 Hi, yw'n gweithio. Fy fawr hynny yw'r parwysgwrn gyda'r hollol? Yn fawr, mae'n amgylchedd y gallwch chi'n gweithio i lawr am ymwneud â'r masgau arall a'u gweithio i ddim yn cyfael â ymddirionedd ymddirionedd. Felly, mae'n gweithio'n gwneud hynny. Rwy'n James MacLeod, dyrech o'r cymuned ar Finos, ac dy'n hynny i gweithio i gael os ymddirionedd â'r hollol ar gyfer cyflau $1. Before we actually start going into the panel discussion, I'd like to introduce Jen and Udam, if you would like to introduce yourselves and let the audience know who you are and what you do and what we're going to be discussing today. Great, thank you. My name is Jennifer Hayes and I've worked for Fidelity Investments. I am also a chairperson of the governance board of the Phin Ops Foundation. I as a member of the Fennops Foundation Governance Board I am actually a practitioner member so I represent the practitioner community. So with infidelity investments I helped establish the first Fennops organization with infidelity and I have helped run it for the last 2 and a half years. pound in. Hi and I'm Udam Daivaraja, headed of Cloud Financial Management for city. At city, what we are doing is building a infrastructure cloud platform Ie, mae gennw'n gweld gwelio am gwrthau papunau, a phatwch yn cyfran o bobl yn amlwg nhw, sydd ei wneud i gwneudio i eirio sy'n gwaith sydd yn styl 좋아하는ol. Yna cyfwyr yma llawer i'r Chelyw Pethau Fellygrwyledig ceisio'r unrhywg llyfrigol. Yn y ceisio, yna ddarfodol. A'r thymau i ddechrau, rydych chi eisiau cwestiynau, fydd arall, nid ei wneud, a chydwch chi'ch cwestiynau yn eisio y gallwn a ddweud y dychwyn, llawd y dychwyn i'r amser, yn oed i'r hoffa i'r ysgulinaid. Felly'r dŵr y dychwyn, Jen, roeddwn i'n ei ddrwynt chi'r hyn yn gwybod. Achos, James. Yn y ddweud, roeddwn yn cyflwystio ar hyn i gyd yn eu bod yn byw ymuncheg iawn, a'r llead o phryg f assassinol, llefniad, proses, greu cant allan ymwiel, of people, three different types of roles within the organization, your financial organizations, your technology organizations, and your business organizations to collaborate and make data driven decisions based on how to get the most value out of the money that out of what you spend in the cloud. The real purpose of Fin Ops and the goal at the end of the day is that when you're spending money in the cloud, it's not about necessarily reducing what you're going to be spending on. It's about choosing the appropriate place to spend your money on, making sure that you have the most discounts available to you as an organization, that you're working, you're managing your workload and your resources so that you're not over provisioning. So you're really talking about it from the beginning of the process which is procuring and deploying resources and doing that appropriately all the way to the financial aspect about it, which making sure that you are not overpaying for those resources when you do them. Fin Ops itself, we think of it in four different ways. We have a set of best practices and principles that we bring to play every time we think about how we want to set up a Fin Ops organization. There are a different type of personas who will play within the Fin Ops organization within your group and there are different phases by which we go through. First, we'll talk about the principles itself. So from a principle standpoint, this team has all of these different personas, the business, the technical groups, the procurement groups and the financial organizations. It is important that they communicate and collaborate together and create a common understanding of how they're going to pursue cloud spend. Everyone in the organization takes ownership for cloud usage. Traditionally, what you used to see is that small groups within different organizations made key decisions about procurement and cost of IT. That no longer comes into play. Within the Fin Ops, I apologize, within cloud-based organizations, what you actually have is that every person who is able to deploy to the cloud is actually able to make a procurement decision for you. So it's really important that you change the culture of an organization and really have it focused on cloud usage and the ownership of cloud usage. We do recommend a centralized team within the Fin Ops community that helps drive these principles throughout the organization and the culture change that needs to happen with it. At the end of the day, decisions are business decisions. They shouldn't be made from a technology standpoint. They shouldn't be made from a financial standpoint. We should be bringing information together that allows the business to make the right decision for the organization itself and really be able to take advantage of the variable cost nature of the cloud. So how do you use that as a business advantage versus it actually bringing more overhead to your organization? One of the key statistics right now is that 35% to 45% of all consumers of AWS overspend over provision on their cloud. So this is really the key underlying principle to the Fin Ops. Thank you very much, Jen. That was a very good overview. So, Udam, now that we have a basic understanding of Fin Ops, can you tell us what value it brings to financial services organizations like Citi? Yeah, I'd say Fin Ops is critically important for any organization looking to operate in the cloud to gain that efficiency. A lot of the things that Jen talked about here, right? But especially for larger organizations, there's a lot more complexity because of different legal entities or lines of businesses or countries. So having Fin Ops, which has these best practices, really helps the