00:00
Mhm Excellent. So welcome everybody to the PNC's digital financial services journey here. Um and I'm really, really excited because this was an opportunity to really get to know a customer who has been able to do some really amazing things with flash back and really transform their environment. So it's not every day we get to talk about how
00:22
flash stack and pure is changing the landscape of our customers and how we're making things so much better. And so I'm really, really excited to welcome scott to the stage and tell us something about yourself. Hi, my name is scott guardsmen, I'm the lead solution architect at PNC Bank. Um we are National Bank as everyone knows and we've been going through a digital
00:43
transformation lately which involves a heavy amount of pure storage in it. Very, very good. My name is Eugene McGrath. I'm the staff field solutions architect here are pure for everything flashback for America's and so I'm really, really excited about this presentation and scott. The one thing I wanted to say is that it's it's very, very difficult to take so many complex
01:03
components to make them so simple and that's something you've done. So it's really kudos to you and your team and what I wanted to do was kind of start about Start and just talk a little bit about the methodology like what was your philosophy and this in this day? zero. Like when your architect in this, what was really important about what you, what you built.
01:20
Our main philosophy was was that did you get into these very, very large environments, you need to simplify things. You cannot have, you know, the lead times that we have, we can't have, you know, complexity for complexity sake and you know, you have to control the cost and control, you know, density because you know,
01:39
every inch of that real estate in the data center, you know, cost you. So those are our philosophy is to try and keep with and make sure that we we we stuck with and I think with our design, we we kind of hit that on the nail. Absolutely, Absolutely. And what I love is that is that the hardware is based on what's best for the organization,
01:57
correct, correct. That is one of our key things is that and we're driving at home a lot more now is, you know, everything that we need to do within the bank has to be what's best for the bank, not what's best for, you know, a particularly one single application or one single individual, but it has to be what's best for the bank. And initially it may not seem that way,
02:19
but it's a bigger long term picture that we're trying to keep with. Absolutely. That's something we preach with flashback all day long is that it's really about what's best for the customer. We look at other solutions out there where you're oftentimes restricted, where you're oftentimes restricted by what that what that product set has or what that can do
02:36
for you. And so building something that's so Greenfield and being able to accomplish everything you have, is really, really fantastic. It has been, it's been great. Let's let's talk a little bit about about these four things, a little bit more. Talk to me about the catalog.
02:48
What is the catalog for you guys? So one of the problems that we had in the bank was that we um we seem to have a lot of disparaging systems that were um not really purchased all at the same time, but you know, and and we had a lot of things that were maintenance wise and took a longer time to do things. So one of the things that we were like,
03:11
we need to build a catalog that is basically a t shirt site, you need small, you need medium, you need large, you know, here you go and out the door, it goes and be able to keep that uh philosophy going forward in all of the different projects that are coming in. So that's kind of how, how this came to be.
03:28
We've reduced, we're reducing costs surrounding it. Um and everything else. Absolutely. And the thing that I love about this, this particular slide is when you talk about maximizing floor space, I think it's often overlooked how important that your your power and cooling and what that how that contributes to the cost of it does. And when you start expanding into multiple
03:48
halls inside of environment becomes even more complex for you to keep that value prop of having the density that you need. Um and that was one of the, you know, as it said in there, one of our methods and our philosophies is how can we drive down the cost and that's a big cost there. And so let's talk about your objectives for
04:07
your data center modernization. So, so one of the big things was we kind of took a philosophy of, you know, any, any resource, anywhere, anytime and we wanted that to be like the main focus of where we were going. So we needed to make it so that any network that we had needed to be accessible in that data center from any host so that any guests can access it.
04:33
And that was our first thing then we said, well how do we take that again now and make it so that it applies to storage and you know, we, we went and figured that problem out at that point and then we had one last thing that it was like, okay, how do we make it so that the bank survives regardless of what happens in that room. And how do we make it so that maintenance does not become a burdensome task and instead
04:57
becomes a simplified, easy to due process and that's, that was our driving factors here. Um we had had an agreement with With my VPs that hey, no time will maintenance take out more than 20% of the entire environment. How do you do that? Well, we got it Based on some of our calculations, we're gonna be down close to 10% of the environment when we do maintenance
05:20
because of this design that we came up with. Um we also made it so that in the event of a problem, let's say it's a power outage or a cooling issue, there's all kinds of stuff that can happen. It does not impact us in the bank. We can keep on going. We could, you know,
05:39
anything that happens. We needed to survive because we have financials or you know, and and clients are relying on us to be up all the time. And what I love about what I love about what you've done here is that this anywhere for all is so important because it's something I preach every single day which is get away from silos, get away from all these complex environments and you create all this complexity by silo in
06:02
all this data, all over your network and by by giving you the flexibility of being able to put that data anywhere That really is a modern approach. That's pretty awesome. And we took it, you know, we took it from a, a concept that that was prior to my coming on board was a share nothing kind of situations that you could control it.
06:21
It's great. It's a great philosophy. It makes troubleshooting much easier and controllable, but you lose flexibility completely. So how do you take that and change that? So we took it and turned it on its head and we made it that you know to share everything but we also simplified it so that troubleshooting becomes a simple task at the same time.
06:42
Absolutely fantastic. And so if we look at day one let's talk about what improvements you've seen within your environment and the impact that those have had. So this was this was a great one that we we talked about. So one of the things I did when we were trying to figure out hey how dense can I make this? Right And I did it was a mathematical exercise
07:02
here really. So I took all of our systems that we had entire the entire bank And I figured out that if I took a 48 U. rack and I jam packed it nopd us no concern about power or thermal or anything else. How much floor space am I using up? So I figured out a rack, how big it is and everything else.
