View Recording: New Year FinOps: Smart Strategies to Create Budget in 2025

Ayman Mageed 0:07 Hey everyone. Good morning. Chinzo Davaatseren 0:12 Good morning, everyone. Amy Cousland 0:15 Hey, man, are you gonna share your screen? Your slides, OK. Ayman Mageed 0:17 I will. Yep, Yep I will. Let’s do it now. Can I get a confirmation please? Awesome. Amy Cousland 0:30 Yep, it’s there. Got it. Ayman Mageed 0:38 All right folks, I am gonna try to. Do some finagling of my screen here, so if you see me looking up here, it’s ’cause I’m trying to look at your chat while I present. Shinzo, if you could do me the honor of answering any chat questions or engaging in the chat would be great. Chinzo Davaatseren 1:00 Sure. I’ll engage in. Ayman Mageed 1:04 Awesome. Cool. We’re starting to get some people rolling in. Give it one more minute. And then we’ll get going. All right. Go ahead and kick off. Hey, good morning, everyone. Thanks for joining us for another set of our webinar series. This webinar is going to be focused on cloud financial optimization or fin OPS as it’s known in the industry and just kind of walk through what we are seeing, maybe some definitions of truly what is phenomen, how can you tackle? The problems that it that it kind of creates and so we will kind of go through that. I have joined so my name is Aiden McGee. I’m the cloud practice leader at concurrency. I’m joined here with me is chinzo. Chinzo happens to be one of our senior cloud engineers and wanted to kind of go through. This topic specifically given that some of us are going into New year looking at financials, looking at budget and want to be able to understand how, how can we take control of maybe the cloud sprawl that we’re seeing in our organization. So first thing that we want to kind of share is actually on the left hand side you have a example of what is very common. If you notice the the time stamp, it’s just this January that this was posted. And specifically what we’re seeing is because the cloud environments allow us to have all this unlimited scalability, this ability to. Get a specific workload or compute power usage with the click of a button. What happens is there are mistakes that can also happen, right? We have situations where you might turn something on. You may not be aware of the parameters that need to be placed on a certain type of compute usage, and you have something similar to this where you have an individual waking up to a spike. In in their cloud spend. And so while I wish them the best of luck in trying to work this out with Google. The main points on the right hand side is kind of what we want to highlight when it comes to your cloud spend. It can be a force multiplier. However, when every when we go from when we shifted from data center operations to a cloud operation centric model, we. Gave power to developers and different users on the cloud that we never gave before. Think about. The procurement process of a new server stack or network stack that had to go through a procurement process, you generally waited three months or so for it to get procured, installed and ready for you to go. And there was gates on how we can manage those costs for we traded that out for the sake of speed and agility. But we also created a situation where now engineers and developers have the ability to spin up net new resources and so. If we don’t have certain guardrails or abilities to track what we are spinning up in the cloud, we can find ourselves spending a lot more than we anticipated. And oftentimes what we’ve seen is that they’re well, well-intentioned. It’s not like someone is specifically malicious in that nature, but what we’re seeing is that they are wanting to. Create a resource they’re not aware of the the cost implications behind it, and then they end up. Over leveraging on on the the cloud budget. Gina, see that there’s a. Hannah. Yeah. Anyway, feel free to throw your questions in the chat and myself for Chinzo will address them as well as we go through this. So I think the most, most important take away here is it’s important for us to leverage cost optimization techniques to make sure that our cloud spend does not get. Over bearing and what we want to do is understand the techniques that we have available to to keep that under control. When it comes to cloud. And the idea of the financial optimization piece of it, we really want our cloud journey to match our business risk and our business objectives, right. And so when we think of your typical ITIL or ITSM based model of people process technology in order to maximize how we’re using the cloud and how we use fit OPS, there’s a component in each one of these categories, right? How do we understand? How do we bring understanding to people so that they understand how to operate with a with a fin OPS mindset? What processes do we have in place that are going to allow us to identify opportunities for cloud optimization or understand how we’re going to refine or actually go about optimizing it? So when we identify a cost saving opportunity, how do we go and execute? How do we validate that is actually something that we can execute on versus something that might cause a degradation in security or a degradation performance and then ultimately technology comes into play not just in the cloud vendor that we’re using, but specifically what tools are we using or? What methods are we using that are that are going to enforce or inform that process or help inform be informed by the process? So what I mean by that is there’s different third party tools that we can talk about later on in this conversation. Or there’s the native tools that are available within Azure that allow you to track your cost. And the opportunities or or potential savings that exist and so try to kind of chat touch on all these topics, we won’t be able to in, in the time that we have, but if there is an interest in kind of chatting more about what your fit op strateg? Would look like. This is probably a good starting point for us to to have that conversation. So what is fin OPS? Fin OPS is not a topic that we have created is it is not a buzzword that that concurrency is created. It’s actually, oh man, it’s about. It’s