OCP 2021 Global Summit Keynote: Easy Edge Its Open and Its Everything - presented by ITRenew

from i.t renew please welcome allie fenn [Music] i'm going to sound like i'm beating a dead horse but it is so amazing to see people in 3d so thank you guys for coming thank you for having us super excited to be back and actually have real engagements with people um we are here today to talk about edge especially edge in the most sustainable way so thank you rebecca for inadvertently teeing up both edge and sustainability for me super important topics and i'm glad to be here with my co-collaborator jim simonelli from schneider who is also driving these things and it's a it's an important part of what we're here to talk about which is that this is a massive open possibility for this community in a very very unique way and it's here today but it requires new kinds of collaborations which is perfectly what this community is suited for so let's dive in edge is massive it's not just a buzzword it's here but it means lots of different things to lots of different people so our thesis is that edge is an opportunity for not just new purpose-built edge micro data centers and other facilities but for every real estate owner globally to participate in the massively distributed network of compute and infrastructure that is going to be essential for us all to realize the positive potential of 5g of iot of smart everything remote everything the world that we live in and it's a hardware opportunity which we'll talk about so what's driving it well there's going to be 175 zettabytes of data created by 2025 according to idc to be 75 billion connected devices we're already up heading towards a billion plus 5g connections up from just 10 million a year ago and super importantly by 2025 gartner thinks 75 of enterprise data is going to be created at the edge that's going to demand a whole new kind of infrastructure that we haven't been we we're not prepared for yet but it's in our fingertips right now and we actually have the solutions we just have to think about them differently what about that infrastructure um fifty percent of it capacity expected to be deployed at the edge by 2025. this is super fast one autonomous vehicle is expected to create as much data every hour as three thousand people immense amounts of data power what is all that compute infrastructure at the edge gonna do well today it's about forty percent of day of power of tied to computers at the edge and sixty percent in in core central data centers that's going to flip and cost is a problem right we got to get all this infrastructure out it's massive amounts of hardware but today the best analysis shows that an average edge platform deployment costs about 55 percent more than a super optimized regional or central data center so we have to think about how do we deploy all this stuff at scale to enable the possibilities but do it in a way that our budgets can tolerate right how are we going to get it out there it's a big challenge we rebecca also teed up that this stuff is massively distributed it's heterogeneous so we have to think about what it what what are the new imperatives that we have to wrestle with well obviously distributed makes it more difficult but environment right so hot areas cold areas dirty areas clean areas small form factors big form factors everything is going to demand you know us thinking about topology in a really different way and managing that stuff in a different way and how are we how are we going to think about this massively distributed you know environment where we don't have the skills we don't have environments are not staffed with engineers right when you're thinking about gas stations roadways cell towers airports smart campuses whatever it is right all of these are opportunities for edge infrastructure and for people to run not only their own applications at the edge but participate in enabling the edge more broadly and also generating new monetization models which we'll talk about today as well but we have to think about what are the complexities the rise from all of this infrastructure in a radically different kind of form factor and environment than we've thought about in the past so back to this cost question if it's more expensive to do it at the edge well wait a minute we're not all hyper scale companies with nearly infinite resources we have to think about other people being able to in in more normal size budgets being able to enable all this stuff so how do we think about delivering at scale with very cost effective resources and then from a business model perspective same thing right we amazon was able to say hey if i build it people are going to come and that worked out really really well but i don't know that that's going to be true in every edge environment how are people going to get conviction to invest to deploy infrastructure not yet knowing who's going to come how are they going to come what are they going to do right so we have to figure out a way to kind of catalyze the edge that that is necessarily requires us to think about collaborations and doing it differently um sustainability this is super important right you've heard some of the stats about energy consumption you've heard some of the sets about data center industry it's way bigger than that right we have a collective imperative to think about the sustainability of everything that we're doing and we are underpinning every industry in the world so sustainability globally is our problem pue centralized data centers we've done a phenomenal job on this and we must keep going great keep going keep going on renewables all that is essential when we think about the edge pe is not going to be anywhere near as good in kind of adapted data center type locations in other locations at the edge but we're probably going to get a benefit from less data transport and the energy that's associated there we don't know exactly where it's going to land but we have to think about it grid mix you can't sit every edge location every edge deployment at the arctic circle and get perfectly clean grids so we have to assume that if we have a globally distributed edge network then we have to we have the global average and grid grid mix and it's not very clean so we got to keep working on that manufacturing is the big one we're talking about hardware this is a hardware community it's going to take an immense amount of hardware to enable everything that we want to do together and it turns out if you look from cdp or any of the other research that's out there supply chain carbon manufacturing carbon is way bigger than operating carbon so we have to think about what the new models are here and we'll talk about specifically re-certified i.t in the it space for ocp racks in common deployment models as much as 75 percent of the total it total carbon attached to that rack comes from manufacturing so we have to think about new ways we're going to argue at i.t renew that we think about it in a re-certified way huge opportunity there all of this is a huge opportunity for ocp this is the perfect community to come together and say hey we don't have to invent a bunch of stuff we have the tools we have the solutions what we have to do is put them together in different ways that enable us to