WEBVTT 1 00:00:03.279 --> 00:00:08.599 You're on board. Caseyaas Inland Express. Caseyaa home, Linda ten 2 00:00:08.640 --> 00:00:24.839 fifty am the station that needs no This here behind. 3 00:00:17.960 --> 00:00:21.559 The information economy has a ride. The world is teeming 4 00:00:21.600 --> 00:00:25.800 with innovation as new business models reinvent every industry industry. 5 00:00:25.879 --> 00:00:30.199 Inside Analysis is your source of information and insight about 6 00:00:30.239 --> 00:00:32.840 how to make the most of this exciting new Eric 7 00:00:33.039 --> 00:00:35.520 learn more at inside analysis dot. 8 00:00:35.399 --> 00:00:37.200 Com, Inside Analysis dot com. 9 00:00:37.200 --> 00:00:40.119 And now here's your host, Eric Kavanaugh. 10 00:00:44.759 --> 00:00:47.399 All Right, ladies and gentlemen, Hello and welcome back once 11 00:00:47.439 --> 00:00:49.799 again to the only coast to coast radio show that's 12 00:00:49.799 --> 00:00:52.840 all about the information economy. It's time for Inside Analysis, 13 00:00:53.119 --> 00:00:55.039 and your host Eric Kavanaugh is here in folks. I 14 00:00:55.039 --> 00:00:57.600 am super psited to be talking to a couple of 15 00:00:57.600 --> 00:01:01.240 experts and in the hardware, and they also know about 16 00:01:01.280 --> 00:01:04.079 software and data and all kinds of technology. And we're 17 00:01:04.079 --> 00:01:06.000 going to talk about a survey. But I've got two 18 00:01:06.040 --> 00:01:09.799 experts from rack Space. Yes, indeed, rack Space, they've been 19 00:01:09.799 --> 00:01:13.560 a powerhouse in the industry for decades, a very serious player, 20 00:01:13.959 --> 00:01:15.959 and they know a lot about what's going on in 21 00:01:16.000 --> 00:01:19.400 this world. Of course, a lot of AI conversations these days. 22 00:01:19.519 --> 00:01:22.359 But Hybrid, we're going to talk about hybrid and what 23 00:01:22.439 --> 00:01:24.599 that actually means. We're going to talk about these experts 24 00:01:24.640 --> 00:01:27.040 and find out and it was a survey that's been done, 25 00:01:27.239 --> 00:01:30.599 so it leaders increasing investment in multi and hybrid cloud 26 00:01:30.640 --> 00:01:35.359 strategies to future proof their operations. That's what the survey reveals. 27 00:01:35.359 --> 00:01:37.799 So let's go ahead and dive right in. So tank 28 00:01:37.879 --> 00:01:41.159 Shield tell us what what did you find in this survey? 29 00:01:41.200 --> 00:01:43.920 What surprised you, what didn't surprise you? What do you 30 00:01:43.959 --> 00:01:44.640 think is interesting? 31 00:01:45.519 --> 00:01:47.719 So thanks er, thanks for having me. 32 00:01:48.000 --> 00:01:50.200 You know, what I found quite interesting, especially in the 33 00:01:50.319 --> 00:01:53.079 in the from a context of analytics and bi perspective 34 00:01:53.120 --> 00:01:55.920 here is that, you know, the proportion of data that's 35 00:01:55.959 --> 00:01:58.680 gone to the cloud has really skyrocketed. You know, if 36 00:01:58.719 --> 00:02:01.040 you look in this survey, it's so that more than 37 00:02:01.079 --> 00:02:03.519 fifty percent of the data is actually now sitting in 38 00:02:03.560 --> 00:02:04.319 the public cloud. 39 00:02:04.560 --> 00:02:05.840 It is across many. 40 00:02:05.640 --> 00:02:08.800 Different cloud providers, it's across hybrid providers, and you know, 41 00:02:08.840 --> 00:02:12.479 we're expecting to see seven percent more going up there. 42 00:02:12.639 --> 00:02:15.039 And the other interesting thing that I saw is that, 43 00:02:15.080 --> 00:02:18.439 you know, the adoption of SAS has really skyrocketed. And 44 00:02:18.479 --> 00:02:21.080 what that really means is that now, more so than ever, 45 00:02:21.639 --> 00:02:25.879 we have data sitting obviously across different clouds insiders, but 46 00:02:25.960 --> 00:02:29.800 also within these different SaaS applications that are running on 47 00:02:29.840 --> 00:02:32.919 their own or customers clouds, which means bringing the data together, 48 00:02:33.120 --> 00:02:35.719 whether it's for the purposes of analytics, all the further 49 00:02:35.759 --> 00:02:39.719 purposes of AI has become an increasingly challenging. 