00:00
Hello and welcome to this month's coffee break. Stop working for your unified storage. Let it work for you. My name is Andrew Miller, lead principal technologist joined today by Jonathan Carnes. Thanks so much for being with me for, for jumping in the boat here today, John. So I was thinking I will pull the music down
00:18
here a little bit as we get going. Yeah, keep it low a little bit as we get start up as always. This is a series. Thank you for joining us this month's coffee break. Uh If you enjoy this or you've listened to previous ones, always want to highlight that we keep all the previous ones online.
00:33
You can find them both at pure storage dot com slash events, the previous ones as well as they actually maintain this Google Slide that has all of them and some actually organized by topic. Uh We found that this this um format actually ages a little better because we try to keep it a little more agnostic and it's not just about the tech, it's about, you know, things that surround the technology and the pure products.
00:51
And then you can even see previous ones there. I do want to make sure to highlight this is the last coffee break before pure accelerate. That is literally next month. Uh next week, I will be out there. I think I'll see you out there John too as well as a bunch of other folks. Um heads up for anybody that wants to join. There is no live streaming this year of the
01:09
keynote. However, the keynote replays will be up quickly on the Accelerate digital landing page probably around 24 to 48 hours. Definitely recommend watching those and all the sessions are gonna be recorded as well, which is actually pretty cool too. But for the last time, the amazing 42 2nd video for Accelerate,
01:44
I, I almost feel like John, I should narrate over this three days, more than 60 live sessions, keynotes, breakout sessions, product demos and free training with certifications. You have to pass them, learn connect and accelerate. I don't think I'm gonna get paid extra for my voiceover work there, but you know, I don't think anyone's counting couldn't,
02:07
couldn't, couldn't hurt. So um as always, there is a gift card drawing. There's a gift cards for attending the 1st 1000 participants to register and attend organize. That's often why people join, you know. So thank you for that. And we will do the drawing at the end. I was actually really cool.
02:21
I was in Columbus, Ohio uh for some E BC uh kind of EB CS on the road or field EB CS this last month. And I actually met one of the folks I should leave him nameless who actually won one of these. And I was like, hey, that's kind of cool. So we do actually send these out and he got it. I'm your host as always, Andrew Miller.
02:37
Uh I don't as you know, usually like to go through all of my background because it doesn't make sense since I'm here every month with you. But in this case, when it relates to unified storage, we're just talking about file and block. I actually started all the way back running an apple file server when I was on the help desk side and I remember how you could do a copy and it would do it server side and not run it
02:56
through the clients and then went over to the it operations side and started with netware. Um RW ce M fa you know, your um if you actually have network certifications and never expire. But you know, so I've been playing around the file stuff for a while. There was Linux in there as well as some windows file servers. And then eventually, you know, some of the storage vendors,
03:14
John, you've been at pure for a little while though. Do you mind giving a, a brief run through of your background and then we'll spend even more time doing it in uh in, in actually, maybe even of your current role because we're gonna wander into your background in section one. Yeah, so I've been APR for six years. Um five of those I have covered uh unified
03:33
offerings um on flash file and block. Um So covering the SNB NFS protocols and uh currently working on file services as one of the P MS because of some of that background that you see there, we'll talk about that more too. But thank you. And uh I, I've been, I've been around it since the nineties. So I started out in Netware like you.
03:55
Uh well, a little bit like you and back in uh Netware four, I believe I started on working on NAS products. Uh that was, you know, yeah, we'll talk about it a little bit later, I guess I'm sure. But yeah, this month pictures of me and my wife at uh I believe that's a dead horses point and then uh my little family,
04:13
you know, I've got three daughters and, you know, that's who I work for. I love the uh I love the train. I assume that's a train that looks like a train train in. Yeah, cool. Hopefully you'll join us next month as well. I do want to make sure to highlight each coffee
04:30
break is great, right? But um Pete Kirkpatrick is a, is a very unique guest to have. I first saw him present for Tech Field Day a while back. He's been here for about 10 years and is the VP of engineering chief architect for our platforms division. So he's done some pretty amazing things around leveraging all different kinds of flash and
04:45
technologies that sometimes you're like, oh, that's under the covers. But this is a little bit of pulling back the curtain of, you know, the pure labs and even the pure hardware labs. It relates to that great Alan K quote. If you love software, you're gonna end up making hardware the right hardware at the right time.
04:58
But join us next month. And I think Olivia if you don't mind tossing the uh registration link for that in the chat, so folks can register. But this month it's all about you John. So you know, Pete's great, but this month you're you're more great than he is. So and I might say that to Pete next month just, just full disclosure.
05:14
So as always, we keep a little bit of the same lightweight agenda and not too much and, and here, you know, going with a little bit of uh it's not snark but what's old is new? Some of these are topics we've been talking about for a while because because we need them. So we're gonna do is first wander through some of John's background actually, maybe the the it version of planes, trains and automobiles a little bit then explore what is
05:38
unified. Really? What does unified storage really mean a blast from the past? You know, a second movie reference in case, you know, that one third tear down this wall, you know, sometimes there's been divisions or different ways to define unified storage, how we're approaching, how we believe we truly unify it.
05:53
Um I was thinking of that as a, as a, as a Cold War Ronald Reagan quote in Berlin John, you were thinking of it as, it's a pink Floyd reference, you know, either way you could decide. Uh I, I think you're, you're cooler there, but yeah. Oh, well, and then we'll wander into some of just what's new in this space that we've been doing specifically.
06:08
But before we dive in, um if you don't mind Olivia kicking off the first pole. Uh so this is, you know, both a little bit of just having fun um as well as diving in. So what is your favorite way to travel? Uh John, do you want to answer that now or later? I, I'm curious, I actually like to travel by automobile.
06:29
So this is, this should be the Americans and this is, man, they're coming. So we'll let this go. And then also, where did your file background start? Because I'm guessing for a lot of the folks on here they've been doing some of these pieces for a while or, or maybe not, you know, please feel free to put in the chat if you're like,
06:45
oh, the, the, the survey doesn't fit who I am. That's fine too. So kicking it off John a little bit of, um, you wanted to walk through kind of your, your history in the file space, you know, some kind of lessons learned and, and principles you, you've brought with you into the role and you were saying, you know, planes, trains and automobiles, you keep seeing the same things over and over again.
