00:00
054321. Hello and welcome to the February 2023 coffee break with pure storage. Like it says they're VM ware and pure. We've done so much together, but we're not done yet. There's still a lot more coming there. I even had someone who was saying like, did you, did you mean a Monty Python?
00:18
Monty Python reference in there? Actually, I didn't but hey, I'll, I'll, I'll take it for free too. That's sometimes what happens when you have fun with titles. There's other ones that come in, but it's more just a reflection of that history, that heritage and how we continue to collaborate and do more stuff together.
00:32
A little bit of housekeeping to start off. But I am really excited to be joined today by David Stamen. David. Thanks so much for uh thanks so much for being in the boat with me here today for the next 45 minutes or so. Oh, no, I really appreciate it. I've, I've, I've been like, I think I've been hounding you for a while now.
00:47
I'm like, where's my time to be able to get on it and talk to everybody and I think we're finally there and it's good. There's been so much great content and I'm glad to kind of talk about everything that we're doing, um, and continuing to do so. A little bit of housekeeping. Thank you as always, this is a series and thanks to the graphic design team,
01:07
just they, they keep doing fun stuff with, you know, whether it's the cups or the branding, et cetera if you want to find previous ones. Because thanks to the solution focus, these have aged better, I believe than sometimes other other approaches. So you can go back and look here. These are now as well on the pure stores dot com slash events page. You can go there to both define coffee breaks,
01:26
but all of the different webinars that we run with different focuses. If you want to find previous ones, you can click down there and look at previously aired events. You find this interesting, you like the format, you like the guests, go back and listen to previous topics kind of thing as well. First month to talk about this, our yearly conference is coming at Resorts World Las Vegas
01:45
from June 14th to 16th. Pure Accelerate 2023 would love to see some of you there. And I I'm guessing I should wear cool sunglasses when I'm there. I'm not sure that just feels right given the image. And as always there is uh there are coffee cards involved as we talked about last month, the 1st 1000 participants to register and
02:03
attend or receive a $10 card. We will do a drawing at the end for a copper edition Ember 10 ounce mug. The one David was just, just showing. That's so that's so amazing. You've got the good background, you've got the mug like you, this is already winning kind of thing.
02:16
Uh Of course, there are some folks here where um we can't send those to you, hopefully, you know, and understand that. But we really appreciate you joining that because especially even on the partner front and pure storage employees, really love having you on as well for introductions for myself. I'm your host, Andrew Miller.
02:31
I'm not going to introduce myself uh because I'm, I'm here every month. But in this case for this topic, we think about VM Ware and Pure. I go back to my early days with VM Ware and thinking about playing with V I two dot What wasn't even V I? It was two esx 2.5 waiting for V I three. So I could have DRS and H A as I was taking my groupwise server budget for an upgrade to buy a
02:53
storage array. It wasn't a pure one but pure didn't exist back then and then actually buy a bunch of one new racket around servers. They were at the time and launch into uh the land of the service console and you did NTP config man and all this stuff and then following it along with VCP 3456. So II I don't pretend that I focus on VM Ware in some ways like I used to but man,
03:11
there's all that history there and still frankly, I still remember the first time I saw VO I think for a lot of us who remember that because it was so cool. David. Do you mind giving a brief background on yourself? Yeah. So here at Pure um as we, everybody noticed it, I guess we didn't even announce it,
03:26
but we got promoted. So um I'm also now our global practice leader for platforms. So I cover all of our virtualization and cloud solutions and kind of act as the way to lead this in the field globally, both internally and externally. So really happy to kind of part partner with the entire ecosystem and see what we can do here at pure to assess with collateral new
03:46
initiatives and kind of all of those transformation projects. I'm not gonna get super deep into the history because we're gonna talk about that in a little bit. But as far as my background goes, um employee partner, um customer um health care uh MSP kind of been all through that. Um And again, you can always find me on Twitter, um github and blogs.
04:07
Um And I'm also a VM Ware V expert, a VM Ware code coach. Um and my background of stuff is, is pretty, pretty wide. So I, I kind of like that is, well, I haven't been doing VM Ware as long as Andrew. Um I've been really diving super deep into it um, since I've been involved and I've been able to kind of understand everything inside and out there.
04:27
I mean, this is now a long history for many of us and Frank. Thanks for the, the compliments folks for the David. Congratulations to you as a global practice leader of platforms. I'm in a lead technology, America's role. We can talk about that later. It's not the focus but you know,
04:41
still enjoying being here at pure and seeing continuing progression and growth of the company and I enjoyed participating in and driving that next month, we will have another coffee break with a surprise, a INML beyond the buzzwords into reality with Pier One. Actually, we end up inviting San Deep from our product management team,
04:58
partly because he's in product management around doing this stuff at Pure, but also because of his history in this space. And so it's gonna be a little bit of a walkthrough through that both that topic at an overall level. And so what he's seen outside of Pure and then what he's doing inside Pure and if there's anything that's buzzword worthy worthy right now, I think it would be a INML no,
05:15
for the title but into our time. So as always, we've got a little bit of a lightweight agenda. Uh, not too heavy. So, like it says, you know, we're not done yet. That's a little bit of a, not too, we've been doing stuff together for so long and there's so much depth there and there's more coming, wanted to frankly start with going a little bit deeper into,
05:33
into your background. David, uh was, you know, but was in calling out some career lessons along the way. We periodically do that with guests that we have on. Uh because we're all, we're all human, right? We're navigating this landscape of ever changing landscape of it.
05:45
Then jumping in a little in the time machine, looking at what pure has done historically with VM Ware, even what V VAS is and some adoption phases there. This is frankly meant to be more educational and here's all the pure stuff. It's, it's looking back in history then specifically new cool things with pure and VM Ware and last but not least a little bit of a look forward,
06:06
you know, what's coming even even more goodness. Hopefully, this feels fair as always, we'll keep it moving, you know, back and forth and please please don't hesitate to put questions into the Q and A. We just commentary into the chat, really appreciate both uh Jason Lager and Tristan Todd joining us today to help with that. You can find them as well on Twitter. They're super deep on these topics as well.
06:25
I've enjoyed working with them over the years and talked about partner back, back around and time at VM Ware and et cetera. So, before we dive in though, uh Lauren, do you mind launching the first poll? And we'll see this pop up here in just a second. So a little bit of maybe a palate cleanser, but even looking back, just curious, what version of vsphere did you start with?