50 00:02:41.199 --> 00:02:42.039 Requirement. Now. 51 00:02:43.240 --> 00:02:45.840 Yeah, and we've also got been like Karra on the 52 00:02:45.919 --> 00:02:48.840 call here and I'll throw this over to you. I remember, 53 00:02:48.879 --> 00:02:50.879 I've been covering this stuff for a long time. So 54 00:02:51.400 --> 00:02:54.479 DM radio turns seventeen years old next month, so I 55 00:02:54.520 --> 00:02:57.240 would do it this a while. And I thought the 56 00:02:57.280 --> 00:03:00.599 cloud would take off twenty years ago, and it really 57 00:03:00.639 --> 00:03:03.080 took off like ten years ago. So for some reason, 58 00:03:03.080 --> 00:03:05.840 it just took a while. I credit actually Second Adella 59 00:03:06.000 --> 00:03:09.879 of Microsoft for pivoting his whole enterprise to cloud. And 60 00:03:09.919 --> 00:03:11.439 I really think that woke up a lot of people 61 00:03:11.479 --> 00:03:13.199 who are like, geez, I guess we better get on this. 62 00:03:13.680 --> 00:03:16.439 But I remember that in our conversations on these shows, 63 00:03:16.479 --> 00:03:19.560 we went from talking hybrid cloud to multi cloud in 64 00:03:19.599 --> 00:03:22.439 about one week because people figured out it's not just 65 00:03:22.479 --> 00:03:24.439 going to be on prem in one cloud, it's on 66 00:03:24.520 --> 00:03:29.000 prem in multiple clouds, which greatly increases the complexity, right. 67 00:03:29.879 --> 00:03:35.599 Oh absolutely. In our SAVE survey, what really came out 68 00:03:35.840 --> 00:03:38.879 was the notion that it's turning into a much more 69 00:03:38.879 --> 00:03:43.080 complex world and that it much like when you're preparing 70 00:03:43.080 --> 00:03:46.879 a meal for your family. If you just buy one 71 00:03:46.960 --> 00:03:50.199 thing all the time, that's not going to satisfy things. 72 00:03:50.479 --> 00:03:52.159 Right, that's funny. 73 00:03:52.560 --> 00:03:56.759 And so when you think about it, organizations providing services 74 00:03:56.800 --> 00:03:59.520 to their businesses, businesses that have different flavors of things 75 00:03:59.520 --> 00:04:02.560 that need to be done, and those different flavors of 76 00:04:02.639 --> 00:04:06.080 things need to be done, you have multiple options now. 77 00:04:06.520 --> 00:04:08.719 And so those options in terms of things to be 78 00:04:08.759 --> 00:04:11.680 done are what we call those workloads. And so when 79 00:04:11.680 --> 00:04:14.039 we think about workloads, we have to think about what 80 00:04:14.199 --> 00:04:17.439 is a workload and where should it be? Much like 81 00:04:17.759 --> 00:04:20.839 you know, do you need something fast for your family 82 00:04:21.120 --> 00:04:23.279 or do you need a big sit down meal. They're 83 00:04:23.319 --> 00:04:27.360 different scenarios, and each of those different scenarios says there's 84 00:04:27.399 --> 00:04:31.079 different options. And what's happened is cloud everyone got. You know, 85 00:04:31.360 --> 00:04:36.319 hyperclouds are wonderful, However one size, they're not fit all. 86 00:04:36.959 --> 00:04:39.879 You know, if you're looking at workloads that have to 87 00:04:39.879 --> 00:04:45.680 be secure, workloads that that are that are that doesn't 88 00:04:45.720 --> 00:04:48.160 need bursting and things like that. Well, you know apps, 89 00:04:48.240 --> 00:04:50.600 you just need to put them on private cloud. Right. 90 00:04:50.639 --> 00:04:53.279 If you need native services that are in hyper skills, 91 00:04:53.439 --> 00:04:57.319 put in the hyper skill. The point is all this 92 00:04:57.600 --> 00:05:00.600 is saying is that to for IT or organizations to 93 00:05:00.600 --> 00:05:04.160 provide the great services to their businesses in terms of capabilities, 94 00:05:04.560 --> 00:05:09.519 you have to give options. And hybrid is a is 95 00:05:09.959 --> 00:05:12.120 a way to think about It's that one, it is 96 00:05:12.160 --> 00:05:14.839 both public and private, and you've got you've got to 97 00:05:14.879 --> 00:05:18.759 be able to do both or else you're not optimizing 98 00:05:19.319 --> 00:05:21.759 the services you're providing to your businesses. That makes sense. 