1:03:13 Webinar

Using STaaS to Navigate the Landmines and Complexities of Storage

Gain perspective on where the storage industry is rapidly headed, and learn new ways to address the complexities of traditional storage with one of the top Storage as-a-Service (STaaS) solutions in the market.
This webinar first aired on April 27, 2022
The first 5 minute(s) of our recorded Webinars are open; however, if you are enjoying them, we’ll ask for a little information to finish watching.
Click to View Transcript
Hello everyone. This is Tricia Ramsey and on behalf of pure storage options I. T. And 451 research part of S. And P. Global Market Intelligence. I would like to welcome you and say thank you for attending today's virtual roundtable titled Using staff to navigate good landmines in complexities of storage.
Just a few quick housekeeping items. Before we get started to ask a question, simply type it in the question box on your screen. We will get to as many as we can during the Q. And a session. The presentation slides are available for download in the resources section in the console and finally the on demand version of this round table will be available shortly
after the live event concludes. We welcome our speakers today and with that I'll hand it over to Henry. Thanks so much for everybody for joining us today. My name is Henry Balzer. Um The research Director of storage at at fourth doing research which is as Tricia said part of S.
And P. Global Market Intelligence. My key role here is to drive all of our research here on storage which includes storage systems software and also the emergence of cloud storage. And one of my other responsibilities as I run our end user research in these areas as well. In which case we do surveys throughout the year and we also do in depth interviews to get a
little bit more insight into what's going on beyond the numbers. And with that I'm going to go down the line with our with our presenters today. Our panelists for today starting with Prakash. Yes, so my name is Prakash. I'm the general manager of our digital experience business unit here at pure driving our subscription and service business for the
company James Hi, I'm James lam in global Head of infrastructure options technology. My name is Michael Cruz. I had a power managed infrastructure and connectivity business here at options. We provide managed services to capital markets, films firms on nearly every continent with a
presence in over 50 financial services co location facilities as we speak. Great, thanks you. Thanks all for joining us today. I really appreciate your participation for today's presentation fairly basic. We're gonna go through a virtual roundtable talking about some various topics associated with storage and service and how that potential new delivery model for delivering storage
resources is potentially gonna change environments going forward, especially because there's lots of challenges and lots of things with that. Many of the customers I talked to are dealing with today and when I look at the list of storage challenges and how we're getting to this transition topics, it's clear that it's probably one of the more challenging times ever.
We start looking at this list of challenges. You can, The thing I notice is there's just so many different areas where we need to make improvements of organizations going forward. The top problem and we will always probably be the top problem is data capacity and growth. That was the top issue for many of the respondents in the survey that we took earlier this year.
And that's not going to change? And to me that's not a bad thing either. I would probably be more worried if the data and capacity in your environment isn't growing because if it's not growing, that probably means that you're not really evolving really quickly or you're not creating a lot of usable data, which is also something that's scary. But I think the big thing that really stood out to me in this survey as opposed to previous
surveys that we run is that the high cost of storage on a cat backs basis wound up being the number two ranked problem that customers are complaining about when we ran the survey in at the beginning of the year. The other thing that switched a little bit was storage migrations, which are always a big issue that actually rose all the way up to third place position. So again, a lot of organizations are struggling
to figure out okay, we're not where should these resources be, How should we manage them? And again, a lot of these things are not just the capacity that we have to manage but where's the best place to run this capacity? Where's the best place for us to optimize these capabilities as we can also imagine. Uh we also have big issues here in terms of the other issues we see are in regards to
disaster recovery and compliance. Obviously there's going to be big issues. We also see issues in terms of delivering high performance and also issues in terms of how do we deliver the right level of how do we deliver the right level storage performance? Especially in these changing environments where you're always trying to react to different opportunities and things like that. So when we start looking at this,
the bottom line here is there's many, many, many problems we have to deal with. And because of these problems and the fact that budgets, budgets for storage are not unlimited. It really shows that we're at a pressure point. Now we need to make changes in order to figure out how to better suit our environments going forward or better suit our spending going forward. And in that regards, one of the things we've
also been seeing as a shift in terms of more organizations looking to use optics to acquire resources as opposed to Capex. In our most recent survey. What we did see was like there's still 42% response. They're going to do an equal mix of both Capex and topics. But the interesting thing that we're seeing is that The majority of people or the large
portion of people 30, which used to be about 31% doing Capex right now is going to drop to about 25% within the next three years. And in contrast on the optics side, we're seeing an increase in optics people going that's going from 26% to 33% over the next couple of years. So we are just starting to see a bit of that shift going forward. And in terms of why this object shift is
happening, A lot of the organizations we asked about, there's a variety of different reasons why they think optics is the way to go a normal one reason for the respondents is 64% saying it's the ability to get greater flexibility by using an object service. Uh, we had another 50% saying it's about faster deployments. You know, some people are saying it's about cost savings though,
we do get pushback from senator. People are the Capex heavy. You think that they can still do Capex cheaper. But for some organizations they still think that there's a there's a potential do cost savings by going to affects other things that they like is predictability and these are budget planning that come with an apex model. And you know, obviously there is also these opportunities to get uh in some cases it is
upper management might be the CFO might be somebody else in this week. We're saying, you know what, we're not going to do as much Capex going forward. We need to do a better job dealing with uh making this transition topics. And with that let's go into the first question here Prakash from your point of view at pure obviously we listed a long laundry list of different storage challenges.
But you from what you hear from the customers, what are the top storage challenges that customers are discussing appear? And how how do you think in terms of your office and your optics modeling peers of service, how could that potentially help alleviate these pain points going forward? Well, you know, I would break it down into three separate categories.
Right? There's the customers, there's the MSP. S and then there's the SAS companies, right? We serve many different types of segments and customers in, you know, a traditional enterprise customer. Um, largely we see, you know, Customers get to maybe 50, asset utilization over the course of their asset.
So there's a lot of stranded capacity and largely it's because of over provisioning and the lack of having the right thing at the right place at the right time. So a lot of the discussions we have with customers are a little less about Capex and apex. That's a means. The apex provides the means to make sure you have the right performance capacity and
availability sls at the right place to meet the right application thresholds. Um, you know, and that's the big challenge that I think customers have its capacity planning, performance planning because you're at the mercy of your applications and your end users. Um, M. S. P. S. It's a little bit different because you're putting that problem on steroids because you have multiple customers.