07:05
Um, but, but do you mind just kind of starting out with where you started as a, as ac admin even? And just, just kind of wander through a little bit if that's all right. Yeah, I mean, I'll try to keep it brief. Um, you know, when I was 14, I was in high school and I had, uh, we had a, you know, old school token re IBM network, you know,
07:26
with an A S 400 server and I decided to figure out how to kind of like, you know, get into that thing. And eventually, uh, I did and I got caught and they, uh, decided to hire me, uh, under a student's learning grant that was going around back in the nineties. Um, and so I ended up working for the school, uh, the entire district actually.
07:48
And we converted from that to network to an I PV four network, uh, and doing a network servers with, uh, you know, obviously Windows nine, you know, 95 at the time. And, uh, doing all that, I mean, basically it was like, kind of a weird experience to have and being, um, at first being a computer network
08:09
technician and then eventually I, you know, by the time I was a senior I was kind of working as the, the juniors admin for the entire school district. It was very interesting work. Uh, gotta, gotta kind of grow up with a lot of technology that way. Um, which is very rare, I think for, you know, a teenager to do that back in the nineties.
08:27
You even then wandered. Uh, and I was even thinking of background actually, I mean, I started in the help desk and then I started doing network stuff on I one it operations but it was around the same generation, I think network four and, um, E directory is still the best directory. But sadly, it didn't matter, you know, some of that I think is out there.
08:44
But you, you even kind of wandered away from tech and then back into the MSP side, you mind continuing there. Yeah. So, um, I, you know, after high school I went on a mission for my church and then came back and, um, I actually got into retail management for some reason. I, I kind of found, uh, I don't know, it was just one of those things that I wanted to try
09:05
out and see how it worked. I did it for a couple of different places, Walmart Circuit City. And then, uh, when Circuit City kind of went out of business. I kind of said, well, retail is probably not where I wanna stay. Um, then I went to M SPS from there on out. So I worked for a couple of different M SPS working for some that worked for Dod uh TSA,
09:25
I worked for Black Rock Financial for a while and then I went into a local Utah MSP that did, um, had a bunch of clients nationwide but, and, and that really you touch almost every type of technology there was in that, in that type of role. Um We were chatting, what I, what I found was fascinating. You were saying,
09:45
you know, just even just thinking about file platforms, so netware, but you work with a whole bunch of different file platforms. And even I think you, you kind of almost had some, some lessons learned out of that, just keep going there if you don't mind. Yeah, we, I think we had almost all kinds of file platforms.
10:01
We had windows servers, we had netapp, we had Compell, we had some Solera. Um Yeah, I mean, you never probably was in the data center. Um And I, and I, I ran the data center so I really had to work with a lot of things. Um And so what I found out though was basically looking at all the different strategies, looking at all the different companies. It really came through that like one platform
10:24
couldn't fit. We tried to do it as an MSP. Like we tried to, like, we came out with a one kind of file server and said, hey, this would be the one for everybody and it just didn't happen. So it was kind of interesting to learn that, like, you know, you could try to make it fit, but then eventually you would hit some sort of,
10:39
you know, maybe you need a scale out, maybe you need a scale up, maybe you needed uh you know, more S and B or more NFS and it, it just came true that like one product usually didn't fit at all. And uh that's actually something I see that pure is, is taking full hold of, you know, we'll talk about that later,
10:57
I think. But really true when I was there, I mean, some of those solutions were scale up file solutions, some were scale out file solutions. They all got created for specific reasons. Uh There were startups that took advantage of those gaps because there was a gap there that they could take advantage of and then either they'd make it or they'd go out of business or
11:17
they get acquired or all of them, you know, kind of thing. That's where plane trains and automobiles come up for me is like, I would work for some of these NASA vendors like network. It was around for a while and then it would kind of crash, you know, like network is not around anymore. Then you would switch over to another one.
11:30
It would just seem like it's pro you know, promising and then it would kind of go away. There was, there's, and there's been some that have stayed true right in that app and stuff like that, but there was a lot that kind of fell to the wayside and, and then what we also had is at times thanks to just, this is just the life cycle of of it companies and startups and products is that the platforms over time would consolidate either at
11:53
a code level or an acquisition level. And so you'd have the same company and maybe it feels like the platforms are built by totally separate companies because they were, you know, it's not good or bad. It just, that's how it happened kind of thing. I I think last piece here maybe is there was um at one point you were,
12:10
you were digging into a a not too fun situation, kind of kind of around all the layering of the technology that happens here. Do you mind kind of walking through that story a little bit without, without any names here, you know, doesn't need to be names on this one. What would you say competitor? Uh But you know, one of the things I one lesson, one interesting experience that taught me a
12:28
lesson about layered NAS products is where you have, you know, a lot of companies put block on top of file or file on top of block. Um And what happened in this case was that we had a company that was doing video surveillance to file systems. And because they, you know, they weren't really paying attention to the lower down layers like
12:49
the block block code, they had actually lost disc disc had been replaced several years ago, that hadn't fully been replaced correctly but kept the file system live. And, and eventually what had happened is we, you know, they, they went to go check to restore video to pull video uh due to a robbery that had happened. It was a financial institution and uh they
13:11
couldn't because the entire file system had been dead for years, but nobody knew because of the way the layer technology worked. And so it was just basically um and it was one of the things that really, you know, stayed with me is like seeing how uh when you have to manage all those layers, it's kind of hard to pay attention to all of them all the time.
13:30
I, I even think they're of a maybe this is cheap joke time but a little bit of uh of uh you know, hey, hey, look ma hey, hey look, Dad Dev Noll has infinite capacity. You just can't ever retrieve anything from it. Right? That's the equivalent essentially you didn't, you didn't know it. I think I'm gonna go ahead and we'll move into section two here,
13:48
John, I I wanna make sure I didn't keep you from putting anything else in. I'm looking deeply into your eyes. Cool. I think it looks like we're good to keep going. So, um and actually I will c let me go ahead and I will end this poll and we'll share it in just a second, but we're gonna start to wander into now, maybe blast from the past what is unified really.
14:07
Um Because there can be sometimes questions about this. Um But let me go and share the results from poll one just in case you're curious. Um it's, it's, it's not an even split, but there's no clear, I mean, planes trains, automobiles, everybody loves all of them trains. Actually, that kind of follows. Um I do more plane stuff frankly because my uh
14:26
family overseas uh but then also we've done long road trips too. It's like the flexibility of just when you want to go. You go. That's the car, right? And that's a little bit of American dream and freedom in there. And where did your background start?