99 00:05:22.040 --> 00:05:23.759 Yeah, no, it makes complete sense. And you know, I'll 100 00:05:23.759 --> 00:05:25.399 throw another question over at you and then we'll get 101 00:05:25.680 --> 00:05:28.920 tank back in this. But you know, observability I think 102 00:05:29.040 --> 00:05:31.560 is a huge play in what we're seeing right now 103 00:05:31.600 --> 00:05:35.800 and finops, which of course is observability because like five 104 00:05:35.879 --> 00:05:38.360 six years ago, you know, it was hard to tell 105 00:05:38.399 --> 00:05:41.759 what the stuff really costs. Now we have very granular 106 00:05:42.160 --> 00:05:45.680 reporting on what the cost is and that allows you 107 00:05:45.720 --> 00:05:49.800 as an IT organization to do what you just suggested, 108 00:05:49.839 --> 00:05:52.040 which is optimize we're going to do this workload in 109 00:05:52.040 --> 00:05:54.040 this cloud. We're going to do that workload on prem 110 00:05:54.240 --> 00:05:56.240 and when you have the capacity to do that. If 111 00:05:56.240 --> 00:06:01.000 you've embraced cloud nativity, right, if you're cloud native even 112 00:06:01.000 --> 00:06:02.920 in your on prem environment, if you've done something to 113 00:06:03.040 --> 00:06:06.360 make it look more cloud native, you can really move 114 00:06:06.360 --> 00:06:11.000 those workloads around. And that gives tremendous agility to the organization, right. 115 00:06:11.600 --> 00:06:15.360 Absolutely, And that's a key driver as technology leaders are 116 00:06:15.360 --> 00:06:18.399 looking for how do I future proof? You know anything 117 00:06:18.480 --> 00:06:21.279 about that, Eric, As you think about your future proofon 118 00:06:21.519 --> 00:06:23.199 you have to think about how do I create the 119 00:06:23.240 --> 00:06:26.959 most flexibility with things that I don't even know are 120 00:06:27.079 --> 00:06:29.120 going to be coming? You know, we don't know what 121 00:06:29.160 --> 00:06:32.000 AI workloads are going to look like, we don't know, 122 00:06:32.120 --> 00:06:33.959 we don't know where data is going to be moved, 123 00:06:34.000 --> 00:06:37.319 we don't know, right, So the question is how do 124 00:06:37.360 --> 00:06:41.319 you then create a way of looking creating a platform 125 00:06:41.360 --> 00:06:45.319 for your company that you're not locked in, right, right, 126 00:06:45.519 --> 00:06:47.920 that you're not locked in, and you create the most 127 00:06:47.920 --> 00:06:53.560 flexibility Because as most folks know or have said, AI 128 00:06:53.720 --> 00:06:57.199 is going to change everything. The footprint of how business 129 00:06:57.240 --> 00:07:00.959 is done is going to change pretty significantly, right right, 130 00:07:01.319 --> 00:07:07.800 And so how so it is incumbent on technology leaders 131 00:07:07.839 --> 00:07:09.959 to you know, I have to drive flexiblit and by 132 00:07:10.040 --> 00:07:12.360 the way, I've got to make sure I do that 133 00:07:12.399 --> 00:07:15.920 in a way that I can. I can not spend 134 00:07:15.959 --> 00:07:17.879 everything in the bank, which which has to do with 135 00:07:17.920 --> 00:07:19.879 the finopps, which has to do with observe a bilak. 136 00:07:20.519 --> 00:07:22.360 We think that you know, if you if you're a 137 00:07:22.399 --> 00:07:25.040 major corporation spending it several million dollars a year, probably 138 00:07:25.079 --> 00:07:26.959 you know there's probably twenty to forty million, twenty to 139 00:07:27.079 --> 00:07:32.319 forty percent of utilization that you could optimize, right, right, 140 00:07:32.639 --> 00:07:36.040 But it's like have you ever read a bill for 141 00:07:36.199 --> 00:07:40.439 your cloud? It's like the phone bill times a million, right, 142 00:07:41.279 --> 00:07:45.519 you have it, right, unless you can actually understand it, right, 143 00:07:46.480 --> 00:07:48.480 you don't know what buttons to change, and and so 144 00:07:49.279 --> 00:07:51.399 you know, like we at a rack space, you know 145 00:07:51.480 --> 00:07:54.079 we we are you know if we are the largest 146 00:07:54.079 --> 00:07:56.240 pin opps service provider in the world. 