06:49
And, and maybe for some of you, it's like I'm here for the coffee card, I'm here because I've heard the word VM Ware. So, you know, that's a solid run at 10%. I'm just gonna leave this open. Uh So, yeah, everyone has time to participate and we'll dive in. So David, looking back at your career, do you mind kind of just starting out with a,
07:07
with a little bit of a kind of a narration? I think it even started for you with comp USA and, and help desk. And then I was thinking of how I started in the help desk. Do you mind starting from there and just kind of wander through briefly? Yeah, my journey has kind of been one of those things like I've always said and I know we're
07:22
gonna get into this, but I'm always trying to figure out what I want to do when I grow. And it, it's literally been this way from the very beginning is I love computers. I love technology. I got like my first one when I was 13. And so, like, one of the first things when I worked because I worked at USA and I was like, the customer service manager, everybody returned stuff and I hated that.
07:39
So I moved to, like, the sales and I had this customer that came in all the time with their families. It was pretty much worked. Like I sold them cell phones every single year because they were free and I pretty much built this really cool relationship. And then ultimately a went out of business and everybody was scrambling to find new jobs and I
07:56
had kind of found one, but they reached out and says, if you ever need anything, reach out to me and I'll make sure to, to help you out. So I kind of worked in a, in a health care industry here and worked at their help desk and really kind of started this journey of what I call it. Um I was there in health care for about eight years starting from help desk to executive to
08:16
VIP support, becoming a network admin and infrastructure admin, a solution architect. And I think that really goes to like my whole background is I learned everything from what the end users are doing and kind of translating that all the way to a solution design. And then over that time, I went to work for a partner doing managed services.
08:35
So even opened up that knowledge across really a whole lot of different verticals. Um I kind of got to the point where I'm like, I really want to be able to focus on hands on and went back to the customer side, focusing on bringing up and kind of modernizing a lot of the platform stuff there. And then I had the, the best chance to work at VM Ware where I did tech marketing,
08:54
traveling all over the world, talking to customers about virtualization and all of these things. And so really from that, um unfortunately, the best worst thing that ever happened, which is the perfect thing to really talk about in this day and age is in January of 21. I think it was or 2020. I was risk. And I thought, what am I gonna do?
09:14
I was laid off. It's never happened to me before. And a couple of folks at here reached out and says, we have this awesome position for you and I've been here since then. It's been three years. Um I think next Monday and I literally have loved it here. So as much as I hated it, it's honestly the best thing that's ever happened to me.
09:31
So I, I know we're gonna kind of talk about that, but layoffs are really big now and I just wanna say like, don't give up it. It, it could always be a blessing in disguise. So let me, let me pull a couple of those themes there because, and this is the goal. You know, we all, we all have histories and, and what can we extract out? That's frankly hopeful,
09:47
hopefully useful for folks listening. So the first one you mentioned is that, you know, I still figure out what I want to be when I grow up. When you said that earlier, when we were preparing, I was like, I use that line too. I, I think I've used it longer than I've known you and you've known me. But it's just like this idea of staying curious,
10:03
look, not, it's not being like we're looking at jobs as stepping stones kind of thing, but we are thinking about what's next. I'm not just here to do the thing, but I'm actually here to learn and build and create more. Sometimes that's even even can be, I'm not here to click all the buttons. I'm here to automate the things, right? You know,
10:18
that can be part of what's next in that. So there's all these ways of that, staying curious goes and, and back to you before I keep going with a couple other themes. So um yeah. No, it's definitely interesting. So, really excited about that. The other piece there and this is maybe two together is um you mentioned a little bit,
10:37
but I just wanna have you pull out a little further if you don't mind just in community involvement and what that's meant for you. And I think even there you mentioned there was a little bit of, you weren't quite sure and you got thrown in the deep end and it worked out. But do you, I kind of play through that a little bit from a community standpoint. Yeah. So kind of the really cool thing about my,
10:56
my early on in that first role, um when I was at ADMIN, I had a, a really good mentor at the time and it's really something that has really pushed me to kind of make mentorship a big part of the career. And I was always heavily involved involved with being mugged. It's been around forever. And obviously, we know they,
11:13
there's been a lot of um challenges there and we're definitely seeing that community grow again, but we had a leader and they said I'm gonna work for a partner and for VM Ware, I can't become a leader anymore. Um Are you, do you want to do it? I'm like, well, I kind of want to, I kind of don't, I, I don't like public speaking and it was pretty
11:33
much like, I know you can do it. Here you go. Um And honestly, like at that point, I hated public speaking. I couldn't do anything. But that was really, I would say a really good launch point into everything else career wise because being in there opens me up to becoming a v mug leader and embracing a whole new
11:52
community from there. It allowed me to branch into like the, the land of being a V expert and working with other VM ware influencers, even though influencer wasn't a thing back then. But understanding how we do and how we network and building this because we always say it's not what, you know, it's who, you know, a little bit and there's always something you can learn.
12:10
So I kind of feel like from there, I got to branch off and present at VM World. So it was always kind of this really big launching point that now what do I do? I talk to 1700 people on a Zoom when I was scared to talk to like two people in a room. At, at one point, I think it's really matured over the past. I don't even know how many years it's been anymore.
12:28
So you're doing great. Last, last ones I think I'll pull here because we so much to go through good stuff. But is this idea and it's actually kind of pairing to up? Um The, the idea of songs I think about confidence without arrogance. And I'm, I'm gonna quote you from, from what we talked about before about.
12:45
It's not what you know, but how you can communicate it with confidence, which, which isn't the same as arrogance is the confidence. You might just kind of continue on that a little bit. Um Trying to figure out the best way to kind of tackle that right? Is the idea here is not everybody knows everything.
13:01
Um And it's part of those things like when, even when you think in an interview, if you don't know the question, someone's gonna know if you're lying and sometimes it's better is I'm gonna follow up in regards to that. And so in some cases, as long as you can express confidence in what you're talking about, it can provide that and especially in our world where we're interfacing with many different
13:20
customers, organizations, large and theoretically one of my challenges in previous roles, even this one is what happens if someone comes into a room that might know more about a specific to technology than you are, there's always gonna be someone that knows more. But if you can relay kind of the requirements and the strategy and the solutions and kind of all of these things, it really allows them to embrace confidence in you.
