Transcript 0:00 There have been engineers, right, that have been working for the organization 10, 15-plus years. 0:07 The, these engineers were extremely successful in the data center and moving to the cloud, um, they basically have to completely re- retrain. Right? 0:14 So if you had a noisy communication protocol, um, then, you know, y- you may saturate your data transfer charges in the cloud. If you are optimizing for sustainability, um, you also often optimize for cost. 0:28 Before you design your solution, start thinking about cost. 0:33 So cost on the, on the timeline is being positioned early on, at the, at the beginning of the design, not, like, once the design is done you try to make it cost efficient. Right? 0:47 Hi, everyone, and welcome to a new episode of our podcast. Uh, today, uh, we're having, uh, Dieter. Uh, welcome, Dieter. Hey, thanks for having me. 0:59 So it's going to be a wonderful, uh, interview, and of course we have our co-present, uh, co-host, uh, Damian. How are you, Damian? I'm fine. I'm very good, thank you. So- It's always a pleasure. [laughs] Yeah. Thanks. 1:13 Thanks for being both of you here. Uh, yeah, Damian, let's, let's introduce our guest here. Hey, guys, I'm Dieter Maisan. 1:22 Um, I have been working in the cloud since 2013 at companies like Google, Netflix, Intuit, and most recently Roku. I'm also one of the FinOps ambassadors. Excellent. That's... It's good to have you here with us. 1:38 We, uh, are, we're very impressed. We look at your site and, uh, one of the things that just, uh, [laughs] 1:44 immediately, uh, pop up is, uh, about the FinOps ambassador and, um, uh, and we have to ask, you know, how, [laughs] how someone can become a FinOps ambassador. 1:55 Uh, I think you need to be volunteered by the, uh, foundation. Uh, it is not something that you can actively achieve. Right? You do a lot of outreach, a lot of public speaking, uh, those kind of things. 2:07 I know, uh, the founder, JR Storment, uh, since 2016, well before the founding of the foundation, and, um, in, uh, in 2016 at re:Invent, uh, JR asked me the question, "Hey, uh, does this cloud financial management have legs?" 2:24 And, uh, I said, "You know, it's like an industrial kitchen where you have, um, you know, six, eight cooks, um, and so it has to be cleaned up every day, right?" Mm-hmm. Um, and that applies to the cloud, right? 2:37 You, y- y- you just don't, never finish, you know? If, if engineers do work, engineers create waste, create anomalies, create challenges forecasting, and so it's an ongoing effort. And 2:52 after you get the title, is there anything else that you do, that you must do or...? There, there is a couple of responsibilities, yes. 3:00 So first of all, you have to be a certified FinOps practitioner, right, um, to become an ambassador. So, um, and in addition, there are ambassador meetings, uh, that we have on a monthly basis, right? 3:13 Um, and there's different, uh, regions. There... We just had the North America ambassador meeting, but there is also, like, a global ambassador meeting. 3:22 Um, and during that, uh, the, the foundation shares information, and they may have projects for which they are looking for volunteers, right? 3:30 And then the ambassadors can volunteer for those and drive, drive these efforts. Interesting. Interesting. That's interesting. 3:38 So also we, we looking through your, [laughs] your site also we, we also noticed that, you know, you, you were th- one of the, you know, people that were doing FinOps, I think the, even before it was called FinOps. 3:51 Right? You were, like, kind of a, an early day, uh, uh, promoter of, uh, financial productions or, or optimizations. So how was it, uh, in those days? 4:03 Yeah, I mean, when I, um, uh, left PayPal and I interviewed, um, at, at various companies, right, I saw there's a, a bunch of positions open at Google, and I said like, "Hey, what's the worst that can happen," right? 4:16 Uh, th- they tell me no, right? Um, so I applied for a Google position, and Google has an interesting, uh, hiring, uh, process where they, um, have queues, um, where, uh, you, you don't actually meet your, your manager. 4:32 The, the queue is completely anonymously, um, uh, interviewed, and then you get in, in a sense, uh, randomly assigned to a job that, that requires that role, right, for which you interviewed. 4:46 Um, I had the fo, uh, fortune where I, uh, was able to meet my future manager briefly. 4:52 Um, and so when I started working at Google, they realized that their public cloud offering is growing faster than they can build data centers. Right? Data centers typically is, like, a two-year effort. 