147 00:07:56.519 --> 00:07:59.319 Right, that's a great way to look at it. That's 148 00:07:59.319 --> 00:08:03.040 a very interesting perspective because what this is old school, 149 00:08:03.439 --> 00:08:06.639 but what gets measured gets managed. I'll throw that over 150 00:08:06.680 --> 00:08:09.079 to should talk. Right, what gets measured gets managed, and 151 00:08:09.120 --> 00:08:12.160 when you can start to see where all these costs 152 00:08:12.199 --> 00:08:15.639 are coming from and align them with the projects and 153 00:08:15.759 --> 00:08:18.759 understand what is the TCO understand what is the ROI 154 00:08:18.879 --> 00:08:21.000 Are we getting return on this investment. The more you 155 00:08:21.040 --> 00:08:23.439 can do that, the better you can manage your environment, 156 00:08:23.480 --> 00:08:26.560 the more efficient you become, the more money your enterprise 157 00:08:26.600 --> 00:08:28.920 makes on the bottom line, right absolutely. 158 00:08:28.959 --> 00:08:32.519 And I think what we're seeing now is that you know, 159 00:08:32.559 --> 00:08:34.840 this entire idea about ROI, which in a sense has 160 00:08:34.879 --> 00:08:37.000 been driven by the explosion of AI because now everybody 161 00:08:37.039 --> 00:08:39.200 wants to do AI and wants to what's. 162 00:08:39.000 --> 00:08:39.720 The benefit of that? 163 00:08:40.039 --> 00:08:42.240 But that is you know, trickering down all the way 164 00:08:42.320 --> 00:08:44.759 to the bottom of the stack, which is the entire 165 00:08:44.759 --> 00:08:47.519 cloud stack, right Because you know, when we talk about 166 00:08:47.840 --> 00:08:50.279 you know, data pipelines, or we talk about storage, or 167 00:08:50.320 --> 00:08:53.159 we talk about computer associated with you know, doing transformation, 168 00:08:53.679 --> 00:08:56.200 it's very important to understand the unit economics of that. 169 00:08:56.279 --> 00:08:58.240 It's not just you know, how much does my data 170 00:08:58.279 --> 00:08:59.639 warehouse cost to run? 171 00:09:00.120 --> 00:09:02.440 How much am I spending on GPUs? 172 00:09:02.799 --> 00:09:06.120 It is what is the unit economics price performance that 173 00:09:06.159 --> 00:09:09.600 I want to optimize for, you know, my inventory optimization 174 00:09:09.679 --> 00:09:12.120 because I'll spend two hundred thousand, But it only makes 175 00:09:12.159 --> 00:09:15.039 sense if I'm actually saving more than two hundred thousand 176 00:09:15.039 --> 00:09:17.919 from the business benefit that's coming from that particular EMIL model. 177 00:09:18.320 --> 00:09:20.399 And you know, one of the other things that that 178 00:09:20.399 --> 00:09:22.720 that sort of listed out in the studies that you 179 00:09:22.759 --> 00:09:26.159 know there's this rise of centralization. 180 00:09:25.639 --> 00:09:28.559 Or COEs, right, because the way we think. 181 00:09:28.399 --> 00:09:30.440 About it is that you know, there are a large 182 00:09:30.519 --> 00:09:34.080 number of what i'd refer to as non functional requirements 183 00:09:34.120 --> 00:09:37.519 where you need to have observability, you need to have quality, 184 00:09:37.559 --> 00:09:40.240 you need to have a standardized way of figuring out 185 00:09:40.360 --> 00:09:43.919 as Ben was saying, what kind of workload works best 186 00:09:44.159 --> 00:09:47.559 in what component in the club, right, is while they 187 00:09:47.559 --> 00:09:49.840 give you a lot of flexibility, they also end up 188 00:09:50.120 --> 00:09:52.960 actually giving you hundreds of options to do the same 189 00:09:53.080 --> 00:09:57.960 kind of thing on different components and having guidance around 190 00:09:58.080 --> 00:09:59.799 you know, I'm trying to build an inventory model, I'm 191 00:09:59.799 --> 00:10:01.399 trying to build a chat point, I'm trying to build 192 00:10:01.399 --> 00:10:04.000 a bi report. Means I need to have these standards 193 00:10:04.039 --> 00:10:07.200 regardless of what actual data is going in. There is 194 00:10:07.200 --> 00:10:11.080 something that's going that is becoming quite key to analytics, 195 00:10:11.080 --> 00:10:13.120 but also all workloads in the cloud. 