Driving business value from a modern cloud-ready platform - BRK2448
And. Not me. All. Right well thank you very much we'll get started today thank you guys very much for joining us my name is jake Smith director, of data center technologies, there's not a lot of us here I would like it to be an interactive session you know we could just go through slides and I don't think you'll get what you want out of that so, if you could move a little bit closer or at least be interactive, that would be fantastic again. My name is jake smith i'm joined here by my dear colleague, almost. Almost. Two decades now we've worked together Chris Peters, he. Is the other half I like to say of the Intel select solutions team the, better half the smarter half I'm, just here to try to make sure the technology actually. Works so with that I'd. Like to first of all thank Microsoft. For, inviting us here I'd, like to thank our Windows operating system. Team led by Andy Vargas who's been an absolutely, great partner, from, day one I'd like to also thank Rachel Frick wherever she may be oh she's right there, he. Was just an amazing marketing. Partner and collaborator, and. Then of course I'd like to thank one, of my esteemed guests but I'll thank him when he gets up here until. Then Chris. Why don't you talk about why we, are driving, business value. From. A modern, cloud ready platform, Jake, thank you very much before. We get into the, content I know, Intel has been celebrating, our 50th birthday basically. I think, you are also celebrating. A birthday today is that correct wow, that's a heck of a call-out yes I am I. Think. I. Think. We're celebrating, your birthday today, thank you it's a very intimate, audience, and that was a very. Big call-out on you. The. Thousands of people watching online thank, you so much it's such an honor to be here on my birthday I truly. And. Those who know me closely know I would never want to be anyplace, else, but. Talking about technology, on my birthday so thank you so much and Jake I think I want to go for another first I don't think there has ever been a time where, people have, sung happy birthday to someone, at ignite on stage so. I'd like to invite our audience. Let's. Sing happy birthday to Jake I will start but please don't let me do it alone. Killing. Me day. To you. Happy. Birthday to. You. Thank. You it's very kind, thank. You all that was very very generous, thank you. Thank. You for, the 23rd. Time yes, Jake asked me to be unscripted, so I dealt. It right back Wow. So. I. Want. To talk and kick this off before we get into the technology, the, question really is why, and that. Stems back to business, so, why are we here we're really to, use technology.
To Create value, for our businesses, and right, now one of the things that is driving, businesses. Around the world is digital transformation. The. Use of data and driving. This as a strategic. Tool and, capability. Inside of our businesses and the one thing that we see all around is how fast things are moving in the business and how, fast businesses, are changing and the, risks of not transforming. And not keeping pace is being. Left at the side of the road and there are lots of stats out there but one of the ones that I continue, to talk to and see this pretty staggering is the, average time that a company spends on the S&P 500, the, beginning of this not. This decade but 50. 60 years ago it was 67, years today. It's. About 15, right. So the rate of change in the rate of technology. Is driving, that because, the disruption, and where, people are from a business standpoint seeing. Their competition, it's not their traditional, competitors, it's, from other industries, it's, from, smaller. Companies that are growing up that are taking advantage of digital, as a capability, and we. See that more and more organizations are, investing. Over. Investing, in IT and IT transformation. From a digital capability, standpoint and. It manifests, itself in the same types of technology. Trends we've seen for years how, do you do more with less drive efficiency, drive growth how do you become more agile and efficient inside of your organization how do you deliver new services how. Do you deliver new capabilities, and how do you better meet the needs of customers. And. I know Jake one heart for, you is security right, so security, is not an afterthought, it's actually an underpinning of this and and so we see that as well so you wanna talk a little bit about the kind of security as a business well, security. Is not just it's, not just the physical security obviously, it's, also not just providing. Basic. Digital security or, security, through I like, to say encryption. Okay security, happens at so many different levels and in a digital world you have to not only secure you. Have to protect but as importantly have to be able to detect, and, so that's one of the things that we feel really really good about this. Program, is because now we're introducing so, many new security. Features Jeff Woolsey, and Aaron Chappell talked about shielded. VMs for. Those who don't, know it's shielded VMs are it provides the capability to protect your virtual machine. Even. When your system gets hacked and then, you take a one step further and you look at a sure, confidential. Cloud it. Provides the ability to, deliver secure, Enclave so, that you can even be even. If you're have, bad actors, on the, platform, that you're unaware of you're secure enclaves, provide, even another level of security so security, is not a it. Is security, without compromise, because you actually you know our customers have told us over and over again Intel. We expect, you to deliver performance Jennifer, Gen and we expect you to deliver the security, without, compromise, and so that is a charter and we are a security, first company as, our. CEO, did outline earlier. This year and, I think it really comes back to this. Business transformations. Digital transformation, is underpinned, by IT, right. So IT is that enable dealing with security issues or dealing the efficiencies, and is the underlying technology, and how do you deploy that and I think that, is another aspect, of how. Dependent, we are in business around, I T and. I T is not just a support organization anymore, actually become the lifeblood of this I think, a great indicator, that Chris is the fact that when you look at even at Intel. Paulette. Oliver who is our current, cio at Intel is a member, of our executive management. Committee that's, really a first time that that's happened to us in about 15, years so you can see. In the late 90s. The, CIO became, a partner, with the CEO and the board of directors, then there was about a 5 or 7 year gap there where that didn't seem to happen now, it seems like the trend is going back where, the CIO and, the seaso the, chief information security officer. Are both, members of the seat most senior executive staff of most corporations it's, a sign of the times, changing, and how IT, is a strategic, tool and I think that manifests, itself in organizational, structures as well but.