5:04 Um, and so they were looking for internal inefficiencies that they could carve out and assign to the public cloud. Interesting. Yeah. [laughs] That's, that, that's awesome. Yeah. 5:18 And, and so I'm curious about, like, how was the transition from, that you, now that you mentioned about the data centers, of have you seen the transition from doing FinOps or doing, uh, finance related stuff from data centers to what you are doing now in the, in the public cloud? 5:37 Yeah, I mean, public cloud, to be honest, is, is everything is new, right? Yeah. 5:41 Uh, whatever you knew before, uh, in the data center we do capacity management, right, where we are looking at, um, three to five-year forecasts, and then we, um, procure hardware. 5:55 Um, and the, the hardware is intentionally overprovisioned- Mm-hmm... um, so that the workloads can grow into the hardware over the amortization period, right, of the three, four, five years. 6:07 What- whatever that period is.And so, for example, at Intuit, we migrated from 26 data centers, um, all into, uh, Amazon Web Services public cloud, right? 6:19 And, um, there have been engineers, right, that have been working for the organization 10, 15 plus years, and, um, were extremely successful, right? There are products- Mm-hmm... 6:32 like TurboTax or QuickBooks that are, um, you know, my CPAs- Yeah... use QuickBooks, right? My, um, accountant is using QuickBooks. 6:42 So the, these engineers were extremely successful in the data center, and moving to the cloud, um, they basically have to completely re, retrain, right? They start fresh. 6:52 So, um, one thing that we did notice is after the lift and shift migration, um, we had, uh, 24% waste in, in the cloud, right? And we used, uh, six waste sensors to determine that. 7:05 Waste sensors are like little scripts that collect information and store it in a central- Mm... uh, repository where it then can be surfaced by, by, by engineering owner or business owner. 7:16 And, um, over the course of four years, we got the waste down from 20, uh, 24% to, to under 5%. Mm. And, um, we, we increased the waste sensors from six to 34. 7:30 So we sort of raised the bar, and at the same time, we, we drove the cloud waste down. No, that's, that's really interesting. 7:38 So you basically made a waste detec- an automate, well, in-house waste detector for, for you, and now you are making a increasing, let's say, the, the suite of, of capabilities for, for that, and that's, that's super- Yeah... 7:53 super interesting. 7:54 And you, so now that you mentioned that, you know, the migration that happened, uh, on this topic, is that now that there is this trend or this debate about the repatriation stuff, uh, so going back to, uh, on premises because of the cost of the cloud is, is too high, what are your thoughts on, on this, on this topic and on this trend that is going on? 8:17 There are, um, a couple of use cases where that may make sense, right? Mm-hmm. 8:22 Um, and, you know, as an ambassador, I do a lot of outreach, right, where I go and, and talk to a lot of people, um, typically multiple organizations every week, right? 8:33 And there are some, um, organizations that were able to move cloud workloads back into a data center quite successfully, but those were niche workloads, Dropbox being one of those examples, that moved from, uh, the simple storage service on AWS to their own storage solution in a data center, which they can manage more efficiently. 8:55 When you think about it, the cloud is, uh, a managed service. Uh, AWS has over 175 products, right? Mm-hmm. And all these products are managed, right? 9:05 Someone else is carrying the pager, uh, when something goes wrong, and so that comes at a cost. Now, economies of scale can, can reduce the cost, right? 9:15 Because AWS has, you know, dozens of data centers with, with hundreds of thousands of virtual machines, um, so they can reduce the cost. 9:25 But overall, I think there is sort of like a 20% premium if you compare like for like in a data center to the cloud, and that is the management overhead, right? Mm-hmm. 9:35 However, um, I also consult with Hennepin County, uh, in Minnesota, and the, the, they found that certain technologies would be, um, very challenging to, uh, implement in the data center, uh, specifically text to speech or, or speech to text. 9:54 Those kind of things, building that in a data center would take years. Mm-hmm. And, uh, in the cloud, it's just a product, right? You, you just use the product- Mm... and it, it works for the most part, right? 10:05 So the, the, the elasticity of the cloud, um, has benefits, right? If, for example, TurboTax, um, i- the United States has one tax date, which is April 15th, right? 10:18 And for some reason, people, they, they sort of file the taxes a couple days before, right? You, you can file them any time, um, after, like, February, right? 