196 00:10:14.120 --> 00:10:16.120 Yeah, I mean, this is it's fascinating. Go ahead, man. 197 00:10:16.799 --> 00:10:18.480 Yeah, And and now I'm just sort of going to 198 00:10:18.559 --> 00:10:20.639 jump in here and what we're seeing like and this 199 00:10:20.720 --> 00:10:23.039 is what the survey said is, uh, you know, there 200 00:10:23.200 --> 00:10:26.000 there's a sort of phase of going through technologists Phase zero, 201 00:10:26.799 --> 00:10:29.159 day zero, day one, day two, day zero, strategy. Day 202 00:10:29.159 --> 00:10:32.000 one is sort of transformation. Day two is the run side, right, 203 00:10:32.600 --> 00:10:35.240 And the challenge is when you start thinking about hybrid cloud, 204 00:10:35.799 --> 00:10:38.240 the talent required to go through all those phases and 205 00:10:38.320 --> 00:10:42.559 run things and run things at scale right, and run 206 00:10:42.600 --> 00:10:46.320 things at scale that are secure in some run things 207 00:10:46.320 --> 00:10:49.480 get reliable. There's a big talent gap which came out 208 00:10:49.679 --> 00:10:51.440 in the survey as well. And so I think we're 209 00:10:51.759 --> 00:10:54.039 going to start seeing more models that have to do 210 00:10:54.080 --> 00:10:59.480 with managed services on the run side because of in 211 00:11:00.039 --> 00:11:03.039 there's going to be you know, my crystal ball says, 212 00:11:03.679 --> 00:11:07.000 companies are going to want to use align with folks 213 00:11:07.000 --> 00:11:09.440 who can provide the managed service for the full stack 214 00:11:11.039 --> 00:11:14.320 side of things so that they can concentrate on what 215 00:11:14.559 --> 00:11:18.240 what what makes my beer taste better? Mm hmm right, 216 00:11:18.759 --> 00:11:21.840 that's funny, right, Yeah, you got to worry about things 217 00:11:21.840 --> 00:11:23.519 and your team members have to worry about things that 218 00:11:23.559 --> 00:11:25.559 make your what is the thing that makes my beer 219 00:11:25.679 --> 00:11:28.360 taste better? Versus the things I need to run? And 220 00:11:28.360 --> 00:11:31.399 so I think they're going to be emerging models regarding 221 00:11:31.639 --> 00:11:35.279 the ecosystem of operations, right that that's going to be 222 00:11:35.320 --> 00:11:36.360 really important. 223 00:11:36.639 --> 00:11:38.840 Well, and you know you mentioned this already, and i'll 224 00:11:38.879 --> 00:11:42.559 throw this to you first, ben AI workloads. There are 225 00:11:42.559 --> 00:11:45.240 all kinds of AI workloads. I mean, people, I think 226 00:11:45.240 --> 00:11:47.000 a lot of folks in the business world are starting 227 00:11:47.039 --> 00:11:51.799 to appreciate that models, deep learning modules can take any 228 00:11:51.879 --> 00:11:54.919 number of shapes. You can have any number of inference layers. 229 00:11:55.159 --> 00:11:58.440 You're going to ingest all these different component parts. You 230 00:11:58.480 --> 00:12:02.759 can stack and layer and design that as you wish, 231 00:12:03.159 --> 00:12:06.120 and there's really no limit to the permutations for how 232 00:12:06.159 --> 00:12:08.919 you do those things. And what I see happening I 233 00:12:08.919 --> 00:12:12.200 think that's very interesting is a company stor are building 234 00:12:12.279 --> 00:12:15.320 up bespoke models for individual clients. I mean, just one 235 00:12:15.559 --> 00:12:18.519 random company I'll throw out there. I'm going to read 236 00:12:18.639 --> 00:12:21.960 safe books AI. And what they do is they build 237 00:12:22.000 --> 00:12:24.440 they take a model off a hugging face, and then 238 00:12:24.480 --> 00:12:28.240 they train your data, your accounting data, your financial data 239 00:12:28.399 --> 00:12:31.360 on your model that is your bespoke model, and then 240 00:12:31.399 --> 00:12:33.960 once it gets to a certain level of proficiency, it 241 00:12:34.080 --> 00:12:36.440 just watches all your finances and says, hey, I think 242 00:12:36.440 --> 00:12:38.840 I found a problem here. Because there is a story 243 00:12:38.840 --> 00:12:42.320 of Macy's where some person hid huge amounts of expenses. 244 00:12:42.360 --> 00:12:44.080 They didn't find out about it until they went to 245 00:12:44.159 --> 00:12:47.240 the quarterly earnings and it was like whoa, No, like 246 00:12:47.360 --> 00:12:50.440 red three alarm fire, that stuff won't go away. My 247 00:12:50.519 --> 00:12:53.559 point is that this is a fantastic use case for AI. 248 00:12:53.919 --> 00:12:56.679 It's not just Jenai. There's a lot of excitement around Jenai. 