The Other thing I think that makes this, worrisome. Is in, spite. Of some of these positive trends. This. Stat is actually pretty interesting is business, leaders see, feel. That more. Times that or not their. Digital transformation, issues are not moving fast enough right, they're not meeting the needs of the business and they're worried and concerned about this and this really puts a. Challenge. On the organizations, and one of the things that's at the heart of that is our, aging technology infrastructure, right so is the opportunity, to become more plentiful, and we'll talk about that more later and how fast things are moving if, we're not keeping pace that, legacy infrastructure, starts to become a. Drag, for Navy days it's an anchor right so maybe not be holding your still but it certainly is slowing you down and. So we see these trends in this capability to take advantage, of this is we, have to break momentum. Of doing. Nothing well. And, nothing, is is. An enemy, of what. Right. Nothing. Is the enemy of progress so, if you do nothing you. Cannot make progress if you, hope that your. Security, issues, will go away because you just hold on to that legacy software, or hardware, a couple, years longer nothing. Will really change so what, we found is, that, the average, upgrade. Cycle, went from five years, to. Seven years to almost. Eight years and, that's. Very very dangerous in. The, last eight years as, an example we have introduced, four new compute, platforms. Each, new, compute, platform, coming with at least three. To five new security, technologies. In. Eight seven, to eight years that IT. Departments. Have not upgraded their infrastructure, there, are now exposed. To. What. The fire I CEO, said this morning eighty percent of the bad actors, in the world and 80. Percent, of the bad actors, in the world according, to the CEO of fire, I our. Nation. States. Very. Dangerous, if. You. Don't upgrade, you, are literally. Allowing. Your business, your. IT infrastructure. And your data to, be put in the hands of others, I. Think. This slide really some of these stats we. Reinforce. That point Jake, right is that the. Productivity, loss, from. Older. Infrastructure. Oftentimes. Is not manifested. In slower, performance, which, it is but, it's manifest in the time you're spending, maintaining. That infrastructure, and the costs of that or, the costs of dealing with security, issues or dealing with that because, the once. You're detected it how do you how. Do you address. That is, extremely costly piece so, being, proactive can. Have a very, solid, piece of productivity advantage. On you so you're spending your time much more positively. In managing, innovation and transformation inside the business and spending, less of your resources. Managing. Infrastructure, and capabilities, well, you you also can't take advantage of all of the great Microsoft. Tools that are out there so, the biggest one for me is that all of the Microsoft tools whether they're hybrid. Cloud solutions, which we'll talk about in a minute or they're on a sure all. Of those new tools run on the latest hardware and you can't use those security, tools if you're trying to go back to the. Nehalem architecture, of. Zeon from 2009. You. Know west mirror isn't going to get you over the hump Sandy, Bridge is only going to get you part of the way there you. Need to upgrade to the latest technology. That Microsoft. Has built so many wonderful, tools around and we'll talk about that in a minute before. We move off the slide though I think the other thing that you, know we underpins, this is how fast. Data and data is growing across the room across the world right, so I think you should have this data point with me a couple weeks ago is 90% of the world's data has been great in the last two years right. Where is that data sitting right, it's most likely in siloed, infrastructure, it's not being taken advantage of it's not being utilized right. It's hard to get at the data and. How, do you put a lens around that one. Of the things that Intel is we start to shift ourselves, as a company, we're, moving to from my PC company to a data centric, company and how, do we use the tools in our portfolio, and that in the, ecosystem. Partners we have to help people. Move. Faster store. More and process, everything to be leveraged those capabilities back to their business and it really ends up starts with a portfolio of technologies and, capabilities but, then it's how it gets delivered and I think some of those pain point gets identified in, if you're not doing that you actually start to hinder yourself, from, your network speeds from your ability to get estores you really effectively, store all that data, cost-effectively.
Really, Start to hinder, the capability, of your business I had a great conversation the, beauty of being at ignite is you have a conversation with customers you get to meet with customers all the time I met, with a customer who's deploying a hyper convergent, environment, his, company, just merged with another they're gonna be 800. Stores they, have to upgrade their entire, infrastructure. And his number one issue how. Do I make, progress. Migrating. The legacy. Data infrastructure. To, a new, hyper-converged. Infrastructure. Thankfully. Thankfully. We're. Here at ignite we've, worked with a great partner, and they have great tools so. Without. Further ado I would like to bring up Jeff. Woolsey, who. Is the. Product. Marketing, principle. Of. Windows. Server, come, on up Jeff they fixed that real quickly I'm not in marketing in any way shape or form oh. He's. Not he the engineering side is the engineering side I apologize. But he is the principal, PM, for. Windows, Server I did get that right that night thank, you, he's also someone, who I have worked with closely for the last decade on Windows Server technology, so, we're going to talk a little bit about what. The Microsoft, Intel relationship, so, important, so, with that Jeff, why, don't you talk a little bit about some of the things that we've done awesome. Thank you very much it's really a pleasure to be here so the the history, of Microsoft and Intel goes back Wow, decades, I mean, yeah a lot of people look at NT four but it goes way back almost as old as me 24 plus years it goes even farther and the history of the Microsoft, and Intel it's a long-standing. Celebrated. Partnership, goes. Back to the dawn of computing, I mean when you really think about it and. We have made some tremendous inroads, now if I stick. Stick. Specifically, with server NT. Really was a watershed moment yeah we had three one and and three foot but Windows NT four, workstation. And server was. The real coming out party well it's all of a sudden a server, became, easy to use it was we. Talk about this very often within Microsoft, before. NT four servers, were the, the vanguard of the high priests, and priestesses, you. Know of mainframes. And all. Of a sudden Intel, and Microsoft said, no everybody, can actually have a server guess, what you you you can actually take a server and you can actually have this in your back office you can have print servers and file servers in fact servers and all these crazy things and in, fact to this day file, servers, continue to be the number one reason why someone deploys a server because, I need to share data with information. And I want it to be easy to use and now, we're adding all of these other things like cloud and and and more to it but NT was you know a big watershed, moment for us because if all of a sudden everyone, could do it and then, we moved, to the world of guess what we went to SMP, servers and I'm not talking cores or threads I mean to sockets, with two symmetrical. Multiprocessor. Yes SMP, servers you know it was a big deal oh my gosh I've got a server with two processors. In it and that was a big deal, and, but we continued to push forward and I remember, buying. A, four. Socket, server I mean I think it was like 15, grand at, the time, war. Sockets, again no hyper threads no cores four, sockets. Okay and that was a big deal and then, all of a sudden we got to this world where guess what you know we continue to innovate we continue to revive we saw the dawn of virtualization. And all, of a sudden now we, have x64. And you mentioned a hail em to me i-i've told I remember the one that was chatting with Jake. A few months about this and I said you know Nehalem, was really a watershed moment for a lot of reasons number one all, of a sudden we had multiple. Cores we had the memory controller on the processor, which really Unleashed performance. In that way you, added a whole bunch of virtualization, extensions, to the processor, up. Until then we had to do a lot of things in software called, shadow page tables, just to manage, all of the address super technical stuff but just, to make virtualization, even work we had to hack it okay essentially, and Intel, we work with him we said this is what we need they said this is how we can give it to you in hardware and make it really really fast, and that.