07:20
We came out with 34 ft by 35 ft and that's literally bumping every rack against everything. And then we took our new design and put it in place. We're gonna we're figuring it to be a nine by 10 area. That's what was 86% if I remember correctly, 86% savings. So not just is it a savings in just the floor space, but that means I don't have the multiple
07:44
halls. My cooling is better, my heat, my my thermals are better. I can spread out equipment, um everything doesn't have to be on top of each other. Um there's a whole, you know, mirror thing. Human capital, which is, which is our most expensive thing goes down because I have less
08:01
servers to do, but more more capacity at the same time. Absolutely. So if you're taking pictures, this is definitely the slide to take a picture. You can see it's even, you know, a money shot for it. Absolutely. And and this is something we're preaching every single day,
08:15
which is because of the small footprint of flash stack and and flash and flash blade. Um it enables customers to really reduce their footprint. So if you think of the environmental impact and we heard charlie talk about environmental impact today, we think about the environment, environmental impact. Just think of how much power and cooling you're saving just by by going to this model.
08:36
It's really fantastic. So moving on, talk to us about the map here, the example that we used for this was a Cisco solution here and we looked at, you know, how dense can I make it uh when we did this, this is about a year and a half ago that we did these original calculations we compared to hey are currently we're using uh a B series blade. We went over to a C series chassis.
09:05
How dense can we make this? And we figured out using the same footprint, I'd have to have three additional racks here of servers in order to accomplish the same thing that we're currently doing right, that I can fit into one rack. So if you look at that, just at face value, that's a, that's a 3 to 1 ratio that I'm removing from the environment
09:29
and that's based on higher end Cpus and everything else. Um, some of the equipment that we have, you know, you have a depreciation time, but some of them run the gamut from, you know, 24 cores to 56 cores. So depending on how you're calculating it, you could have a 6 to 17 to one ratio that you're gonna be removing from the floor. It's a pretty impressive calculation when you
09:52
look at it. Yeah, it's all about consolidation. And so tell me about this, what is this, what is this speaking to? So this was another one where we kind of took this exercise and we said, hey, based on our current compute and everything else that we're doing, how, what can we reduce, how much of a reduction would it be?
10:11
So we took and looked at some of our older equipment that was going end of life. And we said, how would it go if we went and put it into this new design that we have with these higher density machines And I could take out all of these racks and all of this storage Basically consolidated town into three racks once again assuming no heat and power related issues. But I can put it all into three racks with the
10:39
pure storage X rays that we use and make it fly at that point. And the same thing for you know, all of my other components, my V block which is the end of life, my my and my other systems. When you when you talk about impact, you just look at the impact that that this has on the environment. And when you architect for a modern next
11:01
generation data center this is what you wanted to look like. This is this is exactly what that next generation data center it's about giving the that simplicity that flexibility and reducing all that unnecessary floor space, unnecessary power cooling, getting rid of all that and really being more efficient with everything you have. And so let's talk about the results.
11:20
Just some numbers, some facts and so forth. Oops. So this one didn't build up. Yeah that's the one we looked at already. So this one, so let's let's talk about how this has impacted your environment. Let's talk about your sls first of all.
11:40
So with this design R. S. L. A. Um for maintenance for example goes to almost zero because we when we do this based on the design which we're going to get to in a second, it'll make sense. We can have a zero outage time with our system because we can maintenance mode, every single server in a pod inside the system. And from there there's no io going to it,
12:11
there's no nothing. So as I put it, technically, if you wanted to remove the entire server system and dust underneath the racks, you can go and do that because it has no impact on anything else in the environment there. Absolutely. And there's that space reduction that you're talking about. My space reduction based on what we believe is
12:29
going, you know, in the end, after everything is said and done um that this would save us. Absolutely. And so we talked about how that reduction actually affects the company from a financial standpoint and that number is pretty impressive. You have that. Yeah. And and some of those are, you know, it's based around the fact that um four spaces saved,
12:51
not needing additional halls as you continue to build out over time, you know, additional heat cooling inside the facility maintenance times, uh to do things. Um and being able to automate more things inside the environment, which is a key thing that we're doing at the bank. And you ultimately chose Pure because of these two areas.
13:12
Tell us about your experience. So we right now um based on your guys calculations to just so, you know, out of pure one, we're at uh we're at 100% up time in 3/3 year period. We've had zero outages related to a device having a failure or an issue or anything else. That's not to say that we don't have that the device, you know,
13:35
memory goes out but we have zero impact from, we don't even if we don't get a call from them, we sometimes don't even notice it immediately. So 100% of time and our our reduction that we're getting is incredible inside there. As long as you're not compressing your data server side, it just it just crushes that data down into nothing.
13:56
Absolutely. And this is where I get to put my little shameless pitch. I notice here that you're exploring court works with the savings that you save from here for those container workloads. I actually wanted to let you know that this is brand new. Just yesterday, there was a little blog post about this.
14:10
If you want to know more Cisco live next week that we were announcing partnering with Cisco to with I K s and an inter site to make it a lot easier to manage that that cloud right infrastructure. Right, So that's a little that's a little tidbit. Look more. Look for more at Cisco live and we're trying to find ways to help you deploy your containers
14:29
and cloud ready. So so thank you so much for for for participating and I'm really, really impressed with what you've built and it was super exciting to have this opportunity. And uh, and of course, flash back for the win. Right? That's right. Thank you, everybody. Awesome.