So how do you deal with that across a fleet in multiple customers and ensuring that you're running your fleet optimally. And then with SAS companies, it's actually been quite different. The motivation has been um, they, their ability to deliver a SAS service requires growing based on, you know, they're using additional users, acquired additional companies on board
and etcetera. And as they grow, they want that to be a seamless automated experience. So one is management. So how do you automate and embed an O. E. M. Inside that SAS solution management. Number two is monitoring And number three is all of the procurement activity.
So all of those things need to be built into their business processes. So it's seamless to scale with the way their customers grow. So, you know, depending on which customer and which segment 18, you know, we see that kind of wide myriad of challenges. Okay, great points. Yeah, I definitely agree with that.
I mean, again, it's not just about the cost of the storage assets, which obviously no one's gonna say storage assets are cheap, they're they're worth the premium and they have to do that. But there's all these other operational costs and other things we have to think about that are part of the total cost of ownership beyond just the cost of flash or controllers and all
these other things that we think about. Normally when we're talking about storage solution Gordon James with options, how the digital organization currently favored topics or Capex and how do you see that potentially changing the future? Um, it depends on the use case parts for a storage, things that are flexible in that sort of area topics is obviously far more kind to the business.
Whereas things like maybe server purchases where we understand what the life cycle is going to look like Capex still wins for us. It really is the rightful tool for the job. So we enjoy the flexibility of opec's for storage certainly. Okay, great attorney to mike since you're on the business side of things,
I'm just curious when you look from your perspective, has a transition to services made it easier for your product seems to be able to meet the requirements. Has it helped increase the pace of innovation? Because I think the other thing to keep in mind here is you guys are a very challenging sector in terms of being service providers of financial services. How how do you see this transition?
Has it helped you? Yeah, absolutely. Um, so it has been, has made it substantially easier for us actually in a, in a multitude of ways, which is possibly surprising to say because of course our businesses providing managed services. Right. Um, and it's been something that we've that we've been doing for a long time and are quite good at.
But transitioning storage in particular from a complex based model to an optics based model. Um, and even beyond the apex aspect of it but specifically to the menace service that we get from here has been really helpful in a couple of key areas um on the presales front for example it takes the question of storage almost completely off the table, you know, so so typically as a managed service provider when we're walking through a sales
cycle with a client or working through whatever you know what the requirements are, What they think their storage footprint is going to look like, whether that's going to be more heavily loaded for amount of data performance of data whatever the case may be. Um and frequently clients don't actually know 100% of what they need. Right?
And so we'll go through this whole process which takes a lot of time and a lot of effort and will speak out this this environment, we'll build the environment for them or we'll we'll add them to a shared environment that we have already ready to go um and then what we've deployed isn't actually exactly what they need, it's what they thought they needed. Um But then once the service is live you can see that in reality,
you know um there were maybe a couple of things that were missed here and there um Whereas having uh you know storage is a service from pure um really takes that completely off the table and we know what service we're giving them, you know we know we're giving them a super high performance storage, we know if we if we understand the amount of data that they need, that's fine because we can flex into it, we don't have to worry about how we
accidentally bought the wrong, you know, a piece of hardware and all of a sudden we need to make another massive investment too correct for that and add a substantial amount of project delays, sort of, all of that um goes off the table, which is fantastic. Um and you know, and that's true of whether we're whether there's a miss on sort of the data requirements side or on the performance
department side, so um it's been really great in that way um and then from a, you know, similarly, and from sort of a product BNL perspective, our storage costs are smooth and they align very clearly with the revenue that we're getting from clients. Whereas, you know, if you look at uh sort of a legacy storage model, you know, as jay mentioned for things like servers,
um service, which is, it's it's relatively straightforward, Right, because we know exactly how much computer client is gonna need, we're charging them for that, compute um you know, that computes depending on what they're gonna do, not going to massively, you know, scale and decrease and sort of change over time, if anything it'll it'll grow, but you won't have this sort of flexing up and down that you
get occasionally with storage, um whereas with a, you know, a storage model and especially with a storage as a service model, all of a sudden we can very clearly say, hey, you know, as your storage requirements change over the time, we know exactly how that's going to impact both client from a pricing perspective and us from piano perspective.
So it's been really great um, in that area as well. Um, and that's your question around whether it's accelerated the pace of innovation, I would say definitely. Um, you know, we have a storage stand box now where we can explore different product ideas, we can scale up storage to test something out and if it doesn't work, scale it back down, um, we're no longer make needing to make the same
sort of upfront large investment that we otherwise would um, for storage in order to support that. So there's been a lot of value there as well, which has been great to see. Yeah, those are all amazing points And I definitely, I thought the last part we talked about in the innovation part is also really interesting as well in terms of,
you know, the problem with today is it's about speed. Every things about speed. How fast can I create a new workload, How fast can we generate revenue and people want to fail quickly, but you don't want to fail quickly and be stuck with a bunch of stuff you didn't want to buy. So again and that's what resonates in terms of why this is interesting and you know,
why not be able to try these things first before you spend a couple million dollars on something that you want to figure out? But wait, that doesn't quite work what we needed anyway because the business could change that. This model could change or the application can change itself, which actually leads to the next question for James and James from your perspective,
what are the type of workloads that you're running in your environment? And what kind of challenges are you seeing from from those workloads? Um I think we're probably approaching the everything mark. Um so you know everything from high performance databases to file systems to archival and everything in between object file and block
and sometimes a bit of all, so that a fairly wide range of things. So some of the more challenging workloads are obviously more legacy style databases, which yeah, don't naturally scale particularly nicely. You just keep on throwing out that and hopefully it will fix itself out. And we have some very sizable customers doing reverse systematics against databases.
So those are certainly some of the more interesting workloads, but it's also worth noting things like that require some compliance and consistency across sites obviously have their own challenges. So they may not be as high performance on the face of it. But making sure that you can fail over a system within x amount of minutes or even real time between sites is a core business requirement
for certain um certain clients and they have fairly strict sls around them. Oh yeah, definitely. It's I guess that has that been a challenge for you in terms of trying to mix and match and control these various workloads? Um Previously yes, I would say pure in general will do a very good job of covering most of those bases.