at least four years old. I would say the concept of phenot Foundation was created. They’re actually part of the Linux Foundation, and so the Phen OPS Foundation created kind of these frameworks of. Inform, optimize and catching myself on the oh man, I forgot the third one. But there’s there’s three components of of phen OPS, and so when you are. Looking at cloud spend specifically, Soph, is cloud focused and when we when we talk about the word fin OPS we’re talking about, are we spending the appropriate amount of money on what we are receiving in the cloud. So it’s not necessarily just about cost savings, it’s more so am I willing to pay this price for the benefit that I’m receiving? And is there a way for me to continue to receive that benefit without having to pay as much right? And so the best way to explain it is so in NIST there when when Cloud first started maybe many moons ago, there was a definition for cloud and and I I I prefer this definition, I think it was a 2008 definition for cloud and instead that. It is a utility that is used using compute storage etcetera and the word utility stuck stuck out to me because. As organizations, we do a really good job of trying to optimize our utilities. Think our water, our electricity usage, right. If you have a large parking lot outside your parking lot might be on an auto timer where the parking lot lights only come on at dusk and they turn off, you know, at a certain time and. And your water in your in your company. Restrooms might have your sinks. Might have kind of an auto shut off. So it’s not available for someone to just turn on and accidentally walk away from, right? And so we also have auto dimmers auto turn off switches for our lights in our offices. And so we as organizations have gotten really good at optimizing our utilities. And so I’ll go one step further and say if we treat cloud as a utility that it is, is we should take those same practices and apply them into our clot operations. So where are their areas for us to do that? Auto shut off that time based spin up spin down so that we can save. Money. Your your. Appetite for spending on the cloud might be different than mine. I kind of get in trouble for this one, but generally talk about water conservation when I wash my dishes, I keep the sink running and to me I want to be able to continue moving through all of my dishes and you know, wash, put away, wash, put away. And my sink is running. I know that I’m going to pay for that water. Others might be willing or not willing to do that, and they want to either have a soaking sink or or run the water only for a set period of time and then turn the water off. And so that’s a perfect example of as a utility, the differences in what we are willing to spend, what we’re not willing to spend. And so ultimately, we want to be able to maximize the value. Of our cloud technology and what is bringing us, we want to create a cost aware culture. I am aware of my water, my sink being turned on and ask me, continuing to to wash the dishes. I’m aware of it. There is a bill that I get at the end of the month that tells me what my water usage was and if I want to tone down my water usage or or that bill. I I know certain areas where I can cut back. On the consumption and then ultimately we want to be able to have those controls set an aware and transparent for everybody so that we can make decisions as an organization. So it’s not just an it function. While it might own the bulk majority of it. Yours there are numerous CF OS that I’ve spoken to that are this is a topic. This is this is a major topic. I think all of us can agree that when cloud first came out, there was this push to move into it very rapidly. We did a lot of lift and shifts. We did not modernize our organization to be more cloud native in utilizing kind of the best in class scalability characteristics that cloud has available. So we we we might have shifted our data center into the cloud and done nothing and now we’re we’re seeing as we’re adding additional resources and this this really seems like it’s more expensive than maybe on Prem data center. Why is that? Well, we haven’t done those. Those fine tuning, taking those levers to to fine tune and some of those levers are easy. Some of those levers might take some refactoring, some remoditization of the way that our our legacy applications are. Have been have been developed and and need to be redeveloped or or kind of tweaked for where we’re going. And so I will. I’ll leave this kind of plug on the side if you’re interested in more understanding of the phen OPS components and characteristics, please go to the. Phin OPS Foundation website and there’s this is a community driven framework and is amazing what people have gone into for detail and then providing in that space. We can probably put the the link here in the chat. So what are some phenom foundational elements? This is this is actually developed by us specifically on. Where do I start? My like to answer the question of where do I start when I’m looking at fin OPS. I get it. I have some cloud spend. Where what do I need to do? And there’s six foundational elements here. I would highly recommend not. One of them is not more important than the other, but here’s my logical take on where I would start. I think ultimately I want to understand what is my cost, what what is my total cost of ownership. What is my current spend and who’s who’s spending it? You know, one level down would be I don’t want to just know that my cloud is costing me X amount of millions. I want to be able to say, hey, this department or this application or this cost center takes up X amount of my cloud spend. There is going to be a shared services component where maybe the network is shared across all cost centers and so we can talk about how do we how do we show that transparency in a in a equitable manner. But ultimately I want to understand as close to real. Time as possible. What is my cloud spend once I can understand what my cloud spend is, I can start putting in place processes governance. That informs the decisions that I’m going to make. So if I have a biweekly meeting with my. Direct reports to talk about the opportunities