Was A huge moment for us extended, page tables into a lemn thank, you that was a lot of fun and, it changed. And transformed the, way virtualization, was thought of absolute, or was it just two, virtual, machines per, saw per server or per socket it became four six. Eight and, even, today it's. Yeah, well and and and but what's interesting was, once we'd unleash this in the processor, then all of a sudden then we started looking at the rest of the bottlenecks, in the system because, x64, gave, us the big thing we needed which needed a lot of memory and I, kind of laughed because you know 2008, I was a big, server had 16, gigs of ram in it 32, gigs of ram and a server that's crazy my, laptop has that today okay I have 32 gigs of ram in my server and my laptop okay and that, was huge for a server at the time well now that we've unleashed 64-bit. Processors, and all this memory all of a sudden I'm running more virtual machines I'm driving, up my utilization, guess what storage, is now a bottleneck oh I can't I can't get things to the VMS fast enough because I got these spinning disks in there the network is way too slow because I've got heaven, forbid 100, megabit or a couple gigabit NICs in there so that's too slow and all, of a sudden you know it just okay, now we go fix the next bottleneck and so we work on the platform we add bigger pete more pci lanes we add wider. Pci lanes we do more there we, add more memory more controllers, we, add increased, capabilities. Into the processors, now we got multiple cores and we got multiple threads and we continue and this tight innovation, continues, to drive up performance, and the platform, gets better and better and better and the, rising, water helps everybody, here then, we get to 2016, and 2016. The launch of really this hyper-converged, revolution, now that's kind of what you know I've been seeing as people are saying look you, know I've been doing the traditional thing the traditional, thing is I buy a bunch a whole bunch of servers then I buy some dedicated, storage which means now I've got to buy you know Fibre Channel switches and, fibre channel sands, and fibre channel network adapters, and all that which is managed in a completely different way by a completely, different set of people and then, I've got my networking, which is I've got gateways, and firewalls and load balancers and multiplexers, and all these things and there are their own different, devices and then, I've got standard topper rack switches and that's. How we've been doing it and this, whole move to hyper converges, about what wait a minute what if this was just all mainstream servers. Filled. With lots and lots and lots and lots of superfast. Flash cores, and memory and the, magic of software and standard, Ethernet switches and we can give you all of that with the magic of software and hyper-converged, and that's. Where we're seeing the growth right now we've talked about that we looked at the data IDC Gartner, everyone saying you know huge, massive. Double-digit, growth in server revenue, because people are buying these hyper convert systems correct it's a triple digit growth market, IDC. Predicts it to be eight point seven billion by 2022.
We. Are certainly, seeing that growth we, do not actually, dispute. The IDC forecast, by any way because we're seeing that growth but, as importantly, it's, a combination, of choice it, is not just choice on-premise, its choice in the cloud so, the ability to have a hybrid, cloud solution, where you can manage on-premise. Migration, as well. As in cloud migration and back, and forth has been transformative. In the way our customers have had the ability to utilize, this technology but, why do you think. More. People aren't, jumping, onboard, with. The, latest, in HCI what, do you think keeps that, I like to say the laggards, from. Leaping. In you know a lot. Of it is is change. With how you're interact with the system it's humans, that are involved here we're, used to been we've been used to doing things the, traditional I've, got, a traditional. Storage appliance I've got traditional all of these things and now we're kind of changing the paradigm to well, guess what we're giving you a management, console that does all of these things in software and it's a it's a different, of doing things and so humans. We're trying to deal with this change now I will say the hyper, converts also has its place in terms, of scale if you're looking at two nodes, four nodes eight nodes 16 nose hyper-converged, is great but I will also point out when you get to like hundreds, of nodes that's. When you go with the disaggregated. Fabric and I have to remind people when, we when we talk about Asher, Asher, is a disaggregated. Fabric we have nodes. For compute and we have nodes for storage okay there it's not hyper-converged, you can't hyper converge at that scale, can, be very problematic and, so, it's, great for two four eight sixteen, thirty-two you, you want to do that great get, into hundred. Five hundred thousand, no no no you're gonna go disaggregated, you're gonna have separate compute nodes and you're gonna have separate storage nodes the beauty of this is all of this just works in the platform we've built together and part. Of that is having the high speed connectivity. To reduce latency and drive, up performance, things like our DMA and having that in the platform, having. High-speed flash and what work we're doing together around. Apache pass in these in forthcoming systems. To, actually, bring Frash closer and closer to the processor because as we do that we are really reducing, latency and driving up performance, well. You know I always talk about latency, free and for those of you who don't know I I am, a person, who spent my entire career, trying to reduce latency in, intel. Architecture at, every, turn of the crank and jeff has been a partner throughout that throughout, but but there's another reason why, we have an urgency, here, to. Move customers. To the newest platform, what is that urgency and that urgency is absolutely, about end of support and. It's been you know we've been talking about this quite a bit this week and it hasn't been a surprise because quite honestly we, told people two, full years in advance I, watched, this closely with 2003. The last server that wind ups went end of support we gave people about a year year and a half notice and the big feedback was that's not enough time if, you're gonna end up support something give us more time so I told the team and. We made sure we message, two years out from end, of support we started telling people and people in the responsibles thank you for giving us the good heads up so, we have two things coming number one is sequel Server 2008, in 2008, or two on the, large we have a very large install base on this version of sequel server that, in July, ninth 2019. And, there is no better time to be moving that forward because if you are running sequel, Server 2008, in 2008, r2 there's. A very good chance you're running us on very old hardware probably. Compute limited probably, memory limited probably, running on spinning, disks so sequel.