organization go from maybe a place where they weren't so familiar with the cloud to operating very efficiently in the cloud. So I would say having that, you know, I call it a GPS for cloud financial management, right? Because it really helps teams navigate that this new environment that they're getting into where things are very different, right? You have the variable spend model. You want to provide things like the visibility into cost and usage all the way down to the edges of your business. These things are new concepts for these larger organizations. And also the collaboration that Jen mentioned, right? Like that's a new thing because generally you had those conversations maybe twice a year, once a year, when it was time to do budgeting with cloud, you want to have that continuously. And you want that to be a value-based discussion that engineering the business organization and finances having, you know, because your usage is going to evolve and change frequently. So with that description, would you say that Fin Ops actually removes organizational siloes? I would definitely agree with that, yes. That's brilliant. So Jen, who in the organization should be involved in Fin Ops? So we talked about the personas already, right? And that's a really important piece that there are different groups that have different levels of needs out of the Fin Ops organization, right? And when we talk about where it should reside within an organization also, usually what ends up happening within an organization is you have multiple people start to think about cloud costs, right? And how to manage it. The finance groups tend to be very concerned about it. They feel like they're losing fiduciary control over because due to the variable spend and the nature of it, their current controls and their processes aren't really good at actually dealing with on-demand resources that result in variable spend. So you start to feel the finance organizations start to feel anxiety actually about moving into the cloud and you probably have all experienced that with your finance organizations. The technology organizations, of course, they see all the advantages of the technology itself, but you'll start to see the CIOs and the executives within the technology organizations begin to be concerned not only about the budget and the cost of it, but over provisioning of resources too, and how you can actually start to leverage the information. And so what we see is that in general, in many different organizations, technology and finance is the place that these start to funnel up with each other. And you can actually see in different organizations the phenops either reside in the finance or the technology groups. And they will tend then to lean on what type of activity that they want to focus on first. If they're in the technology organizations, they will lead to resource utilization and usage and how to control that portion of it. And if you see them bubble up in the finance organizations, they would tend to think about how do we do budgets and how do we do allocation of resources or cost back to the appropriate business units, how do we reduce that cost? They'll be really focused. The best type of organizations probably bring those collaborations together. And actually what we did within our own organization was had co-sponsors from both the finance and the technology organization to help manage and make sure that we didn't go one extreme over the other to break down those silos. That's awesome. So we were talking earlier, both myself and Adam were speaking about the similarities that phenops could have to an open source program office where you have different people from across the organization coming together in order to remove blockers from that aspect of the organization. Would you say that phenops is similar to that type of self-organizing coming together of different stakeholders? It really is. And it's one of the big principles of this, right? And Uddan talked about how it breaks down these silos. And that's what you want to have happen. One of the things that when we start to educate people on phenops that we focus on first is actually creating common language between the two different groups because one, there is a completely new language that comes up with the cloud versus on-premise and technology resources. And two, you need both the finance organization and the technology organization to understand what each of them needs in order for this to be successful because if you don't have collaboration on this, you know, we have seen places of where, if the finance organization tries to control things too much, right, they try to limit the amount and the usage of innovation and the usage of the cloud-based resources. And that's not actually what you want. You want them to be informed and understand the value that these things bring and then they can bring their set of skillset to help reduce cost in these different areas too. So, being part of digital transformation and, you know, seeing how Agile and DevOps actually transforms a big enterprise in terms of culture, would you say that, you know, bringing phenops into an organization also, you know, has an unintended consequence, you know, of changing the culture of an organization? I would agree with that. I think, I mean, the collaboration piece, not to harp on that, but that really is core to phenops and that, I think, then leads to people having value-based discussions in their organization, right, because we're always strapped for resources, always, you know, having too many things to do in engineering, but if you can have this kind of data-driven