for cloud savings and what they’ve done. Action items that they’ve taken on there. I can then have my monthly meeting with the CIO and or the CFO and talk through. Here’s what we’re doing. Here’s what we’re projecting. Here’s the savings that we provided. Oh, by the way, these savings can then go and fund. Other components of our of our organization might be able to take you know 200K in savings and go get that additional incremental headcount that you’re looking for or you can take that 200 K and fund a project that maybe would have been delayed till a. Of the year. And so this is this. Is gotta watch my words here. This is not. A savings realization because of the way that we like that cloud is an OpEx fund. But most CF OS will agree that it is definitely cost avoidance. And so if I can avoid. An X percent of savings or of of cost. I can generally negotiate my way to using those using those funds for something that would be more a higher return on investment. If I had, then if I had just done nothing. And so kinda go back to the foundational elements. We’ve got transparency first. Process and governance gets refined. Once we have those numbers because we want this to be repeatable. We we don’t want this to be just a one time Band-Aid fix. We want to have a process where we revisit this set period of time. We have some controls. Maybe we are coming out with policies that state. You as an organization can only spin up the following skews of virtual machines, right? Anything outside of these SKUs require some sort of authorization, some sort of exception to the policy that allows us to manage our fleet much more efficiently. That kind of goes into the people tools component of this where our people need to be aware of our processes. Our people need to be aware of having kind of that cost conscious mentality as they spin things up. Our open source tools out there that as you’re spinning up a specific VM through code, it’ll actually populate what the cost is by hour. And so in that in that scenario, you’re actually empowering the developers directly to be able to see what the cost of what they’re SP. Up is at the console as opposed to waiting. You know deploying it, forgetting about it and waiting for the the shock of the the. Price tag. And so there’s a there’s a people component of who’s involved in this process who needs to be aware of the work that we’re doing, and then how do I build this into our day-to-day culture? The tools I mentioned earlier we have in Azure, for example, we. Have like the Azure cost advisor and cost analysis tool that gets us a portion of the way the the other portion is. What am I going to be reporting off of? Am I going to use native reporting technology? That’s Azure for example. Or am I maybe gonna take some of that, connect it through an API or an export and get on Power BI where I can slice and dice this a little bit better? Maybe I don’t wanna put in the effort of building my own power Bi dashboard and so I wanna go and procure one of the leading tools out there right now to be able to give me a little bit more granular approach to how I’m reporting this. So there. There’s a plethora of tools out there, some that help for the reporting aspects, some that will actually with some automation conduct the optimization efforts for you now. I’ll practice that. I would not want that tool doing something in my prod environment without testing it, but for my dev and QA my test environments. To me, it’s almost a no brainer to be able to have a tool with set parameters that’s kind of operating up and down. That that scope that I’ve given it and allowing me to save some money even when I’m not actively managing it. And that kind of leads us into that DevOps architecture space when we talk about scaling this out, we want fin OPS to be something that isn’t always done manually. There is ways that obviously we’re going to go into the console and we can just turn things off manually, but what we really want is some sort of automation set in place that allows us to, you know, maybe it’s not automating the whole end to end, but maybe. We have automation that gets us to awareness, gets us to better transparency and then we can have those conversations as an organization to say, OK, here are the cost centers that are spending the most money. This cost center makes sense because it’s our major application that we run the business on. Awesome. This application does make sense, but I see a increased amount of spend in the dev QA environment. Is that expected or do we need to do something about it? Maybe we can save money there and so. It really comes down to aligning our fin op strategy with things on the right that we see having a cloud centric it strategy, having value management at the forefront of how we’re deploying cloud workloads and then ultimately how does it align with the business strategy because if at. In 2025, if your IT organization is not aligned with the business, that’s probably the first conversation you need to have. And so we generally want to be nesting our IT strategy underneath the business strategy and understanding. How the components that we are driving towards in it are supporting each one of these business objectives that we have for the year? So what are the major components of successful fin OPS? I kind of went into it on the foundational elements. These are some additional areas that. Oh, here it is. Inform, operate, optimize. That’s the fin op. Those are the three fin OPS core characteristics from the FIN OPS Foundation. But as you go into this, I won’t drain the whole slide as we go into kind of the fin OPS mindset. We really want to look at these again 6 components, the collaboration, how we’re driving business value to the business. How we can bring ownership to everybody, not just the the director of infrastructure, but actually bringing it to the individuals that are deploying those workloads. How do we get good transparent data in near real time? Is there a fin OPS team? We we have a slide later that talks about, you know, do we do a FIN OPS team? Maybe we’re not the size where we can have a dedicated team for this, but we can definitely have. Some additional duties