Is You, know it, was fine in the day but. You're gonna get easily a 10x if not way way way way more bigger performance. Improvement, whereas vice shall II would say a 12-point 2x we'll get to that and. Then of course we've got Windows, Server 2008, r2 and 2008. So this is another one about 50%, of the install base for Windows Server is on 2000, era 2 it's extremely. Popular OS we get it but at the same time we also have to tell everyone it was a great oh s I know I, was right there at Ground Zero working, on it it's a fantastic OS, in its, day that. Day was 10 years ago a day, is not right, now ok, I would not be telling people hey let's go install 28 2008, r2 and another, big message I want to be super clear about it and Jake's already discussed it but the last thing I want to see if someone goes out and buys a brand-new, intel server brand-new, select, solution and puts 2008, r2 on it you're, not gonna take advantage of any of that great work that they've done in hardware and the reverse is true as well I don't, want you by taking buying, brand new 2019, and sticking us on your 10 year old server because, you're not gonna have the hardware that we need to do what we need to do for security and platform, and performance. And scalability as, well so these are two big dates for us and we are telling everyone honestly, you need to be on a seek and destroy mission if you see. 2000-2008, in the title whatever the product is sequel, Windows Server we got it we got a movie we got to get you on to a modern, secure platform and we have to do it by 2020, so the. Point is and we're over emphasizing, this and we'll overemphasize, this over and over again until the, date Cap'n's. Is that. Intel and Microsoft are, committed, to helping you modernize, and getting off of that end of support but, Chris there's there's, other issues associated. With in to support the customers need to think about. That. Was good, think. Of microphone anyway it goes back to it we talked a lot about the the risks of old hardware right but the risks is old software is some. Of the things that I know you're gonna talk about here in the details is the functionality, capabilities. That have been hit vast with every generation of, release of new software coming and the security, benefits than those in being will take advantage of that so this is again shows, the downside, risk as opposed to the upside opportunity which, I think we should talk about in the next slide but again it you get to higher maintenance costs the same things we see with all the hardware you see with old source software.
So Again not to belabor the point let's go move to the next slide which really I think talks about. I'm. Dying to get to this one this is the one I've been I've, been waiting for this one for a long time we, only worked on this for a clarity there are so, many, benefits here, that I'm sorry that the slide is that the text is so small but you can see just how many benefits, there are when you compare, a platform, whether it was running Server 2008, or even Server 2012. R2 okay. If you compare that to a modern, Windows Server 2016, and 2019, it's. Just there's really no comparison so, take a look at the moment for the hardware scale, 2012. R2 the maximum. Hardware we supported, at four terabytes of RAM and 320. Logical processors, now, for a lot of people now you may go it's not a big deal well, we support in 2016, and 2019 up, to 24, terabytes, of, RAM, that's, not disk that's RAM, in the system is up, to 512, logical, processors, and the, reason why we support that by the way I get that ass all the time is Jeff how many people ask for that very, few actually the reason why it's in there is because as you're sick needed, it you'll. Notice between 2012. And, between. 2012, and 2016, was a big jump because I went to the Azure team I said what do you guys need I, don't, I'm not sure no customers, are asking for much larger than 2012, and they said oh no no no no no Jeff we, run the biggest workloads. On planet, Earth I want, you to go find the biggest server shipping, on earth and whatever that is support. It okay. That's it and that's, 24, terabytes of RAM and 512, logical processors, then. At the same time they said and by the way while you're working on it we need to support the biggest baddest VMS known to man which, is why we have VMs at support that's not a typo 12 terabytes of ram per, VM per. VM and, the, reason why is because azor offers, these so. If you want to run big scale up workloads, big databases, big sequel big sa P Hana I guess what the best place to do it, is on, Windows server and of course this is all supported up in Azure and up to 240. Virtual, processors, per VM so. Those numbers are monstrous. Numbers, and this, allows us to take advantage of the massively, rising core counts and on the latest Intel Xeon scalable. Solutions well and in the latest we've, now gone from 8 cores, all the way up to 28, cores, we've added a vx520. 8, cores per socket, I'll take that please yes and and as, we, add Intel. Obtain, datacenter. Persistent, memory we'll be able to go up to 3 terabytes, per. Socket, or socket 28, cores, 3 terabytes, avx-512. New, security. Instructions, new, advanced, encryptions, next-generation. Instructions. Lots. Of other features not not to mention increased. And lower, latency, in all of our virtualization technologies. So, we're reducing latency x' gen / gen almost 4 x + and the reason why the memory support is so important is when you start getting to these higher core counts you got to be able to feed those cores work and you need a lot of RAM to do it otherwise, you can have courses sitting around waiting for stuff to do because hey guess what you didn't be actually populated, with enough memory and so thrilled to see these massive, memory support now in.
These Systems the other thing I want to point out real quick is the security, aspect, so the work that we've been doing we've been talk a little about security. Shielded. VMs a. Lot, of these platform, capabilities require. Platform. Support, we need TPM, we need euphy we need secure. Boot there's a whole bunch of things that we need enabled, in the hardware to, make shielded, VMs a reality, and it's really simple with, a shielded, VM of VM is a black box ok, and I can basically say this shielded VM it. Can only run on this server or this rack of servers so. Even if the bad guy Edward Snowden, breaks into your datacenter, copies, it onto a USB stick and takes it home he, can't start it up he can't turn it on he can't mount it he can't get inside it why because when, this VM turns on it this, VM can't turn on because it doesn't have the keys to even decrypt. Because. A shielded, VM the very first thing it does is it says I'm a shielded VM hand, me my keys. Nobody's. Going to give it the keys and the keys will only run any known healthy. Host and the, way we get that is attestation, via, TPMS. And all the security work we're doing on a platform which is why we emphasize TPMS. And so this is why it's really really, important, when we talk about the upgrade, path that our customers, need to take you. Talked about it earlier it's, insufficient. To do, nothing they, do, nothing means. That you're putting your entire organization, at, risk whether, you're a profit generating, organization, or not you've, put your organization, at risk a software-only. Upgrade, path, insufficient. A hardware. Only upgrade path. Insufficient. What. Is the sufficient path for both got it on both you really do you, really have to do them both hardware, needs software the software needs hardware they look, it's like chocolate and peanut butter okay they go great together you really need them both because they both have dependencies, on each other and it's, a processor, it's a platform, it's security, capabilities, in both locations that feed off each other to deliver the, best highest performing, reliable, and secure solution. But. There are business, benefits, oh absolutely. So let's, talk about that what are some of the business benefits that we see. Digitizing. And going to digital transformation, but the benefits of going into a modern IT give. Use the opportunity. To, realize some of the business benefits for revenue growth better. Margins better capabilities, better services, which, ultimately turn, to both top line and bottom line growth avoiding. Some of the cost advantages, or disadvantages of. Maintaining an old an older and aging infrastructure. Jeff. Thank you very much for joining us. It's. Been a lifelong friendship, my friend I hope it continues because, as. They say when, the performance, is this good why would it stop, so. Jake let's try transition, now I think one of the things that we've set up is. How. Do we accomplish this now we have the the need the imperative, moving forward but the question is how do we take advantage of next slide of these. Capabilities, and we look at the, amount of rate of pace of change we saw that on on Jeff's, slide where the frequency.