Certainly when you augment it with a bit of service providers software you certainly do get a better some of parts. Um So we find that generally speaking, working with pure on most projects works pretty well. Um This can even go to the stage where we start talking, they'll always start discussing with the end client directly and pure to see where maybe workloads can be optimized and something
that we found pure don't shy away from it all very interesting. Side effects come from those conversations normally all positive. Great stuff. Thank you. Was there something you wanted to add? Precaution? No. Look, the workload dynamic are very interesting
because um every workload has its own you know cops profile etcetera. And one of the capabilities that we've been produced over time is using our data intelligence, you know we collect probably about 20 petabytes a year of usage like just around how people are using we do application classification. So you know whether it's an oracle database or S. A.
P. Or sequel server or you know um Epic and health care. Like we have a wide degree of figuring out application profiles and optimizing and sizing application profiles based on the data intelligence we collect that allows us optimizing for those application workloads across that spectrum. So, you know, this is one of those miles driven
arguments. Right. The more customers we have, the more data collection we have in terms of that where we can create these application classifications, um, the better we're going to be and we can localize it to different types of customers and segments as well. Okay, great stuff. In the next section, we're gonna talk a little
bit. I've been mentioning the term stands but I've been remiss because we didn't really just find what exactly storage as a service is. And you know, just from my perspective, I think the interesting element of this is historians service Yes, it's not the next model, obviously that's very different. But to me it's more than that because if you look at these things and how they're actually
being delivered these services. It's a service where the storage vendor themselves actually own the equipment are responsible for the management of equipment. And that's a big that's a big change because because of that a lot of these lifecycle management burns that we think about in terms of how they impact the total cost of storage in terms of installation, system maintenance capacity expansion,
hard to refresh all along down the line, Those things wind up becoming more the vendor's responsibility and hopefully allows you to offload some of your staff to be able to focus on other things there. I think the other key benefits obviously as an optics model, these type of offerings in the status space should provide elasticity to allow you both grow and shrink consumption based on your needs.
So you're not stuck with a bunch of storage just because you happen to have a big workload that month. If that workload goes away, you should be able to cut down your consumption, especially if you don't need high performance storage after the workload goes away. I think the other element here is because of costs and whatnot.
We need to make sure that there are various levels of performance tiers that are available to match what's going on with workload needs. In our most recent surveys, what we asked about is we asked a couple of questions about the customers, about how they felt about stands today and where they think it's going for the organizations that are already trying it out or we're thinking about trying out stars
remarkably. 43% of them already said that you know what we strongly agree that status is going to replace our existing storage systems, going forward. which is that quite high considering how early we are in this model in terms of stats. The other things we're seeing is when we start looking at the drivers for status we had 61% responded saying that they were using it
because they thought it would give them more ability to rapidly deploy resources when they need them. Which is obviously going to be a big thing for options and and other volatile environments where they need to be able to Respond to requests as quickly as possible. We had another 40% saying that they thought that optics was going to help them reduce spending and create some some cost savings
because again being able to offload some of those operational type costs and put them on the vendors side instead of on your IT staff side. And then we also saw that you know which we're talking about a little bit more 49% were pretty happy with what they were seeing in terms of the amount of buffer capacity the status providers were providing to make sure that you know their workloads don't don't don't overrun
the amount of resources they have there. We're looking at the negative side. We still see about 17% who are are not doing it because they're they're worried about the organizational change impact of this and again while you know it's great to talk about technology and things like that at the back of the line of this there's a lot of people involved too and some people depending on what
what this impacts in terms of their responsibilities and their power with the organization. You're going to see some pushback in regards to that in terms of moving stats. We also see some people that, you know, they'll argue they think that they can do complex cheaper and in some cases maybe that works. But you know, by and large what we're seeing
from organizations, a lot of them are looking to get more flexibility And potentially reduce costs by offloading some of these objects concerns that we talked about earlier. And another issue we see is there are some organizations 37% were holding back because they had a little bit of some concerns in terms of the security and compliance impact of these things. But you know, I mean looking at these things
and I think we start looking at this too many stars, even though there are some security concerns, it's me it's not that's not the end of the story. I think potentially a lot of organizations, especially when we're thinking about ransomware and some of these new challenges are coming about. A lot of organizations are looking for help actually on the,
on the security side. So to me, I think there are potential ways that we could use stats in the future to be able to take advantage of getting some of that expertise and some of the off load management to actually potentially help on the security front there and with that, I think I'm going to go here and we'll start with mike on this first in regards to your deployment, were there any compliance or security concerns that came up
while you're evaluating storages and service so definitely no concerns. There were definitely uh that said of course there were a lot of things that we had to look into to ensure that we would be able to use storage as a service, as an underlying platform for our financial services clients. Um one of the key questions going in that that was critical for us to answer correctly and
that helped, you know, allay any any potential concerns that could have been there were around whether Pure would actually have access to the data. Right? Um Generally speaking we don't even want to have access to our clients data. We certainly don't want anyone else have access either. Um and that was key for us to answer during
this sort of investigation phase um and as part of it, you know, we were able to run through and go yes Pure doesn't have access to the data. There's no you know, possibility for them really to access our client's data so we can take that sort of risk of of um data access off the table and that was really key to be able to do um some of the other key requirements that we had going in.
Just sort of as a baseline requirements for security. Generally speaking where you know, um could the storage do encryption at rest which it could um didn't have good snapshot functionality, which it did um but most of the really, you know, the the key thing for us, once we were able to take the question of is there any risk that we're allowing a third party to have access to client environments or client data?
Once we could take that question off the table um as a managed service provider and say no, you know, still even using this as a service, it's still going to be entirely within our control and no one else is going to have access to any of that that you know, critically sensitive client information. Once we're able to take that off the table, the decision got much easier. Right. And at that point we were focused much more
heavily on, you know, what are the key features from a security perspective that this is going to add um what are the key requirements this is going to meet in order for us to be able to deploy it to meet our client's requirements for for what we do with their data. Um And Pierre definitely meant all of those. So that was pretty good. And there's also been some recent developments um which you know,