14:37
So actually what's funny here to me is SMB in Windows. That makes sense. But net is still in number two. That's pretty crazy. So and kind of cool actually to be honest. So, ok, so then I wanna actually let me make sure here. So. Oh yeah, actually before we go to before we go
14:56
to. So let's actually continue maybe with you were you were actually alluding to this already John a little bit of maybe kind of a brief history of Naz because you mentioned file on top of block block on top of file. But it even really started back with I wanna say Linux and Uni and Windows and of course net then kind of came into that but you might kind of chat a little bit where where Nas came from
15:18
and even then why people wanted unified as well. But back to you. Yeah, I, I mean, obviously file storage has been around for ever right before. It was even shared storage and, and, and surprising enough, people probably still have some of those files floating around because you never delete a file, right? Um But I think the um the term uh NAS,
15:43
you know, started out when, when you have shared storage, right, that came from window servers being sharing across networks. Um that where some of the other, you know, Linux servers that would share out. So when you got in, you know, when you got the sharing protocols like CF and NFS, uh you know, like B one and things like that. So those allowed us to really share the data
16:02
across and then where the term unified comes from or even what I would call multi protocol rays, people saw they want to consolidate, you know, they had multiple window servers, they had multiple things that was, you know, the complexity of trying to manage all those upgrades and availability. They wanted higher availability, they want to consolidate workloads together because what
16:25
they would often find is you'd have these servers running off of block data, but they would need shared storage and those would be on two separate devices. So a lot of times it made more sense, bring it all together, um condense it all, consolidate it all and it made life a lot easier to manage. Um But I mean, many times, you know, early attempts to combine protocols usually ended up
16:48
being where you still had to manage both to manage one or the other, so to manage, manage your block as well because technically they just kind of married two devices together. And so that, that's where I, you know, I think that's kind of the history, I mean, it's, we should probably go a lot deeper than that.
17:06
But generally speaking, that's what I think really kind of uh you know, you grow from single, you know, single files, file shares, NAS heads and then to unified or multi protocol. And I mean, and some other pieces where I'm I'm kind of host but kind of presenter. You know, I, I think about both use cases, you know, sometimes you have use cases that they
17:25
need both block and file, whether it's the boot volume, one thing, the file data, you need something else. I mean, it's all being consumed by the same thing we're trying to do from an application standpoint or an internal customer standpoint. Uh There's also sometimes I think the the rules of, of cross protocol doing like this accessing
17:42
the same file, the NFS and SMB where my, my, my historical joke is the first rule of that is Crofts protocol is don't do it. That's the 2nd, 3rd and 4th rule. And then the fifth rule is like, look very carefully about how the systems map permissions. Don't forget that because we're gonna come back to that joke later.
17:59
I'll, I'll contradict myself. Well, I think that is um also a big part of consolidated storage, right? Because as I, I remember back when um I used to play around with web, web servers, a lot, a lot of people would host web servers on Apache servers which ran off Linux most of the time and they would edit http files on windows
18:17
machines because that's what everybody had, everybody had a Windows machine. Um And so all the ID ES and everything else were built for Windows. And so you need that multi protocol access to the HTP files to be able to like do that work. And, and then that created his own set of problems. But a lot of, you know, um up and coming solutions, we're building that type of model
18:37
where you could access the data across both protocols. The other piece I sometimes think about here. And I've, I've, I've used this kind of concept personally and when I've looked at various pieces of the parts of the industry landscape is what I think it was a design center, like when any given product was designed, what was it originally designed to do?
18:57
10, 2030 years ago? Um What capabilities and features did it need as far as like the baseline stuff, you know, before it could launch at all? And then even what were the prevailing technologies, open source libraries, things that you could use to build it if you will.
19:12
And the further you get from the design center, usually the messier things get, you can feel it, you know, over time you strap more and more things on top or you kind of more for it, we can tell it really wasn't made to do this and that's not a critique of any given product. And sometimes you see products with exceptional foundations where they both were able to draw the balance of seeing further into where the
19:31
product might go versus what it had to be to get launched, you know, kind of thing and then make it as a company. So sometimes when I look at this space, the Unified space just well, you and I John having spent a lot of time with different products here, you can kind of feel how a good bit of what's in this space for unified storage,
19:46
maybe multi protocol storage. Like you said, it's pretty far from its original design center and what it was built to do nothing wrong with that. It's just, it's just reality. Ok. I think we are into, uh, poll number two, Olivia, if you don't mind kicking that off and we'll jump into section three.
20:04
Um, so there we go. Awesome. Thank you. So curious as we're gonna move into, uh, you know, kind of our, our approach to this are curious to see here, uh, who you currently use for file storage. Uh because this may be um unified, it may not be, that's even something that's some interesting sometimes as you have unified or multi protocol platforms,
20:23
but you only use it for block, you only use it for file, even it can do both. We see that especially at larger customers. All right. Um And then candidly, you know, we're curious, you, you, you're joining us today. So we're never gonna be too pushy, but we're uh we're interested to know how many of you all
20:36
are are in, are looking at this space either from a, I'm attending a coffee break for a free gift card and interesting enough conversation in the afternoon or you're actually like, yes, the our specific intent here. So we'll come back to that poll in a second. So let's dive in John two.
20:52
how pure is approaching this now, now, obviously, uh tear down this wall and how pure truly unifies storage, whether it's a Ronald Reagan quote or a Pink Floyd reference kind of alludes to a little bit of how we're approaching it. But there is a piece here of, you know, where we started and because we didn't start with file on flash array, we're talking about flash array.
21:17
But do you mind kind of narrating John through a little bit of a little bit of the history there because we had an acquisition. It's not the focus, but it kind of informs out the decisions that we made because we made some hard decisions along the way to do it the right way. So back to you, if you don't mind to start through that.