And Cadence is increasing, and we look ahead right, now and you look at the the breadth of Intel technology networking. Storage compute, processors. You look at the breadth of the portfolio, Microsoft, is bringing with, Azure stack with, Windows Server 17. And sequel. Server and. When. Server 16. Moving to 19 the, big question, is how do you put this together efficiently. Effectively to. Go from that lower left to the upper right quadrant as efficiently. As quickly as possible and that really, comes together is how do you put hardware and software together for. Solutions. And what we really have thought been thinking about and started, launching last year is this concept called Intel select solutions and what. This was in the next slide please is how. Do we take advantage of the deep Intel, silicon. Capability. How, do we take advantage of, working, in partnership, with software optimization and, bringing hardware and software together in, configurations. That have been tested proven and. Verified. With a deep engineering, commitment, across the ecosystem between Intel and Microsoft. And. Aimed, at those workloads, that are incredibly. Important, that we're running our businesses on today or, we see high growth opportunities, like the hyper-converged, infrastructure, that Jeff mentioned right. And are, typically, very difficult, and time-consuming for, IT organizations. To figure out on their own or the ecosystem, for that matter to, figure out on their own because everyone is trying to climb the chain to digital transformation, and services, from, a value-add standpoint, so we thought well what can Intel do with this and so we launched, this program to build these. Infrastructure. Capabilities these, solutions that we call reference, designs and document. Them from the ecosystem and Jake I know your team is deeply, involved with development. Of those references asked you talked a little bit about the methodology we're, applying would, love to so we spend probably, seven. Or eight months on every select solution, working very closely with Microsoft. To ensure that we are optimizing, each and every portion for, those of you who are deploying project, Honolulu, you've taken advantage of.
Those. Capabilities, all the telemetry that you see is a reflection, of those optimizations, that we worked with Jeff. And his team how, does that actually end, up in performance, in your shop we then run that in benchmarks, where we benchmark, our components. And the. And the components, that microsoft, recommends, at the highest level, and we push that very very hard to ensure that we can deliver a level, of performance, that when you purchase an Intel select solution, you have a baseline, performance level, that you receive and you can in fact buy an enhanced, or plus version, of Intel, select solution that will deliver a 30%. Or better performance, increase over the base version. Of select solution, and then. We provide peer, review across, the industry we don't do this on our own and say yeah it's just marketing look at the documentation, we, provide full documentation, on how to configure how, to create performance, and how to replicate, the performance that we created, every. One of the select solutions, has full documentation. On how to configure and how, to deliver, the baseline. Performance of, a select solution, and you, can actually, tweak each and every select solution deliver even enhanced performance beyond, that. No. We won't and we won't because Jeff won't let us number. One and. Number. Two we won't because it doesn't make sense so. Earlier. Today we we announced a. Well. Actually earlier, Tuesday, we announced a record, number, using. A technology, known as Intel. Data center of persistent, memory it, was a thirteen, point seven nine eight million. I ops okay, that included, using Intel obtain technologies, the highest level of performance that. We have ever ever seen with. A hyper-converged, solutions, with Windows storage the, storage space is direct. That. Is the, high end of the extreme ok, not, every customer, is going to need 13, million I ops to run their sequel server environment. To run their office 365. We, have that capability as, Jeff. Pointed out Asher. Is requesting, it so we're capable of building it however, we give you the configuration details, so that you'll know to configure, your systems, to the level that you need to and that full documentation. Allows you to do that ok, I don't necessarily need. 32. Nvme, drives at 8 terabytes each. I might, need 24. I might, not necessarily need. 512. Gigabytes, of memory per. Socket maybe. I need a different little combination maybe, I need a different way of interleaving. That memory to deliver performance all of that is available in, the documentation, of a select solution, so, that when you work very closely with our OEM and our channel partners you, know you're, getting a fully documented, fully, validated. And verified, solution. Great, thanks Jay can I think what when. We look at what a select, solution is it is a premium infrastructure. Configuration. That is built for a particular, workload great so it's not one-size-fits-all. It's, very uniquely tuned we actually choose the benchmark in partnership, with Microsoft. Or with the ISV so, that's again it's not just Intel, picking what's gonna work for our stuff and showcase ours but what, really is going to make this workload shine and. The, best part is you, can see this all on the floor today at ignite, yes. And, the key thing from an end-user benefit, standpoint, of what we're trying to achieve is how do we accelerate, time, to market, accelerate. Time to innovation and accelerate. That time to growth in business transformation. And that's, done through. Reducing. The time we're spending evaluating. The, work that we're doing between Intel, Microsoft and then the OEMs who, we asked to meet, or exceed this, recipe, so again, they have a tremendous amount of engagement innovation. In technology at the system's level and solutions level we, want them to add on top of this we, asked them to do is to follow the recipe and meet the minimum, standards at work doing. For the industry and. Therefore because we built it we've tested it and we verified it we know how to deploy it therefore. The less. Time guessing, or less time trying, to figure out what to deploy what, to choose less time trying to figure what to deploy and knowing, that it's gonna work inside your infrastructure ahead of time and what this does is it starts to build an ecosystem, capability.