249 00:12:56.840 --> 00:13:00.120 That is one whole category of very interesting technologies. But 250 00:13:00.200 --> 00:13:02.480 AI models have been around for decades. 251 00:13:02.399 --> 00:13:03.279 Forever, forever. 252 00:13:03.399 --> 00:13:08.879 This is extremely efficient at understanding particular business use cases, right. 253 00:13:08.720 --> 00:13:11.840 Ben, Oh, they absolutely are. And as you said, Eric, 254 00:13:12.039 --> 00:13:15.200 you know, if you look at models, it's you know, 255 00:13:15.440 --> 00:13:17.639 the essence of these models is you you have trained 256 00:13:17.639 --> 00:13:20.960 a model based on the data set right right, and 257 00:13:21.000 --> 00:13:23.679 you can make that day SAT really big, which is 258 00:13:23.720 --> 00:13:27.759 you know, general chat GPT massive, or you can focus 259 00:13:27.799 --> 00:13:33.159 on your specific data set right right. And so I 260 00:13:33.200 --> 00:13:35.360 think you're I like you bringing this out for folks 261 00:13:35.360 --> 00:13:37.639 to understand, right, what is that? What is the thing 262 00:13:37.679 --> 00:13:41.159 you're trying to do right right? That's right because based 263 00:13:41.159 --> 00:13:44.799 on what you're trying to do generates a whole lot option. 264 00:13:45.120 --> 00:13:48.519 The challenges is most folks just go to hey, I 265 00:13:48.600 --> 00:13:51.960 just do chat GPT, right, and so all the different 266 00:13:52.080 --> 00:13:56.039 nuances associated how do you optimize get lost really fast? Right, 267 00:13:56.120 --> 00:13:58.600 And so unless you're actually talking to someone who understands 268 00:13:58.639 --> 00:14:02.080 the nuances but also understands think about this er the model. 269 00:14:02.240 --> 00:14:04.240 If you look at the value chain of like from 270 00:14:04.320 --> 00:14:07.879 data to models to the decision, that model is just 271 00:14:08.000 --> 00:14:10.240 a little bit. It's like five percent of the stack 272 00:14:10.399 --> 00:14:14.320 tech stack related the delivering value, right right, And everyone's 273 00:14:14.320 --> 00:14:17.519 getting really excited about that. But guess what, folks, that 274 00:14:17.600 --> 00:14:20.600 other ninety percent underneath underneath the water line? Did you 275 00:14:20.679 --> 00:14:21.480 That ain't correct? 276 00:14:22.000 --> 00:14:22.320 Right? 277 00:14:23.480 --> 00:14:27.759 It doesn't matter, It doesn't matter. Yeah, So that that's 278 00:14:27.919 --> 00:14:30.440 really that That's what you know, when we think about 279 00:14:30.519 --> 00:14:32.639 hybrid clond what came out in the service, you got 280 00:14:32.679 --> 00:14:33.559 to worry about. 281 00:14:33.360 --> 00:14:36.679 That stuff, right, Well, yeah, go ahead, go ahead to that. 282 00:14:37.039 --> 00:14:38.399 I just want to remplain that, you know, that's a 283 00:14:38.519 --> 00:14:41.200 very good example of what we mean by workload of 284 00:14:41.240 --> 00:14:43.799 were because if you think about it, the training of 285 00:14:43.840 --> 00:14:45.960 a model or fine tuning of a model, et cetera, 286 00:14:46.120 --> 00:14:48.240 this requires a large amount of data, a large amount 287 00:14:48.279 --> 00:14:50.000 of compute, but for short periods of time. You know, 288 00:14:50.039 --> 00:14:52.080 you may do it daily, weekly, Whatever is the cadence 289 00:14:52.120 --> 00:14:55.120 required for that particular workload, which definitely you know, for 290 00:14:55.240 --> 00:14:58.000 which things like public clouds tend to be much better. 291 00:14:58.320 --> 00:14:59.559 But the actual usage of. 292 00:14:59.480 --> 00:15:01.120 That model, of the the influence of that model is 293 00:15:01.120 --> 00:15:04.639 a much more predictable compute activity where the rest of 294 00:15:04.799 --> 00:15:09.440 influence becomes much more important. Uh, And kinds of things 295 00:15:09.480 --> 00:15:13.559 tend to be much better in private cloud or you know, 296 00:15:13.759 --> 00:15:17.000 self hosted kind of environment. So you know, even within 297 00:15:17.039 --> 00:15:19.639 a particular use case, thinking about different stages, what's the 298 00:15:19.720 --> 00:15:22.600 right place for that to go, whether it's public cloud 299 00:15:22.639 --> 00:15:25.039 one or more public clouds or it's private cloud, actually 300 00:15:25.039 --> 00:15:26.000 becomes quite important. 