and our components that are tagged in. So you have a persona of who is tackling these types of functions and then taking advantage of cloud the way it should, the way it was intended to. I mean, we’re in 2025 still talking about how do we prevent? You know virtual machines from being on when they’re not needed to be used. How can we automate the turning off and on of a virtual machine so that it can save us money? We only pay for what we use. Are there ways to move away from AVM? Go into containers. Is there a way to even move away from containers and use specific serverless functions or event based execution and so this is this is that that crawl walk run component of your cloud journey and PHN OPS need to just like security needs to have a component in that. As you continue to. To mature in your cloud execution. So I’m gonna. I’m gonna pause. I’m gonna let Chinzo kind of walk us through some of the capabilities and things that we want to start tackling as we as we start having the conversation of like what is our fin op strategy. What are some areas that I need to hit on so that we can be successful? Chinzo Davaatseren 23:52 Hi, guys. Can you hear me? OK, awesome. Ayman Mageed 23:54 Good to go. Chinzo Davaatseren 23:56 So, like Amy said, Phenoms is a journey. It takes organization to call and then walk and run. Been up since basically in in the business side of it, it’s combining business and financial and technology teams together and I’ll you know, being in this industry, we’ve seen a lot of people try to implement phen OPS governance in their organization and technical aspect. We can apply policy. And technology and. Other rail guardrails to monitor this, but we have to start measuring. How are we spending and how’s our infrastructure running? So we need to review. Monthly consumption rates and then we need to monitor how are, you know, shared services coming back. So I can share some of the stories that you know being in the industry. The CEO would ask from the CTO, OK. We want to expand our region to different region. How much will this cost me? So this is an application reservation application and this question a simple question. If you want to expand into different region, how would we look like? What’s our budget look like so? If we have a successful fin OPS, governance or journey. The CTO, the product owner will be easily sated. Currently we are consuming this much amount and then our user base is 1000 people. If you want to increase it to 2000 people, this would cost us. How much right? So we can go to the next piece. The important aspect of our cloud optimization is not only technology, but also people who are raising like tracking and reporting on this related. Expenses. If you go to next slide. Some successful roles in Fin OPS operation, the product owner who’s actually owning the application or the infrastructure, not the CEO level product owners are the critical aspect of, you know, Metro phenoms practices. And when we also do the infrastructure architecture, how much, what type of virtual machine, what? Type of. Storage account are we using? Is it being? Is it being utilized at all time or cold versus hot? Storage systems and then part of counter and infrastructure architecture and cloud system architectures are the dedicated people who are who can, who can achieve this cost optimization on daily base level. But they will report up to the the finance and business aspect of it and. There’s a accountability comes within the organization. So how much are we really spending? Is this appropriate resources to allocate efficiently? Are we using our technology? Are we using our resources efficiently? These are the questions or these are the aspects we can monitor with with the roles. So these roles are not complete brand new roles. These are existing roles within the organization, but we just add the the saving mindful saving in within the organization and having more control and accountability. Df each organization or each team to have. What is your utilization? Why are we spending this much and have? Control and visibility for spending. So there’s a few ways we can start tracking our spending and then by department, by region or either workload. Once we start. Tracking and reporting and spending, we can forecast if you were to add, like I said, 200 more clients or 200 or expanding into different regions. How does it look like so mature maturity aspect of it is not only we start keep tracking of it we report. On it. Now we can forecast into the future. There’s two ways to do it. Cloud cost allocation and tracking a visibility of. Either on the subscription level or on the management group level. Can we go to the next slide? Amy, please. Ayman Mageed 28:40 Yeah, I think that’s a good point is that, I mean, I’m counting about 12 boxes here and that doesn’t. That doesn’t mean that we need in order to do fin OPS. We need to have 12 more people on the team. Chinzo Davaatseren 28:44 Yep. Ayman Mageed 28:50 I wish right and so. There is. There’s something to be said about some of these roles that currently exist, and the fact that we. Are shifting the mindset of not just oh, we’re trying to spin up another VM, but training people to to ask that that question very similar to. Hey, your CEO just texted you asking you to go buy a gift card, and they sent you a link to to Venmo them or whatever it might be. We have been trained, hopefully to know, OK, my CEO is probably not texting me, and they’re probably not. I prob. Shouldn’t touch that link that was sent to me. And so there’s that moment of pause where you identify information, you review it, and you say, here’s what my action is going to be. In the same way, when we say, oh, it’s been up another virtual machine, let’s put another domain controller out there, whatever it might be. Maybe we need another firewall? For whatever reason, and we can stop and ask the question of do I need another resource? Is there a way that I can do this? A different way is there if I do need another resource, do I need? A. A virtual machine that I can mine Bitcoin with? Or do I need a virtual machine that? Is specific to what I anticipate the workload is going to need from a compute