10:27 Because this is when your, uh, tax records, uh, come in- Yeah... employee and, and, and those kind of things, right? 10:34 So you can file them anytime, but most people, like, file them, you know, one, two days or on tax day, right? 10:41 So, uh, Intuit had this huge problem that they needed to have the data center capacity for the tax day, um, and then the rest of the year, this was unused, right? 10:52 And in the cloud, you can simply scale up maybe two weeks before tax day, right? 10:57 And then you're ready, you have the capacity, and then after tax day, you can release that capacity, which also means that you are not going to be billed. 11:06 So that elasticity, uh, has, has the benefit that you can scale up, but it also has, in a sense, the, the opportunity that, um, if something is unused, you need to turn it off, otherwise you're still incurring a cost. 11:22 Yeah. Right? No, that's super interesting, interesting. And yeah, it's a great reflection that, uh- Yeah... 11:27 because we, I don't recall when we were talking, Damien, about this, uh, regarding, like, load testing and, you know, peak kind of things, for, for example, uh, retailers, like with these Black Fridays- Mm... 11:40 uh, you know, Christmas events, and this is something that definitely, uh, needs to be evaluated before, uh, considering a reputation. This is, this is an interesting example. 11:50 [laughs] First time I think that we talk about taxes and [laughs] Yeah. Yeah, yeah. Something what nobody wants to think. It's interesting. [laughs] It's like... 11:58 [laughs] Of course, they will leave it to the, to the last day. Yeah. But, uh, but yes, elasticity is a good, a good example of how the cloud can help you. 12:08 And of course, if you wanna move that into the on-prem, of course, you w- always wanna, need the hardware to beUh, supporting on your top, on the, on the, you know, the, the, the worst performance, let's say, right? 12:23 So you're wasting a lot of money. Yeah. Uh, a- and of course, I think the on-premise we, uh, we talked in this show several times about the clarity, 12:32 the clarity the, the cloud offers you, and I'm not sure you, you can even get to the point of that clarity on the cost clarity, right? Yeah. Uh, on, on on-premise. The- Yeah... 12:45 there is a lot of ways that you don't know, right? Yeah. And, you know, there's also surprises, right? When you have an on-premise workload, the data transfer is free, right? 12:56 Um, but you move into the cloud, all of a sudden you need to pay for the data transfer. 13:00 So if you had a noisy communication protocol, um, then, you know, y- you may saturate your data transfer charges in the cloud, and that is a common thing that I see when, when people migrate that they sort of forget about certain things, right? 13:16 Um, some of the virtual machines, they don't come with, um, storage, uh- Mm-hmm... the disk storage, right, or virtual disk storage. 13:26 So you, you, you buy the virtual machine, but it's like you can't run an operating system on it because it does not, uh, have ephemeral storage, so you need to a- attach elastic block storage and, uh, that has also an additional cost, right? 13:40 So that is another common mistake that I see that engineers do when, when they, um, size workloads in the cloud, right? That they forget those things. Yeah. 13:52 Um, moving on to, to that topic, like, what are your, your recommendations when you w- have to do an exercise of sizing workloads in the, in the cloud? Like, what do you recommend? 14:04 Because I'm sure you have done a lot of, uh, you know, sizing and estimation exercises, so what are your main recommendations for, for doing that kind of things? 14:14 You know, that, that may sound weird, but, um, one of the best things to get your engineers to provide high quality work or deliverables in the cloud is to, um, ask for a certain percentage of your engineers to be, uh, certified. 14:33 Mm-hmm. Um, it gets... You know, at, at Intuit, we had to mandate that, uh, 50% of the engineers need to get some, uh, any AWS certification, right? 14:43 It doesn't matter if it was professional, associate, specialist or those kind of things, right? Um, just because, you know, once you have enough certified engineers in the team, it will cross-pollinate, right? 14:55 The, the, like, a, a, an uncertified team member can ask the certified team member for advice, right? And, um, that, that is, it's really the knowledge that is driving the quality of the work, in my experience, right? 15:09 And it, it sounds a little bit barbaric that you just mandate 50% get, get certified, right? But it, it helps the engineers, right? They are more employable with their- Mm... ne- next employer, right? And they, um, 15:26 are, uh, able to help, uh, and, and they know when they need help, right? And all these cloud vendors, m- maybe except for Azure, right, they are very welcoming, um, to provide solution architects for projects, right? 