 Back to for the second session of the track today. We have a great presenter for you presenting the open-stack way So I'd like to welcome to the stage Bill Franklin. Hello welcome to The open-stack summit. I got it. We've got people still coming in so if you can move that way that might be great I have a couple of questions just since you guys it's day one for you You're gonna be exhausted at the end of the week So if I get you to move and be a little bit more active it may be more comfortable for you For how many people in the room? This is your first summit. Okay. How many people in the room have been to at least five summits? Okay, so for those of you that it's your first summit and drink lots of water. It's exhausting It's a marathon. It starts in the morning. It goes to late at night if you're here for the design summit Really drink a lot of water because you're gonna be in rooms with lots of developers doing a lot of talking and often discussion start You know, there's a few developers in the room sometimes sometimes they add add things and it'll start at 7 in the morning It'll go until 9 or 10 at night So we get a lot done at these summits, but it's pretty involved One or two other really short questions. How many people are here because they work for companies that are interested in cloud computing? I mean most people maybe I don't know But how many of you are here because you're actually developers who want to work on or hack on Open stack itself Okay, so a number a number of you guys are building the infrastructure A lot of you are working for companies that are gonna either install the infrastructure or consume the infrastructure and run applications on it so what I'm gonna I'm gonna briefly talk about today in the I think I've got about 40 45 minutes left is Talk a little bit about the the landscape of Enterprise IT what's going on in enterprise IT right now and kind of where it's going Open stack and cloud software, but specifically in the enterprise right so so if you are You know here from a small research facility or something else, you know I'll be talking about how your facility might be using it, but I'm not necessarily gonna be talking about advances in open stack You know from a research perspective. That's one of my talks later I Want to talk a little bit about what we're doing at HP inside open stack and how we're trying to deliver Open stack solutions to the enterprise and then just briefly touch on if I have time at the end sort of this whole maturation of the open stack ecosystem for the enterprise so what's what's actually going on in Enterprise IT right now and and while trying to read white and blue on the slides at the bottom is really hard so IT has really become very very integral to business success more and more of Really is hard to read these more and more companies really need to Innovate really rapidly and quickly and change what's going on Give you an example of that. How many of you here from Europe? How many of you here have heard of Europe? Okay, that's good Imagine, you know grease exits the euro and we and we wind up into a two or three take two or three stage currency market Okay, the first investment bank that puts up the currency swap systems will probably Capture anywhere from two to four billion euros a month In terms of fees and benefits and other things associated with basically running that currency exchange Okay, so so that's a real-world situation in as rapidly as you can innovate to the changing times as to what happens You get to capture, you know revenue from that. Okay Another thing that's going on which is represented by here There's more than there's more people registered for this conference I think they're coming from more than a hundred and forty different countries, right? We all work for a very global workforce. We all have a very global customer base HP does business in more than a hundred and eighty come countries, okay? So you need to be able to cooperate with all those people and collaborate and The digital data explosion has just been absolutely mammoth, right? It's been it's been really huge and Then on the technology side, there's massive explosions with lots of things. We're seeing Sort of over the next two years at two to three times growth in global cloud infrastructure We're seeing a massive amount of growth in intelligent edge services. We're seeing anywhere from 50 to 80 percent increase of new cloud applications or cloud applications that deal with with big data And we're seeing a huge expansion of both cloud apps, but also cloud development And when I talk about cloud, I'm talking about dynamic elastic infrastructure. Okay People have talked about, you know Dense data center. Let's virtualize every application. I have I'll put a web front end on it and I have a cloud I'm not talking about that in the Second paradigm or second revolution of infrastructure computing you wrote your application If you wanted it highly available You clustered the underlying infrastructure if it wasn't running fast enough you bought faster machines, you know assuming you write You wrote toad cut tight code if you're in memory data set got too large you put more memory in it So you affected scalability reliability size by modifying the underlying infrastructure in the cloud world a cloud Native application needs to operate a different way. It's stateless reliability and other things are built into the app not the infrastructure So we consider sort of cloud computing when we talk about it at HP. We're really talking about this third revolution of computing So finally in this in this third revolution of infrastructure computing or the new style of it If you want to think about it that way people are really striving for no vendor lock-in They're striving for portability of their workloads and they're really concerned