decision-making, then you really end up having, you know, better decisions made for your business. And to build on that too, you know, one of the nice things about what Agile and DevOps brings to an environment, brings to organizations, right, is their ability to increase velocity. And finance decisions have traditionally been very long in the making, right, budgets are done once a year, you do these very long studies to see whether or not you should spend money. The cloud, as all of you know, the transparency of cost is amazing, right, and the real timeness, so we tend to get, you know, updates even during the day about our usage. They build based on sometimes seconds or fractions of seconds of cost, and so there's this amazing amount of data that is given in near real time that you can make decisions on. And so one of the greatest things about FinOps is really moving that whole-time table forward and really thinking in a very agile organization of how do you bring finance into a really agile organization and make good cost decisions based on near-time decisions. Absolutely. I remember previously we were talking about accountability and how FinOps actually brings accountability into teams, you know, who are actually developing and engineering, you know, with cloud-native solutions. Can you tell me how that is different from a CIO who's doing upfront budgeting, you know, compared to implementing FinOps, you know, where you've got near real-time data, you know, and the ability to adapt fast to, you know, impact and change within your application? I mean, I would say the iterative approach emphasized here is really key to that, right, because as you're getting this data, you're iteratively being able to, you know, make changes to your plans, right? And as a... And I understand, like, there is an importance to that upfront planning, but that planning also needs to account for the fact that your business needs, you know, change, your engineering solutions might change. So having this visibility, I think is super critical, you know, for this kind of change when it's happening in your organizations. Yeah, I would also say that one of the ideal situations, right, is unit-based metrics and economy with the cloud, and it's really being able to understand and make good decisions on real information about the cost of something. So traditional data centers, they tend to, you procure and you're having to take those costs for years, if not decades. With cloud, of course, you can, they're on-demand resources, and depending on the way you purchase our eyes and that you're really talking about short-term costs. And so one of the most, the proudest days that I had within FinOps was meeting with a squad who was doing development work, and they actually measured the cost of their resources and their unit transaction before and after a release, every single time, before and after a release. And so they understand what that release caused from a cost standpoint if they gained an efficiency out of it, which most of the time they did gain an efficiency out of it, but a couple of times it increased their cost and whether or not there was enough value that was being delivered from that release. And so their work times, when they can back out releases because the cost was not enough valuable, was value. So this is actually where you want teams to get to. That is a perfect union of the technologist, the business need, and the finance behind it, right? They're, and I've got to say that most engineers, once they understand that they can actually impact cost decisions, become very engaged with FinOps, you know? So that control that they've never had on-premise is great. And that leads to what the CEO needs. That leads to what those product area and delivery needs is understanding what their products are actually costing. And when Jen talked about the granularity of this cost, you know, the cost data that comes out of the cloud vendors, like part of this is also to adjust that granularity to appeal to the different audiences that look at this data. So the finance folks may not necessarily care about the granular details of every instance or every database or the network charge that happened yesterday, whereas the engineer really does care about this. So it's important to understand the layers in this and aggregate to the right level so that people can get the right value out of this data. Right? So when you were talking about a squad measuring the book before and after effects of making a change, is Finland something that you would put like on a screen, you know, with your squad, like with agile metrics to measure, like burned down charts, you know, and other cycle times and lead times for engineering? Is it something that can count towards your success? It is. And so, you know, there are different, you once asked about the challenges, right? There are different opportunities and challenges with this understanding and breaking down cost to the right persona, to the right person, so that you're showing them what they actually need to make the decisions that they need to make to take their ownership of usage. And so we use a external vendor for our data platform because the amount of data, if anyone's seen the AWS files or the GOOTWIDE is a tremendous amount of data. And so we take that data and we present it in ways for each of these different personas and what they need to do. And that should be a different part. So again, what we would like to see in their natural routines is that they check their usage, they check and see whether they've cleaned up their development environments, whether or not they're