15:42 Have that person sit with your team for, like, two, three weeks and provide advice and insights into best practices, Well-Architected Framework and those kind of things. No. Uh, it's really interesting. 15:56 I, I haven't thought about that on, on... Because in my company, there is definitely, um, some incentives on to that. 16:03 I know that there are companies that even give bonuses to people that get certified, so there are different programs that are, that are interesting in, in that topic. [laughs] No, no. Yeah. Yeah. 16:14 Bonuses is definitely better than layoffs. I- No, for sure. [laughs] [laughs] For sure. Definitely. Also, I, I, I... It might be when I was, uh, learning and learning. 16:28 You all, you all the time there are new things, and when you do the certificate or recertifications, you only found out there were new things that are going on 16:36 and, uh, and new, uh, optimization or, or another, uh, key features that, that might help you in optimizing or helping the, the workload, uh, not only, uh, on the, on the, you know, on this, on the economic level, but also- Yeah... 16:55 on the efficiency, right? When you work efficient, in any case, you know, the pillars, or any, any pillars that you choose will help you eventually economically. Yeah. Exactly. Exactly. 17:05 And this is where Well-Architected comes into play, right? 17:08 E- every cloud vendor has, um, something that, um, you know, provides best practices and guidelines on how to, how to work in the cloud, how to build things in the cloud. Totally. 17:21 And, and regarding this skill set or this training on educational level, uh, what do you think are the skills or the, you know, the, the education that someone can get to become a good FinOps practitioner, uh, especially here from, from the technical, uh, perspective because b- we here are bo- coming from, from the cloud world. 17:44 Uh, but from, from that side, what are your, your recommendations on, on formation and education and on, and skill sets? 17:51 You know, that, that is a, a very good and interesting question because the engineers, they know how to write code, right? Yeah. And that includes things like-Um, even, uh, how to use a GitHub or GitLab, right? 18:04 How to use, uh, libraries, right? A lot of the code, uh, uses public, uh, uh, open source libraries, right? Um, which are maintained outside of your organization, right? 18:16 And could, uh, uh, have a security risk with them, right? Mm-hmm. Yeah. We just recently had a situation of a, um, Linux compression command line tool that had a, a, um, a vulnerability exposed on purpose, right? 18:32 Um, so there, there... And, and then the engineers are also familiar with other tools like, uh, PagerDuty, for example, right? So that they- Yeah... they can integrate with that. Um, and observability tools, right? 18:44 Like a Datadog, um, Splunk, New Relic, um, those kind of things, right? 18:49 So the, the engineering skill set is not just really focused on, on writing code, right, but there's all these other systems that they need to become quite proficient in, right? 19:02 And I would say that the, the cloud is yet another system that they have to become proficient in, right? Mm-hmm. 19:11 And as I said, um, there's over 175 products when I checked the last time, and, um, just in, in Amazon Web Services. The same is true for Google Cloud, the same is true for Azure Cloud, right? Yeah. 19:22 And, um, depending on where that engineer is working, they need to know, uh, maybe a, a dozen, two dozen services, and need to be familiar with them quite inside out, right? Quite- Mm-hmm... and know the details. 19:37 So for example, a virtual machine on, on AWS works differently from a virtual machine on, uh, Google Cloud, right? 19:45 Google Cloud is much more fungible, uh, when it comes to CPU and memory, while, um, AWS has certain, uh, slots that, that you can get. Right? 19:54 And then, um, there's also differences with, with, uh, additional ancillary services, right? Like the simple storage service on, on AWS, and the, the lifecycle management, um, when... 20:09 And then on, uh, just Google Storage, Cloud Storage, right? Um- Mm-hmm... 20:14 for example, on the AWS side, when you have, uh, an aged out object, um, that is, uh, aged to a colder, less expensive storage, right, uh, it gets rehydrated into a staging area, right? 20:28 While on, on Google Cloud, Cloud, uh, Storage, um, the, the object is available instantaneously. However, the aging out is a 365-day time period, right? 