21:32
Yeah, I mean, we go through the full history. It's, you know, this um file service is not our first time trying to unify it on flashing. We did do uh window servers before for a couple of years. Um And that really did, you know, like there was quite a few customers found that helpful and useful and they were familiar with that product uh enough to be able to make you make
21:52
that be fully fledged. And we had a great, we had really good success with it and that just drove to the point that we, we needed something more. So we had to make decisions whether we built everything from scratch or did we go and get in, you know, find the company that had protocols tax that could work for us. Um We found a company out in Sweden called um if you read the word and you know Spanish,
22:15
it looks like it's, but they call it enough in Swedish. It does mean green computing. Um but they were a software defined um hardware gnostic nas product. Uh They ran the OEM software behind IBM Spectrum mass. So they had a mature um protocol stack. Uh both SNB and NFS,
22:35
they had cross protocol access, they had many of the features of an enterprise mass solution. So um we did acquire them and then we took the decision, we had to make another decision at that point, whether we take their software and put it in a VM and just give it a block volume, call it a day and say, hey, we're done, it's a fast route to market, there's an advantage there but then have two different, you know,
22:56
U I to, to manage it and whatever else or do we actually look at it and say, how does this software can interact with purity on flashing? And it turns out the reason one of the other reasons we acquired them was because their software was layered, they kept their management and their storage software separate from their protocol stack. And so what we did is basically took their
23:19
protocol stack, threw away most of their storage because we already have storage code um through the management because we already have management code and basically integrated their storage stack directly into purity right alongside of the ice and fiber channel protocol stacks, we didn't have them layer on top of each other. And this allowed us to do a lot of cool things. Basically, it says purity had all these really
23:44
nice things that have been developed over the years, like high availability, non disruptive hardware um changes where you can go from, you know, data in place, hardware upgrades, uh you know block volumes that are thin provision across the global pool of storage, all of that with file, keep going, keep going. So like this, this slide is perfect.
24:04
Uh So we, we also could take advantage of global D dup across the entire array, which means which means we're not limited to like, you know, the pool of storage that is, you know, in some vendors where you have a pool of storage that's directly assigned to file, you can only dedupe within that pole. Um No, that you can dedupe file against block now and block against file.
24:24
Um You can also take advantage of the entire global pool storage. So as you increase your raise size, you can take advantage of that for both block or file, whichever one you need to, then we looked at the, the the user experience for file out there. And a lot of these times these um vendors have developed features over time by different teams or different. Even they've had to integrate different
24:44
companies in or they stole, they, they acquired, somebody took a piece of their uh their code and integrated into their. Now, the UX is all look different. We wanted a very simplistic UX. So we came out with a design where you know, everything looks the same. It's, it's, it follows the simplicity that you see with your block. So we have these things called managed
25:04
directories you see here, um Every file system becomes a root manage directory. So you never, you don't manage file systems, you manage the root managed directory of a file system and then you can manage any, any, any directory, seven layers below that, that you want to um you can create those on the array and you can apply policies and all the policies are managed the same.
25:27
They have roles and they have managed directories they attach. So when you learn one policy, you've learned them all like it's not hard to pick this up from day one and just start going with it. It's very easy to use. Um All the configurations are reusable. So you don't have to keep on reconfiguring things if you needed. You know, somebody comes up and says,
25:46
hey, I need a new share for the HR department. Well, I already have a policy for hr Bam. I just create a share and I don't have to rethink all the configuration or reask for all the configuration. I, I love the policies especially because sometimes we have this balance we have to make between A U I being simple or working for small use cases and being able to scale out well and policies obviously,
26:06
I mean, they had to be a good balance in between or they're not super complex. We're not penalizing folks that want to run a little bit of file along with their block or have it for a more limited use case. But it also is built so that can scale out effectively from a management overhead standpoint, even for potentially really large environments kind of thing. So I love that the um the other thing I think
26:27
of here actually, and this is now just a little bit of a, aside from a kind of a philosophy, product development standpoint is this is now going back to where you started John because I maybe I was gonna put this in earlier. But you know, this is kind of planned but not totally as I think everyone joining us knows is when you, when you look at this space, there's kind of the the truism for maybe a Silicon
26:45
Valley or a development standpoint that uh creating a good block platform, a solid block platform takes maybe 3 to 5 years. Uh You've got to have amazing developers do all the work, get it out so customers can find what are the most common corner cases harden it up. OK. That's 3 to 5 years. Give or take file, file is messier, right? Uh It's about who controls the file system is
27:06
in control of the client is in control of. So and so the number there often that I've heard is in the 5 to 7 years because even for instance, we say SMB or we say SMB three, well, what set of individual features do you mean inside that? Right? Because SMB three is really just like a super set of SMB two. So all that stuff has to be individually
27:23
developed and it takes time and all of those things have corner cases and bugs, you have to sort out and like all this stuff, right? So at this point, we chose, as you mentioned, John to do this kind of the what we think is the right way long term versus the fast way we could have just put this on top, but the right way long term because it's preserving, you know,
27:44
like it says here, the I mean, the pure experience, the capability appear on flash array and I'm I wanna be a little cautious in a public forum saying this. So I realize that of I, I believe we're the last company really investing in this area and we think all the benefits of Evergreen um and non disruptive migrations and upgrades and all that stuff it applies to block, it applies even more to file file migrations are so much more
28:10
annoying than block migrations kind of thing. So that same value even more powerful. I think the last thing any system wants to do is a file of migration when you have to worry about the permissions and the cut overs. And you know, if you lost the CEO S file somehow in the migration, I think being able to say like this is migrating to pure file is going to be the last migration you need to do.
28:33
That's, that's a wonderful thing like and even then when you do hard upgrades, I mean, a lot of people claim to be able to do hardware upgrades but they don't do data in place, hard work grades. This is data in place, hard work grades, which is, you know, truly non instructive in that way. And I think that's really important to understand is like you're not,
28:49
you're not, you're not gonna spend days and weeks trying to plan out hardware Prades. Uh you get the same experience on file that you've had them block on, on, on, on flash a rate, you know, being able to go up from one model to the next. Make sure here are the I I see one comment about the slides not sharing. Are you still able to see those John or uh or do I stop here?
29:10
You can see it? Ok. So uh Ulysses, I think it may just be you but I always appreciate the comment. I wanna have that be a problem. One last thing here before we go to use case John, I, I made a joke about cross protocol and I I think it was Calvin and the chats that he agreed with like, don't do it, you know, kind of thing.
29:25
But I think even when, when we acquired comp we got an acceleration of all the protocol features, that's a lot of work, real value there. But they'd also done something unique on the cross protocol front too, right? They did. So copyright solved that um you know, the hardest use case to support on file normally is
29:45
cross protocol because you have all these permission issues and you have to decide who's doing what. So when they developed their code, they actually went all in, they, they designed it for cross protocol, starting out, it actually wasn't designed for one protocol, the other, it was designed for truly cross protocol.