That Is first of all grounded, in the validation work at the OEM does the. Certification, work that the ISP, does the, verification. Of leading technology, and performance from Intel and then, as often delivered, we hear this later by, solution, providers, that are adding services, and capabilities inside your business so. This is really a very collaborative, brand, as we market with this and it's been very exciting for us to see some of the momentum grow just after one year well I'm. So excited because we've delivered, the performance, and you know like, I said it's my birthday I love, to talk about technology, but, I like to talk about performance and the reason I like to talk about performance is because you need performance. To deliver solutions to your customers and the, joint solutions, that we've developed with Microsoft, or Windows Server 2016, Windows, Server 2019, sequel. Server 2017. And V next, and Azure stack these, select, solutions, provide, us the capabilities, to give, you. End-users. And IT departments. And cloud deployers, around, the world the, capabilities, to know that there's a minimum performance level that you should expect there's a minimum performance level, that we will deliver and. There's a maximum, opportunity for. Innovation a, good, example we see on the show floor today is Fujitsu, Fujitsu, has adopted the windows admin Center and taken, all of that telemetry, and capability, integrated, into their overall system structure. To provide differentiation. For their customers, and allows, them to build differentiation. On top, of Intel. Select solutions, with Microsoft. Absolutely. So, let's move a little bit and discover, what's. Inside the, hyper-converged. Infrastructure, for Windows Server software-defined, storage and. Jeff mentioned this before, the amount of growth that we're seeing here in the amount of opportunity, and why, we're seeing it is pretty, much because of cost right, dealing with all this data how do you cost-effectively, store it manage it inside, your infrastructure we're going to hear an example of. A real-world customer. That really has embraced this, in later video so we want to keep moving to that to that point but it's a tremendous opportunity from, the business capability, and that's what's driving this, growth for hyper-converged, infrastructure. From. A select solution standpoint, again some of the ingredients you see on the on the very far left or scalable. Processors, are Intel SSDs, and our Intel obtain SSDs, and. In, high-speed Ethernet combined, with the latest software. Capabilities. From Microsoft. With Windows Server 16, and very soon Windows Server 19, coming, out and. You can see a example, of the types of configs and if you look deeply at these configs, you, will find as used technology and if ways so. That I'm, the base configuration we're, featuring our nvme, drives for the Intel and SSDs but, in the plus configuration, to get that additional performance, gain at a neutral, price performance. Advantage. So we're not it's. A minimum, performance. Advantage over the base configuration of, 30 to 40 percent yep, and here. This is where we feature the app the Intel obtain SSD so again it's featuring, that the advanced technology, is one of the ways to get to a higher performance result and. It's not just the advanced technology, it's it's actually the work that we've done with partners as well that makes it also very compelling, the work that we've done with data on and, Howard low I want to thank Howard Lowe and his team and the work that we continue, to do with Doug Fisher above.
Career Kirk's Calgon and the team at Lenovo is very exciting, and the. Emerging. Presence. Of Huawei, has been very very pleasing. To both Microsoft, and Intel as we, begin to adopt these new premium. And Plus features but, the reason, that we get these capabilities, is because Windows, 2019. Comes with some really cool new features we've, got new virtualization. Capabilities obviously. We have new UEFI. And security, capabilities, that Jeff talked about but we have new networking capabilities, we've added our DMA, so now we have increased, not only in performance, but we have a direct, memory mapping, we have new storage capabilities, because we've, not only added, octane, SSDs, but we've added this new capability this new feature which, we showcase called Intel. Obtain. DC, persistent. Memory or data center persistent, memory which is a new memory architecture, not only delivering, persistence. Ok, but it actually delivers the capability, to do memory and storage mode to overtime, so, we actually are really excited, about what these hyper-converged. Enhancements. Are bringing and it allows us to say, to you confidently. If you, upgrade, with. Your current architecture, the current shipping, product, the Intel, Xeon scalable. Processor. Architecture. Family, if you, upgrade you can actually do in place OS upgrades, from, Windows Server 2012. Through, r2 and, Windows Server 2016. That is pretty impressive you can take a seven-year old software, hour, and you can upgrade in, the current environment which gives you a lot of flexibility, and I didn't talk about containers. That even talked about integration, with orchestration. And server fabric, and some of the other things just because, we're running out of time and we do want to show how customers, are taking advantage absolutely. So, this next slide is you know again these are the building blocks so this is the new hardware capability, Jake, just talked about the new software capability, lots, of tremendous power and capability speeds. And feeds and capabilities, inside the, compute memory networking, storage I call this our you know the pantry, right we had these are the tools that we have available inside. Of our solutions, but, really when we start talking about how do you put together this. Hardware. And software together, upgrading. From that lower left quadrant or the upper right quadrant the, power and capability really does get featured in what, does it look like inside the end-user environment from performance and capability standpoint it's. Also something that we want to alert all of our customers. Partners, friends. To is Intel is more than a CPU company yes we have added 7x the number of course but. As Jeff pointed out we, had to move beyond that that wasn't enough okay. Yes we could give, a mesh, architecture to, get to 28 cores that allows, us to add Numa, features that nobody else in the industry has yes, we can do that in the cpu yes we can add security, features an acceleration, and virtualization and security, absolutely.
But. That's no good without the memory, capabilities, so we've added the memory capabilities, and yes, because you can add this caching. Capability, with obtain, DC persistent, memory yeah we can show you unbelievable. Performance, numbers record-breaking, but that's not enough so we had to add obtain, datacenter, SSDs, as well so, now you could actually move, to a tier of hot data that delivers, performance time and time again but. That wasn't enough because if you can't move that data from, one server, to another or, from your storage, your, legacy storage, infrastructure, to your new storage. Infrastructure, how are you going to do that and you need 10, gigabit, and beyond, Ethernet, connectivity and we're able to deliver that as well, so now we have a platform, of, Windows, a platform. Capable of running really, any Windows, environment, on the planet that has reduced latency. Across, four architectures. Of pause compute, memory. Storage, and networking and I can't wait to show you about fabric later. Alright. So this is my lower, left to upper right quadrant right so you'd look at the old hardware look at the old software left-hand. Side, 20. DRA 20 servers, 4, racks then, you move that into how do we new software new hardware, capability. To the same level of capability and performance. At, 1/4. Of the, cost and infrastructure, right so 70%, boost. In the savings, with respect to what, it's costing, you to manage and maintain this environment again. That frees up resources to, go drive the innovation so this this is the proof point behind the lower left to upper right hardware. Software together upgrading, and the capability, have. But. You, know you, get this innovation, because you're delivering, unprecedented. Performance and I'm gonna talk a little bit about Intel, obtain technology, we, use obtain technology, it was part, of a joint, relationship. Developed, with micron, many many years ago a long-standing. Relationship of. Innovation, and. Product. Development but now we're delivering the lowest latency high, skew s highest endurance, performance. On the, planet up to 60 drive writes per day I wanna put that in perspective if, you have a 1.5. Terabyte drive, that's. Capable of 60 Drive writes per day you, can write. 75. Terabytes. Of data every. Day for, five years under warranty. Never. Been done before folks. Every. Day for. Five years, seventy-five, terabytes, of drive writes that's. Spectacular that's, what often gives us and this, is just the beginning this. Is just the beginning of a next generation of storage and memory infrastructure. That removes, bottlenecks, and creates, new layers of opportunity. And. Options. For our customers, but. You know what enough. Let's. Let cosmos, and for. Those of you know cosmos, Darwin you're very very fortunate I know I am and, I'm gonna let cosmos, tell, us all a little bit about what, it means when all these solutions come together. Organizations. Around the world are looking, to hyper-converged. Infrastructure, for, their next wave of data center modernization. And there's. No better way to, go from hyper-v, to, hyper-converged. Infrastructure. Than. Windows, Server 2000, 19 which. Includes hyper-v, storage. Spaces direct and software-defined. Networking everything, you need in one familiar. Product. And license, but. More than the affordable, price or, the familiar management, tools what, customers, tell us they really, love most is the. Performance. Two. Years ago on this stage we. Showed a demo of a single, storage basis direct cluster, providing. Hyper-v, virtual machines. With, nearly, 6.7. Million, random. I/o ops or, storage, operations, per second, that. Was. With Windows Server 2016. And hardware. That's now two years, old. Today. It's. Time to see if we can beat that record, the. Demo. You're about to see is the, result of a deep engineering, collaboration. Between Microsoft. And our friends at Intel and it. Will show the very latest in storage. Innovation. Intel. Obtain, DC, persistent. Memory. You. Can be forgiven if you aren't familiar with opting it's pretty new stuff an. Opt-in module, like this one goes. In a ddr4 memory slot but. The storage it provides is persistent, meaning, you can power down power.