20:43 So as you access an older object, it, it is ins- available within, you know, 10, 15 milliseconds. However, you have to pay for it to age out again, right? Yeah. 20:53 While on the AWS side, um, it may take some time to rehydrate, and then it stays within that staging area for 30 days, after which it, it gets cleaned out. 21:04 So the engineers really need to understand all those intricacies, right? When, uh, you look at Elastic Block Store on the AWS side, there is, uh, GP2 and GP3 options, right? Um, now you can switch between GP2 and GP3. 21:18 However, many engineers are not aware that there is a cool down period of four hours for switching back. 21:26 So while the switch is in most cases seamless, right, and it can be done online, you don't have to take the volume offline, um, if you decide to, to make any other changes, you need to wait four hours before these can be applied. 21:41 So the engineers really need to understand, um, how these cloud services work or you will get surprises, right? 21:48 For example, we had a situation where an engineer used a Terraform script, uh, to instantiate a vendor solution. It was, it was a, in a sense, a SaaS solution that was self-hosted, right? Mm-hmm. 22:00 And when, when, whenever they had a configuration error, they just tore down the, the, uh, cloud workload and instantiated again, and they did that, you know, 20, 30, 40 times. 22:13 What they didn't realize is that they started orphaning those, um, Elastic Block Store volumes, those, uh, virtual disk volumes. 22:22 And that kept, you know, because the entire system was supposed to be a, a, a SaaS solution, so I think the, the volume was, like, two terabytes in size, and they had, you know, after a month of, of trying to install that software, they, they had like, you know, 160 terabytes of orphaned volumes lying around that they were simply not aware of, because from their point of view, they were just running a Terraform script. 22:47 Yeah. They aren't dealing with the consequences of doing that. Yeah. Definitely. Yeah. No, that's, that's an interesting use case. 22:53 And, uh, and I do agree that, uh, for doing optimizations and for doing, uh, FinOps work, you definitely need to know inside out the, the cloud, um, services. 23:06 I, I, for example, and I talk with Damien and with the people that we talk about, that, for example, I am much better in GCP than in AWS, and it doesn't, it, it cannot... It translates only in the generalistics- Yes... 23:19 like system designs, uh, cloud knowledge a bit. 23:23 But then when you are actually using the tools, the, the, the proficiency that you have on a day-to-day basis is nothing compared, uh, when you are used to do a work with a provider all day, uh, for every day in, in your work life than, for example, if I now have to switch to AWS on, on a day-to-day, I will have to stay rolling, you know? 23:43 Uh, and then if I am on GCP, I am super comfortable. It's, it's, I totally agree on that point.Yeah, agreed. I, I have the same observation. We actually use all three clouds- Mm-hmm... um, in different proportions. 23:56 The majority is on AWS, but some of the data engineering workloads are running on Google Cloud, because the, um, data analytics offerings that come out of the box seem to be, um, having a, a more breadth and more depth and- Mm-hmm... 24:13 higher quality of the features, right? Um, and so the engineers, they, they have to relearn, right? They, uh, instantiating a virtual machine on AWS works a little bit differently from GCP, right? 24:25 And then and, and if you go to Azure, um, you can run something with, uh, DTUs. Those are processing units. 24:32 Um, and, uh, something that uses processing units, for example, cannot use a reservation on the Azure side, right? So a- again, it's, it's different. High level, yes, they're all virtual machines, right? But- Yeah... 24:45 um, how they actually work differs from cloud vendor to cloud vendor. Yeah, yeah. Exactly. It also, it also impacts the, the automation side. Yes. 24:54 'Cause when you are cl- shutting down i- in AWS, not the same shutting down in Azure, when you need to deallocate, you know, to save. 25:01 And this is a very import- important, uh, topic and very important, uh, item that people need to know, not only the, not only the DevOps, but also the FinOps, right? When they're, they're analyzing all the, 25:15 the optimizations that can be done in an automation, um- Yeah. For example, a, a, a, a good example here is on the Azure side, if you have a database, right? 25:26 And let's say you are an organization that only works from, like, 8:00 to 6:00 PM, right? Mm-hmm. Um, and y- you're not really international. You don't really have engineers in other countries working, right? 25:36 On the Azure side, what you do is you actually reduce the compute units for your database, uh, to, to the point where, um, it, it is still running. It could still do, like, one query a minute, right? 