00:06
Hello, everyone and welcome to the pure storage, automotive point of view. My name is Peter Eicher and I will be your tour guide as we explore the automotive industry together. Well, before I start, I wanna say a brief word that because I am talking about the automotive industry, I will be making reference to various uh auto manufacturers throughout the course of this presentation.
00:28
But because I happen to mention uh a manufacturer that is not meant to suggest that those uh that that manufacturer is a pure customer. Now, we certainly have many uh customers in uh the automotive world. But my referencing anyone is no indication if they are or not a customer. So I just wanted to make that point clear, but now let's get started.
00:50
Well, it's certainly not your grandfather's auto industry anymore, right? Uh It's an industry that has seen a dramatic change just in the last couple of decades, probably more than a change uh in the previous 100 years. But let's take a quick turn through kind of what you know that that journey of transformation that the automotive industry has
01:09
gone through. Now, we started of course with a an analog vehicle which, which was probably about 100 years of, um, automotive development, uh, where they had little or no digital functions. And then we began to see the intelligent vehicle. Now, as I understand it, the first vehicle to have a kind of computerized,
01:27
uh, ignition system was Volkswagen in, uh 1968. But what we think of as the intelligent vehicle really started to come around in about the late 70s and early 1980s. And by the 1990s, it was pretty much standard in all, all cars. They were all being built with onboard software with digital functions, monitoring systems. And you started to see digital uh displays like
01:48
in the picture and so forth. That brings us to today where we have the connected vehicle, which is basically taking the intelligent vehicle and connecting it to something outside the automobile itself. Now, the first connected vehicle was actually General Motors in 1996 with their OnStar system. I mean,
02:05
not counting third party clunky car phones that you would say uh earlier than that, but that wasn't part of the car itself. Uh But on Star was the first system. But what when we think of the connected vehicle today, that kind of thing where you would go out and get internet services and so forth began about 2004 where uh BMW released one of the first uh connected vehicles. And now that's again,
02:28
pretty much standard stuff but where that's all going is a full mobility ecosystem. Now, we're part of the way there already, but there's still a lot more to be done there. And that's where the car, your car is gonna connect to lots of different things. Not just the internet to, you know, play music or something, but it's gonna connect to the dealership to the service department,
02:47
uh to other vehicles on the road, to different uh partner companies offering services, uh different platforms, uh things like smart cities, for example, we may even see the car kind of talking to the road infrastructure and it may be forced for example, to slow down in certain parts of the city because the the infrastructure itself is talking to the car and telling it you can't go beyond a certain speed,
03:10
right? Lots of very interesting things certainly being planned and a lot of this will be dependent ultimately on the rollout of five G and later six G data services because of the uh much improved latency of those services services, you can do kind of near real time things uh or you will be able to now where we are today in that connected vehicle. When you, when you start to look at the I T
03:32
infrastructures of automakers, you know, they're, they are needing to optimize today for that connected vehicle development uh which is where, where we are. Uh But they also at the same time, need to prepare for that fuller mobility ecosystem, which is coming along uh fairly rapidly. And now the new world of uh software and A I Driven Automotive is really uh where things are
03:52
headed. Uh mckenzie, the analyst firm makes the point that they see uh you know, they do a lot of studies of the automotive industry and they are saying that in the next 10, 15 years or so, uh what's going to separate one manufacturer from another uh is going to be their ability to develop software and leverage things like artificial intelligence, machine learning, high performance computing,
04:14
all those things that kind of roll together under A I. So it's the software that's going to separate uh winners from losers in this industry. And again, it's a highly competitive industry and you see that, you know, the software at A I taking roles in everything, right, in autonomous driving, uh and smart manufacturing, right,
04:30
trying to reduce defects, trying to reduce problems on the assembly lines. Uh electric vehicles are far more computerized than internal combustion vehicles, even though they are both connected, there's still a lot more computerization on the electric vehicles that connected everything we've been talking about. Even the buying experience for users is becoming uh more software driven,
04:50
uh the service experience and so forth. And ultimately underlying all this are goals around zero emissions and, and sustainability which are being driven uh partly by government regulations, but also the manufacturers themselves have really taken a strong stand uh in favor of this. But there's a challenge, right? What I call a challenge of technical debt,
05:12
uh when it comes to that need to move forward with software development. Now, the auto industry is interesting in that it's really split completely between what I would call the Legacy, uh O Ems and what I would call the new kids on the block, the all electric type manufacturers. Now, when you look at the Legacy O Ems, there's a, a number of,
05:31