operate in these much in radically different environments if we do that it's open we have all this possibility we can do things differently it enables breakthrough total cost of ownership it enables huge financial opportunity and i would argue it enables us collectively to think about how do we leap to the fore how does ocp immediately jump those growth stats that bill put up on the screen we can go bigger we can go faster if we think about ocp equipment in all of these edge locations let's challenge let's let's let's take this landscape right it's a huge opportunity because we do have the solutions and we have the community members we have to be communica you know collaborating across everything from ot to i.t power to cooling to orchestration of of you know capacity and maximization and so forth which we'll talk about all here because collaboration is is the key um what are the critical imperatives i've touched on some of this i don't have to tell anybody in this room how much more difficult it is to deploy manage service hardware in highly distributed and heterogeneous kinds of environments when we think about those environments it becomes very clear that we have to think about the integration of i.t and ot we have to think about all in one we have to think about easy we have to think about that across the life cycle from deployment to to management to end of life integrated service models a single hand to shake public cloud taught everybody that it should be as easy to go click click click small medium large and i'm up and going but there's inherently a whole bunch of workload that's never going to run in public cloud environments and we have to deliver that same user experience and it's harder at the edge but we have the ability to do it we're going to talk about that today to make this real in terms of is this happening today with partnerships no one company can do this by themselves and then the last thing is we have to think about well not the last thing second last thing power efficiency how do we get smarter about capacity utilization in these environments that's a big challenge and we have some opportunities there too and then we have to think about both the cost and the sustainability piece of this we have to challenge old status quo models we cannot assume that anybody has the budgets or the the governance capability and and appetite right now to just trash the planet as we do this right so again there's a huge opportunity for us to rethink business models generally but one of those is how to think about so how to think about sustainable hardware and of course we want to talk more about that as we go with that i'm going to turn it over to jim siminelli from schneider who's going to come and take you guys through a use case that's very real that's happening a couple of examples in how we're making this real as a community [Music] all right thank you ali so my job here is i want to get more concrete here i want to share a bit of how we can tie open compute and edge in some real use cases so with that i'm going to do this by talking about two use cases from two companies relatively small companies that have common challenges as well so the first company is apc towers this is a a classic tower operator that is the is going to deploy new equipment for their network and are considering as deploying that new equipment is what more can they do with it beyond their core operating model the second is integrated roadways is building out smart infrastructure for smart highways and smart cities now both of these companies have some things in common they have great property near highly densely popular densely populated areas they have great network access and they each have their own core business models but they're thinking about the challenges uh and there's really two challenges that come to mind is you know first is gosh if i'm spending capital in deploying compute at the edge are there new business models that i can do beyond my core to improve my return on investment the second challenge is as we've talked about is that it's complex we're not all you know big hyperscalers the idea is gosh how do we deploy uh compute software physical infrastructure in a standardized scalable way from dozens to hundreds to thousands so those are the challenges that they also share as well as the opportunities and we've been working with them and thinking about this and the question is is can it be done can these challenges be overcome and we think that the answer is yes but it isn't easy you know we all know that any capital investment comes with associated risk both of these companies have lots of risk factors based on their own vectors their own types of investors but they sort of come down to three areas that are there they're they're collective high risks the first as i've mentioned before is that these are not large companies they do not have expertise across multi-domains they know their domains quite well when you start getting into uh compute when you start getting into new types of software um and how to service and manage this at scale that's a challenge for them so the fundamental question is how do we make this more simple the second is none of these companies or none of us i would put forward have a crystal ball uh you know we don't have smart cities everywhere we don't have smart corridors everywhere we do not have uh you know compute everywhere on every rooftop etc and while they have confidence in their core business models they are looking for ways to perhaps de-risk that a bit are there new revenue streams that while they're deploying the compute that they can actually take advantage of without having to be experts in that model in the last as we talked about all these companies have investors and those investors expect these companies to grow in a responsible way which means they really have to tick the sustainability box as well so how are they addressing this in the real world and what are they doing today so i want to first address how they're de-risking or enhancing their return on investment and both of these companies have decided to deploy hardware no surprise open compute hardware and not just to run their core business applications but they're opening up to um to deploy new revenue models for them as well and we'll talk about how this is working out but within these same systems that's driving their core business they are allowing compute uh to be available because they have great network they have great locations near densely populated areas that they believe there's some value there and if they can turn it into a new revenue stream for them all the better and there are these two streams that they are working towards the first is opening up compute for ultra low latency application deployment and the second is ultra low latency cdn deployment those are two streams that they believe are going to be value valuable both to the market and are willing to invest and open up their hardware for this type of application we'll talk about how that will work in a minute the second piece of this of course is physical okay now if i can do that and i can wrap it in some standardized scalable simply deploy architecture now they've got something so they've got their core business they are opening up