301 00:15:26.000 --> 00:15:28.200 And that's I think something that is captured in this 302 00:15:28.240 --> 00:15:28.600 as well. 303 00:15:29.480 --> 00:15:31.679 Yeah, and that is important. And you know, I've said 304 00:15:31.679 --> 00:15:34.639 this for years now and I feel somewhat vindicated a 305 00:15:34.720 --> 00:15:38.279 throat of event. First my mantra was that the rumors 306 00:15:38.279 --> 00:15:41.759 of on Prem's demise have been greatly exaggerated. What do 307 00:15:41.840 --> 00:15:42.639 you think, Oh. 308 00:15:42.559 --> 00:15:46.360 My god, Oh my god, that is Think of that 309 00:15:46.399 --> 00:15:50.720 like it is an urban myth. It is like the 310 00:15:50.759 --> 00:15:56.279 biggest urban myth that has been perpetuated on us, right right. 311 00:15:56.600 --> 00:16:04.799 This is because the thing on the workload you need, Yeah, yeah, 312 00:16:04.840 --> 00:16:06.919 I mean like if I'm taking if I'm driving my 313 00:16:07.000 --> 00:16:10.759 kids to school every day, right, my Honda Civical work 314 00:16:11.200 --> 00:16:12.080 just fine. 315 00:16:12.240 --> 00:16:12.840 Right right? 316 00:16:12.960 --> 00:16:15.480 Or or my Honda because I have a bunch of kids, right, 317 00:16:15.960 --> 00:16:19.960 I don't need my sports car right right. And so 318 00:16:20.639 --> 00:16:25.240 you know, when you start looking at the the economics 319 00:16:25.559 --> 00:16:29.720 of this, it starts looking like, you know, maybe that 320 00:16:29.799 --> 00:16:34.200 truck is a much better value and that's really the 321 00:16:34.279 --> 00:16:37.759 key word, is a much better value for what I'm doing. 322 00:16:37.759 --> 00:16:40.759 So and we are the survey will show the amount 323 00:16:40.799 --> 00:16:43.840 of folks think about repatriation. 324 00:16:43.759 --> 00:16:44.200 Mm hmm. 325 00:16:44.720 --> 00:16:46.919 Right. It's kind of like buying, like buying a boat. 326 00:16:47.080 --> 00:16:48.480 You know, you get really excited to day you buy 327 00:16:48.480 --> 00:16:51.840 a boat. Then you realize how much it actually cost, right, 328 00:16:51.960 --> 00:16:54.679 and you're like the upkeep the upkeep right, and you're like, 329 00:16:54.759 --> 00:16:56.600 all right, the best days of a boat or is 330 00:16:56.600 --> 00:16:58.200 and when you buy the boat and when you sell 331 00:16:58.240 --> 00:17:00.519 the boat, right. 332 00:17:00.080 --> 00:17:03.320 That's pretty funny, especially if our hurricane is coming in 333 00:17:03.399 --> 00:17:06.359 right then you're like, oh no, oh no, this is 334 00:17:06.400 --> 00:17:09.279 a problem that can't get insurance. 335 00:17:09.279 --> 00:17:12.680 Exactly, which, yeah, which goes back to Eric, like what 336 00:17:12.720 --> 00:17:14.880 are you trying to do? What is the workload that 337 00:17:15.039 --> 00:17:18.279 matters for you? And worder and be clear about the 338 00:17:18.359 --> 00:17:21.680 options because you know what, you don't want to overspend 339 00:17:22.039 --> 00:17:24.160 because with AI, like you know a lot of comps 340 00:17:24.240 --> 00:17:26.599 like I talked about CIOs, I go, hey, I wonder's 341 00:17:26.640 --> 00:17:28.480 AI thing. But you know, I don't have my much. 342 00:17:28.519 --> 00:17:33.240 It didn't go up right, right, customer customers aren't paying anymore, 343 00:17:33.920 --> 00:17:36.920 So who's going to pay for this? Right? So that's 344 00:17:36.960 --> 00:17:39.359 where when you start looking at the delivery costs and 345 00:17:39.480 --> 00:17:44.559 unit economics, Innovation costs money. Right, I've got to figure 346 00:17:44.559 --> 00:17:47.079 out how to optimize my current environment, which means I 347 00:17:47.119 --> 00:17:49.960 actually need to you know, I've got to do spring cleaning. 348 00:17:49.960 --> 00:17:51.799 I've got to do my workload to wear so I 349 00:17:51.799 --> 00:17:52.880 can pay for this stuff. 