25:49 But it cannot handle the workload at peak. And then in the morning, you basically increase the processing units so that the, the database is able to handle the, the daily workloads, right? 26:01 Whi- which is something that you can't do on AWS or GCP, right? Um, so again, it's, it's, there's differences, substantial differences, and you need to know how to take advantage of the elasticity of the cloud. Yeah. 26:14 That's super interesting. 26:16 I mean, the, what, that's what we are also seeing is that coming from the, let's say, FinOps engineering perspective, so people that are used to do, like, more hands-on optimization such as the one that you mentioned about automation, that you actually need to have the cloud knowledge to actually implement those, and that's what we are seeing. 26:33 Like, what, what we try to do is to increase the content and increase the awareness of that. 26:39 It needs to be some technical content, uh, regarding that, uh, so that people are, uh, better in these practices, you know, in being able to, how to reduce the, the compute, uh, you know, capabilities automatically from 6:00 PM. 26:56 So you need a script for doing that. Yeah. And how do you learn to do that script? 27:00 And that, that's content that I think it's, it's, it's left to do in the, in the FinOps perspective, but it's something that definitely it's worth, especially for, like, more engineering FinOps perspective. Yeah. 27:12 And, and I agree. This is why I think that certifications are so immensely helpful, right? Mm-hmm. Because you, you do get formal training, right? 27:19 And, and so- some of the training that you receive, at least the foundational things, they, they will st- stay with you, and then you know wha- what you need to watch out for and where to get more information and when you need to get help, right? 27:34 So if you, if you have a big project coming up where you need to start a new workload, you know that, uh, there's a solutions architect that I can possibly invite to my project, and we can, you know, whiteboard a couple of approaches of how we want to do that and get the feedback from the solutions architect. 27:52 Yeah. And other thing I was, uh, the other day, uh, it was, uh, at a meeting in, in Berlin. 27:59 I was, I joined remotely, and I was talking about, uh, trends for 2025, and one thing that I come to mention when I was drafting the, the presentation is that there is, uh, and what you mentioned about certifications, is that there are specialties more or less in every pillar that, from the Well-Architected Framework. 28:19 So there is networking, um, there is security- Data privacy... uh, all, all this stuff. 28:25 Do you think it should be, or it should happen in the near future that there is a certification of, and a specialty in cost optimization and in sustainability? 28:34 Because it seems like there are two topics that are, you know, forgotten in that, in that area, and I think it's super interesting, interesting that the providers could, uh, give a, a certification on that. 28:47 You, you actually, you touched on a really, um, important point, right? Uh, in a sense, yes, those certifications are required. 28:54 And uh, interesting story, I actually worked with AWS on a cloud financial management certification. Um, however, 29:03 because it would directly compete with the FinOps Foundation, um, it was decided that not to proceed to offer this, right? 29:12 Because you can get a FinOps practitioner or FinOps engineer certification from the FinOps Foundation. Um, and around sustainability, um, maybe there is something similar, right? Mm-hmm. 29:23 Similar organizations where you can get sustainability certifications. Hopefully, they are technical enough, uh, for the specific cloud vendors, right? 29:31 A sustainability certification for AWS will look different from, from GCP and from Azure. Um, but o- overall, I think, yes, that is important. 29:43 At some point, uh, you know, I also have a thought where, um, cost and sustainability, there's a lot of synergies between the two, right? Yeah. 29:52 If you are optimizing for sustainability, um, you also often optimize for cost.Um, region specific, right? If you move something to a, uh, uh, more sustainable region, maybe it's also cheaper at that point, right? 