James can probably speak to him a little bit more detail or Prakash around things like safe mode, which has allowed us to add um you know, really valuable new functionality for clients um without necessarily needing to, you know, change the cost of service that we're providing or the level of service we're providing anything like that. So um generally speaking on the security front, you know,
there's been a lot of value add over time as well from pure as we have been adding on various bits of features, which has been pretty great to see. Yeah, that's great. I think that kind of leads into the question I had for Prakash in terms of, you know, you hear from the customers every day and you're trying to convince them to move this model what type of security features disappears the service
have and not just security features. What are the things in terms of the delivery and the service itself? Uh do you guys integrate to make sure that your customers can feel, you know, feel secure with going to your offering? Yeah. Look, I think, you know, not to reiterate my best point, but obviously, you know, we don't look at anyone's data so that's,
you know, the first element that you have to take off the table. Um We look at how the system is used, right. Um The second thing is um more of a SAs based concept. The number of way to secure. Wait The # one way to secure our landscape is standardized and reduce your surface real estate.
The more diversity you have in your landscape, the harder your security problem is so having a service here that is standardized. Makes the problem infinitely easier to go ahead and monitor compliance to your security policies. That's kind of the first step, it's not a feature, it's just a principal right then. Kind of the next step is um,
a capability that we've had built into our product for quite some time. Our storage has a stateless controller architecture where we could do software updates with non disruptive, non disruptive lee with no performance impact. So because of that capability, if you think about, you know, I don't know, the Linux operating system right now has 2 - three critical vulnerabilities that
come up a week. So like, you know, the people that say my change window is six months or a year in storage. That's the most asinine thing in the world right now, given what's going on with critical vulnerabilities and systems, you need to be able to get daily software updates.
So, you know, we are in the process of rolling out like today, like we're doing upgrade checks and customer environments to make sure that like these upgrades don't change host mapping or confederation or like, you know, all of these configuration items that customers make, we can detect and ensure that, you know, you're not preventing yourself from upgrading and then,
you know, we're moving to a place where we expect to have a higher velocity of software updates for our vulnerabilities in a customer landscape from vulnerability management. Um, you know, I think Michael mentioned safe mode as well. Um we know ransomware is the thing, we see it, you know, you saw it with the colonial pipeline, you see it happening in our industry right now.
So ensuring that you have protection policies and you do policy based management to ensure. Okay, I don't just have one copy of data, I might have air gapped to that. I have two copies. I physically gapped at what my R T O R P O is. So we've introduced a capability of data protection assessment which allows you to look at the lens from a compliance capability because security is one of those things where
you know, you can, how security you want to be is how much you want to spend At some point. You know what I mean? You could have 20 copies of data where the third one's the 20th copies sitting in a vault underneath the mountain. But you know, like there's a high degree of flex in terms of what you do. But first step of standardization. The second step is like seconds to
remediation based on detection of an issue. Not minutes, not hours, but seconds. And the only way you can do that is by using the data to drive more of a salt. Like, you know, think about it this way your IOS gets updates on your phone, Why shouldn't your storage system?
Right? So that's the approach we're taking to securing the customers landscape. Yeah, they're all those are all amazing points and I definitely agree. I mean it's crazy. I think maybe half of my presentations easier on ransomware has so much fear and so much uncertainty and so many changes that still have
to be made there in terms of how do we do immutable storage and how do we make sure we protect the safety net and how do we make sure the virus more resilient. The challenge of storage managers becomes more and more and to me I think it just I definitely feel for the for the I. T. People who have to do with that and actually leads to my next question for James and James before I asked I asked this question to you.
I mean I think the main thing to me is I can see from your perspective as well because I've been covering storage for Well over 20 years. I went through all the sand training zones, unmasking building all these sort of environments and whatnot. But I mean when when you start coming in with this new service model in terms of staff, was there any organizational resistance to change from your IT staff when they saw this
coming in, what was your what was their take on this when they were introduced to peers and service? Um I. T. Staff generally were very receptive and actually got behind very quickly even though they realized there was going to be a significant amount to do to move off the previous vendor, I think the bells whistles, features and the fact
is we went to a standard model, we just took a very, very standardized approach which obviously makes everyone's life simpler in the long term. Um was really good um like anything when you turn around and effect you tell your client base or your tams to tell the client base that we're migrating their entire storage solution globally. You know, it obviously takes a little while for
some people to start running with you, but again, it was after you start getting some pretty good traction, which we did very early on. So it's the 80 20 rule, right? You can get 80% of all your work done in 20% of the time and then you get 20% of the struggles that take 80% of the time, but once you start winning and start delivering sort of even to some of those sort of slower
clients to migrate some new solutions on the new storage, then you start getting some real pace behind you. Um so yeah, generally speaking of the uptake was really good um Senior management, we're very, very behind it um which obviously makes most things a lot quicker, but um yeah, I think the, certainly the initial pure integration team did a
great job of sort of softening the blow, getting people up to speed and then get then from there on it just sort sort of moved along very, very quickly. Okay, one quick question for you also with the extra time or the lessening of some of the responsibilities. Uh what what do you, what do you think your staff will wind up doing more of in the future? Will there be more focused on some adjacent
things like compliance or will they pick up other skills or what, what's your initial impression there? Um more holiday time? I can't argue with that too. No. So, um, moving away from effectively fighting fires is a great thing.
It gives a lot more time headspace to start thinking about new things. Cool things you can do and in reality we've just reeled off three products in the last year and a half um, under options in general, I think it's growing up over 30 or 40% year on year, so it's great not having to hire more storage guys to just stare at, you know, discs that go read every soft and swap them out.
Um, so yeah, being able to do that. And also as you pointed out, the, with security means patching of storage arrays and stuff becomes more frequent occurrence or has to be. So where we were stuck lumbering around with maybe 23 year old firmware sat on arrays in a previous vendor, we're now obviously pushing much more close to real time and that's
obviously as a direct response to the security landscape changing and also additional features which we see all the time. Um, so unfortunately not as much holidays I would like, but they are certainly happier people and there has been a marked decrease in sort of the ones and similar issues. That's more to do with stability of the
platform rather than the past approach or it starts to approach. Okay, great. Let's get into the next topic. As we talked about, there's a lot of different things you can offload with status and that's one of the survey questions we asked about last quarter as well.
Was one of the things that organizations would be willing to offload to their staff provider. Uh, no surprise that 24 7 support break fix was very high up there. We had 42% who said they would even pay additional to get that type of capability. We had 41% saying security hardening. Again, security comes comes as a big issue. It's definitely something that a lot of organizations want help with.
