those types of things are built right into your work. So that's that's what's exciting there.
36:47
And, you know, you got customers, you know, if hopefully everyone saw the interview with Roblox in the first week, you know, they got a skit that theirs was a scale issue, right? They it wouldn't COVID came in and everyone, everyone's kids, if everyone knows Roblox, right, everyone's
37:04
kids got sent home. Right? They they gained a year's worth of users in one month of new users. And they basically said, if we didn't have, you know, containers and Portworx to scale, we we would have fallen down, we basically would have crashed, because it was just too much at once. So they that, that
37:24
automatic scaling is essential to them, being able to, you know, do awesome lead through COVID. And, you know, IPO and all the stuff that happened, you know, because of that, right? It really made a name for them like this, this thing is awesome. I think my kids, just yesterday asked for more roebucks. So I
37:42
mean, they're doing well. Nathan Hall: Yeah, that rapid scaling is a common thing that we see with with Portworx. And I just think of another you know, one of the largest cellular providers here in the US that relies on Portworx for when a new iPhone comes out, and that
38:01
surge in retail sales that the capability to explode out scaling very quickly, so that you don't drop sales or kind of slow them down. That's the perfect use case. No, we we did talk about API a lot already. But j, I'm curious in terms of where do you see some of our most popular and effective
38:23
integrations with orchestration and automation frameworks? Jon Owings: I think the most used and most asked about one is Ansible. Right? It's, it's, it's just a great tool. It has so many, so many ways to do everything from networking, to deploying apps to deploying Kubernetes. And having that
38:43
integration into our arrays just fits right in for a lot of customers that asked this question, I've actually had a bank asked me this, like, I see how easy it is to manage one of your arrays, like with the GUI, it's super easy, but we want to buy 500 How do we manage 500 at a time and Ansible? Does that
39:02
right? It's one of the it's probably the one of the most popular frameworks. I mean, we have SDKs for Python and go and those types of things. And if you really, really want to do the, you know, do it yourself, you can write like some people like hey, I'd rather just write the Python myself. But um,
39:20
Ansible will do just about everything you can from the COI, or the gooey of flasharray or FlashBlade. Nathan Hall: It Well, you know, as the father of two young boys who were home for about the past 14 months, there is a part of me that wishes Roblox hadn't figured out how to scale so
39:38
quickly, reliably. But maybe would have been even worse. So, at this point, I want to open it up. Thanks to both of you. This is a great great summary in terms of you know, how we're here to help some of the things that you know, we're we can integrate with some of the trends that we're seeing and I
39:58
know that we have certification programs running The portfolio and we we partner with lots of key Alliance partners on ways to train and augment your skills. I'd also say just us, as field technologists a huge part of my job, our job is going out there to customers and helping educate and helping figure out well, you
40:16
know, is this possible? How can we make it possible? and learning how to get there together? So we're always happy to, to directly engage and kind of help on that career journey as well. Actually, it looks like we did have another question come in here to the chat. So how DBA and I see how to DBA and IT
40:36
admins jobs are going to change as we see a trend of having multiple skills. So just looks like you know, how do we see the DBA and IT admin shops changing with this trend of multiple skills? And, Marsha, maybe you could start with that answer. I'm not
40:53
Marsha Pierce: gonna say that. In general, they always had overlap areas, right, where they had to work together. And I always joke that, you know, DBAs have great communication skill. So we haven't always explained our infrastructure needs to our IT admins in the best possible way. We know we need certain
41:12
things, as we are going to containerization data, some of that has gotten lost, because people stopped worrying about how everything was laid out. And this is where I talk about really getting with your IT admin and understanding that storage strategy that you're still going to need, right?
41:31
Because you have stateful stateful data that you're going to have to create, you're going to have to talk about how you're going to integrate that into your overall infrastructure plan for copy data management, aka Dr. Just like JO was talking about. And then the other thing I think that people need to also
41:49
start talking about, and it's it comes down to more communication between the groups and working together, is that one of the things that we saw with lots of spinning up of, you know, containerization, and databases and services, is that we actually started to get in silos of data, right? Somebody went
42:06
and developed something, and they spent it up in one cloud provider, as a Database as a Service, it was a unique database, but all of our data was sitting someplace else. And how do you get that data then to integrate? And because there's shims to your deeds for that data? So what do you have Jon?
42:24
Jon Owings: Well, I think I think we're gonna see more people become like a data arc, not like, like, I'm not a sequel DBA anymore. I'm a data architect, right? That's up leveling, right? You're no longer a vSphere admin, you are like a cloud architect. So there's more there's, I think
42:41
more people are going to be seeing things from an architecture standpoint, because a lot of the hands on stuff, maybe maybe automated, and there's still going to be people that are going to have to, you know, racks back servers and install OSS, and you know, that some of that some of that's
42:56
going to be automated. So there'll be more people that are going to, I think those that embrace this change are going to get that more architectural type view strategic view of what's going on when it comes to data infrastructure, utilizing the cloud, whatever it's going to be right. So I think the those are
43:17
going to be the ones that kind of rise to the top. Nathan Hall: I agree. Great. Well, thanks, Marsha. Thanks, JO. Looks like we are just about up on time. So I don't believe we have time for any more questions. If there's any that we've missed, we'll happy to be able to reach out directly. I
43:33
want to thank you both for for joining me today. And thank everyone for attending today's session. I hope it was helpful and we hope you have a great rest of accelerate it by our canvas
  • Video

From NoSQL databases to containers to DevOps automation tools and AI, there's a general shift from specialists to generalists driven by the diverse ecosystem of technologies available for solving complex IT problems. With technologies evolving at breakneck pace, tighter development cycles and integrated IT teams have never been more important. In this Pure Storage industry panel discussion, Nathan Hall, VP, AMS Systems Engineering, Marsha Pierce, Director, Field Solutions Architecture, and Jon Owings, Director, Cloud Architecture, discuss the major trends shaping the world of IT professionals.

07/2024
Pure Storage FlashArray//X | Data Sheet
FlashArray//X provides unified block and file storage with enterprise performance, reliability, and availability to power your critical business services.
Data Sheet
5 pages
Continue Watching
We hope you found this preview valuable. To continue watching this video please provide your information below.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Your Browser Is No Longer Supported!

Older browsers often represent security risks. In order to deliver the best possible experience when using our site, please update to any of these latest browsers.