13:41
And as we're leaders in kind of handling this, they need to make sure that who they're talking to is someone that they can trust and become a, a value partner in their organization. The phrase I think of there sometimes is knowing what you don't know. I think of this for myself too and then being, being comfortable with that and, and a degree of humility in that hopefully there because we know a lot of stuff,
14:02
but we don't, we don't know everything. Final piece there is um It's just kind of a comment. It's a time back to the community. I love David. The comment about, you know, best, worst thing that ever happened to me. I think John Jonathan commented in the chat about it.
14:13
Love seeing yourself and Jace mccarty come over here and this is where from a community standpoint. I mean, there is pay it forward stuff. We don't know exactly how it's gonna work out, but it was, it was those community relationships that ended up helping you land it pure even through what wasn't the most fun way to get here in a way. And we're glad you're here.
14:29
So, no, I appreciate that. OK. Section number two. And actually, I will go ahead, I think here and I'll just, I'll just go ahead and close the poll because we had a whole bunch of people respond and then we'll dive into the time machine and let me share the results back out here.
14:44
So hopefully everyone can see this. What version of vsphere did you start with? Uh So we, this is actually a pretty solid spread only 4% saying what's VM Ware. So hopefully, um III I hope you enjoy the time here because we're not going to do a lot of VM Ware basics here. This is V basics, uh 101 stuff, but uh we're pretty evenly spread between kind of the three categories and then a bunch of
15:07
folks that aren't VM ware admins and they are just here to learn. So very cool, very cool. So time machine time, we thought about this a little bit how to approach this because there's so much that pure has done historically with VM Ware. Any given item that we might do, we do might be new to you because you haven't looked at it
15:25
kind of thing. So what I want to acknowledge that uh about oh a year and a half ago now, roughly I had Cody Homan on what we talked about there, you know, ain't nobody got time for managing storage. A history of simplifying VM ware storage is still relevant. I I didn't go back and listen to the recording but I don't think anything in there is stuff
15:41
that doesn't apply anymore and especially some of Cody's history and background is really cool to hear. But we want to do a brief review of some of the general things that pure has done for a long time and some of these you may not know. So I think David, do you mind kind of taking us through the core of in this case, really flash ray but even pure for VM ware workloads and,
16:01
and what stands out to you? Yeah, and I kind of really like this. As I mentioned, part of my history was I was a pure customer and the best thing that kind of showcases the ideal storage platform is just simplicity and expense. Um This was probably like 68 years ago, is we had our SD drop off R A and said, I'm going to lunch when we get back, we'll configure it.
16:23
And 15 minutes later I replied, it's already up and running. I'm, I'm already running workloads on it. It, it is truly simple and a lot of that comes down to a lot of things we're gonna talk about. It's when we think about up time. Right. We don't have to worry about trading off balancing stuff between storage controllers and balancing it there.
16:40
It's all done automatically. We don't even have to worry about that. We can withstand that performance through both failures and maintenance and again, through that uptime failures rarely happen. But what's important is all of our features are just enabled by default on the array or the options are kind of hidden that you don't need to think about that.
16:56
So think about global data reduction, then provisioning duplication, compression, those are things that are always on. And so when we create volumes, all you have to do is give it a name and give it a capacity. There's no other knobs to turn. There's things that you can set such as like protection for replication and snapshots,
17:13
but it's there by default, you don't have to worry about those. And the best thing is is you don't have to worry about maintaining hours of managing software updates. It's all delivered directly through the cloud through our support organization. And we're even driving that one step further through our cloud delivered updates.
17:28
But as we think about this from a VM Ware perspective, what does it really mean? Um We are a technical partner with VM Ware, which means we're always doing joint development, which means everything we try to do, we try to provide native and by default. And IVIS, we have a best practice guide, but it's more of a best practice recommendations. Why did we choose these settings and what happens if you were to change them?
17:51
So our multi pathing is default in the hypervisor. Our CTP policies are default within the hypervisor. We work with S A for our balls and all of these services. And so it's kind of an important thing is that when you set up our array, there's really no best practices to set, it's just there by default.
18:07
The last thing that I kind of want to focus on is performance because that's always a big thing when we think about having storage for these environments. And sometimes when we think about best practices, customers ask me what is your best practice for a volume size? Like we're using two terabytes today. And ultimately, it comes down to it's whatever you want.
18:26
It's a business level decision from a flash array perspective, whether it's one terabyte or 64 terabytes performance is exactly the same. We tend to see customers carve it up based off business requirements. This one needs to be snapshot of this one needs to be replicated and we'll really dive into from a VMFS perspective. How do V balls actually change that?
18:46
And so it really is this ideal scenario, right? Flash ray X is our latency optimized flash ray C is our capacity optimized but still very performant. There's a lot of integrations there though because VM Ware has it's more than just what's here, right? There's a good, there's kind of a big portfolio there. This is a really big portfolio.
19:05
So I'll kind of build that out here for a second, right? So if you really think about VM Ware as a whole, it's really drastically matured when we used to think about VM Ware and even V bugs and these user funds. It was what does vcenter and vsphere do. But over time, we brought in V Balls and VM Ware cloud foundation and cloud director and
19:23
now there's Tan, there's Horizon and now the whole a suite of stuff, right? If you haven't realized that they got rid of, um it's no longer um it's not, not be like cos then you realize now it's a um and so really what it comes down to is it's not just coordinating at the storage, but it's also putting out all of these things and making sure that we have this full
19:46
ecosystem. And again, as a technology partner with VM Ware, we have the ability to develop and kind of drive a lot of innovation with these solutions to make sure that we're collaborating well and continuing to drive what we do very well. And what's also important is flash array is a block storage array. Well, we do have some file protocols. They're not there for VM Ware yet.
20:08
But what's important to understand is whether it's fiber channel is right, we can maintain all of that really good at leveraging existing investments with vivas. We can also adopt those off of those things. And as we see a lot of innovations and technology drive, especially around mbme over fabrics, we're also gonna be seeing a lot of options change there.
20:28
So with the latest version of purity, we now support MBME over fabrics with TCP. Um And our great solutions team just put out a bunch of content. So thanks to Nelson for that, we can always link that after the fact, but a lot of customers are looking to adopt that because you don't need really any specialized hardware, right? You can use your the off the shelf hbas.