So I'm not surprised by that as well. The installation deployment another 37%. So a lot of people doing that. I think the one area here that stood out to me in terms of why they didn't want, why some organizations didn't want uh, staff vendors to be doing the Decommissioning of hardware and honestly that winds up coming in some various very specific use cases we talked to,
you know, some intelligence type agencies and very highly secure compliance type agencies where they don't want any vendors coming in there. I think that's why you see a little bit of an outlier there in terms of 90% of organizations didn't want the status vendors being responsible for taking out the old resources and shredding them and whatnot. So that's understandable there.
But in terms of, I don't wanna spend too much time on the slide because I think there's a lot of things behind the numbers of, we should dive a little bit into going back to James James when you start thinking about the operational challenges that you have at options. What was the biggest operations you had operation challenges that you had prior to go into pairs of storage? Pure pure as a service.
Um, well I think there's actually two breaking new ground would probably be my number one pet peeve and number two is turning to senior management and saying, please, can I have some more. Um, I think both of those stand out for me. So breaking ground is a big one, effectively right, sizing and array for an environment that sales guy comes to and talks
to you about is a very difficult and somewhat whimsical thing. So yes, it's going to be the next big thing. I need all of the storage available full time. So what do you do go out and buy a million dollars worth of storage equipment. Probably not the best shout, but being able to actually deliver a relatively workable or commit to a workable size and know that we can scale very rapidly with pure
support up. It makes life an awful lot easier and makes that whole conversation that it kills the conversation in. Almost all use cases number two, which is the I. O. Or you know how much storage um obviously previously you're buying a new shelf, maybe some more controllers you have to come with a business
reason why I. O. Has gone under a certain amount and yeah, it's never a small number. Right? No one ever wants to see that too often appear at the end of their desk a couple 100,000 for you know, some some something that no one really wanted to buy. So yeah, those are the two things I would say.
Yeah, when I was reading the case study that you guys did in regards to your use of I thought one other thing that really stood out to me really, really stood out, you mentioned the impact on reducing storage support tickets. I think it was like You eliminate like 95% of them. So how much of a relief was that not not seeing tons and tons of these tickets coming in every
day. Um Good, less conversations are unnecessary. You know, it's one of those things and absolutely you did speak previously about having multiple storage types and all that stuff sounds great. In reality you're not going to get any problems if you just put people on really fast storage,
um really fast storage is there to say sorry for really badly written applications more often than not. Um the cash that sits on the back end does say sorry really well for, you know, largely written queries, um, poorly set up monitoring tools, you know, there's a lot of stuff that if you just put the right storage in place, it says sorry, you know, where not enough RAM being allocated to a
system, these things you say sorry for, but also gives you the time to turn around to the client and say, hey, by the way, your system is wrong sized, let's approach it this way and get it there. Um so yeah, it's amazing how many um, how many problems don't turn up if you just throw the damn thing.
Okay. And Michael, obviously you're, you're dealing with financial services companies. I work at one. What we're crazy. So when you look at your workloads, how dynamic are the workloads that you have today? Is the resource consumption within for your customers as you look at resource consumption,
Is it fairly predictable? So it is and it isn't. Um and what I mean by that is within any individual client, it tends to be relatively predictable. Right? So a client who has a given use case for storage or for, you know, processing or whatever it is that they're doing is going to
have that, you know, use case and that workload fairly consistently throughout the duration of the you know throughout the duration of of what they're doing with us. Um Obviously over time they may add other workloads and that other things that they're doing. But typically financial services, especially capital markets financial services tends to be when you get into a production
environment relatively consistent and stable, right? You know the number of change windows you have is limited. Um The number of sort of dramatic changes to how you're going to operate uh for any given thing is is fairly limited. So within an individual client um you know clients tend to be relatively consistent and stable in terms of what their workloads are
going to be and you know but between clients that can vary greatly. Right? So we may have some one client that uses you know a large amount of data and it mostly just sits there and they kind of have a little bit of io but not that much. And then you've got another one that will you know dramatically ramped their storage and then drop it back down at the end of the day.
Um Same thing with with I. O. Or or anything else that they're using as part of that workload. Um So from client to client it does tend to vary fairly wildly. Um And really where the challenges because it's you know once you know what a given clients um profile looks like. It's fairly easy to support. Right?
The challenge really comes um when you're first on boarding that client cause you have no idea what their work looks going to look like. Um And frequently you know back something I mentioned earlier, they tend not to either because frequently we're we're speaking with someone on the business side who has a business need that they need solved and they may not um understand at a detailed, intricate level exactly how you know,
storage there storage performs or exactly how their their um CPU performs or anything like that. Um So uh one of the really valuable things about using um you know pure as a service as a result of this is so one as we mentioned previously, right, we don't have to get it exactly right when we're first on boarding a customer um we can get it a bit wrong and then correct over time um without any impact
to uh you know, timelines for delivering the project without any impact to the customer in terms of their performance and and the service of the experience. Um So that's great. And then you know, the other one to to the point that James mentioned is we can guarantee a consistent performance profile for what they're giving them for what we're giving them.
So we don't end up in a place where you know, we brought a client on board and all of a sudden, oh no, you know, we need another three different devices do three different things because we need more data and we need more performance and we need X Y Z. Um and our cost has suddenly ballooned and we've completely misjudged, you know what the cost of providing the service will be all that sort of goes out the window um
which is great. So so uh we're much better able to service um you know, sort of the unknown footprint that client will have when they come on board um and then once they're on board of course, you know, in a fairly short period of time we learned what their, what their usage profile is going to look like and we can ensure that we're um servicing that appropriately. Okay, great stuff James from your perspective,
when you look at your experience with stars so far, do you think stats will replace all of your existing storage systems in the future? I see it probably not including tape, but yes, I would probably, I would say so. The storage anyway. Okay, because you got to come up with tape as a
service now. Yeah, I think that's not on our roadmap anytime soon, but you know, um it's surprising how many times we've gotten arrested for it. Yeah, it comes back. It's never dead. Just just get longer. Yeah, that's about it to you for you again. How is pure helping customers transition
appears the service, you know, from traditional storage systems. Um Well uh, there's kind of two or three elements. One is uh, largely we're seeing different segments of customers at different places in the maturity of their journey. Right? The first places, you know, we're working with