you know, companies listed there with the date that they were founded and they go back a long time, right? Da was actually going back two centuries to 18 90 Ford and General Motors in the early part of the, you know, 19 hundreds. Uh and you can see the rest. Uh So the thing is that all these companies
05:47
were started before the era of shooting even even began. Uh So they all originally had mainframe computers, they were probably using a S 400 or now, you know, I B M I series on their manufacturing floors. And in fact, a lot of these systems still exist, right? But when it, so when it comes to moving to uh the new world of things like containers and
06:09
micro services and so forth, they're stuck in a way with a lot of legacy I T infrastructure that they have to maintain while moving ahead at the same time. Now, the new kids on the block largely didn't have that right. They were born long before, long after the era of mainframes and, and so forth. Uh So they began with much more modern
06:28
infrastructures. Now, when you look at the, you know, the birth dates of some of those companies like Tesla and Rian and Karma, et cetera, they did come around before the era of containers, which started around 2014. Uh But still, they were, it's much easier for them to shift gears to that kind of development because they don't
06:46
have that whole legacy infrastructure even. Uh for instance, the data models on the legacy side, you have a lot of monolithic database infrastructures oracle sequel server. Uh in DB two, we still see uh out there uh in some of these manufacturers, you know, databases that have been running for decades and contain uh just vast amounts of vital information, but those aren't so good when it
07:06
comes to trying to do new kinds of software development. Of course, the legacy players have their own advantages, right? Vast dealer networks or millions of customers, right? Name recognition. Uh I mean, on the new kids side beyond Tesla, a lot of those names aren't particularly
07:21
household names at this point. Um So you do have kind of a balance of advantage and disadvantage, but we're gonna, we're talking very much around the software development. So this is where, you know, the legacy of manufacturers really have a lot of challenges. So what are some of those challenges of modern uh software development?
07:39
Well, it all comes down to the the fact that in order to develop software in an agile quick responsive way to consumer demands, you really need to move to the new applications and, and data services that are out there today, which means containerization, right? Open source databases and, and, and and data services, you know, COCA MySQL, Cassandra, a zookeeper, you know,
08:01
there there's just so many of them uh elastic et cetera, you need to move to things like object storage. Well, part of the problem there is that those aren't necessarily compliant with your existing I T standards and processes because you've been doing it for so long and you have so much legacy infrastructure, this can hold you back or make it more challenging because the unfortunate thing is
08:22
you can't just walk in tomorrow and unplug your D V two database and switch to something else, right? Uh It's a very long process to, to migrate data, to move to a new system. Uh There are also a lot of intensive resource demands, right? You have more data than ever. Uh everything that's happening in automotive
08:39
autonomous driving connectivity, et electrification, et cetera. It all generates data and data in vast amounts, right? So that data has to be processed within uh you know, time windows that are forever shrinking due to competitive pressure. Uh And traditional storage architectures often can't keep up with this and then you have increased complexity again,
08:57
especially for the legacy vendors because you know, you have to maintain that legacy infrastructure at the same time that honestly competitively, you have to start rolling out things like micro service architectures, right? So you have to do both of these together. Uh and your I T skill sets may not be ready for that because it's different, right? Even things like backup, disaster recovery,
09:16
data migrate, movement, data migration, it doesn't work the same in a containerized world, the way that that you know, that you've been used to doing it. So these are a lot of challenges. Um And when you look at artificial intelligence, it adds even more challenges into the picture. Uh For example, uh that first data point is from uh the V M ware.
09:35
They did a survey and they found that only 53% of A I projects make it from pilot to production, which is another way of saying that about half of A I projects fail, right? That's after spending money, time, resources, et cetera, the project ultimately never even gets deployed, right? That's a very high failure rate.
09:52
Uh And that comes out of the same things we've been talking about that, you know, that those new A I infrastructures are difficult to manage, difficult to integrate with a legacy infrastructure. It needs that more modern infrastructure, things like automation, repeatability become very difficult, you know, even if it worked in,
10:08
in in a pilot often fails in production. And that ability is to scale right to scale to the massive amounts of data and the constant stream of data that's coming in. And then with uh you know, with sustainability, you have another layer of concern is that while you have these massively growing amounts of data, how do you store and process more data than ever while dealing with power uh consumption restraints,
10:34
either the need to eat, reduce or at least maintain your current power levels. How do you add, you know, petabytes of storage every year while maintaining that you can't just keep buying the same stuff, right? It's not gonna work. So there's a lot of challenges out there. Uh you know, that make, make a lot of these things are really difficult to roll out.