30:08 And, um, th- there's a lot of synergies. However, at the, at the foundation, I don't see, um, the, these merging. Um- Mm-hmm... 30:17 you know, rather the foundation may offer a, uh, sustainability certification, but I'm sure that there is many other organizations that already offer something like that, so it wouldn't make sense to compete with it. 30:30 No, that's, that's super. I think I wanted to, to know the thoughts because when I was drafting this, I, it was come to my mind like, yeah, why there isn't a, you know... 30:39 There are these four pillar or well, these six pillars, and there are only, I think, uh, three, four, I don't recall the, the number of specialties now. There are a huge amount of them, and now with AI, even more. 30:49 But yeah, that's, that's really insightful. I think, uh, people will know now the, the, the reason behind it, definitely. Yeah, I think that everybody, uh, actually thought about it. I, I, I... 31:00 When I finished one of the first, I think was in AWS, one of my first, I asked the same, "Why don't you have a badge for, uh, uh..." But I didn't, I didn't get so deep. 31:10 I, I never [laughs] I didn't get, I didn't know the answer back in there, back in those time. But, uh, yeah. Well, I guess that, uh, we are now at, uh, now that we are in Christmas time, 31:24 uh, it's a good, uh, question for, for you, Dieter, to tell us what do you think or what do you wish or what do you think will happen next year in the FinOps? 31:35 What do you see will ha- what, what will develop or how do you see FinOps developing next year? I mean, FinOps, um, let me, uh, actually, uh, look this up because we have a roadmap. 31:46 [laughs] So I can tell you exactly what the focus will be. So, um, implementing focus. Focus is the, um, cl- uh, cloud agnostic, uh, billing, uh, format, right? Mm-hmm. 32:01 So that, that's going to be one of the, uh, things that, uh, next year we are going to focus on. FinOps scopes. 32:08 So scope could be, uh, FinOps for the data center, FinOps for the cloud, FinOps for SaaS, FinOps for AI, uh, those kind of things. Uh, cloud sustainability is another focus. Um, FinOps for AI. 32:23 We are actually thinking of a, uh, ce- certification for that as well. Um, then FinOps aware product decisions. So this is all about shifting left, right? Before you design your solution, start thinking about cost. 32:40 So cost on the, on the timeline is being positioned early on, at the, at the beginning of the design. Not like once the design is done, you try to make it cost efficient, right? 32:54 Um, and the last of the six items that we are going to focus on is the public sector. There is a, a lot of, um, you know, cloud usage in the pl- public sector. 33:05 All the schools, all the libraries, um, every city, every county, and every state, and then the, the federal government with its military, right? They're all using cloud, and there are some restrictions. 33:19 Like, for example, the budget is approved on an annual basis. So you get like X- Mm... amount of million budget, right? However, when you buy a reservation, you make a commitment for three years. Yeah. 33:30 So now you are unable to buy these reservations because technically you don't know what your budget will look like in year two and year three. 33:41 That's, that's super interesting, and thanks for, for the insight on, on the roadmap. Um, yeah, definitely public sector is... 33:50 I'm seeing, uh, from, from the subscription base that there are a lot of people from universities, schools, and counties, uh, which surprised me at the beginning. Uh, but definitely I... 34:01 When I started to dig deep into the news as well with the, there was a, a contract from, from a large consulting company with the airf- I believe it was the Air Force, uh- Mm-hmm... of the US, so... 34:14 And also I've seen several things about the US strategy was one of the numbers, and, uh- Mm-hmm... the home department of the UK as well, uh, they, they are starting, uh, working on the FinOps. 34:27 So yeah, definitely interesting things on, on the public sector as well. Public sector and, uh, big companies as well. They, they, you know, all products are very strict with yearly budget. 34:39 I think they, they, they have the same issues. Yeah, and it is very interesting. Thank you for sharing with us. Absolutely, yeah. Yeah. 34:48 So, uh, again, so, uh, again, for me it was very insightful, and Dieter, it was a pleasure to have you here with us. 34:55 Uh, looking forward to see you again in the future, sharing with our more, more, more ideas and, uh, stories and, uh, and, and roadmap. And, um, again, I hope you had a good time. Absolutely. 35:10 And, and thanks for inviting me, and happy to talk again. Yeah. It was a pleasure. 35:15 Uh, Dieter, we will le- leave all the links, uh, regarding your, uh, you know, your website and everything that you do on, on the description of the, the podcast. 35:24 And, uh, we hope we, we could have a, an future episode in, in, in 2025 for sure. Of course. Looking forward to it. Okay. So then, uh, both of you have a Happy New Year because we are on the date. 35:38 And, uh, see you, see you next time. All right. Take care. [outro music]