a lot of MSP s specifically like options to ensure that that that is the best way to go and consume as a service offerings because obviously you need more than just storage. And the MSP model is something that we see a lot of our peers a service business in and we encourage a lot of customers to look at, you know, the MSP approach to delivery.
So that's kind of, you know, pretty big on our agenda. The second thing that we do is a lot of education in terms of what the description differences between a subscription and the service because you know, a lot of people boil this down to Capex versus apex. But like that's a subscription in a financial offering.
A service is about labor reduction and sls so, you know, there's a education needed in the market around. Okay, well what is the value of usage based payments versus what is the value of saying? I'm signing up to guarantee that I'm getting this level of I ops and if I don't, it's the vendor's responsibility and the vendors risk for solving that problem.
Right? So in a service model that means that it's not just about, I bought the hardware versus the vendor bought the hardware. It's, I am buying a performance obligation and the vendor may need to ship multiple boxes to meet that performance obligation or one that's up to the vendor to decide how the vendor is delivering. So that approach like, so you know,
there's a difference between just, hey, I bought a box and I'm paying based on utilization. I'm delivering the service and the market maturity and understanding of that I think requires a lot of evolution. And that's where we're spending a lot of time and that's why we're having this conversation today. Mm hmm.
Okay, great stuff. Finally, we're getting to the last topic here. Last topic. We're going to talk about buffer capacity and how much is enough. And I'm seeing a lot of different perspectives based on the different types of companies. We talked to some being a lot more conservative than others. But when we did our most recent survey,
what we saw was on average, we had about, when we asked organizations that they'd be comfortable with. On average, we had about 74% though, we did see a fair number that would deal with a lot less than that and some outliers that wanted a lot more. And I think beyond the numbers. What was more interesting was the conversations we had with customers when I talk about the
conservative customers. Health care was a big one. And for some of them, a lot of them because they're there, how that mindset, how important it is and How compliance sensitive is and how they can't lose any data for those organizations. They wanted to have big buffers, they wanted to have 70% or more depending on the volatile their environment and they also
didn't want to lose, they didn't want to be too close to the margins because that conservative nature because they didn't want to have problems because of that. In other situations though like this financial services customer we talked to, they were fine with a smaller buffer capacity because they realized they understood what their environment was going to be like and they knew that you know,
even with the catholic system you might have a ton of capacity when you first start out but when you're getting to year 345 of that uh that system that's when you do get normally to more of a 10 to 50% capacity left. So they're they're already used to running with less buffer capacity there and from there regards they they felt that they didn't need to have a lot of buffer capacity to be able to deal with their workloads without having to
worry about spike ups blowing out the amount of on premises capacity had there. I'm starting off with Prakash Prakash, how much buffer capacity is pure provide and what steps even more important, what steps does pure take to make sure that customers don't run a run out of that buffer capacity Well so you know that that's one of the most difficult questions? The answer is I'm going to give you the consulting answer is it depends.
The only way to solve this problem is you have to take a look at all the variables that go into this one is what is your supply chain, volatility and delivery? S. L. A. Of what it takes to get gear from one point of the planet to the others. And last year, even amidst covid, we were able to get gear On average
delivery under four days anywhere. So you know, we were able to outrun even the supply chain disruptions last year and maintain that and we build that into our planning the second piece is what is the volatility of the customer Because if there's rapid growth, if there's not, you need to maintain a higher buffer.
So like the dumb answer I could give is 20% right. We, you know, that's what we write in our documentation but in all practice you can't offer a service without actually monitoring the utilization in building a data science model that says, okay, here are the variables that go into how much you need to go ahead and meet your service level obligations of ensuring that the buffer capacity is there based on that
workload in that environment. Right? So and that is highly tailored and the good news is we, you know monitor the environments to go ahead and set that on. What makes sense in every environment. Okay, great. So Michael from your perspective as a customer,
are you confident in the amount of buffer capacity you have and have you ever has options ever gone to a point where you got close to running out of capacity? Um definitely confident the amount of buffer capacity we have, we did get close to running out of capacity once that we've never run out. Um and that was entirely on us. We on boarded a very large client very rapidly
um without considering necessarily what the impact would be, um and and got to within, you know, relatively close to the edge of our buffer um but pure came, you know, showed up at the site the next day with a pile of additional data nodes and stuck right in and we were right back to where we've been. So um it was a little bit of poor planning on our side on our side in terms of very
rapidly or more rapidly than we initially wanted to onboarding this customer, but you're had our back forth and happy data or had the capacity on site within a day, which was great. Um so that's, that's the only time we've ever come close otherwise, um it's it's very well managed. Yeah, that's the amazing thing is, I mean the reason why I bring up the buffer capacity is
that's really a safety blanket type issue and I think the faster we resolve that thing and get that out of out of the question, I think the easier it'll be for organizations to move over to stas and these type of more flexible models because again, a lot of organizations don't want to just buy tons of tons of storage and have sitting around, it's also used. And that actually gets the next question with
James, I mean James now that you have this more flexible model, how are you planning to reinvest some of the money that you're saving from stas you to improve your own environment? Um Sorry? Um I think the most I would say probably the biggest the biggest cost saving is probably in people's time that I see anyway.
Um ultimately we don't run a piano inside the inside of my section, the engineering section as well. So I think the biggest thing that we're seeing is um saving in natural human time or human input as it were. Um I think what it has actually allowed us to do is reinvest that in new products, new deliveries, getting faster time to business and all that kind of good stuff.
Okay. I think in the case that you guys mentioned other initiatives you're doing there as well in terms of some of the automation efforts to have those been going very well. Um well it's a mixture of automation and good architecture, I think I'm gonna say that automation fills the cracks where architecture doesn't sit, it shouldn't be the answer to every single thing because much like
an air fix kit, if you're when you glue it together and the whole thing is full of automation, it just means the kids going to look terrible. Um So things like we are a We are a very sizable user of the technological Vivo, which pure and VM one obviously pushed into market. We have been I think the number three global or something silly like that um user um that we
find reduces the amount of touch time automation and similar that is required just to good integration and good auction at the end of the day. Um Just the the same same can be said for a few few things where automation, I think some of the stronger parts of the automation, have you seen the best part of the information on physical services? Probably has really made the biggest change
there. So things like automated and physical server deployment with Pure is certainly a very nice thing to do rather than our previous SAM vendor. Yeah, I think that's all great stuff. It also harkens back to some of the things we talked about earlier in terms of yeah, not only, hey, let's offload staff so they don't all burn out and quit. Let's also be more proactive and it definitely