10:52
So how can pure help? Now, what we're gonna do in a in a few minutes is talk about specific use cases that pure has rolled out with uh some of its automaker customers, things like things like manufacturing uh autonomous driving and so forth. But I wanna touch on here kind of the high level view of how pure can help. And and the first thing is that we have a
11:13
modern data architecture which is unlike a lot of our competitors now in the same way, you know, I split the legacy O EMS from the new kids on the block. Pure is like the new kid on the block in terms of data storage. Whereas our competitors are mostly the legacy, you know, they all came out of a disk based architectures uh pure,
11:32
never had a disk product, right? It, it was born as a uh all flash system and because of that, we can do things or we have done things very differently. Really. We've been down to the bits and bytes level. Now that's a whole side discussion. And, and there are a couple of great blogs on our website about this talking about uh uh how we have better science and it's really
11:50
interesting and valuable stuff, but not for this conversation. Uh But the bottom line is that our uh difference in our science in our architecture allows us to get the performance. We do the incredible levels of reliability. We have to do things like proactive support, that's A I driven and so forth. So having a fundamentally different architecture is uh important.
12:12
Then you have performance issues around high performance or, or low latency, but there's two sides of performance, right? There's latency and there's bandwidth now, latency is critical for workloads like databases, S A P which is still, you know, very prevalent in automotive. Uh that's where you need to process transactions very,
12:29
very quickly. But then you have bandwidth issues which are for things like A I and analytics or log of management or log monitoring. Uh They're fundamentally different. I like to to use another car analogy, right? Uh, latency is like a drag race. You're trying to get one car from point A to point B as fast as possible.
12:46
Whereas bandwidth is like a 10 lane highway, you're trying to get as many cars down the road as you can without running into a traffic jam. Right. And these are two fundamentally different problems to solve. Uh and it becomes even trickier when you need both at the same time, right? So, uh you know, that's an area again where
13:02
pure can help uh with our, our various product lines. We've always been very simple. We've always focused tremendously on, on, on ease of use, ease of deployment. Uh I'll talk about that when we talk about manufacturing. Uh the fact that our software is very cloud and container friendly, right? You, you may be in a phase right now where
13:20
you're virtualized, you're using like V M ware or whatever virtualization platform, but you're moving to containers. So you have to move to containers. Now, we do great jobs with both. So we can make that transition easier. And finally, there are our history of continuous uh innovation. One of the nice things you know, that our customers really appreciate is that all our
13:39
features, all our software features are included, uh you know, at no extra charge, we're never up charging you for a specific feature. So anything that we release, you have access to uh and in fact, you can upgrade uh the software non disruptively. At, at any point, you can even upgrade hardware non disruptively,
13:55
which is uh becomes critical in some scenarios. And we've also innovated around buying models because financing can be very different from one company to another. But do they prefer, you know, Capex spending or do they prefer Opex spending? And we give you models to satisfy both whatever your particular procurement procurement needs are, you know,
14:14
we we can help you, we even have our own, you know, financing and so forth. So what is the point of all this? Why are the auto manufacturers so interested in software development and artificial intelligence? Because they're all targeting data driven outcomes. And if you look at like in, you know, the uh annual reports of a car manufacturers,
14:35
you'll see a lot of the exactly what I have on this slide here, right? Of course, they're always as any business does trying to accelerate revenue growth, trying to grow profit margins, uh trying to get faster time to market because it's such a competitive industry, right? You have so many uh other players in the space,
14:52
whoever you are, right? That they're, they're doing all the same stuff or they're trying to do the same things. Uh there's this movement towards high growth subscription services, right, where you want to start getting that annual recurring revenue due to services you're providing to your customers and the bottom line of all this is really to enhance customer
15:11
experience and brand loyalty. Because one of the challenges as an, as an automaker is if it's time for me to get a new car, it's very, very easy for me to go to somebody else. If I have, you know, if I have currently a car from manufacturer A, I can move to manufacturer B pretty easily, right?