30:01
And actually, when we first integrated them into flash ray, we at first we only supported cross protocol as well because that's what they code supported. Um And we've, you know, since developed the ability to have individual protocols supported. But um that's, that's, that's truly where they build it at. And so it really means that we have probably one of the best implementations for cross
30:20
protocol there is and it's really easy to manage. Like you don't actually have to sit there and do a lot of granular management on the cross protocol for file services. It's like you just set it up. We use ad as our user mapping service. We don't, you don't have to manage like separate user MAPP or services and you know,
30:36
worry about, you know where that's at on AD R range. Um You just configure it using um UNIX mappings and windows and we just automatically tie to that. It's really easy to set up and monitor. I did see a great question. Um I'm probably gonna mess up your name. Uh Good John and thank you, I think is responding to some of those.
30:55
Thank you. We have Manish and Olivia and Kara here to help with the questions we will do as always. I didn't say this at the beginning. We'll do Q and A at the end here, we'll stay on to do kind of lab Q and A but please put them in the Q and A box as we're going along as people have been. Thank you. But this is not inherently a competitive flash
31:09
blade. We're gonna talk about that next, but let's talk about maybe finish the section up. John just kind of mind walking through the overview of the use cases because in some cases file or NAS is almost a little bit. I don't wanna say intellectually dishonest, but it's like that can mean so many things, right? So uh it may may kind of highlight where we
31:28
focus the unified flash ray file on. Yeah, like I said in my experience, like um you, you, you wanna, you have certain devices that do really well at certain things and So we definitely have use cases where we, we shine through here and these are the use cases here that we do. Um And a lot of, you know, a lot of the sub use cases will fall under general file applications.
31:48
But um we've had, we've had file service, we introduced file services originally in 2020 we really just focused on user shares back then. Um But we've since expanded out uh in our, in our latest uh user share, you know, use case that we've started to support is NFS data source on VM Ware. And, and this isn't just like, you know, a dummy data store. This is with the VA I offloading for NAS with
32:13
certified data stores through VM Ware with all the plug-in support that you expect from pure. So, uh we wanted to give people, you know, the pure VM Ware experience on NFS data stores and uh I think there and there's, there's some stuff we'll talk about what we did special there dive in. Yeah, but just as it was the sense sometimes for folks, it's like, am I hearing file? What does that mean?
32:32
What are they really focused on? OK. Let's uh I'm gonna actually, I'm gonna first close out and share poll number two because that's what we always like to do as part of this. You know, if you're kind enough to respond or actually frankly if you're not. But, you know, in this case, we had over 800 people to respond. So, thank you.
32:46
Uh We'll end this poll and we'll share it back before we do poll three. So uh always fascinating to just see kind of the, the, the landscape of who joins. So um John, hopefully it makes you feel good. 13% of the respondents are using pure for file storage, you know. Hey, that, that feels good.
33:03
I think that might. I have No, I I I'm not sure what our market share is for file storage, but either way I'll take it um 6% are still using netware. That's pretty amazing. Maybe you're joking. I don't know where you just like a sitting there forever. I'm not touching it and then there's a healthy mix of other stuff and it is also always
33:19
fascinating to me to see that uh you know, about a quarter. It's, it's roughly split evenly looking now 6 to 12 months, 12 months or longer. Any uh any comments there thoughts that brings to mind John before we keep going. No, I mean, if you're looking uh please give us uh talk to us e and see what we can do for you.
33:36
Uh You know, we can look at your workloads and see if we're the right fit. We definitely don't, you know, we're all about customer experience here up here and making sure you have the right experience. So that's why it's good to talk to your se about it and just make sure everything lines up. Then Olivia, if you don't mind launching Poll three before we jump into the final section
33:55
here. So you can see this one, this one we kept relatively short. Um You know, frankly, we're curious if, if you're coming to this, do you even know that flash rate could do file? Because there's a long history of amazing block storage on it. Um As well as um your VM Ware environment, which is where we're gonna dive into.
34:10
What do you run that on today? So going into section four, which is candidly, what's new. Uh There was a big announcement from a pure marketing standpoint within the last month or so. John, I want to say you did some recordings for that folks if you can go hear John all that and think man and other folks too,
34:28
but there were really kind of two big things to focus on there. One was, well, we've had file on flash ray kind of unified or you know, on, on flash ray for a while. It was digging diving further into the extension of the use cases and the focus there as well as what you just mentioned uh you know, NFS data stores for VM Ware.
34:48
So, so let's actually just start out with the use case thing because I always appreciate the question. We wanted to take this head on and oh yeah, I'm gonna do this one in Turbo mode, a little bit of um peer has flash rate and flash played and they compete with each other. Right? No, actually they're very intentionally created separately just and this isn't,
35:05
we weren't making this up, John when we planned this like your story of being on the MSP side and seeing like all these different products, not curse Peer, it's not a million different products. It's two specifically, but it's because of the different types of requirements. It's a mix of difference in capacity, large and small, small and large files of file count uh
35:22
that can drive major things from a meta data and concerned currency standpoint of the use cases that drive difference in I ops and capacity and bandwidth needed and latency, right? There's all these different things that go into this. And at core, there's two architectural choices, scale up and scale out flash array focuses always has focused on a scale up approach
35:45
leveraging evergreens. So you can swap things out and grow and change controllers out without the pain of migrations and disruptive upgrades, right? That's what flash ray does. So we often see, you know, like we showed before quite a few use cases that really make sense there, but there is overlap in the use cases.
36:01
So we covered the su actually a little bit uh a couple of months ago, Justin Emerson was Flash blade e the launch of that. But we often see there are some use cases that fit better toward a scale up type approach or maybe even a unified sense. You want it all in one box or one type of platform some that fit better for a scale out. I'm not gonna read them,
36:20
you can read them here. But then often frankly, the platform choice can actually depend on the scale. Is it a little bit of this is a lot of this is the main purpose of the box. This is it gonna grow to crazy bandwidth. You say I want 100 gig of throughput on flash blade, not super hard, an easy equation. You say I want ded Dulic application.
36:38
Oh, that's not what flashlight does or really any scale out system does. Well, due to the nature of scale out I kind of thing. Well, that's over on the flash a side. So this is where please engage, oh, you can reach out to me and John obviously, but engage your pure account team because there's only two of us for the world, the country.
36:53
So you know, um but we actually help him walk through that and having a more intelligent discussion hopefully than just like, oh, all we've got is a hammer. Everything looks like a nail. There's just one platform. No, there's actually a bunch of different requirements that we ought to map to intelligently to help you be successful.