about driving down their their lowest cut their total cost of ownership Okay, if you deal with the financial market sector, what is the most expensive thing for them right now? It's actually capital As a result of both the layman collapse and a number of currency collapses The cash reserves that banks need to hold now are anywhere depending on which country you're in and what type of operations they do Anywhere from four times as large as they were 10 years ago to in some countries in some regulatory spaces 180 times more cash than they needed to hold before So for a significant amount of the financial industry and financial services the cost of capital to them is one If not two orders of magnitude greater than it was 10 years ago So when we talk about driving down total cost of ownership, it's always been a snag word that people go Okay, everybody wants to run run more cheaply But if I put it into real terms for you you deal with certain industries now for which they're available capital is far far different than it was 10 years ago and And that's sort of the business tone of what's going on also as machines become more and more efficient and more and more More and more powerful power consumption has gone up real estate has gone up So the desire to get smaller data data centers highly more efficient But also be able to flex when you need demand burst to other things firms are really changing how they work Right, so it becomes an internal service provider at a company okay, and The other thing that it is happening in industry is a term that HP actually started a long time ago Which seemed to make complete sense for me long time ago three and a half years ago in the cloud world Which is hybrid hybrid cloud computing? part of our fundamental core belief is that Enterprises will not move all to a public cloud. How many of you work for companies that have been in business for more than 50 years Okay, how many of you guys still have workloads in those companies that run on a mainframe? Eight or nine of you are raising hands. How many of you still have workloads that run on Unix servers? I Didn't say Linux. I said Unix Okay, so my point behind that is that stuff didn't disappear It's not going to disappear overnight. We're going to work in it We're going to work in an environment if you work for any company that's over a certain year a certain age Or that it's over a certain size you're going to have hybrid workloads. You're going to have Check clearing again in banks often part of it runs in a mainframe part of it runs on some Linux servers part of it Is web-based all those things need to come together right you may migrate some of those workloads to a cloud You may re-implement some of them, but you're not going to change them Okay, one of the largest banks in the world largest retail banks in the world 65% of all the code still running in production is written in cobalt and runs on a mainframe Okay, now I'll combine with that and say that 90% of all the new code that they write is written in one of two Languages and either runs on an open stack implementation or runs on a variety of other cloud technologies All right, so they are living in a world where they have both Legacy systems they have stuff running on their own private clouds internally their own private elastic dynamic Infrastructure, and they have some stuff that they've moved out that runs on Cloud technology managed for them by a third party not a public cloud. It runs in a single single tenant environment So that's that's what we talk about hybrid and what's happening now is almost 30% of All of the enterprises in in this particular frost and Sullivan survey, which was I think about 600 different companies 30% of them are actually either cloud is crucial to their business or a percentage of their workloads are already starting to run on cloud So this is a technology that is rapidly being adopted We are shifting to it and enterprises are using it and they are using it in a hybrid like manner okay, so All being most people in the room are engineers or have some engineering background or at least took up math class Or a science class at some point in their life So one of the things we do in sort of engineering and science is you know We get down to first principles and we talk about you know that classic word definition of terms So I'm going to define a few terms. I already talked about cloud elastic dynamic computing Infrastructure we talk a little bit about when I say traditional apps Right, I'm talking about the traditional way that we used to build applications, right? We would decide we're gonna have some applications We either write them in-house internally Let's not even get into how long that is Monty already gave you some illusions to that earlier But we write some applications or we purchase some applications. We want those applications to run So we need to provision some equipment for that often that means we need to actually go out and buy some things So we procure them we might have to talk to other services We need to get those services to give us access God forbid you want to do a table scan on a big installed Oracle database and a financial service So you got to get permission to do that you got to get all that stuff lined up Then you have to configure all of those resources Then after you've got all the machines and you've got it all configured And you've got all of the permissions to talk to all the different things you want then you actually get to install the software and then you actually have to integrate all of it and This stuff takes often weeks to provision and this is being carefully generous when one particular company I talked to from the period for which they decide they're gonna do something and they get the budget approved It takes 18 weeks to get through their procurement