or or memory perspective, right? And so having that thought process built into all these individuals and the the people within our organization that are operating with this technology will will incrementally change what your trajectory for cloud spend is. Let’s say you’re spending increases 10% every year in cloud because of just natural. Business growth, that’s great, but your cloud spend does not need to increase 10%. At the same pace as the business, it can actually have been flow depending on what your Technology Strategy is. And So what we’re looking for is if I can take 1%, two percent, 3%. That’s a significant difference over the horizon of our cloud spend. I will go to the next slide. Chinzo Davaatseren 31:13 So there’s a few immediate with saving the opportunities there. Microsoft provides long term commitment based savings you can. Downsize some of the resources and then, like Amy said, snooze some of the resources. VM container pods doing off hours. You can turn it off or scale it down. During weekends and holidays or during if, let’s say your business requires you to have 24/7 availability during busy seasons, you can scale it up with a reserved instances, reserved instances, early committed. And then you can save a lot of money with that. Also you can do automation how you deploy and how you turn off some of the resources using eks or other other tools, but this is the immediate cost savings that you can apply within your organization. Today, however, there is a midterm or long term opportunities. So the tagging comes very handy for better cost insights and optimization. Are you tracking it within the department? Project workload, even geography. What are the spending looks like? And there’s ways to apply tagging either. It’s on a post deployment of the resource or doing deployment or even how you sunset some of the virtual machines. How much did it cost you before? And then why was it turned off? At the beginning of the presentation, Eamon had very interesting case where $50 monthly cost was going up to $70,000. How can you prevent this within the organization and set up reports and alerts based on the the cost metrics or CPU metrics that that alerts? OK we are seeing high consumption. On this compute, what’s happening in this department or what’s happening in this resource and team, you can track and have transparent visibility into their story. It could be that they’re migrating into new environment, or actually it’s a business case that a lot of clients are potential clients are logging into your system and purchasing. You know your inventories, you’re purchasing some of your products. Another part of way to improve in long term and mid term cost saving is develop a optimization cycle mode and have a team meetings. How can we improve and optimize some of our cloud economy? Can I just put a a single policy and then hope that everyone follows it? We actually follow up with the people. How are we doing? How did we look like from the last quarter to this quarter? How we gonna do in the next quarter and then this will give us give us an opportunity to identify some some ways to save money. Do we really need this virtual machine? Can we move it to? Pass platform as a service or Saa’s software as a solutions and is there better way to run this code? How how can we run this code? In must optimized way, how can we deploy our code base and application in the cloud native way with the with the saving in our mind, right? So these are the few techniques that we we have the third of techniques, techniques, techniques. Technique you can use to utilize. Have a have some cloud savings in there. Ayman Mageed 35:11 Yeah, I think this is a good point to add that when we talk about PHN OPS, I know we’re talking about cloud spend, but this is an opportunity to get everybody engaged across it. And the business in a certain way. I’ll even go as far as saying, like when we see this. Executed appropriately and and the best executions that we’ve seen. It usually brings in some sort of ticketing process alongside it to track all of these opportunities. And so if you’re using ServiceNow, for example. You can create when you identify these opportunities you can create because we have good tagging, we can create tickets that are assigned to that specific cost center or that owner of that workload and have them review the potential savings opportunity. Have them come back and address that to. The Fin OPS team and say yeah, that is a valid viable savings optimization. How do we go after that or that’s not viable because it’s going to cause a performance degradation in the application for whatever reason it might be. And we can then create kind of a a register of the opportunities that are available and and what we’ve normally seen is infinops, when you’re executing fin OPS for the first time and then and then kind of continuously, you can generally find 15 to 30% savings in. The first. Tranche of work after you get through that first Band-Aid fix the the savings do Peter out. But I have a really good friend and a big financial organization that is running a fin OPS team and is telling me that he’s continuing to find 8 to 10% savings, 8 to 10% savings is a lot of money. Over the years, and so his ability to. Address those savings continuously through automation through. A. A ticketing system. He can track the people that are assigned to an optimization. He can follow up because we know we’re all busy and say hey. Not that we want to put someone’s name in red, but ultimately, if you’re reporting up to your CIO or your CFO and you’re saying, hey, we have a potential opportunity to save $1,000,000 across these different cost savings opportunities. Here are the people that are ***. To those. Opportunities or who are impacted by that. It creates conversation and the conversation can be as simple as. Hey Chinzo, can you justify this 100K in savings that you’re not going after? And if you can, then, great. It’s it’s a positive net net value to the business. And if you can’t, we probably need to go after it and and attain those savings. There’s a whole other topic that we can talk about, which is like the gamification and or the cost realization