Back Up again and the, data you wrote is still there an, unlike. Regular Ram obtain. Modules, common sizes up to 512. Gigabytes, and, this. Is fully natively. Supported, in Windows Server. 2000, nineteen including. With storage bases direct. The. Performance you can get out of one of these modules is. Pretty impressive. But what's more fun is an, entire rack of them it's. Actually a reference design that Intel and Microsoft have, been working on together, 12. Server nodes running, Windows Server 2019. Each. Packed, with a future Intel, Xeon scalable. Processor, persistent. Memory and nvme. Let's. Take a look. We. Start on the new Windows admin center dashboard, it's especially for, hyper-converged, infrastructure. Here. We see an overview of the cluster, we. Have our 12 servers are. 72, drives and over. 300. Hyper-v, virtual machines, they're turned off right now but we'll, come back to that in a minute, to. Round things out we. Have 14 volumes, which give us nearly a hundred terabytes, of usable storage capacity, and you. May have seen on the front photo this environments not even fully loaded we actually have additional, PCI slots and drive bays available, for expansion. Let's. Look around a bit here. Are those twelve servers from Intel and. Here. Are the 72 drives, we. Can group by type and see that we have 24, persistent, memory devices which we'll use for caching. Now. You may notice the size shows us 768. Gigabytes, so how is that possible. Well. If we click in you, can see we're actually interleaving. Multiple, physical modules, together at the memory controller like. I said this is pretty new stuff going, back. You. Can also see those 48 nvme, drives for capacity, and if, you still think of flash as small think again because these nvme, SSDs are 8 terabytes, each. Back. On the dashboard I want. To introduce you to two charts, the first shows. Us the total cluster wide number of I ops that VMs are getting and. The second shows us io latency, as measured by the file system in Windows so this is not device latency. All. Right it's. Time to get this show going what do you folks think are you ready to see some awesome performance. Okay. One last thing if I could just get an extra digit up there perfect. So. We have these 300, virtual machines that are all turned off each. One when it comes on is going to run an open-source Microsoft, tool called disk speed to generate load, so.
What Do you say we start turning them on let's. Start with just the VMS on node 1 as, they. Come on and we can time-lapse forward here we, can see the eye out shoot, through the tens and hundreds of thousands straight to 1 million, 1. Million for K random read I ops just with the VMS on node 1 this. System isn't even breaking a sweat, let's. Turn on the VMS note - you. May start to see a pattern here, those, VMs are gonna request another million, I ops and they're gonna get it two, million I ops. All. Right, are you guys ready I'm ready let's, turn on the rest of the virtual machines and as, they come up we'll go node by node you'll, see that no matter how many storage, requests, these VMs make this hyper-converged. Cluster, can satisfy it through. 3 4 5, million coming up on our previous record 6 million I ops and continuing. To increase through 7 million 8, million 9 million and here's, a big milestone 10 million 11. 12, 13. Million. I ops, folks. This is our new record and we, believe the, industry record, for, any hyper-converged. Platform. The. Pace of innovation here, is really, exciting, in just. Two years we. Can do double, the AI ops and by the way last time it took 16, servers this. Time we only needed 12 and, this. Is only the beginning there, is so much more in development, for both Intel. Obtain DC persistent, memory and storage. Spaces direct, thank. You very much for counting AI ups and millions with me during this demo on your way out please be sure to grab your account I UPS and millions laptop sticker and have an amazing week at ignite back to you Erin. Well. Thank you cosmos, that's a wonderful video and, we just felt like it was important to show you guys because it's really, difficult. To. Help. People understand, well how did you get that performance level and it. Is that deep engineering, work it is the time and it. Is fantastic. That, we, have such a great partnership, that, we can show it to you guys and we want to make it available to you and by the way because. We do peer review on everything, we're, very open with all the documentation on how we got there we're not afraid of that number if anybody contests, it please come on forward we'd love to show you how we got there but, how do we get there well we get there because really. It's about reor connecting, the storage memory as well. As how the software, addresses, storage memory traditionally. You think of it as DRAM. You start with DRAM and you scale as much there Jeff, talked about being excited, to have 32, gigs. Of. RAM. Available when. We introduced Nehalem, well now we can introduce persistent. Memory capabilities. Not only improving, memory capacity, but it is actually a new programming. Model called, pmfs. Or pmm depending. On who you're talking to and we work very closely, with. Andy. Rudolph and a. Jack. Vargas, on our team with, the rest of the industry this thus, NIA group, to create an industry standard, we created an industry standard that Microsoft, endorsed that micron, endorsed, the many people in the industry have endorsed, including.