350 00:17:53.279 --> 00:17:55.039 Right. Well, I mean this is the you're big of 351 00:17:55.079 --> 00:17:56.920 an excellent point, and we'll dive into this in the 352 00:17:56.960 --> 00:18:01.400 next segment, folks, because spring cleaning, right, just understanding what 353 00:18:01.480 --> 00:18:03.880 you have, where it is, who's doing what, Where is 354 00:18:03.880 --> 00:18:06.079 your data? How much does it costing you to store 355 00:18:06.160 --> 00:18:09.400 this stuff? What is the overall plan? It's very difficult 356 00:18:09.400 --> 00:18:12.640 to know that from a large enterprise perspective. It's very 357 00:18:12.680 --> 00:18:15.640 hard to sort of piece together all the components to 358 00:18:15.720 --> 00:18:18.400 really understand what you're looking at. But that is a 359 00:18:18.519 --> 00:18:21.480 huge issue right now that companies need to do, especially 360 00:18:21.720 --> 00:18:24.240 as this ALI era dawns. But don't touch that toll. Folks, 361 00:18:24.319 --> 00:18:26.039 be right back. You're listening to Inside Analysis. 362 00:18:26.440 --> 00:18:38.039 Respect, Welcome back to Inside Analysis. Here's your host, Eric Tabanac. 363 00:18:40.720 --> 00:18:43.079 All right, folks, back here on Inside Analysis talking to 364 00:18:43.119 --> 00:18:45.440 a couple of experts from rack Space. Have a big 365 00:18:45.480 --> 00:18:48.799 IT survey that just came out. Very interesting information. Not 366 00:18:48.799 --> 00:18:51.480 not a whole lot of big surprises for me when 367 00:18:51.519 --> 00:18:54.240 I look at the market and understand what's happening here. 368 00:18:54.279 --> 00:18:56.759 And let's face it, the shiny new object now is 369 00:18:56.839 --> 00:18:59.400 AI and everybody wants it. They want to leverage it, 370 00:18:59.599 --> 00:19:01.920 but you got to be careful. Number one, how much 371 00:19:01.920 --> 00:19:02.559 does it cost? 372 00:19:03.039 --> 00:19:03.200 Well? 373 00:19:03.279 --> 00:19:04.759 Really, number one is what do you want to do 374 00:19:04.799 --> 00:19:06.960 with it? Like what is the business case around it? 375 00:19:07.240 --> 00:19:08.960 Number two is what it costs? And then you can 376 00:19:09.039 --> 00:19:11.119 kind of figure all that out. And finnops is a 377 00:19:11.200 --> 00:19:12.920 huge part of that. And Ben, I love how you 378 00:19:12.960 --> 00:19:16.400 said rag Space is the biggest finops provider because that's 379 00:19:16.440 --> 00:19:18.400 what you're that's part of what you're doing is helping 380 00:19:18.400 --> 00:19:21.799 people understand what does this stuff cost? And once you 381 00:19:21.920 --> 00:19:25.680 once you define your business objectives, then you really got 382 00:19:25.720 --> 00:19:27.599 to focus on what's it going to cost us, how 383 00:19:27.599 --> 00:19:28.839 are we going to do it, where's all that going 384 00:19:28.880 --> 00:19:31.359 to come from. That's where a rack space wild comes 385 00:19:31.359 --> 00:19:32.240 into play, right. 386 00:19:33.200 --> 00:19:39.319 Absolutely, what we so we manage over billion, billion and 387 00:19:39.400 --> 00:19:45.039 a half dollars of cloud spend with our customers. Within 388 00:19:45.119 --> 00:19:47.920 that context of that, what we then do is provide 389 00:19:47.920 --> 00:19:52.599 a service to those customers that that uh, that there's 390 00:19:52.599 --> 00:19:55.279 a finnops service that essentially once a month, we go 391 00:19:55.319 --> 00:20:00.759 over your bill, right, whether it's it's Google, you're w S, 392 00:20:01.640 --> 00:20:04.759 we go over your bill right, and we say, hey, 393 00:20:04.759 --> 00:20:06.720 this is happening, what's going on here? What about this? 394 00:20:06.799 --> 00:20:08.599 You might want to turn this off? You want might 395 00:20:08.640 --> 00:20:13.759 want to move things here because it's like, uh, depending 396 00:20:13.759 --> 00:20:16.160 on the configuration. You know, you know, it is really 397 00:20:16.200 --> 00:20:18.240 easy to buy resources on the cloud. 398 00:20:19.240 --> 00:20:21.400 That's too easy, right, that's the problem. It's too easy. 399 00:20:21.440 --> 00:20:23.680 It is so easy, right. 400 00:20:23.720 --> 00:20:26.680 And like your kid buying squish mellows with your credit card. 401 00:20:28.160 --> 00:20:29.119