20:46
But where this is important is reason why it took, I wouldn't say it took us a little bit longer to do that. Was that even from a VM ware perspective, it wasn't there. If you think about in VC seven, there was NVME over fabrics but it was no va I no clustering, no shared disc. It was pretty much a raw performance volume
21:03
which met some requirements but not a lot of them. And so now it's kind of the right time for the right protocol as we kind of think about this. And then when we kind of think about um the overall strategy of what we're doing from a protection, I always like to say is while we have the right protocol for the right job, we also have the right form of data protection for that.
21:29
So whether it's the most basic of doing local snapshots on an array for ransomware and, and availability, we have the ability to do that when we think about how can we globally protect our data. And again, whether it's through anything, we have the ability to do our replication, whether it's synchronous, whether it's asynchronous, whether it's periodic, whether it's continuous,
21:48
but really tied in, in a very efficient manner to have customers provide these low RP OS and low RT OS, whether that's between their private cloud, whether it's between their, their public cloud um or even their hybrid cloud or whether it's something sitting in a in a coo or an MSP like the tier points or the equinox metals of the world. And even the ability to tie those in with all of our backup partners through our integrations
22:10
to make all of those backup utilities operate even more. And then taking that one step further to bridge that to the public cloud, whether it's Aws or Azure or even our clouds, not functionality. But the best part about this is a unified visibility through pier one. So it's kind of really one of those really great things that when we think about modern
22:28
data protection, there's many different ways and it's always the big kind of question is the business continuity or disaster recovery. And I feel like our whole MD P um is all about kind of encompassing all of that together and delivering that to our, our customer base. So last question before I kind of go ahead, you know, II,
22:48
I was gonna just make a comment about ESX top, but I feel like I jumped the gun on you. So keep going. Yeah, I was, I was clicking on the wrong slide, but that's why it didn't happen. I can make it. So, so anyone who spent time staring at ESX top with all the different things you look at and your head hurts.
23:03
Hey, we got something for you. Yeah. So virtual machine analytics and I'm probably gonna uh get in trouble for saying this because I always like saying it is VM analytics is probably one of the best free value add tools that you have as a pure customer, but you theoretically don't need to be a pure customer. Um So we'll kind of follow up with that what this actually is.
23:21
So, so virtual machine analytics is essentially the way that we can correlate physical to virtual data from your VM ware environment to your storage. So if we think about this. This is kind of a cut-off screenshot a little bit, but we show the disk to the VM to the host, to the data store, to the volume to the array.
23:37
And as an admin, you always know that someone comes to and says my VM is slow. Um And then it's normally the storage team pointing fingers at the infrastructure team pointing in fingers at the network team and then the security team and nobody really knows what is actually happening. And so with VM Analytics, you can actually search for your VM that's experiencing an issue
23:57
and see the breakdown of latency or performance at each individual layer. And so if the VM is showing high latency, but the host isn't and the volume isn't that right there. It really shows that there's an issue at the VM level. And so the infrastructure teams aren't scrambling and saying, hey, this is a, an application driven workflow that needs to be
24:14
engaged. If we're seeing latency somewhere more upstream, then that kind of shows that target. So it really depicts all of this. The really cool part about this is while the volume and array information gets flown home directly from our pure arrays, we also will collect the data store down metrics through
24:31
VCENTER. So if you're using V balls or RD MS or NFS or any other third party storage, we'll still show that correlation of this first part. Um So Tristan just posted a link to all of this as well. So it's really a great way to troubleshoot. It is only seven days of historical data built in, but a lot of this can still be mapped into the Aria operation management pack that we kind
24:53
of talked about earlier. It's meant for real time troubleshooting. I sometimes describe this as someone who walks in your office and fastest past to not my fault, talk to somebody else or hm, can you come back later while I look at something that kind of thing? Real time troubleshooting and with some history?
25:07
Thank you, Lauren. Do you mind pulling up? Poll number two, we're gonna go into a little bit of a Vival overview and uh man time always goes fast. It's crazy. So, uh poll number two. Are you using V VAS today? And the reason we're going to do this first is literally David.
25:22
Next is gonna give a bit of a viva technical overview and then I'll give a little bit of frankly kind of adoption history. So I'm not even gonna read this poll at you because it's relatively easy to read. Are you using vivas today? Go look at the questions that the the answers David, do you mind continually just kind of walking us through basically where V VAS came from and why it got created and what it does
25:41
and this is to be real, this is pure, purely no bad times intended education for folks out there about Vival. So please, if you don't mind. Yeah, definitely that I saw that Craig does, did ask a question to the host and panelists, please make sure you actually put those in the Q and A so they can be tracked and documented shared and we'll, we'll love to connect with you for a bunch of those things.
26:01
Um And so really this whole concept of is that we think about this kind of this quote, I kind of like it is with a you're provision storage for infrastructure instead of the applications. And this is really one of those key things, especially when we talked about provisioning storage and data store sizes and kind of all of these items because the way storage was essentially originally designed and if we,
26:25
we can even take this to a container land is how we can help optimize that. But that's for another day, it's when you have AES X I host, um you map store to that individual host and that one host uh runs multiple V MS. But in the end, that single singular data store is a single container for all of your storage, which means if you have many different V MS, many different disks,
26:50
there is gonna be kind of that mapping of all of that visibility. All of those data services are really locked down to that bottom. And so it's very tedious and kind of hard to manage because if you want to be able to assign capacity, you have to grow that entire data store. If you want to manage performance, it's out of that data store.