15:28
There's no, nothing's really stopping me. Uh But if you start to get the, your customers tied into the user experience, suddenly that changes I I I use the example of, you know, iphone versus Android. Now, I'm not taking a stand on either one. Uh But the point is they are fundamentally different experiences,
15:47
right? If you look at an iphone, the way it works, the software experience is very different than Android, some prefer one, some prefer the other. But what, what a lot of people do do is change, right? You don't, typically if you're an iphone user, you tend to be one for a very long time, if not forever.
16:02
And the same similar similarly on the Android side because the user experience is so different, it locks you into a particular operating system, right? This is what the car manufacturers are trying to get to, to the point where you're so happy or so tied in to the the software experience of that car that if you move from, you know, vendor A to vendor B now you have a whole
16:25
different software experience. So even if you like that car better, you like, maybe it's cheaper, maybe you like the styling, whatever you're gonna have to give up something that's become very, very part of your daily existence, right? So it's difficult to do and that's why this focus on software is so critical to the uh car manufacturers.
16:43
So now let's look specifically at some use cases where we have helped some of our uh auto manufacturer customers. Now, speaking of that, I will mention that eight of the top 10 global automakers are pure users to one extent or another. Uh Again, I'm not gonna call out specific names, but uh we'll leave it at that.
17:01
So let's look at some of these specific use cases where we have helped uh manufacturers. Uh One of those of course, is manufacturing, right, very critical. There's a lot of talk around industry 4.0 moving uh you know, even though manufacturing is already highly computerized, it's getting even more. So uh A I is being used for things like, you know, have cameras looking at everything on the
17:21
production line for spot defects or spot issues, you know, in a way that humans can't, they just simply can't have that kind of a scope of view and that's all A I driven. But when it comes to actually deploying storage in a manufacturing facility, one of the most critical things is storage reliability, right? Uh One manufacturer where we have started
17:41
actually deploying our, uh, storage into their manufacturing centers. They noted that a down time, uh, of a production line costs them, uh, between one and $2 million. Now, can you guess the next part? Is it one or $2 million a day? An hour? No, it's actually 1 to $2 million a minute.
18:01
Right. That's a pretty big bill that runs up. Uh, if your production line suddenly has to stop because you had a storage problem, right? So reliability is really important, but not just that it's also that ability to upgrade your storage software or hardware without bringing anything down because a lot of these production lines are 24 7,
18:22
right? Uh So this particular manufacturer uh put our storage through all kinds of highly rigorous uh testing. They were pulling out flash modules, you know, unplugged, turning off controllers trying to do all kind kind of stuff uh to make the storage stop working and we pass all their tests uh and hence are uh in the process of rolling out across uh many,
18:45
many uh of their manufacturing locations. Of course, performance and, and data management are aren't important, right? You always have to be able to handle uh the required performance levels, but it was also that ability to innovate into their existing automation framework. In this case, it was a, it's an answerable environment.
19:03
Uh and pure has always been an API first API friendly uh storage platform. So, and we have lots and lots of uh tool kits for the different automation tools and so forth, terraform whatever. Uh So it's very easy for you as a user. If you have an automation framework, it's not difficult at all to insert pure into that. And of course, another big aspect was energy and space savings.
19:26
Now, this is a big problem in data centers. But when you're in a manufacturing plant, you know, it's not a data center, it may have a data center in it. But typically those are space limited, right? You can't make them any bigger, you can't bring them more power, you have what you have yet, you also have to deal with, you know, massive amounts of data,
19:42
data growth. So pure is really helping uh this manufacturer in in because our uh our products are so dense that you can fit so much more data. And we have such great uh data reduction technologies that in the same amount of rat space, you can fit vastly more uh data or just shrink the amount of rack space you're using. And also because our power consumption is about, you know,
20:05
ballpark, 80% less than uh most of our competition. We're also helping them with that, you know, at the same time, you're trying to roll out more things like GP US in the manufacturing plant which are very power hungry, right? We can help you dramatically reduce the amount of power, at least you're spending on storage. So lots of really good things we do for uh auto
20:25
manufacturers in the manufacturing space as well as any other manufacturer that have a lot of similar uh concerns. Now, another area is improving the customer experience. So we talked about, you know, the need for the software already and we're helping one of our uh manufacturer customers exactly with this by helping them with their uh movement to
20:44
containers and micro services by using our port works uh product, which is a port, which is a native data management platform that really helps free up your I T team from a lot of the kind of day to day maintenance that they would normally need to do because port works, takes care of that for you, things like data protection, uh data migrations, um load balancing,
21:07
performance, balancing, all that kind of stuff. Uh even migrating between say on prem and into public clouds, you can do all that with port works. Again, it's a whole separate side discussion. I'm just touching on it here. Uh But it really helps, particularly if your I T infrastructure team is not that well versed in the container world, right? It becomes very important.