All Of our partners, this gives us the first generation. Of a new memory architecture. To really change the way we, view software, programming, models we, now, add obtain. Datacenter, SSDs. Ok. And you saw some of those SSDs, okay these SSDs, are capable of huge capacities, and amazing. Performance, 60. Drive writes per day performance, allows, you to write over this drive over, and over again which is incredibly, important, particularly. When you're trying to run a database or other, types, of operations, along, those lines and then you could take our 3d, NAND SSDs. Cosmos. Were reiterated. Seven, point nine 8, terabytes for, an SSD yeah. That's our first version we've, got more coming ok. So we are adding, massive, capacity for the warm tear at, the nvme, level, interfacing. As close to the CPU as possible, so the performance will be consistent, over, time and, now, you can add a huge. What we call efficient, storage here in this, - and 3d nan space where you have really cheap storage and you can add huge, capacities. Over and, over again, and we, do not believe that spinning. Rust and I apologize, if you're offended but you need to take your spinning rust out of there and I won't be but. You're. Still going to need hard disk drives. 90%. Of the world's data was created in the last two years. Many. People have forecast, that by 2025. Every. Nine to thirteen weeks we, will be doubling this, of the internet today. Want. Everybody to step back and put that in capacity, the, Internet today is roughly depending, on who you believe between, 47, and 48 yottabytes. It's. Not petabytes, it's not exabytes, it's, not zettabytes. Yottabytes. Okay. So we're talking about the capacity to generate between, 200. And 250. Yottabytes. Of data by. 2025. This, is an incredible, number for us to think about I always, like to joke with my friends in the industry I say hey, we're two fab short of a dozen. But. It's it's not really a joke when you start talking about the capacity and the growth of information, in. Our industry. But. What's coming I. Think. Would basically when. We talk about where, we're going next year on select solutions and capability, is taking advantage of that you're not more excited than I am, no. I, don't, think I can be. So. Now I learned something. Good. Is. We're, gonna be integrating, in that, obtained. Data center persistent memory technology, in the next generation of our select solutions that capabilities across this and you'll see this theme across. All the solutions that were featuring here today as. Well as the products that ecosystem were building up and around yeah. And for those of you don't know about, two years after Jeff, and I worked together on hyper-v, I was, fortunate enough to begin the process of beginning to work on obtain DC persistent, memory and. So many of the people in the crowd with us today I want to call out Jack Vargas in particular not. Because he's great but because I, care, about him but because he's awesome and he spent a lot of time trying to solve a very difficult issue ok, persistent, memory is not easy. We have spent almost 15, billion dollars so we're clear we, have partnered. With a, great. Partner, in micron, to bring it to market and now, we're. Both delivering. Next-generation, solutions. Here for, obtain, data center persistent, memory obviously. We have a lead here, that. We are very excited about to show the industry, and the reason we're so excited is the performance, so, talk a little bit about what that and sequel yes. So again sequel, is the next solution that we have inside. Our portfolio, with Microsoft, and. I think there's some really interesting things, here as we start thinking about the analytics we're talking about data all around the data outer we take advantage of that for, digital transformation, and there's two capabilities, we're thin as I've been looking, at and seeing the converge is the, online real time processing, of data which is critical, for, transportation. Airlines. Manufacturing. How do you keep your database as you're ingesting all this information, up-to-date in real time right, because if you're not updating real time then you're making bad decisions on the data that you have and if you're taking in this data how are you using it more real time so transaction, processing, the other aspect, that sequels use it for is how do you take the data that you have in storage and draw, insights, in that through, data warehousing capability, and as. I've you know looked at especially at ignite learning, that more and more these, two types of capabilities are converging, right, the goal is to be able to do real-time insights, and get, the insights, that you can see to make decisions, more real-time whether, that's at the edge for an IOT device or, whether it's the day that you're ingesting, inside of your data center and capabilities, and that is really where I see the roadmap going and, what sequel is doing, in our solutions, today so on our side of our portfolio, for sequel we, actually have two very different workloads for online.
Transaction, Processing and data warehousing with the goal to try to converge them as we go forward and. We have a lot of partners who are working with us here as well don't we Chris yes, we do I think this is this is one of the first select solutions that we did launched with the Z on scaleable products, back. Last year in 2017 so you see a whole breadth of, select. Solutions, built on the sequel or. Solutions. From the ecosystem and partners that are building on this but, also I think attests, to the fact of analytics. As a key foundational, element inside of any of the businesses, is seeing this and so again across the board data on Fujitsu, Lenovo, HPE. Quanta. Super. Micro and we went are already verified we have a whole other suite, of technology, and capability partners that are continuing, to build on something before it very exciting it's, so exciting. Because these are available right now you, can go down on the floor right now on the demo and you can see these solutions, on the floor right now it is, not. Only exciting, for that reason but. It's exciting, because gosh. Darn it these, are really, transformative. To the way businesses, can be run you can now begin to do not, only just business inquiry you can do advanced analytics, and now we can start doing machine learning we could start doing artificial, intelligence, and we're, building on a portfolio of, a modern data infrastructure, capable, of doing modern, data analytics, one, of the things I think is interesting in this and goes back to each solution, is different right, so on the hyper-converged. Infrastructure, it, was a four node two socket four, node cluster of, servers and the sequel it's a two socket building block right so clearly is you need to add more transaction, processing, capabilities, you're adding more and more nodes you're integrating that across your eco inside. Your infrastructure, so again these are building blocks from the select solutions that are foundational, elements, to build on top of than around from the ecosystem standpoint, as well and, in the decade-plus you and I have worked together how many times we've been able to release a new. Platform, that was delivering, 12x performance, Jen over gen not. Very often very rarely, so. Not, only that we're delivering a consistent. Experience and, so I'm really excited about that because we're now a consistent. Experience from, what you're deploying on-premise, to also what you get to take advantage of in Azure and. That's really important, because. When. We did, early, surveys, and I do the the, I'd, like to say the unofficial, Jeff Woolsey survey, how, many people in the cloud are gonna do hybrid cloud and a hundred percent of the crowd raises, their hand the, the official, data that we have is the 81 percent, of enterprises. Around the, world have. A hybrid cloud strategy. The. Unofficial. Is in a roomful of 3,000, people I saw three people raise their hand that said they had a cloud only strategy. So. Hybrid is the model going forward and having, sequel, be able to perform at the same level on-premise. That it doesn't asher is a very big deal and, we've, also added a whole bunch of security, and that is another, part that we we probably spend, not enough time talking about but, the new TDE, capabilities. Along with our accelerated, encryption, capabilities, have now p
2018-10-05 06:37