27:07
And again, for us, it's not really an issue, but it really comes down to array services. If you want that data store, say you want like that volume one to be snaps shotted. Well, you have to say the entire volume of a vmfs perspective needs to be snaps shotted, which means you might be protecting more data than what you want. If you think about this in an Srm world, we go ahead and think about you might need a whole
27:29
bunch of protection groups and a whole bunch of data sources. And there's a lot of sprawl. So when we think about how EMS kind of works, right? It's, it's still that many volume to that host mapping, but it's still difficult to troubleshoot these issues is find out where is everything defined VM Analytics can help with
27:45
this a little bit because we have that mapping from a capacity level and a performance level. But it's still not truly visibility. The V MS don't have access to those data services. And so this is really where we see the benefits and I'll kind of build this out here of where Vals really help. Um When we think about a vmfs data store with a line, the data services are at the lun level,
28:06
which means you have all of these workloads with the exact same SL A if you have a noisy neighbor, well, that one VM or that one volume can impact everything on that entire data store. But as we're trying to track growth, uh what happens from a performance and capacity, it's at that lung level, it's not there. But what's also important is this compliance drift of when we actually look at a volume and
28:28
you apply a policy to a VMFS data store that's not at the storage level. And so if your storage admin changes that policy, you really don't know that it happened at the VR level with vivas in the integrated storage policy management that becomes front and center and provides you a lot of that flexibility. And so when we kind of think about where vivas come in,
28:49
right is think about Vivas as an integrated RDM and we kind of think about RD MS as a bad word, but it's not bad in the sense, it's a rod of ice mapping. It's, it's an integrated, yeah, it's a goal. It's an integrated management framework that allows you to virtualize your arrays to be able to enable a more efficient optimized centered application.
29:09
And that's kind of VM Ware's definition and we'll kind of showcase what does that actually mean. So these here virtual, right? Um Simplified storage management operates troubleshooting because it's not managing storage anymore. You're managing your data at the VM level becomes granular storage. That's both policy based managed and all of the
29:29
data services are still accessible and offloaded to the array. We're not gonna go super deep into everything V balls because that would just be a, a much longer story. Uh We've done really deep with deep debar or webinars on this with our team so we can always link and follow up with those. But when we think about what is a virtual volume, we're gonna cover this at a high level.
29:49
A V again is a well orchestrated system of a raw device mapping. Not an RDF. Why these are beneficial is because there's no file system on the underlying disks, no VMFS and NFS, which means it provides a whole lot of flexibility as we're having to migrate workloads around. And this whole data mobility platform,
30:07
we don't have to worry about if we take that VM and migrate it from VM ware to Hyper V or from Hyper V to KVM or from this to that or to the cloud. Um It truly becomes the guest os is the data file system and you don't pre provision anything right? You say I want a disk with 20 terabytes, it provisions that disk and maps it to the VM. It's integrated and storage policy based
30:28
management. As you say, I want it to be snapshot it or I want it to be replicated and the arrays policies are exposed and allow that and then the replication management and un map a whole lot of benefits here. Um I'll kind of talk to this and, and kind of see if there's any other thoughts on this, Andrew. But the reason why pass through un map on V
30:47
Balls is very important is when we even think about performance and deploying. Um Traditionally, when you use vmfs, you have to think about thin or thick provision in Ks, right? It's a, it's a complicated question, even though it's two choices, we always say if you want performance go thick, if you want capacity efficiencies go thin.
31:06
But if you want the best of both worlds, there wasn't really an option until now. V balls are a raw disk. So you get all of the performance, but it's a thin provision disk. So you get all the capacity efficiencies. So things like database servers that might have a high turn of rate and you want to do in guest on map are really great candidates for some of those.
31:26
I think if you don't mind, David, I'll, I'll wrap up this section and uh jump into adoption history. There's, there's so much good stuff here. We could even go into like it shows there about testing without replacing coexistence. It's very easy, it can be easy to move into them. And then like you were saying, it lets it be policy driven,
31:43
application centric, et cetera. I know I'm almost accelerating this a little bit. So I wanna make sure I don't keep you from putting in any other key items you want to mention here? No, that that's kind of important. This kind of just showcases that whole infrastructure centric approach to, to application centric approach.
31:57
Yes. Yes. So if you've been in this space for a little while, you've heard about Vivas for a long time, I think my first preview into them was maybe 10 years ago. I might be making that up. Someone can go look on Google and play stump the chump and be like, hey, you know, so one thing that we've realized and actually we first
32:15
did some of this when you and I, they were presenting a tech field day uh at VM Ware explore. So some of this you may have seen before, but, you know, some things don't change. It's still, especially the history. So often when we look at technology, uh we have the idea of an adoption curve.
32:29
OK. You know, so you've got tech enthusiasts, visionaries, prime Minister, et cetera. And you know, there's this idea of like, you know, eventually you get to uh early adopters and you get to mainstream and late adopters, et cetera. What we saw with Vival interestingly enough was early in.
32:43
There were some bumps along the way and we're gonna hit on some of the challenges that we've seen and that we've engineered for really well later. But there was almost like there was this gen one of val adoption and then there was almost like a reset because there were some challenges with Vasa provider and stability and some other pieces, you know, this is going back 10 years ago, like all these great goals,
33:02
the core principles and ideas that we're talking about have been the same, but some of the underlying implementation and the capabilities, even at the time, storage wise, wasn't robust enough to handle it. So it was almost more like we had kind of gen one Vival adoption and then it was more of like, OK, let's reset and almost kind of start over.
33:19
And, and the reason that I bring that up is because sometimes we'll have folks who have looked at Vivas a long time ago and been like, it didn't work for me. I just want to highlight there's a new set of capabilities and al almost a redefinition of Vival implementation there. But what we saw then was that as we built this out originally,
33:39
Viva is a standard that if you didn't just uh uh apply comply specifically with the way it was designed. Um If you and you went colored outside the box a little bit, you'd actually have problems. So we saw, saw, you know, there was a standard and it wasn't both flexibility and stability. We're almost kind of going against each other this way.
33:58
I know I'm being a little bit high level industry. So, you know, the analogy is that we had, you know, gen one and gen two V vas, if anyone remembers this from like as a kid, you know, you put this together, we got it and you glued it together and then it was done, it wasn't changing, you know, kind of thing and you broke something off.
34:12
Well, that's sad. So the goal is as we've been working on this is that we want to get from gen one and two V VAS by a lot of work, a huge amount of work around scale performance at scale because that's a different thing and stability at scale. So this is the performance of Vasa of managed snapshots, et cetera to get to something that's more like this,
34:33
maybe call it a gen 2.5 V vas. Actually, my only presented this. My, my boss Jie Wallace was kind enough to send this to me. I've actually, sadly not finished this and I need to but where it's, you know, way more flexible. Yeah, we're still putting the thing together and that we want to continue to be the best Vival and VM Ware storage platform because
34:52
according to VM Ware pure is actually the most deployed platform for virtual volumes. Amazingly, right. Let's go ahead and uh Lauren actually, I'll, I'll do it here. I will close up the first poll, second poll and share that back. Uh David, this is just interesting, you know, we're seeing are using vals today.