their hardware and the the access to their equipment for new services to address their revenue model now what does it physically look like so what it physically looks like is as you might expect it starts with standardized open compute hardware that's a platform that they know it's scalable it can handle many different applications many different services and running their core services as well as their new services on it all this is packaged in an easy drop in place data center that provides the physical security the environmental certainty the ability to be scaled dropped in place remotely managed all in one space on their network property that they have in place and this is really the foundation of how they're looking to grow their new investments in the future to succeed with the growth of their of their companies now with that i might say okay now how does it do this okay this is the thing how their their risk um how they're looking to de-risk and open up new revenue streams and physically how it's being done how did it happen so it's happening through a syndicate this is four companies coming together sitting down with these new companies looking to take a bold step in edge and the network edge in particular and putting together a solution that we think makes sense and can scale so the first company is schneider electric the company that i work for and we build data centers of all size from large data centers to small edge data centers you see here and we're providing what i would consider the the anchor integration we're pulling together all of the physical systems we're adding the software um we are providing the the muscle to deploy at scale the remote services to ensure that these sites can stay up and running wherever they are in the wild the second company is ali's company it renew and this is really kind of a key element here it is all built on open compute platform open compute software uh or excuse me open compute hardware that they are really doing and something with with a unique dimension they are choosing it renew because they are going to deploy second life hardware and that really ticks their sustainability dimension as well no new carbon had to be embedded in these systems it dissolved second life plenty of compute plenty of opportunity for their applications just makes sense the second element there though is it really hits a price point that makes their business models tick so it was very easy for them to make this decision to go open compute and also choose the path of it redo second life hardware as well the third company is a little less obvious but when you think about the wild in these edge and network edge sites oftentimes they're pretty power constrained you might have 10 kilowatts that you're trying to maximize core business application core workloads and new business workloads how do you make sure that you are maximizing the utilization the compute utilization of that site subject to the power constraints of that site and virtual power systems is a company that has software that runs on premise that allows this this uh understanding between the workloads and the power constraints to ensure that these sites stay as maximally you maximally utilized as possible in the last company is probably the the key company that unlocks these new value streams if i have compute that's available how do i take advantage of this if i'm not a cdn expert or a distributed application expert well well edgex is a new company that is uh that enables that and what they're doing is they have um an operating system that runs locally that enables the the the execution of distributed applications in cdn applications they also have cloud orchestration that enables these applications to be loaded and located where they are and where they're needed they also are the the broker or the interloper if you will between the people who need these services and the people who have the infrastructure and manage the the revenue flows to make sure that everyone the value chain is treated as well so this is really kind of a great example of small companies taking an investment that's core to their business that has edge computing on the network edge and also opening up these new revenue streams which we think is the model for the future of edge computing you cannot deal with smart cities and the iot of things where they're having compute everywhere and to put that computer you need to bring in people who have property network access who can really manage their core sites their core business as well as grow their own revenue streams itself so with that i want to bring ali back and have her wrap up and conclude the whole big picture here thank you jim thanks ali so i'll be quick here we've talked about community we've talked about tco we've talked about sustainability all of this is enabled by open it's a massive opportunity for ocp to move now as i mentioned a lot of those stats explosion of data all this this is now it's happening now we have the imperative to do this sustainability sustainably when we talk about that in the i t renew context it's all about recertified hardware it's about circular economic models it's about thinking about everything from the power to the hardware and embodied carbon and so forth this community as you can hopefully see in what we just talked about is very uniquely suited to pull the pieces together as i mentioned this is not this is not a lot of new invention we are all doing a lot of innovation in the community day in and day out this one is really about collaboration and business model innovation and thinking about things differently to capture the opportunity and that opportunity is massive so there are an infinite number of potential edge network sites right anybody there's real estate can do this you can run your own apps you can participate in the network you can drive monetization you can enable the broader opportunity already today two and a half quintillion bytes of data i just love that word generated every day in 2021 idc says by 2024 we're gonna spend 250 billion dollars at the edge in just infrastructure and on the sustainability front if i take start with the one and a half trillion dollars of spend that's expected to be spent on iot by 2025 and back out this is just you know kind of quick math if i back out the hardware portion of that and then i back out the server portion of that and i apply our life cycle analysis of ocp equipment specifically and the carbon that's embedded in ocp equipment the potential opportunity here is 45 billion kilograms of avoided emissions in terms of carbon so this is a huge opportunity as we go capitalize and catalyze on this edge opportunity and deploy all the hardware that's critical and essential to you know everything tech enabled going forward to do it in the most sustainable way so i don't know what edge means to each of you uh probably something different but come talk to us we'd love to to figure that out and help all of us collectively go enable this opportunity in the most cost effective sustainable way as a community thank you very much appreciate it [Applause] [Music]

Channel: Open Compute Project

Views: 659

Keywords: OCP Global Summit 2021

Id: 9248mSkv7gg

Length: 21min 27sec (1287 seconds)

Published: Wed Nov 10 2021