37:09
That's the goal I leave out there before I keep you keep going, please. I would I would just add to that. Like, you know, in the comment I made earlier is also that normally when you have like two platforms, they, they seem completely different managed. Uh But when you go work on flash flash blade, I guess what you gonna,
37:25
you, you're gonna manage your file system level. That's how flashlight handles things, but you're gonna handle everything through policies. So you're gonna see very little uh you know, a AAA known effort that, you know, flash blade and flash blade are aware of each other in the US and the user experience and that we want you to be able to be able to use both in your environment for whatever needs
37:46
you have. And then that way, you know, you're, you're using the best file product for the use case and you're not trying to shove things into, you know, silos um and into wrong places like you basically can take advantage of flash, right? You can take advantage of flash blade. And I think that's really nice that uh peer has done this,
38:05
that they realized long ago that we needed both. And uh they both need invest heavy investment to be able to do that. It's based on the underlying requirements that we've both seen over our careers because sometimes we've tried to split the, we've tried to, we, we've gone, tried to fit the square peg in the round hole and you're like,
38:22
oh, I can make it fit, but it's gonna be painful every month to deal with the the operational challenges of it. So one part of that announcement was focused on the expansion of flash ray file use case or even some I manager were beating the drum about it a little bit more, you know, kind of thing to be honest. But then also there was a very specific piece
38:38
about launching support, not just as a me to checkbox kind of thing like, oh we do that as well, but some real intentionality around thinking about NFS data stores for VM Ware, which for me is actually I started with I for VM Ware and then I did NFS for VM Ware. This is years back. You can guess the platforms. But do you mind walking through some of what we
38:59
looked at from a kind of a challenge and landscape standpoint there? Yeah. So obviously we, we identified within the marketplace, a lot of pain points that people have been saying, listen, these, these pain points come from customer uh real customers like this is just something we can, you know, went out on our own and said that existed,
39:17
these came from multiple conversations with um you know, people using other vendors. And a few of these is, you know, data store size for n because people have limited file system size like 100 terabytes or whatever. Um 2 56 terabytes, they, they, they had to really plan out their storage needs and it felt,
39:36
um, I, you know, as I would see this, I kept on thinking this kind of reminds me of a game where you play with blocks and you would always wait for one single, you know, straight line block to fall down and something would mess you up and you would, you know, so Tetris basically, and I, and I said, man, it feels like you're paying Tetris within these limits.
39:53
Uh The next thing that they, they wouldn't offer is they wouldn't let you manage the V MS on the array, like you would just have to manage the data store. Um If you want to do anything with the V MS, you have to go through VM ware and you couldn't, you couldn't really like a snapshot, for example, you had to take the whole data store like them in any like V MS you don't really care about.
40:12
You still got those. So you got like this snapshot bloat that always occurred. Um Then lastly is when you were trying to like set it up and everything. Like I went through several setup docs and they were like several pages long and you had to hop between, you had to hop between one interface and to the array and back to the interface and back to the array and double check everything
40:30
and make sure it was all working. And I was just like, if we're gonna build this, why can't we just do this much easier. Why can't we make this a less complicated solution than you know what exists? So that's where we came up with a pure differentiation for this. And so keep going, you're on a roll.
40:48
We, we, we because our systems are unlimited, we have unlimited data source site. So no more in that game. The next thing is uh we decided to make every V MA managed directory. So using that, um we got a couple of different things. You got to be able to see space and capacity and performance data on each VM.
41:06
You also got to take VM snapshots. Like you didn't have to take a snapshot of the entire data store you could, but you weren't forced to. And then lastly when we built out the plug-in support, we're just like why we know how to make data stores and we just automate this. And so we made a kind of a one stop shop plug-in experience where you don't have to hop between interfaces to like set up your data
41:30
store and manage your data stores and connect a new one. You can just go over and tell it. Hey, here's the array, here's the ESXO I wanna do, it'll create the NFS export, set all the rules up everything it needs to do, that's all automated there. So it's like why live in the password, which it's like we don't know what to do for you.
41:47
We know what to do for you. Um And it, and it gives you lots of options there because you can have a data storage, a file system. You can have a data storage system to manage directory within a file system. You got lots of options. But the key is like, I love this question, by the way.
42:02
Um I thought it was good all the time when I thought about, uh you know, playing a Tetris game with your data stores, the only winning move is not to play that game. If anyone wants to put in the chat, I'm not gonna say this one. that is yet another movie reference. So feel free to put it in if you,
42:16
if you know what it is, if you're everyone's still paying attention. Hopefully bam that was fast from Calvin. But other folks keep chatting and I'm still not gonna say it that way if you're not saying the. So the other piece I think of here John is that if for folks who joined us back in January, I wanna say we had David stemen on and we were talking about a lot of different VM ware things because that's a lot of what he focuses on.
42:34
I think it was January recently. You can go back, folks can look at the some of what you're talking about. Almost feels a little bit like vivas what we do over on the block side. Do you mind kind of drawing out that connection a little more. Yeah. So is you're managing things at the V MD K
42:49
level, right? But we wanted to give you a taste of that by letting you manage things at the VM level. So we don't go all the way down to the V MD K level. It doesn't mean we couldn't in the future because obviously we could make every V MD K and manage direct as well if we wanted to uh let us know if that's something you want.
43:05
Um But you know, it lets you see this granular area of control. It lets you take control of your VM ware if you are using them as data stores, if you want to continue using different stores, this is not saying like move off our, you know, block product to the file. This is more for people who have this environment, they need to keep this environment or they have a use case where this is required.
43:28
We want to give them the flexibility and let them be able to do what they want. So, and these two examples, for example, I was talking like, hey, you could have a file system as a data store or you could even have a subdirectory as a data store because we don't restrict you. Um We let you do whatever you want there and you can do the same thing with both,
43:43
like you can have V MS underneath there. Um And we automatically do all that for you. So if you, once you've decided where the data store is at, um And you've got to set up into ESX. Um when you're creating new V MS in that data store, we'll create the manage directories for you that's all automated. And then the only thing you would have to do is
43:59
go decide how you want the snapshot policies to run. I think there's one piece there that you, you mentioned, I want to make sure to highlight is this is us embracing folks that want to use NFS for VM Ware. We still have an amazingly strong set of capabilities um market leading around VM Ware block integration, all these pieces and we're going to continue pushing that further.
44:19
That's been longstanding. We're also offering to people that say, you know, this really makes sense in my environment. I have specific reasons, et cetera, but we want to do it in the pure way versus just like uh Yeah, sure. You can use it for NFS. We did just enough Q A to get on the HCL or their compatibility with VM Ware.