launching things within spot usages or whether they're covered under reserved instances and things like that. So they can make decisions that are based on what they need and you want to see this in a fairly consistent routine, whether it's a weekly routine, you know, or something even, you know, depending on the persona how often it needs to occur. That's fantastic. I would add just one point to that. It's also the trend that that's important. You want to look at this data instantaneously, but you also want to see is my team trending in the right direction? Because it is overwhelming when you see that, hey, 30%, 40% of my spend is being wasted, but you want to understand, is it going in the right direction or the wrong direction? Because then, you know, it's not always possible to get down to zero for, you know, optimisation like savings, at least in an efficient manner, because you're prioritising other things, but you might at least want to know that you're not making the problem worse as you keep using the cloud. And I'm going to just add one more thing, you know. One of the things that is really important here is anomaly detection. So trending is actually really important in this space. So you understand when you're going to see increases in spend and when you're not, and what your normal trend is. So anomaly detection is really good, and there are a million different reasons of why you could actually have anomalies happen. And so one time, we had someone misuse a service, actually wrote the coding wrong and was calling the service a thousand times more often than you needed to, which just generated this huge increase in cost, right? And then it ran over the weekend, and of course, we continued to pay for that resources. But by having anomaly detection there, right, based on the trends, you go through and you detect this and you can cut that off as soon as possible. Finance is worth, you know, when people were moving to the cloud, they all heard stories about, you know, millions of dollars worth of bills that they didn't expect because of resources using that it wasn't required. But I will also say the second reason is that unpredicted usage of resources is actually also being used as indicators of possible data breaches, too, and breaches within your environments. And so having that anomaly detection tells you when your resources aren't being used internally or externally by the right people, so. So briefly, now that we're into the last 10 minutes, can you tell us, you know, how fidelity and also how city are actually using Phinops and maybe some of the challenges that you've actually overcome whilst using it? I feel like I've talked a lot, but so fidelity itself, we are a privately held company. We have 16 different business units-ish and each of those different business units actually maintain their own budget. They maintain and direct their own technology resources. And so when you have this type of model which you get traditionally with these large enterprises, have a lot of these different business units, you want to be able to continue to give accountability where accountability needs to happen, right? But you also want to work together as a group in order to get the most out of some of the, like, the pre-purchasing and other sort of decisions that you're making. And so what we decided to do was a hub-and-spoke model, right? So you have a central organisation that we have, it's a virtual team. It is made up of people from each of those major business units, the larger business units. And we actually work in a contribution type of model actually within it. It is led by both a person from a business unit and a person from the centre so that we get both of those perspectives. And what we do is we talk about all the different opportunities that we have and assign them out to the business unit that is most close or would benefit the most from doing this. And so if we brought on an additional cloud service provider, whichever business unit was going to be using that cloud service provider would be the one to take the lead and then contribute back to everyone else. And so this model has worked really well with us because it, again, one of the principles here, right, is everyone takes ownership for their cloud usage. So the business units still have to take ownership for their cloud usage. And they have to take ownership for implementing the right sort of policies and processes in place and that yet contribute back to the larger enterprise as a whole. Yeah, I mean a lot of the same at CD2. We, organizationally, so I report to the head of public cloud. So cloud financials has a seat at the table for all the things that we want to drive. But from an implementation perspective, I'd say we're doing the visibility aspect first, right? That's first and foremost in this. I think if you can't give visibility to your engineers that are using the cloud, you cannot drive accountability, right? So we're tackling that first. We're also tackling kind of the budgeting process to sort of say that, look, we're in the early stages of this. We can't expect applications coming into the cloud to know how much they're going to spend in 2022 in like July of 2021, right? So instead we're going to a more iterative approach for that of get into the cloud, start using it, start building your solution, incur costs, then from there project out to what your production workload might look like and what other regions might look like in terms of costs so that you can get a better, more accurate estimate and a forecast for your usage. So, yeah, obviously there's also optimization related things. We