feels like that's the benefit as well. I mean uh you know, as much as I enjoy doing storage, learning about storage architectures, I don't think anybody really wants to do farm work changes on the weekend? I think we want to get past that and you know have you know, it'd be nice to be able to offload some of these things that are important but not
necessary, strategic and going forward. I think that is one of these key things. If there's any takeaways you get from this really, it's what can you do with that opportunity? What's the opportunity you have when you can free yourself up from these different types of burdens And with that we're going to open things up here so that the audience can ask some questions here as well.
Um Tricia do you want to ask, tell the audience how to get in questions? Yes, thank you, Henry. Um as a reminder, if you have a question, you can put it in the question box on your screen, we have a few already. So let's get started. Here are first question um yes, my organization is considering a change to an on demand consumption model for storage,
but some in my management team think it's too expensive in the long one. But what are your thoughts on the costs and how to justify it? That's for Prakash. Yeah. Look, I'll give that that wanna go largely, there's uh data is the answer right? I think what you need to do is you just have to build the model and you know, what does the model entail when you're actually going ahead
and saying that? Well, first, what is your existing utilization of your existing assets, right? Um, to what is the labor associated with, you know, all the operational overhead with, with running automating, scripting, you know, and then three is risk, right? One is cost, the other is risk, What is the risk of not standardizing in the world that
we're in today? The risk of not standardizing is huge. So all of those things going to, you know, build building the model and over time you want something that gets better over time, not something that ages over time. So with a service, it's built on a foundation that gets better over time with an asset. It's, you know, the fine wine doesn't get
better with age. Right? So I think you, you know, need to understand the economics of building building in, you know that over time and you know, almost always when we have that discussion with customers, it always works out positively great stuff. Um, actually, just to jump in on that briefly as well.
Um, you know, when we look at that, what the cost of storage is, service has been to us versus the previous model. Um, it is actually cheaper in many ways. Right. If you look at legacy storage models, um frequently, uh, legacy storage model will have, you pay a huge amount of money up front for a piece of kit and then you pay a lot of money for support and you get three years of
support and once your three years of support is over all of a sudden the renewal of that support contract has blown completely out of control. Um, and it's almost as though you're purchasing a new kid again anyway, which is not what most people model for and, and you know, to, to comments that we made earlier around, you know,
how do you write size the size of the environment that you're building? That's really difficult when you're building, you know, when you have to purchase individual devices that you may or may not use um, to their full capacity. So when we actually, you know, took a look at it from a holistic cost perspective,
especially when you consider the costs of, of people hours that are reduced from your own infrastructure team, no longer needing to, to support the infrastructure and the storage itself. Um, your costs are actually your cost profile overall is much better when you switch to a service based model, you know, it's definitely a great point because again, that Capex model,
that's the painful part when you get in and when you get out, when you get out you have to that 3 to 5 year cycle where I don't think anybody wants to deal with unwrapping and taking out tons of equipment and all the other things that I think that to me is that big benefit for stas is helping us reduce some of that lifecycle burden, especially because it's so painful getting in and so painful getting out and upgrading to
these new systems going forward. Uh Patricia, next question, Yes, this will be our final question. Uh can you give us a little more detail around the operational impact of using Pure as a service on a day to day basis? How would things be different if you have not chosen a stash product to address your storage issues?
I think that would be James. R Mika, um I think that would be James, although it looks like um James, are you still there? Yeah, so am I, I have a feeling so I'm just knocked over the wireless router. Um No, can you repeat the question? Only call the first film? Mhm Oh no problem.
Can you give us a little more detail around the operational impact of using pure as a service on a day to day basis? How would things be different if you have not chosen a stash product to address your storage issues? Sure, absolutely. Um day to day basis because we obviously have a fairly sizable fleet um normally we have a catch up with pure
storage management monthly, so that is kind of the day to day touch touch point. We also obviously receive um statistics from the user units to make sure that we are kept abreast of consumption. Um so and then ultimately as I think as previously mentioned it when the system is requiring an upgrade or more um furthermore buyouts,
throughput or, or so and so forth. It's just a matter of scheduling time with pure actually we normally get reached out to and say, hey, time to upgrade um starting in time for them to get site and do the extensions really. Um That's obviously somewhat different from say running out of IOS apps on the traditional unit where you would have to go and talk very kindly to your CFO um about them, opening the purse um and then you have to do the
dance with the standard sales guy um to get a number for them to then arrange something in several months time. So there's a, there's quite a bit of difference that we've, we've noticed it's much easier process day today. Also the tooling that obviously has to have been built out for cure to support this fleet of products works incredibly well from a consumption point of view,
an endpoint consumption point of view. Um So things like Pure One is, is a very, very good tool, I've got lots of time for it. Um So yeah, no, I think day to day operations within the pure ecosystem are very, very pleasant. Um Including upgrades actually,
I would probably say talk about upgrades and extensions, so the ability to move from say a basic 50 all the way up to the next 90 are three with it without making any huge changes. It's a seamless upgrade. Software upgrades are seamless cage upgrades are seamless and, and certainly when we're understanding that this is actually a long term partnership where
we are paying for a service, which is the performance level. Um, and we don't really say, we're not saying this is what the chassis is, that it's going to be delivered in, but we're saying this is the storage, you know, we are expecting this S. L. A. Um, it doesn't mean that our understanding of when, when we will need to upgrade, when we need to make changes,
when we need to make some big old rip and replace is that that changes quite significantly. So as long as the slots are being met, we don't mind. That piece of hardware has been around for x period of time because as long as the escalation because like as long as it's supported, that's, that works for us, we're not going to get to five years and arbitrarily
stripped something out and cause a sizable headache if it doesn't need it at the same time. It's also been proven that doing upgrades between units is seamless. So even at the, at the end, I'll tell you what, we've had some really fantastic some upgrades previously between different vendors and it is not a slick process in any way, shape or form. Um, again, that also goes back to,
I suppose how pure view. Software updates. Software updates are meant to be done. Software updates in legacy sound world everyone is scared of doing which means they're not going to be easy to upgrade at the end of life. It's a circle right? That's exactly where you get to.
Yeah we want to have that speed but I mean you don't lose your job if you update your iphone a ready upgrade goes bad. It could be a very bad day all around definitely Tricia. Alright well thank you everyone. Um that concludes our roundtable for today. Thank you. Henry Prakash James and Michael for a great
discussion. As a reminder the on demand version of this event will be available shortly on behalf of pure storage options I. T. And 451 research. Thank you for attending and have a great day. Thank you. Thank you very much.
Thanks everybody.