21:29
Now, you're also uh the the end result of a lot of is that that is enabling greater developer productivity because it's a code Brunetti is native, right? The developers can actually manage their own storage resources via code they can simply request a, you know, amount of storage from the infrastructure import works will give it to them based on different performance S L A s and so forth.
21:50
It's really makes life a lot easier. Uh Instead of, you know, putting in a ticket for X amount of storage space and waiting three weeks, right, your developers can just get it whenever they want it. And ultimately, you know, port works works with any uh infrastructure, you know, uh public cloud or on prem storage. But if you use it along with pure uh all flash
22:10
storage, you actually get a bunch of additional advantages and of course, our typical performance and reliability. Uh and again, that's a, these are all drill down topics for discussion. But uh we are really helping this particular um automaker who is specifically using this to develop their in car. Um you know, user facing applications to try to
22:30
drive that brand experience. Now, of course, you have autonomous driving and connected vehicles, autonomous driving, assisted driving, all these things, these are super data, data driven data uh hungry and this is where you see a lot of GP U use, right? From companies like NVI, which is a partner of,
22:46
of yours. Uh But one of the problems that organizations have is that, you know, they roll out these GP U resources but they can't feed them fast enough like a GP U is a hungry beast. Um And if your storage can't serve up data quickly enough. You sit there with a, a lot of unused cycles,
23:05
right? A lot of unused compute cycles, which is wasting money and wasting time. OK. So we had one uh partner with, I can't name is, is NAV Info, which is a company that does uh mapping, you know, digital mapping for autonomous driving. And they had exactly this problem. They had this,
23:21
you know, uh GP U environment and it was sitting there only running at about half of its potential because the storage was the limiting factor. So they moved to our pure uh flash blade product, which is super high bandwidth. We talked about that earlier, right? It's a, you know, it's the 10 lane highway of storage uh that allows you to have tremendous amounts
23:39
of data, uh you know, into the system out of the system being processed within the system. So this has allowed them to maximize uh those costly GP U resources. This in turn reduces project timelines, right? If you can do more at the same time, uh and in shorter amounts of time, if you can process the data faster, everything goes faster.
23:59
And because again, we're so api first api uh friendly uh flash plate fits very easily into existing, you know C I CD pipelines. Uh So as part of your whole uh end to end delivery uh process, you know, we we can fit right in there and just really supercharge uh your performance. Then of course, there's a transition to electronic vehicles.
24:21
And the thing about uh electric cars, as we mentioned earlier is they are very software driven, they are much more uh software driven uh than internal combustion cars. And everything comes down to algorithms, right? Product advances, product enhancements are very algorithm dependent. And again, this is similar to what we were just talking about with the GP US,
24:41
right, the ability to process lots and lots of data quicker gets you to better uh results. You know, one thing I learned recently, which I thought was sort of fascinating was around uh car charge times. Now, one of the barriers to electric car adoption is uh a lot of people are uncomfortable with the charge times, right? Because you're used to going to the gas station
25:01
filling up, you know, a couple of minutes you're done. Whereas you might have to wait, you know, hours to charge your car. In fact, you can increase or decrease, I should say decrease the time it takes to charge a car via a software update, not even changing any physical components in the car or the charger. If you have a better algorithm that you come up with and you update the software in the car,
25:24
your charging time can decrease, right? So those are the kind of things that get delivered by better smarter software development. Uh And of course, you know, there, there, there's so much going on at once that you need to have a very large data repository because you also get things like digital twins or virtual twins, right, that are being used to help design these cars and those are super,
25:44
super data intensive, but you need data from all parts of the car. So having siloed, you know, data lakes uh isn't gonna work, you need one giant, you know, data reservoir, so to speak. Uh and we can help you with that as well as well as planning for the unexpected, you know, talk about our various consumption models uh because it's difficult to know at the
26:02
start of a project how much storage you're gonna need. So with our, you know, buy as you need uh our pay as you go consumption models, you know, you can avoid those either buying too much stuff and letting it go to waste or buying too little and then being constrained in a project, right? And then of course, energy and sustainability as we've been mentioning along the way,
26:22
right? Uh going back to our discussion about how our technology is fundamentally different. This allows us to do things that other companies cannot do. And that is why we, you know, we have the full studies on our website. If you want to take a look, our flash array family, for example, uses approximately 80% less power than competitive
26:41
flash systems. That's the same thing about disk if you're still running disk, uh you know, the, the power savings are, you know, as much as 10, 10 times, right? And the difference is that, that science, we have our own direct flash technology that we develop. Uh And we can, uh you know, because of that, uh we can deliver the kind of density and power
26:59
efficiency that uh most of our competitors can't. Uh and we reduce e waste because of the, the way you can upgrade a APR A in place. We've eliminated the notion of that forklift upgrade where every three years or four years, you know, you take the whole story a basically wheel it out and bring it to a, bring it to the dump, right? Which is ultimately almost what happens.