want to integrate these optimization recommendations sometimes from the cloud vendors, sometimes from third-party tools directly into the developers' SDLC processes, right? Them being there in a tool is great, but then if nobody looks at it, then they don't get action. So instead if they can directly be incorporated into their development process, then these things get prioritized just like the rest of their engineering work. So those are some of the ways that we're... So I'm going to have one more thing about it because I think both Citi and Fidelity did it very similar, it's two, is that down at the bottom here there's these different domains and there's best practices from Fennops for each of these different things. But there's also something at the top that says crawl, walk, and run. And so what we decided to do was not tackle all of these domains at one time. We put our focused energy first of all of where we could make the most impact of the smallest number of people because culture change takes a lot longer than that. And so for us, we came through and said, okay, understanding cloud usage and cost is fundamental to all of this. But really then going to cloud rate optimization and organizational alignment was the key pieces for us. So we spent the first year and a half really focused on those things and doing it really well. And then we started to go into the next domain area which was cloud usage optimization, which is actually harder to do, involves a lot more people and the return amount is smaller. So, Jen, within the last few minutes, you mentioned that Fennops is actually an open source framework. Can you explain what this means and how organizations can get involved and maybe ask more questions and learn more? Definitely. So I'm passionate for an advocate. I reached out to JR who helps really, and he's out there who helped create this foundation because I wanted to participate in it, right? So two and a half years ago, three years ago, I was googling and did cloud optimization and that his book came up in it and then we started to learn more and more about the practice and become engaged. So Fennops is part of the Linux foundation. It has 4,400 active practitioners right now and it is an open source, open contribution type of model. And so from those 4,400, we have, I'm going to hopefully get the numbers right, we have 18 different sigs that you can participate in and we have a number of different open projects right now. Fennops does this great thing is every year they take a survey of all of the practitioners and they ask what is the thing that you're struggling most with? And based on that, they open up a series of projects and so some of the projects that you could participate in right now is how do you get engineers engaged to doing cost optimization, to doing resource optimization? How do you do shared cost and cloud container cost Kubernetes? How do you actually take that environment and get it back to the people who are actually consuming it? So they really listen to the practitioners and they allow them to drive the next set of projects that are opened up and then all of those practitioners are able to do it. They also allow certifications because one of the other goals is really to promote this practice and the practitioners themselves and so they've developed a certification, there are multiple other certifications being developed right now and training and personas. So it is really about supporting this community. That's amazing. We now have just under a couple of minutes to take some questions from the audience if people want to put their hands up and maybe ask the panel. Alisundri. I get it, thank you, you have failed. So the nice thing about the transparency of those bills is that it has everything in there. I will say that when we were doing business cases of moving to the cloud there were costs that we didn't assume when we were creating the business cases and then when you start to see the bills you start to understand what it is. And then you have to, for us, we look at the different costs and our best tool to use is the engineering, the architecture community when really dealing with those costs. So we talk to them and bring information to the table to help them then make the determination about the right way to design an application or to design a transaction so that that cost gets reduced and yet doesn't impact the service level agreements or anything with it. So our goal is never to tell anyone, no, what you're doing is too expensive. Our goal is to have a meaningful conversation where we're bringing data to the table and then ask people to think about things in a different way. A lot of other things that happen on the cloud that you don't get on-premises there's so much more access to different types of services and service level agreements that sometimes architects and engineers come in with one viewpoint of how they're going to solve a problem. And by having the discussion you're opening them up to all these other things. Well, maybe I could use this other service or maybe I could archive this or those type of things and really pushing the architecture to use a lot more of those cloud-native services. That's amazing. We're out of time now. Thank you very much for that question. Jenfrann, sorry, we're unable to take that, but I was going to actually ask, maybe if you can hang around outside. I will definitely go. Maybe we can take a few more questions. So thank you everybody for being here this afternoon to learn more about Phinops. Thank you Jen and thank you Adam for being here and enjoy the rest of OSSF. Thank you everybody.