cycle even if HP is really blazingly fast they get hardware in maybe four weeks and Then it takes them and this is the one that I found astounding it takes them nine weeks to get it off They're loading doc into a data center and get it available So the period I just described is you know It takes less time to actually have a baby than it does to actually get this hardware For that particular company, okay So you can only imagine why people wanted to move to a scenario where they could automatically provision things So what kind of happens in the cloud world just again my sort of definition of terms Let's talk about cloud native apps. These are the applications for which, you know, it's somewhat stateless They run in this elastic dynamic infrastructure So the developer actually gets to build the app and the application and specify how it's all gonna work All right, they then push that code out into the cloud for this is a DevOps model I could use Andresen's term, you know in infrastructure as code, but it pushes it out there and you instantly provision the environment and then the cloud if it's a well-written application and it takes advantage of all of the native cloud services It can automatically scale some of its resources. It can do a lot of Resiliency restart it can do a lot of self-healing and a lot of other things This environment is usually instantly provisioned and the application winds up being managed as a service by the application developers themselves in a traditional DevOps model Okay When you do that, okay, we're trying to extract the development experience away from Needing to basically configure and run all of the infrastructure Okay Now how many of you out there actually are programmers or started your careers as programmers or have some amount of technology behind it okay, so I Started my career a really long time ago And I my first job out of school was actually I was a Unix kernel engineer hacking on writing device drivers for a living So I spent the first five or six years of my career writing system level kernel code My point behind this is that you know my view of the user land application developer with somebody who called the system call interface I then moved along in my career and started managing an applications development group and found out that nobody calls the system call Interface they all sit in a much higher level my point behind this is that the application developers of today particularly in the cloud space They're not writing code all the way down necessarily at the open stack API level They're writing code at a slightly higher level of abstraction. They're writing code often at a PAS layer That's cloud foundry pivotal everything's that live in that world or they themselves write an abstraction if you deal with a large bank or you deal with If you heard the Walmart keynote or or others people write an abstraction above that infrastructure So that they write all their applications to that and this cloud native app model supports a lot of that type of capability So you wind up with cloud native apps They don't assume node failure. They avoid local storage. They're stateless Often there are multiple processes and where they need to deal with state Excuse me. They're talking to servers and services that are Reliable and they're doing and speaking to those through service connections and those connections need to be able to be transient Just as we have sort of a open source manifesto and a number of different things around agile development and others If you're really curious about this way of building cloud native apps There's kind of a manifesto in this space called the 12 factor app You can go and you know Google it on the internet or go to 12factor.net, but it speaks a lot about this concept of cloud native application development so if You want to move to the cloud and you want to start building applications that live in this space Today we kind of live in the traditional world where you have it ops. They're over here Usually they have people guarding all of their equipment and it's behind fences and they run all of that And you know I go back to the early days of Unix and usually the sys admin guys have big hiking boots on and jeans and Carabiners with large amounts of keys and all sorts of stuff and Those are the people who manage the network and the storage and the operating system Partially the virtual machines and then developers lived way up higher than that and often they were in different buildings And you know the stuff didn't work very much with each other now. What's kind of shifted is That IT operations winds up really building a bunch of automated application services Compute as a service storage as a service all the things that open stack has been building for the last four years and Then developers are consuming those services and sit on top of it and they orchestrate or marshal How all of those services work together? Okay, we are shifting one of the things that's happening in open stack as Monty was talking about earlier And some of the other people will be speaking about later today We're shifting open stack a little bit as a community not not necessarily our product directions Open stack as a community is starting to shift to the aspect of we now need to cultivate Application developers to write the applications for our infrastructure. We built really cool infrastructure But infrastructure without apps to run on it is not necessarily very exciting Unless you're an operating systems geek and you're just interested in about building the operating system So open stack is doing things. That's partially why we have the product working group I was co-chairing with Chris Kemp and activity around the developer ecosystem So that's another thing that's happening. I think the ISV Ecosystem thing is on Thursday So please