27:19
Uh But with ours, because you can replace individual components, you can upgrade even into a new controller model, a new controller generation without, you know, throwing the whole thing away. So it's not only does that save you from having to do a lot of data migrations, but it saves you a lot of e waste as well, which is again, is very, uh you know, high on the radar of most uh auto manufacturers.
27:41
So wrapping it up, let's, you know, I, I really haven't talked much about product but I just wanna kind of put all the things in place, you know, and on this slide, we have across the top, you know, a lot of what auto automakers are concerned with right now, things like machine learning and high performance computing A I uh the rendering side, which we haven't talked about another area where, you know,
28:00
performance is, is important. And then traditional applications, right? You still have those oracle DB two sequel giant stacks out there and lots of them. So you, you, you know, you, you need to maximize that uh as you know, until you, you transition to something else. Uh So those are kind of a lot of what is going
28:18
on on across the top and on the left side, you have all the data sources, right? And there, you know, we talked about a lot of this autonomous vehicles connected everything, but we haven't really talked about supply chain or simulations or robotics, right? There's just so much data being created so many data sources, right? And those can all feed into our stack of
28:37
storage, object and file based uh storage, a block, I should say block, object and file based storage with our management framework across the top, all leading to you to have uh better outcomes around better flexibility in your I T department, reducing that complexity, increasing performance, uh improving product quality and so forth.
28:58
And finally, to wrap it up with a more specific look at our actual product line, right? So this is all our uh our products up to the up to the minute here, I'm not gonna drill into all of this, but just quickly the you know, the flash array X and XL family are real high performance, very good for those transactional type workloads, super reliable.
29:17
You know, those are the arrays that that that manufacturer that I spoke about is deploying in their uh in their production facilities. We have flash A AC which is a high capacity all flash storage. So this is kind of meant more for like a tier two workload or a or a large data repositories or data protection, right, as a backup repository, for example, and that can really help you replace a lot of
29:39
spinning disc. We have our flash blate S which is for very very high performance file and object storage, that's your 10 lane highway. So for things like analytics and A I and so forth, we have that massively parallel processing that really can speed up some of those critical uh projects you're working on. And then our most recent project just released uh about a week before this recording was uh
30:01
flash blade B which is capacity focused file and object storage. So this is very good in conjunction with some of the other products. Uh it can be great as a data protection repository replacing you know, those racks and racks of spinning disk, you're currently backing up to at a, at a equivalent price point and a much lower cost of ownership,
30:19
you can replace that with the flash lady and also a lot of the A I uh type workloads don't all need high performance. A lot of what you need is just big performance, lots and lots of room for, you know, billions of objects kind of thing that may get called on occasionally but aren't necessarily uh requiring the highest end of performance. So you can have for example, a flash blade doing your high performance workloads.
30:43
And then the flash blade e deployed for that sort of longer term warm coal type repositories, which you occasionally access but don't access all the time. And again, all of these come with exceptional energy efficiency. Uh particularly if you're replacing disk, you know, you can, you can take out racks and racks and racks of disk and replace them with a,
31:03
you know, uh a couple of couple of flash blade. Uh and you're done, right? And then finally port works which we talked about, which are co kubernetes uh data service platform, which is a software, you know, only solution which can help you uh in your journey to uh Kubernetes and micro services and containers, which as we've seen uh multiple times throughout this presentation is
31:25
going to be uh extremely critical uh in the future for all uh auto manufacturers. So I hope you found that uh interesting and and lightening, you can visit our website for uh all our various uh collateral information around uh what we're doing in, in the automotive space. And otherwise I thank you very much for listening