of fin op savings and. As an organization, we have seen some organizations will have a central location where all those savings go, and then the the finance organization re appropriates those funds for better company projects or or efforts. We’ve also seen it where if you find it and you save it, that money goes back to your cost center. And so we have. We have legitimately seen teams be able to hire net new developers because they were able to find. 300K savings, whatever it might be in their environment. And so even if. Even if you don’t care about the the savings aspect and being a good steward of your organization’s funds, finding the finding a way to negotiate getting an additional head count, or that tool, that shiny new tool that you’ve been wanting to purchase but you haven’t been able. To get the funding for and being able to justify that by here, I’m going to make some cost cuts in cloud area. We’re still going to be business as usual, but I can use those funds. To to maximize another area is a game changer for some organizations. So, so chinzo kind of mentioned tagging and I think we want to go back and and talk about this because tagging is the first step to getting good transparency. If you go, I can’t. I mean, I’m not. I would ask some people to raise their hands here and and say how many of you have workloads in the cloud that are named. Based off of where they’re located, what type of resource it is, is how it’s used, what application it might be used for, and that’s I’m not gonna say that’s bad. That’s actually standard practice. The difficulty that you run into when you have a a resource, a naming convention that gives you all that information. On the on one hand, it’s really good because I can look at resource and say that’s in Arizona, that’s on Gov Cloud and it is dev or it is test and that’s awesome. The issue is all of those things that you just identified. If they’re not labeled as tags as well as in their name. You lose the ability to start slicing and dicing your data from from a data integrity perspective. And So what tags allow us to do is I want to say the running number right now is on Azure. You can have 50 tags assigned to each resource. I really hope you don’t have to get to the point where you need 50 tags, but these are there’s at least five or so tags that you would want to add to. A resource to be able to help you slice the numbers to be able to get better insights for your organization. It’s gonna be what is a business unit aligned to this. Who? Like what type of environment is this? Production, dev, test, et cetera. Is it tied to an application? Could like cause. Imagine if you can tag it with the application and from this point on you can sort all application based resources and get a total cost of ownership for that application. ID. Maybe there’s a specific, yes, it’s tied to the finance department, but I need to be able to understand who is the resource owner of this. Maybe there’s a business owner and a technical owner, so maybe the business owner is. This is SAP so and so runs SAP. But if you have a technical question about this resource, the person that managed the technical aspects of this is this person. If there is a project. That’s ongoing. You can tag like awbs code or. The project name so that you can roll up all the project costs of that specific project and that’s generally like a a temporary tag. You can take that tag off at a later time once the project’s complete, but it allows you to have accurate billing information you. Can show back or chargeback to different cost centers within your organization. So you can go back to the. Engineering organization and say here’s how much your R&D environment has cost by the way. Or you could even say, hey, we’re going to throw this in your budget and next year you’re responsible for maintaining this environment. Let us know if you’re planning on increasing your spend in the cloud because you’re you have a big feature coming out or you have a big release that you need to account for, and so without this tagging structure, you don’t have the ability to enforce governance have account. Across your teams, a lot of automation that we build is based off of or or automation and or visualization that we do is based off of tags. And so just like any. Thing power bi effort. You’re gonna want various different ways to pull insights for your organization. Chinzo Davaatseren 43:15 Yep. Also on that workload intakes, the automation can be built on the traceability side of it. Ayman Mageed 43:20 OK. Chinzo Davaatseren 43:24 So when we have a new workload coming in, we can create a new tag, OK. These are new workloads. What we want to do, the application or migration, data migration intake and we can create a new tags on it. Ayman Mageed 43:43 Yeah. And so because right now majority of us already have a cloud environment and so there is a practice, there’s a process that has to get put in place for tagging the existing workloads. But then we want to have a process in place that allows us to put tags on workloads as soon as they are created. And so you can do that using Azure policies. For example, you can do it on like an audit mode where if someone. Deploys something ahead of time. They deploy something. They don’t put a required tag, then it gets flagged. Then you can go follow up with them. That’s the least restrictive way of doing it. There’s other ways where you can actually require and kick back the deployment of that resource until the appropriate required tag information is on there. And so that’s how you start to enforce appropriate data hygiene going into the cloud. The harder part is most likely the exercise of looking at your existing workloads and doing the tagging exercise. Probably better done on Excel first with an Excel export and then importing those tags into. Azure that gives you a better landscape of what you’re doing. I saw comments. I’m gonna take a moment. Chinzo Davaatseren 44:57 I mean, yeah, we have a question. The part of the struggle is that tagging does not translate into actual invoice. Ayman Mageed 45:02 Yes. Chinzo Davaatseren 45:03 Microsoft, for example, to reconcile. Ayman Mageed 45:07 That is true. There