come to that if you're interested in how we make open stack more attractive to ISV's and software as a service companies But as we move from the today world to the future world on the cloud native apps a couple of other things also wind up happening So if you if you want to build a pass layer or an abstraction or make an environment where enterprise IT builds these applications You you have to have both a combination of infrastructure as a service and this has type of abstraction It has to support a broad Environment right a polyglot of of environments Java. I'm an ex-sun guy So Java is first on my list Python Ruby even dot net and a whole slew of other things and People want to be able to move things around in a variety of different ways Right, so they're looking at open stack. They're looking at cloud foundry from a past perspective They're looking at Docker. They're looking at different types of virtualization strategies different types of orchestration strategies So where we're really moving to is this capability of enabling cloud native applications with a rich set of support for for different environments and ensure portability and flexibility and What's also happening though in parallel with all of this is companies are shifting more and more from using traditional proprietary software products To a broad adoption of open source okay, and 30 years ago 40 years ago Right all software was pretty much proprietary Companies wanted it that way when we sold them software there were escrow agreements There were all sorts of other things in case anything happened Right and and that provided a huge amount of lock And you had things like Linux you had things like Apache you had things like the GNU project you had a variety of things and a bunch of things happened that shifted so Enterprises now are more interested in getting open source technology than proprietary. Why they don't have vendor lock-in They don't deal with complex software escrow agreements. They don't deal with a lot of other things But it's also much easier to find people who know how to program in all of these technologies and administer all of these Technologies and when you have a problem, there's a huge community. You can go to to get information about it So we're seeing this massive massive adoption of open source. It's really been going on for the last 20 years so that a Combination of both this migration to cloud and movement to cloud it being hybrid and open source are really the three things that frame HP's overall strategy of what we're doing. So what HP's done is Our vision is really to build HP healing on which is a portfolio of cloud products and technologies from infrastructure as a service platform as a service AWS solution built around eucalyptus a variety of other technologies and we're really building HP healing on based on Open-stack technology is the bottom from an infrastructure perspective In a developer led completely open source based distributed world Okay, and we're focused on on doing that To be able to deliver really what we feel is open source that's enterprise ready For companies so using the drivers that caused enterprises to choose open source Excuse me cost savings portability Don't eat two handfuls of cashews before you speak Most of the drivers around open source as I was saying cost savings portability Really, it's an open standards type way of doing things Avoiding an aspect of being locked into anybody because you can somewhat switch from one vendor to another if they're all based on the same open source technology And then what we've done at HP is add to that 24 by 7 by 365 global support Fairly rich integration space Done work in areas around security maturity and building out the overall ecosystem HP actually is one of the people who leads the security working group and activities around Open-stack Monty didn't allude to this earlier But we do a huge amount of the work of running all of the continuous integration and the smoke testing the tempest environment and everything for the Open-stack community So if you ever do a put back to Open-stack, how many people in the room have ever committed code to the project? one two three four five six, okay All of your code all that put back goes through a whole bunch of infrastructure that HP hosts and Takes care of for the whole open-stack community So we do a lot of this work around part of this But those benefits provide not only to the community, but also to our customers who consume this All right, because it allows it allows us to do things so what we've done Is HP is built a product called HP healing on Open-stack, which is based on it's a standard Standard distribution of Open-stack that all of you are pretty familiar with And then we built on top of that a development platform based on cloud foundry called HP healing on development platform that kind of marries a combination of Open-stack Combined with cloud foundry Can block can combine with all of the things that an enterprise would want Running a distribution like this on top of things that are all based from open source And from the drivers of cloud enabled tech of cloud enabled application development So it's really kind of this nexus of these three activities We are currently shipping HP healing on Open-stack. I think it's a one dot one dot one and cloud foundry our dev platform also did a recent release and When Tim who I saw sitting somewhere He'll be up in a while with a number of customers of ours That'll talk about their engagement with HP and some of the technology I've always found I can stand up and tell you all about what other people are doing or I can have them stand up and tell you about what they're actually doing. They're far more interesting than I am anyway So what are the things that we felt were crucial to delivering into these so we've