Henry Baltazar

Research Director, 451 Research

Prakash Darji

Vice President & General Manager, Digital Experience Business Unit, Pure Storage

James Laming

Global Head of Infrastructure, Options Technology

Micah Kroeze

Senior Vice President, Product Management, Options Technology

Using STaaS to Navigate the Landmines and Complexities of Storage

Engage in a thought leadership discussion with Pure Storage, 451 Research, and Options Technology on the future of storage. Gain perspective on where the storage industry is rapidly headed, and learn new ways to address the complexities of traditional storage with one of the top Storage as-a-Service (STaaS) solutions in the market. Hear real-life examples directly from a customer on how they are applying STaaS to uncomplicate their datacenters.

Reduce Risk with Evergreen//One™

Downtime is not an option. Look to Pure to ensure you’re always ready to meet capacity demands.

Explore Evergreen//One
Continue Watching
We hope you found this preview valuable. To continue watching this video please provide your information below.
This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.
Meet with an Expert

Let’s talk. Book a 1:1 meeting with one of our experts to discuss your specific needs.

Questions, Comments?

Have a question or comment about Pure products or certifications?  We’re here to help.

Schedule a Demo

Schedule a live demo and see for yourself how Pure can help transform your data into powerful outcomes. 

Call Sales: 800-976-6494


Pure Storage HQ

650 Castro St #400

Mountain View, CA 94041

800-379-7873 (general info)

Your Browser Is No Longer Supported!

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