is a component that is the ever elusive line items on the invoice that are not as easy to allocate directly to a resource. There are some tools out there that have claimed to say they’ve cracked the code. I’m. I’m not sold on that. Really, what ends up happening is you go through this exercise and you probably can get. Probably a good 85% of your organization can be allocated to a specific line item on there. What you start running into is when you get into egress and transfer of data line items or like shared services like network groups, that such where. There’s some techniques that most people will use in the industry, such as they’ll they’ll either take that shared service cost and they will, or they’ll allocate it as a shared service cost. Sometimes it’s like the unknown bucket and then they’ll split that unknown across all the remaining cost centers. That’s the like baby step number one. To do that the other way that has been done is by consumption. So if I know that. A cost center or a department is using 60% of my cloud consumption. Then I allocate 60% of the shared costs to them. And then if you can specifically pinpoint certain network security groups or network, sorry, appliances, then the IT organization, the IT cost center might take on those costs. They might say it takes on this because they’re providing a service to the rest of the business. Here are the remaining. Costs that are are split out, so I’ll agree with you, John. It’s not a perfect science, but it’s more transparency and clarity that we’re able to give our leadership than we’ve ever been able to before. And so it’s a step in the right direction. Chinzo Davaatseren 47:08 And our next it’s a great take away to the next slide. How does the leveraging Phen OPS tool within the Microsoft looks like? Ayman Mageed 47:20 Renza go for it. Chinzo Davaatseren 47:21 Oh, sure, so. Microsoft is has their own native tool cost management. Dashboard where you can see how much of your consumption, what services and what location and which of the groups are using how much. And then there’s a you can declare it budget for a subscriptionbased. This is the budget. It won’t go up or it won’t go. It cannot go maximum of this. You can establish those kind of control and also there’s a you can exported this into regular basis into power BI or other formats. I’ve been playing with. The the copilot on Azure on how we can see a detailed messages on the cost management side for so applying. ChatGPT models on top of it, you can ask questions. Hey, how much is team a consuming at this moment? And how does it look like? What was the historical rate of this team using this? And then you can present this evidence to the finance and business teams and then it will help you make a decision easier and faster. Those cost alerts. You can set it up in each quarter or month and then you can even export it to compare those to the. Compare it to the invoice from Microsoft. So you can run a ChatGPT on top of. This is the invoice I have received. Can you please compare it or try to allocate what was the resource? How much was it used? So that gets you the 85% easily with the ChatGPT. It’s not a manual anymore. Ayman Mageed 49:11 Yeah. Chinzo Davaatseren 49:11 Do you have anything right? Ayman Mageed 49:14 No, I I think this is you can only do so much with the tool that Azure provides you and so this is a really good use case for it. Chinzo Davaatseren 49:20 Yeah. Ayman Mageed 49:24 There’s in private. Well, I don’t think it’s private. I think it’s public preview now. There’s a there’s a way to do cost allocation in the tool. The only downside, and this is across the board with any cloud provider. There’s no way to create your own cost allocation and have your bill reflect that, and so your invoice will always be that jumbled, massive jargon ’cause. That’s how the cloud providers track their skews and and their services. But you can create your own cost allocations in a dashboard. And it tries its best to take components from. From the the different resource groups. The only downside, in my opinion there and like many other tools, is that it doesn’t apply those tags to those resources indefinitely. It’s just a kind of a data overlay that’s using the filters that you’re that you’re providing. I have not tied this with fabric. I had a conversation with someone. John, this is great that you’re bringing it up. That has mentioned that you can actually bring various sets of data into a model and create that same the same individual that I was mentioning earlier. Buddy, that’s at a financial services company. He actually was using to to the leading tools in in the fin op space and decided instead of using that tool he’d rather create. He’s on AWS, so he’d rather create a, you know, like Athena based model that. He has a data lake with certain sets of information. That’s the API calls that are going into the data lake. And then he has a virtualization, a visualization component that allows him to query the data lake and get those insights for his company. So he’s able to get what more granular information. So he was still in the middle of developing it. The only downside with that I think this is like what? There’s two down sides with that one the GUI in the user interface is not as slick as some of these companies that have millions of dollars focused on developing that. That’s one that’s minor piece. The the second piece is that you put in a lot of effort maintaining. That as an application, let’s call it. There’s a lot of effort that goes into it. And then what ends up happening is the cloud provider tweaks or changes one aspect of their API and your model goes all wonky. And you have to go sit there and troubleshoot why it’s doing that. And so it does create a little bit of overhead unless you’re aware of how everything is running and kind of stay up to date on it. Chinzo Davaatseren 52:10 Yeah. On addition. Oh, go ahead. Sorry. Ayman Mageed 52:10 ‘Cause it good and before. Chinzo Davaatseren 52:14 On addition to that, I was just wondering so service Fabric standalone you can deploy service fabric on the Windows Server and then that will give you a cluster either on premise or on cloud. That