44:35
No, let's like actually do it better. I think with that John, we are actually right about on time for folks that stayed with us. Thank you. We'll do the drawing in just a second. Hopefully, it feels like as always, we went through a good bit with a little bit of history and the landscape and even our core architecture and then some of the new stuff
44:53
because there's truly like most recently innovation going on here. Anything that I left out John that you want to highlight before we move over to the, the drawing and Q and A, I'm not letting you go yet, but you know any final thoughts I would say uh you know, come visit us and accelerate. Um So man, who you mentioned is the director of he's great rock star guy.
45:11
You'll see him down accelerate. I think both of us have some talking sessions come and see us. Um Let us know what you need. Um How we can help you guys out. If you're interested in file services, please reach out to your account team and let them know and we'll get you on the road to using it.
45:25
Um And we wanna make sure, you know, and if you see um you know, something you want, let us know as well, you know, we're always on your feet, the product over time, we got lots of cool stuff coming on the road maps. Um This is a heavily invested product line for peer. So lots of good stuff coming in the future as well.
45:44
We, we do truly see a lot of opportunity here like I said, we believe it may be one of the last if not the last truly investing in this space. And by the way for anybody who had joined us last month, you might have noticed that the that the title changed. You might be able to guess at some of the reasons for that.
45:58
But there's still a little bit of an Easter egg in the graphic there. So, you know, you can have fun with the ice cubes dropping into the coffee there. But to wrap up, thank you, John. Let's share the pole. Let's do the drawing. Then we'll kind of let our hair down such as we have it and uh do some open Q and A.
46:12
So let me end the pole here and I will share it back. So, have you heard about file services on flash? So this is cool. 73 quarters had already heard about it, but at the same time, that means a quarter of the folks joining us today had not heard about it. So, hey, hopefully you feel like you learned something and maybe you're,
46:28
you have some interest here. And then do you run your VM Ware environment on NFS today? So that's decently split. And if anything actually um block only is in the minority there with either NFS or a combo in the lead. So it's kind of cool to see for drawing. As always,
46:49
we end with a drawing for AM 12 and ounce travel mug retail value. Oh, I think it's 100 and 30 or maybe it's 100 and 99. You can go look up on Amazon but you won't find the one on Amazon that has, you know, the cool logos and stuff on it. Of course, that the winner today is Brad A from Texas. Thank you so much for joining us,
47:07
Brad, please as well. Make sure to join us again next month with Pete Kirkpatrick. We're gonna talk about it. Now, if you love soft, you're into making hardware and pretty interesting, cool hardware. We're gonna dive behind the curtain a little bit of Pier's hardware lab.
47:21
Uh you know, super secret hardware lab too. A little bit of a look behind the curtain with that John. I think we are into uh you know, we can kind of relax a little bit. We basically stayed on time as far as what we were hoping to do. But for everybody, we target about 45 minutes for the content side.
47:36
Let me just kind of relax a little bit. So Manish has done an amazing job with keeping up with answering questions, but there are some new ones coming in and I think if you don't mind John, I may um I'll kind of turn the music up here a little bit because this is where it gets a little relaxed, you know, keep it low and uh may even go through and uh answer some of the kind of pull up some of the questions that he has already answered just because I
47:59
think they may be worth sharing with everybody here too. Uh The first one that I think I wanna talk to you about is uh well, this is probably worth just re highlighting because we spent a little bit of time on it. We went fast. So, is this competition to Flash Blade? Um I think you're gonna say yes, and you have fights every day internally with your,
48:19
you know, corresponding product managers or, or, or no. Is that not how it works? No, actually we're like me and are best friends with the product. Uh We meet quite often to uh talk about the product lines. Well, I think they're complimentary not competition.
48:35
So the idea there is that uh is there use cases that both can do equally, you know, in certain scale sizes equally well. Yeah. And, and that comes down to consolidation, you know, ideology is like, again, what do you want to consolidate onto what platform? Because like I said, a lot of times you're not just doing one of these use cases,
48:53
you got multiples that you need to, you know, house. So which ones do you want to do? Are you doing user shares with some packs data? Oh, we'll probably look at fa are you doing packs data with some archive or some uh A I? OK. And then we wanna go to flash there because those are the good mix. The idea here is that whatever your spectrum of
49:12
flash file or is we're there for you? We've got a place to house it. Uh We're gonna give you the pure experience, the evergreen experience. You know, I can't tell you when I first started a pure, when we first started getting the file, like how many people told me, hey, can't, you can't we just get the file experience, the pure version of file experience
49:31
for doing this. And so I think that's where we're headed. And um we got lots and, you know, lots of future, you know, future stuff that we may even consider you. Obviously, you don't see file databases on here because we do such a good job on the block. So you might wanna block, but there'll be lots of opportunity here and we just wanna make sure
49:54
everybody has a home. There's even a little bit of kind of history I want to put in here too of when uh Flash Blade started. And now this is about what you're doing as a startup. So Flash Blade was a startup within a startup. OK? So flash array started a couple of years later, Flash Blade started as a startup within a
50:08
startup. And when you're doing that, you look at where there are gaps in the market that you can do development against and drive into and get a couple year competitive advantage and lead. That's how you make it as a startup. Um So Flash Blade looked at areas that frankly weren't being served by the classic file market to be honest, um around crazy concurrency and scale and throughput and other pieces that
50:27
flash play does really well to actually kind of almost like a little bit of a category creator product. What we're doing with flash array is taking the core goodness of flash array into the general purpose file market. That's not diminishing it, but it's a general purpose where frankly everybody here has an incumbent in this space, you're already using something else for it.
50:45
Whereas flash play times came in like where you needed something you didn't have anything yet. So you don't necessarily start out as a startup by going into a crowded space. Now as you get larger and you play out all the benefits of Evergreen and other stuff, then it makes sense to invest there, which is what we've done. So there's even some kind of thoughtfulness about the product uh timing here,
51:01
just it, it's good product management discipline if you will. Yeah. And I think, you know, and, and the relationship we have between the product teams is very strong so that we can make sure the user experience is just through to your feels similar like the same company at least. Um There's a question from um about NFS V four and even,
51:24
uh and about, you know, kind of FS logics, user profiles that's gonna get a little bit specific. But anything you want to comment there about either kind of NFS and N FX logic user profile, some of these always, by the way, will be hey, reach out individually because we have things coming. We don't say timelines publicly, but you mind commenting on either of those.