35:15
Uh 17%. Yes, for the majority uh specific uses. We're almost at a halfway i point of people using it. And then there's a mix of people that aren't uh if you don't mind, Lauren, go ahead and, and uh sharing the third poll that would be great. And we will uh dive into our final sections here today.
35:32
So if you're not me, kind of, it should be interesting. I was expecting that to be a little bit more of a, an, an imbalanced ratio as we kind of thought about this question. And there's a lot of stuff here and I, I know there's a lot of these questions of, I know that probably the people that said they're not using it or the RD MS for Life are
35:49
probably kind of following into this. I need active Dr and active cluster support. We hear you. Um It's not a nice thing. Um We're working very closely with VM Ware to deliver those solutions. Um So stay tuned. Um Always consult your pure account team or your VM Ware reps to kind of give you any updates in regards to that as well.
36:10
So let's talk about what's new and cool uh specific or the new coolness. That's what I'm supposed to say. That's the title. So there's some pieces here around Vsphere, there's some pieces here around V VAS and then even a little bit of philosophical, you know, API S versus file systems. So do you mind taking us through some of the what's new?
36:27
Yeah. And this is kind of showcases that we're not done yet, right? There's this whole evergreen, not evergreen, but this whole development of continuing to mature it. Like our vsphere plug in is really kind of that, that first point of extensibility, it's bringing in all of these features that you don't need to manage storages sometimes like it
36:44
is, I get into a call with an account team or with a customer and they're like the storage team is like, what am I gonna do now? The VM Ware guy can manage all of this and it's not really our goal, but our goal is to allow you to do everything to manage the storage right through the Vsphere plug in. So that is if you want to um and again through here through powershell or through any of those
37:02
tool sets, but right click a cluster and say configure host, it'll configure the host on the array and all of that. Um Right, click it and say create data store, it'll create a volume, knock it to your huster and provision it. You can create and modify all of your data stores and snapshots and things like un map.
37:18
But we really take this one step further because if you think about what we've done, um part of it is being not done yet is with Vsphere eight VM Ware defecated the local plug in. So if you're, if your customer today and you might be familiar with having to go to an array and install the plug-in from there. That was what we call the local plug in the four dot X and V four.
37:39
And there's a lot of challenges with that local plug in. So VM Ware deprecated that with a 0.0 and if you look at our new five dot X plug in, that is our new remote plug in. But it also has provided the ability to bring a whole lot of feature sets to the plug in that our customers have been asking for. So role based access control, remote architecture, the ability to do point in time
37:59
recovery for both and BMFS um and kind of providing all of these insight that you just cannot get natively, that we're able to provide that in there. And there's even more and more stuff coming into it. So we're really excited to kind of bring in. I know there's a lot of conversations around NBME and O TCP and stuff. And again, it's coming,
38:20
it's coming, we're not done yet. There's always a whole lot of development features you want, let us know. So it's not just been all roses and I and I appreciate this is where we want to be transparent about some of the challenges we've seen uh back to you. Yeah. So as you kind of mentioned, Andrew,
38:40
there was really a lot of complexity around the initial setup and sometimes that, that bad taste in your mouth from an initial setup. A lot of the times back in the day, Bosa was a BM you deploy, then you have to do all of these things to get to run and be stable at here. We kind of always have it as a native service on the array.
38:57
But challenges as again, we've seen in the Q and A are certificate management and fail over timing and performance and all of this. And if we think about what are some of those challenges that have kind of stopped adoption, it was a lot of those. And so we kind of reset really in like 21 and 22 is you didn't see a lot of features come out. But what we really did was focused on scale and
39:18
its performance at scale stability at scale and making sure that it truly was ready for that prime time that our customers can come to love and adopt. In that case of all of those customers say I'm running everything you don't have to. But you can, once we kind of thought about that, we've looked at how has this also matured over time, right? If we look at that 2020 to 2021 kind of 22 mark,
39:42
there wasn't really a lot of features, but after that, we've been working with, with additional features, additional scale and kind of bringing that on because not only is it just us making a really large investment in in this, it's VM Ware as well. And Jason Mas is a really good resource at VM Ware to kind of listen to and see his webinars and around be balls.
40:02
But again, he's a really big proponent of this and we're definitely gonna see this growing from a VM Ware perspective, just not pure but hopefully other storage vendors to make it a better taste in their mouth for all of our customers when, when we talked about and I went by pretty quickly in the previous slide about talking about scale performance of scale stability at scale.
40:20
One of the things that's fascinating that often we don't think about actually gonna go back here. I'm saying real fast is that um I think we do a double click is that um often you think about the data plane and the management plan. So there's the data plan, the data flowing across and there's a management plan. V usually the data, the actual protocol calls, if it's going in a file system,
40:41
api interactions versus file system interactions are just a different thing. API interactions are actually sometimes heavier or they weren't built to have the same type of file operations passed across them. So we actually saw things around, you know, managed snapshots that were relatively demanding on the VS A provider.
40:55
So we had to do a lot of optimizations around almost our management plane. This was stuff that we saw as pure that we had to embrace and, and we did that goes to why, you know, we're the number one deployed platform, but this is where and this is now a little bit of a uh an industry commentary if you see people saying, oh, they have checkbox Vival support and they aren't acknowledging some of
41:14
this stuff. It may not be as robust and well engineered under the covers as you might want or hope because we had to un unp piece a lot of these things that vals and Vasa provider uniquely and differently stress from a storage ra standpoint and solve for it. That's why, you know, like it says number one deployed platform section number four. Yeah, please.
41:34
Yeah. And one thing is we'll kind of address this one question like this. How does the size pier present to VM ware been fixed on either side? And it definitely has. So, one of the kind of considerations is when we created a V container um on an array, we had one per array.
41:47
And ultimately, um we can blame Cody for this. I don't know if he's listening but hey Cody, um and they said, how big of a storage container should we create? He's like, I don't know, nine petabytes. Um And it was kind of this running joke because there was no max size and that was the max size that VM ware supported. Um So that's why it was nine petabytes come to find out.
42:06
It wasn't really the max size because SDRS and content libraries can only go two petabytes or lower. Um So we're kind of changing the default sizes from that. But two things that are kind of coming is one is you could have always contacted peer support to resize that container. It wasn't an end user facing change, but it was possible.