spent a lot of time working on Issues around security On issues around configuration as Monty was Monty was showing you earlier some of the challenges of Standard Open-stack, so we've spent a lot of time really focused on this Some of our newer newer releases include things around encryption of data and transit a lot of visibility around some of some of the Transactions and logging that's going on one of the other things that HP is capable of doing is we have a wealth of Enterprise class applications out of HP software Things around security things around networking things around identity things around storage also significant things around Application development application lifecycle application test So we've done a lot of work on integrating HP's broader software portfolio into this so things like specifically around Arc site the HP's product CSA Ops where a lot of the other technologies that HP has many of those technologies are probably already in your enterprise and part of our hybrid strategy is not only Just the fact you're gonna have hybrid clouds But bridging to the technologies that you're already using in that space meaning if you have a bunch of unix machines and Linux machines and they're being managed by HP technologies now You should be able to add cloud technologies and have those same management tools be able to work with them as well So we've tried to focus on being able to extend those and that goes to a lot of our security products I've already mentioned some of the some of the cloud management aspect But more importantly, we've also done a lot of work with storage How many of you are familiar with cinder? Okay, so as you know cinders highly pluggable so a lot of our Three part product products that have HA capabilities and other things We've tried to make those storage products work really well with with cinder. We've done a lot of work around Not only some of the DDU capabilities some of the HA capabilities and some of the encryption capabilities We also do a lot of work to ensure that our product works well with a variety of different things If you were at the keynote this morning you saw or come by the HP booth, right? HP Helion doesn't just run on HP hardware It also runs on other people's hardware it runs wins with other people's storage so we've tried tried to work on ways of extending that and We try to work on increasing the support for all of them So we work on we work on that We also support a number of different hypervisor technologies and continue to move forward trying to support all of those and then finally talk a little bit about what's going on from an an ecosystem perspective If you Harken back a small amount sort of in history back to earlier days of Linux in the early days of Linux as Linux was being developed Originally it was a set of people who were interested in building the operating system right and Linux was really successful for three for three reasons right small amount of computer history One Leonis and others picked a an open-source license that most people understood Some some enterprises weren't necessarily all that comfortable with but everybody kind of understood it And it met sort of the needs of both IT companies Enterprise companies Universities and software developers so it was a license that most people understood The second thing was Linux was almost kind of a movement. It had a vibe a I'm trying to think what the hip term is But you almost had like the equivalent of developer raves that went up which are now I think the current buzzword is meet-ups But you had an environment where people really got together and wanted to advance this and it was as much as social Phenomenon as it was a technical phenomenon and the third thing that made it really successful was Four companies five companies right at the early stage of Linux HP Intel IBM One or two others Invested money in that movement and that activity and helped create a foundation around it We've only seen that happen two other times in the history of open source One is Apache and the second one is the other one is OpenStack, okay, but in all three of those instances You created an environment where people participated in it and they were excited about it But how could enterprises tell that it was starting to mature and become sort of fit for prime time? And and where that comes in is when people start not only enterprises looking at consuming it But people in that ecosystem start consuming it meaning when you see ISVs start talking about okay Well, we currently run on Amazon, but our next aren't the next release of our product is going to run on OpenStack or people start looking at OpenStack that way you have services companies that operate in it. This is my Eighth or ninth OpenStack summit the first one. I went to was I dropped by the one that was in Santa Clara But the first one I was actually And and in Boston, but the first one I was registered at was in San Francisco in San Francisco There were about 800 people there and almost everybody showing there or Boston there was about 300 Everybody's showing there down in the vendor floor. We're basically all all vendors trying to talk to each other Okay, now when you walk through the vendor floor here You're gonna find consulting consulting services people have applications that run on this stuff people that make logging tools make systems management tools Make other things right now you are in a situation where you have customers that are using this But when they get up on stage and talk about it, they're not talking about how they took OpenStack and they built it themselves They're talking about how they are using OpenStack and they're using it with these three other vendors And they're collaborating this way and they're and they're using this public cloud and this public cloud my point behind it is that