51:40
I can't comment on NFS before because we did announce it at our launch um that we just did for file uh NFS data stores. We will be um it is, it is something we are gonna have this summer. So I'll be looking for it in the uh August time frame for NFS that we pour out one and what we'll be reporting. Um So that will be there and that will also come with support for data stores on 4.1 as
52:06
well. Um As for FS logics, user pro Yeah, the user profiles is something we do. We, we do roaming profiles very well for good use kits for us. Um I usually just say talk to your account team and then they look at sizing, see what flash Ara model you need and you know, things like that scale up file solution. So different models of flash ray have different
52:30
um file, I think there with user profiles or sometimes a BD I discussion, but not necessarily, right. It may be related to that. Um Yeah, evidence logic comes different couple different flavors that whether you're using it with Citrus, using it with Microsoft R DS. If you're using it for virtual hard drive with the app cache or if you're using it just for
52:51
the roaming profiles, there's some different things there. Um, you know, different ones have different, uh, things that you want to talk to your account team. Uh, next one, I think there is a question. This is just a very specific one, but I also want to highlight too that there is a, the question is what is the individual file size maximum on top?
53:10
There is also as well. There's a great, and this is absolutely available to your customers. I'm not sure if it's actually available publicly on the internet because it wouldn't necessarily be, but it actually shows all the maximums on flash ray file because every product has maximums and we want to be upfront about those kind of things.
53:23
So um I I I'm I'm actually gonna play stump the chum here, John because already answered this and maybe you see the answer, what, what's the individual size limitation? You win? Ok, cool. Um And we do, we do publish those two customers on them.
53:39
This is about being um just being real. So as you're, you're starting out in a, in a landscape and competitors may be looking for stuff, but we want to be very transparent to our customers and you ask us, you know, because any product has maximums, of course, they do. I think all, I think all the uh if you went and looked at the,
53:53
the mass limits for, you know, competitors, you would find that they also sell in publishing customer. It's kind of a general role. It's not like we have a lot of limits. We don't, we don't really have, I mean, like file system size, we don't really have a, we're not a limit there. So the box, but we do have limits.
54:13
It's actually like the block side of flash ray too. Like the file system, you can have one file systems. It's the, the maximum size of flash ray, you can have one block volume. That's the maximum size of flash array. Uh which is kind of crazy to think about sometimes.
54:24
Uh maybe the last one here I think in case there, there's always more streaming in which is we've kind of leave this open and, and multiple folks saying, you know, thanks and they appreciate it and good kind of thank you. That's actually encouraging. Sometimes if you're please always feel free to reach out to me directly by the way because like the these,
54:38
some of what we cover is based on your feedback, we're probably doing some polls again around that again in the next month or two. Uh possibly last one here, John is from Calvin. Um He's kind of trying to follow the managed directory concept, the managed file system and asking if it's similar to nested mounted file systems.
54:57
And I'm not gonna, won't wanna kind of have you do a full deep dive, but it feels like, you know, just since we're in Q and A mode, a little bit of another discussion around managed directories and file systems might make a little bit of sense if you don't mind, it could be kind of like net amount of file systems to a certain extent. Um It's, it's just um,
55:18
less complex. It really is just, you know, you are identifying folders within a file system that are important that you want to do stuff because obviously you're gonna have all the rest of the folders and you can have folders in the same level of that file system that are not managed directories. Um So to an end user managed directories doesn't look like any other folder.
55:39
Uh But I think it's probably less complicated than just file system. That would be the general sense. I mean, there has we want to have some layers of a layer of distraction but then where we can apply policies on but not more than needed. And uh even following that thread for a second, uh help me out there as we talk about the policies, there are different, there's different policy types.
56:02
I didn't mean to change that slide. Apologies. Let me get back here. So you know that you can do snapshot policies, export policies. That's, you know, the mounts for NFS or SMB quota is and then those can even go multiple layers deep too is right? If you don't mind, keep, keep going there if you would any policy at any layer.
56:17
So we if you start with the route. Uh the root directory being level one, we can go eight levels total. Um So, and then you know, why does you need to go? But at that point, you know, you could have nested quarters all the way down to that at eight level if you needed to, for whatever reason, and that can actually change how your
56:38
mount points look. So if you're, you know, carrying out that sub boulder and you want the mount point to be different than the file system looks like you can do that by setting another quarter there. And we will always show um the client when they manage mercury, whatever its hard quota sizes is what the mount the DF will show or,
56:57
or the window map drive will show the sizes. So let's say you have a file system that's limited to 10 terabytes. If I didn't put any sub quotas, all my shared folders would show us 10 terabytes. But then I could put sub quotas of like 500 gigs or a terabyte. And now those map drivers would show up as a terabyte even though the file system can help
57:16
them. Um So you do a lot of things and we do um We did account for key, you do a revision for some reason. Let's just say you forgot to increase domestic quota, but you forgot to increase your parent quota. We would dynamically show the map driver space as the most restricted quota to that to that, that that man work that way your end users
57:38
never have to worry about knowing how much space is available to them in a she well, and obviously through this session for those of you know, if there's still a good number of people with us still, thank you for stan some of what you were just walking through. John is is when I kind of gave the summary of like policies and the other pieces can help you
57:54
get it to serious scale while working out operationally because there's some there's flexibility that you can adjust things around as needed without being like, oh I made this decision two years ago now I'm really stuck. I've gotta do crazy redo it. Some of what you just hit on is what gives kind of that the heft beneath that statement. If you will and many of these things you can
58:14
change on a fly without you know, interrupting them. You don't have to disconnect a policy to edit it, for example, so you can keep your share alive and edit divorce. Um You don't have to like, you know, we all do all that or if you have that policy attached to multiple directories, you could detach it from one and just update the one directory for
58:31
community you um or you can update all of them by just the policy itself. So lots of ways to be able to um extend yourself into bigger configurations without having to do a lot more work. Basically, the whole title of this thing, letting the file server work for you instead of you working for the file server.
58:50
That's the whole idea of policies where they came from. So that I think we are uh we are at time, John, thank you for being an awesome guest as this frankly. I know you would be, but you know, thank you. Please make sure to join us next month again with Pete Kirkpatrick about, you know, behind the curtain appears hardware labs.
59:08
And thank you so much from the entire coffee break team. Olivia Emily, Carol Manish. Thank you. Have one of the questions John. Thank you for being a great guest. I hope to see many of you next week at Accelerate and if not to see you on next month's coffee break. Thanks. Have a great day.
59:24
How you doing, buddy.