42:23
But with the latest versions of purity and 641, we now have the ability to do multiple V all containers um which now means that you can carve up multiple containers and not just have one large one, but also what will be coming um Very soon. Well, it be be the ability for end user quotas on those containers as well. So you can make it 10 terabytes, five terabytes and maintain it all yourself.
42:46
So yes, that challenge has definitely gone away, especially for customers who kind of want a tendency approach in regards to that. So there's more coming and this is where for being real. Uh This is a public webinar. We can't have road map discussions here, we can with you individually under ND A, you know what a surprise but wanted to hit on some
43:04
of the themes that we're looking at. Of course, there's more Vival engineering today than ever before, but there's, there's kind of three main categories. Do you mind uh briefly hitting on those David? Yeah. So number one stretch to support, I think that is also the number one question we've had in this QO A and thank you to Jason for just
43:20
putting this really good link in there. So as you mentioned, it's not necessarily an US thing. It's a joint development is the reason why we don't have active cluster support for balls is because it's not physically possible today from a VM Ware perspective. So as part of this joint engineering effort, both pure and VM Ware,
43:36
as well as probably other vendors are looking at the ability to support stretch storage support. If you are at VM Ware explore. Um You're at 22 we actually did uh uh a live demo, Alex Carver and Jason. Um and we actually did a tech preview of this. So we actually put a link to it. We put a blog, how it works, there's a video and all this great content.
43:55
So fundamentally, it's there, there's still some stuff under the covers that VM Ware needs to do. Um Again, we can't talk any timelines, but if you want more specifics, consult your VM Ware reps for any type of VM Ware specific road maps, or you can also work with us to have these more detailed one on ones um based off that, that we can have under NBA, it's always fun having those conversations with all these
44:15
people, but we have to be careful with what we say there, it is performance and scale more work there, right? There's always, there's always there right is you give something there and you always want more and more and it's for a good reason, right? So performance is always one of those things, skills, always one of those things, right? Especially from Balls.
44:34
I, I have all these B MS in there and we're kind of to that point where people aren't really asking for a lot more, but we know that we can give more. And so a lot of that is really happening when we think about these, it's not just performance of like the I OS, but performance of backups and replication and all these things. And so it, it really kind of ties into a lot of
44:54
that managed backups and stun times that we've really done a lot of, but it's making that better over time, what we talk about with certificates and other things that are going to become. Um And really the reason why that is important is if you're on any of our ways of safe mode scale and objects around, especially around snapshots becomes a challenge.
45:13
Um So there's a lot of stuff in flight to kind of help address that and this scale is just one of those things and for those with us, thank you for, we, we are as often happens, we go a little a minute or two past the 45 minute mark. That's cool. There's still a drawing we haven't forgotten about in case someone's like, hey, it's 45 minutes.
45:28
Nvmeof man, this is even impurity 6.4 0.2 when we first put this slide on, I think it was in a what's next? And it's a what's now maybe for uh TCP rocking fiber channel around for a while. But uh anything else you want to add there before we wrap it up? No, I think we can, we can jump to the more info because everybody's been asking us questions.
45:46
It's a really great resource for, for that. So, um but yeah, if we jump to the, the more info, um Tristan Jason. Awesome. Thank you so much for all of this um youtube channel. What's new QR code to our platform? Guide? Everything we talked about is in there. Um Links to documentation,
46:03
links to Balls, links to the MFS, our best practices guide BM analytics, um All of that as well. Um Still seeing a whole bunch of questions come in. Hopefully we'll stay a little bit after and address those offline. But the most important thing now is the wonderful drawing. So I wanna make sure you're here to give props
46:19
to yourself. David Cody, hoster, Alex Carver, Nelson Elum Jason Lager, Kyle Gross Miller Tristin Todd. If I've left anybody out, there's so much great content there and there's a whole team of people more than that behind this. It doesn't just happen even as we're, we're the ones up front talking. So David, we went through a ton of stuff thank
46:35
you, like I said for being in the boat with me. And we've been kind of like as always like we, we plan it out but not precisely. So we are in the home stretch into a drawing in Q and A. So from the drawing, I need to make sure to look back here, Diane Kay from Illinois. You are the winner of an ember mug value $130 a kind you can control with your phone. Please make sure to join us next month where
47:00
I'll have Sandy Bond as we'll walk through A I and ML and even some of the application that we're doing is pure, but like this even says we're not done yet even here because the goal is that we hang up for another 10 or 15 minutes because man, there's been so much great Q and A and uh I think Jason and uh Tristan have been typing their fingers off a little bit, which is great.
47:20
Maybe I will. Uh I'll pull the music back up here and uh dive into David. Yeah, I was gonna kind of take into some of the the questions. So some of these might be easier to update live. So um J Pendergraph asked multiple evolves managed by customers. Can they be multi size and dynamically allocated?
47:38
The answer is yes. Um Multiple containers, multiple sizes um from there. Uh um As far as Dave Whitaker's question, how many V centers can connect to the same array. Um Ultimately, there's no specific limit, it's bound by VS A and all of those limits. So if uh the easiest way to consider this, if you're using the self signed Vasa certificates,
48:01
you can have one sso domain connected up. If you want multiple non linked V centers to talk to an array, you need to use signed certificates. Um So one quick note, I forgot to do the final poll. So I want to toss it in here. Who is, who is, who is I curious about the results? And we'll keep doing Q and A.
48:17
So let me uh share the results there, which is um 9% don't be a benefit. Understood, you know, if you're not using vials today, why not? We're not here to push it hard to push it, push it on you. But because we still do a really good job with vmfs across the board. So that's actually an all right answer for peer
48:34
to be honest. Uh Some, a solid quarter is understanding it and then a mix of the rest there. Uh Yeah, hopefully like this lo and our helped customers understand it, find out that they're not complex and scary. You're not locked in and you can do Yeah. So object limitation is still a big consideration, but again,
48:52
we're not done yet. So we're definitely gonna be scaling that up. Cool back to questions. I'll let you keep reading where I can read them to you. Yeah. So, so scalability, right? Always looking at a support site, each array model has a different scalability and it's growing with purity. So there are certain limits right now with
49:08
Excel support. I want to say with the latest purity, I think it's 5000, no 2000 BM and I think 20,000 Boes. Um So definitely we're going up and those are gonna continue to climb. Um And then one last one from Mr Luke um shared vas can they be increased in size while online? Um And I would say not yet. Um Again, it's not enough thing.