is that a Sign or evidence of maturation is a rich growing ecosystem It took Linux about eight to ten years to get to that degree of Maturation it took a while everybody adopted the you know Apache web server Web server, but it took a while for everybody to get to the stage where all of those tools were really rich and There were things for managing and others the case of OpenStack We've really been running in earnest for four and a half years And we're already seeing that we started seeing this Certainly in Atlanta, which was a year ago in terms of a lot of people there If you saw Concast on the rolling living room, they've done a huge amount of things with this but where we're starting to see it is in a huge amount of Sort of partner spaces validating third-party hardware working with different bare meddling provisions working with different types of Technologies that are all coming in and the companies are collaborating with each other we also see it in the aspect of Very rich activities that HP has been doing around its own life cycle management capabilities But also how it plugs into others. So we built things around Centralized logging for cloud health. We built things around Parallels automation, but we've also built things around our stuff working with with other people How many of you are familiar with TSA net? Nobody in the room so if TSA net is the backbone that whenever you do enterprise software, right? You'll get you know Oracle running on something that's running on something else that's talking to some automation problem And if you're a big bank and you're dead in the water and it doesn't work You pick up the phone and you scream at the first person who answers the phone But the problem most likely involves four separate companies working together the thing underneath it that allows us to transmit Problem cases and everything else to different vendors is TSA net. So if you go look at TSA net's website You will now see tons and tons of open-stack providers that are in there So the ecosystem is reaching a level of maturity where you can call you can be running HP Helion on You know HP hardware. We're talking to a third-party storage device. You call us with a problem with Helion We treat we we realize it's the interaction between us and that storage vendor We have the capability of routing that trouble case to them now and others. It's reached that degree of maturation So it's become a very mature environment So I want to close with Ten things you probably don't know about HP Helion development platform and open stack one and hopefully you got it from at least part of my talk We have a fully featured pass So platform is a service that's really ready for enterprises and it's running on top of what we view at HP Certainly is as a need as an enterprise ready packaging of all the open-stack technology And some of the speakers that Tim has I think will give you some evidence and proof points for it We have a fairly tight integration between cloud Foundry and open stack because frankly Those of you that are old enough to remember how to write in C How many of you have how many of you actually wrote applications where you talk directly to the system call interface? Okay, well actually I should say how many of you are actually old enough to write applications in C um But most of you probably wrote to an abstraction at a layer higher than that and that's really why we're focused on this It comes hopefully I've at least been able to exemplify to you that it's it's kind of flexible and I've talked to you about the importance of Portability being able to move workloads from a private cloud to a managed cloud to a public cloud and more importantly The ability to move workloads across different things in open stack and we focused a lot on making it easy to administer and Open stack is pretty difficult as Mark said this morning, you know, you shouldn't need a PhD to be able to run this and And he's pretty true. We've tried to focus a lot on a modern developer experience I mean I make jokes about C, but nobody really writes much in C anymore unless you're an operating systems geek, right? Most people are probably writing in Python writing and writing in Ruby writing in a variety of different languages Their developer experience is much much different than it was, you know, when I started my career The type of support that developers want is drastically different than the type of support that sys admins want so we've spent a lot of time with our Platform as a service organization Building the types of support things that developers need from a support perspective HP is pretty much noted for providing 24 by 7 by 365 World-class developer support So I don't think there's much question around that. We also have a pretty rich professional services organization product and deminification and and as a last comment and sort of a tribute to Mr. Jobs and just one more thing This maturing ecosystem, you know, the one more thing things should have gotten a laugh at least from one person and So the final thing and I guess it's the evidence for me that open stack is really here is This maturing ecosystem and as you walk around the conference particularly those of you were here for the first time Look at how sort of seasoned this has become not in terms of how slick the presentations are because they're not very slick but in terms of Who's actually here and you should be able to go from the bottom of the stack all the way up to the top so on that Thank you very much. I want you to Have a wonderful summit and We are very committed to open stack We are very committed to the open-source nature of all of it and thank you for your time And I wish you a wonderful summit come by the booth come by the the pub crawl this evening and certainly come by the HP party on Tuesday night. Thank you