way you can. There’s ways you can tie. How much is it really costing you? But you’re losing some of the throughputs that Azure provides if your premium? Fabric services on the on using using their. Their infrastructure. There’s way of doing it as well. Ayman Mageed 52:52 Sorry, mom. Yeah, that’s actually a very interesting proponent. We haven’t had the opportunity to go delve deeper into that one. So I guess a good segue though is whatever method you use for your for your reporting or your cost transparency mechanism. Ultimately, there’s certain metrics that are fairly standard that will help guide your your insights and and here are some examples. Most of these can be visualized. It’s important to understand that it’s just like any data. It’s garbage in, garbage out. So we wanna make sure that we’re we have good data hygiene, but cinzo kind of alluded to some of the questions that can be asked and answered. When you have the right data and the data allows us to answer questions like what’s our what’s our monthly cost, how? Has it changed by month? What are the top resources that are causing this? Spend one of their resource groups. That’s that’s easy. That’s like elementary of of the fin OPS piece because we can get that off the cost management dashboard in Azure. What we really wanna answer is who’s that resource group tied to what application is it serving? If you’re an organization that relies on digital enablement from your cloud? Applications. Then I wanna know what is the revenue that’s tied. To the spend that we are doing, that’s like that’s the next level of. I think I think the go to example is usually like think of like Goodyear Tire, right? Is there a way to get to the point where I say for every tire the the cost, the unit economics of a tire is $1.20 of cloud spend and so for every tire that we roll out of? This manufacturing line it costs us $1.20 of cloud spend. How can we get it down to? 118 per tire that 2 cent savings across X million dollars of our X million attires that we roll out in the manufacturing line equates to this savings. So we we can get to a point where it’s never going to be exact, right, but you can get a. General estimate of how much is our overhead. What benefit are we achieving from it? Where are the portfolios that we currently have the most cloud spend and is that a suitable amount of spend for what we’re receiving or are there or are there ways to potentially look at the ROI of refactoring that application or the way that it’s architected to save to? Squeak out an extra 30 percent, 20% savings, and then from an operational perspective. You know what? What resources have the appropriate tags? How many non compliant resources are out in our environment? How do we go about making them compliant? How do we understand? SKUs that are maybe our reservation, our reserved instance strategy is to ensure that we have only these five SKUs in VMS because we’re throwing, we’re buying reserved instances against all of these Sku’s. And so if we have people spinning up incorrect skews of Vms, there might be one of these five Sku’s that’s more compatible. Or closely compatible to what they need. And if not, we give them an exception to policy for that specific SKU, but ideally we want everybody building in these five SKUs because that is where we get the most savings as an organization because trying to manage reserved instances for 1510, eleven different SKUs is is. A big pain. So you want to try to standardize as best as possible. Now I’m not speaking in absolutes, but this is just kind of ways to make life easier for yourself. And then ultimately. What is the utilization of our resources average. You wanna look at both average utilization and peak utilization ’cause, there are times where a component doesn’t get used until the end of month. Maybe it’s used for end of month building or in a month invoicing and accounts receivable and. So we need to be able to be aware of, you know, is there a way to spin this down for a period of the month and spin it back up as we anticipate the seasonality of that peaks and valleys or is this something where we just need to? Keep this on. Maybe it’s something that can be a function Azure function instead of a a server that runs all of this is assuming things that I can’t answer for you today, which is is the application or the function compatible with that type of new technology, or does it require some? Sort of server based. Effort. I know we’re coming up on time. This deck will be shared with everyone. Here is a very high level. Don’t don’t share this with these guys, ’cause. They’ll probably come back and send me a nasty gram. But these are some high level features and functionalities of some of the top tools in the space as it pertains to how they compare with Azure cost management as well. And so cloudability now is actually cloudability and turbonomics are now owned by IBM, so they have kind of. Taken over both Cloudability as a cloud management software turbonomics as an automation tool that allows you to execute automatically some of those cost saving opportunities. Flexera cloud health big. Big competitors to that tool set and then cloudchecker. It’s kind of an up and coming emerging tool set most of these tools on the right hand side of Azure cost management have some sort of model that is based off of your total cloud spend anywhere from. 2 to 3% of your cloud spend. But the benefit there is if you’re able to get the transparency of the data and get after saving 10 to 15% plus more of your cloud spend, the tool begins to pay for itself. But you really need to have some sort of process surrounding that, or else you’re just adding another tool that’s not going to be utilized appropriately. And so as we wrap up, thanks everybody for your time. There is a survey in the chat. Please fill it out there. It gives us feedback on the organization itself and kind of the webinar, but also if you want to learn more or are interested in any of these fin op strategies, feel free to throw that in there so that we can reach back out and continue to have that. Conversation. So I appreciate it. Thanks again and have a good rest of your day. stopped transcription