49:31
We're working with VM Ware. Um No particular road maps or thoughts on that. But yes, that is a challenge, especially when you're using shared discs. Um even shared VDK have that same, same issue. So that could have been that 2% that said um Long Live RDM just for a couple of those perks that you still have to kind of um consider. So awesome. So the status of NBME over fiber interop
49:56
ability, um that was a big one. So NVME over fabrics, Rocky V two has been around for a while. NBM over fabrics. But fiber channel has also been for a while NBME over TCP or VMFS is here today with Purity 642 and NBA over TCP with B Balls will be um a future conversation as well. If there's things that you wanna dive in more
50:20
specific to those particular timelines and statuses and details. Again, we're always more than happy to jump on a call, one on one and provide that information. I'm gonna pull one out of the answered section just because I was scrolling through that to see if there was only one that we wanted to highlight.
50:36
Uh There was one from Gary um saying, can you Air Gap V Balls? And man, I can't resist a good Air Gap ransomware commentary. So, hey, but the what made me think of was actually a post by Cody Homan that I put in the in the zoom chat about the case for vals and ransomware. Um which is actually kind of a funny topic because it's a case for vals to,
50:53
to help with ransomware. I'm pretty sure that's what, what coding means. But it's about the idea that pure has um call it permissions air gap capabilities that's safe mode, you know, it's air gap capabilities. And then from a Vival standpoint, it makes the unit of recovery more granular.
51:08
So you can actually make a case for the with the pure array. You can do a a virtual air gap, a permissions air gap and V vas can fit within that because safe mode is an overarching set of capabilities in the array. Um Gary, if you're curious, I will actually put that in the Q and A back to you, go read that post that I put in the chat or other ones there because there,
51:25
there is, there is a relationship there, although it's not necessarily one that I, one that I lead with necessarily. So, yeah, and I think a big part of that too is defining air gap and, and dark side. So there's kind of air gap. It's the requirements for V VAT is that the array needs to talk to your V center and talk
51:43
to your ESX I and doesn't need access to anything else. So if your management network is air gapped and those three can communicate together, yes, you can go. But if it is completely air gap where nothing can talk to nothing, then there could be potential architecture ways around that. But it, that is that we didn't pull it up of when not to use V VAS um but it's pretty much
52:02
active cluster, active VR and those air gaps are non. Um what is the word, non um stable networks where connectivity is there? Because again, Bosa and all of that is really that and I'm looking through the rest, there's lots of thanks as always, it feels like uh people enjoyed the session, which is great, you know, hey, you, you lived up to your billing David.
52:25
Thank you. Um Jason or Tristan. Is there anyone else anything else? And I mean, we're, we're in kind of relaxed to let our hair download even if you want to come off, any questions that you want to highlight or comments that you wanna put out there on the topic because, you know, thanks Andrew. Um There's been some really good questions.
52:43
Um I mean, I've, I've got a background in VD. I there's been a lot of questions about uh Vival and um support for VD I environments. Um things like instant clones and integration with citric provisioning services. I just say that uh we're working actively this year to kind of refine some of our architectural guidance. Uh The best practices around Citrix and VM Ware
53:04
Horizon. So I just ask everybody to stay tuned and keep your feedback coming because we're gonna build some really uh good modern VD I content. Um And you'll start seeing that this year. Awesome. Thank you, Jason. I see you off mute too. Yeah, I nothing here.
53:20
I was just gonna chime into that. I, I was getting ready to respond to the virtual volumes question too. As far as I know, I'll have to double check. I didn't even think instant clones supported virtual volumes unless that's changed recently. So that might be something I the question came in anonymously. So um for whoever you might want to dig into.
53:39
So yes. So there's definitely the way to kind of utilize instant phones for that. There's a lot of flexibility around that. So one of the benefits of something we can do is obviously scale when you're buying group limit hits, there's some benefits. So if we think about um the way we kind of do it for containers. Um We utilize the concept of a first class disk,
53:57
think about a first class disk as an independent. Um We didn't really go down into the detail of the V but when we move a VM to a V, we create a configuration Vival um we create a disk for the OS A disk for the data and then one for the swap file. And when you use like a a first class disk, it's kind of its own independent container,
54:19
which means that no matter how many first class disk you create, it shares a configuration. And so that's kind of that use case where you might have say a database that you want to be able to clone multiple times, it's not attached to an instance, but it's just the volume that you wanna go to something like that can be looking at how we
54:35
can tie those into the instant phone functionality to kind of get around those object limits. But again, we're working on those. So again, 2000 BM si think is today the limit. So 2000 volume groups on the array, if you're doing anything more than that, we'd love to hear from you to see what are your requirements to see what we can do for those particular cases.
54:56
I think we were at the end of the questions there was, there was so much we were going through David, I sometimes play the game at the end of like, man, I wish I had said that earlier, but we were off to the races. Uh Anything else you want to highlight before we uh before we bring it home and close up for the month. Um I don't, it was, it was really um exciting
55:14
to do. And again, we're always here. Um We're active on social media. Find any of us, tag us on pure um We have code dot pure storage dot com where we have links to a lot of our automation resources. Um There's also a slack. Um You don't have to post questions specifically around automation and code.
55:31
We have channels around V plug in evils and all of that. It's a really great way to interface directly with customers, partners, employees engineers and really just communicate with us in, in that standard media. So always a good thing that you mentioned code dot peers storage dot com. Um Go check out all the great things there.
55:48
That's what you can have when you have an API first platform that's simplified under the covers and then you care about automation. All that stuff is what comes out of it that please make sure to uh actually one last note from Matt uh about how do we get on the slack. You go to code dot pure stores dot com. I want to say it's in the top, right? If I can see it in my head, right?
56:06
It's, it's right. It's right there to the link. So yeah, we just search for pure storage code slack. There you go multiple ways. Not too hard. Thank you all so much again for joining us. Please make sure to join us next month for a session with deep walking through A I and ML and how we're turning buzzwords into reality
56:25
and both, you know, both general background of what he's seen with this being real in the industry and then even what we're doing specifically here at Pure with that. Thank you, David. Thank you, Tristan Jason Lauren and to all of you for attending the end of this month's coffee break. Please join us next month. Have a great day.
56:44
Hi, everybody.
