 Let's go ahead and get started. We've got a slide up here just to showcase some of the other things that we've got going on here. And as IBM, we've highlighted one of the sessions that we've got going on tomorrow. We have the 240 right here in this room. We've also got a booth where you can get more information about our products or offerings. You can get some golf balls and hats as well. There's a lot of different things there. I'd like to thank everybody for being here this late in the afternoon. We'll just call this the pre-party. So break it out if you got them. But no, we're really excited to have you here and we appreciate you spending the extra time. You know, it's been a long day, but there's been a lot of good information that's put out throughout the day. Today we're going to get about 40 minutes of a little bit of a shortened session. We're going to cycle through three speakers. Myself, Alex, I'm a product manager for the Smart Cloud Foundation products. We've also got Andrew Trossman, a distinguished engineer who's going to talk to you about OpenStack. And then James Peng, who's here to talk about the products that are available as well. So we're required to put this up. We're going to talk a little bit about a product called Smart Cloud Orchestrator that we're very excited to announce later this month. So our lawyers make us put this out because it's got that coming soon label to it. So my parts can be really short so we can give you the technical details which I know everybody here loves. But I just want to give a quick overview of what we're doing at IBM and how our cloud strategy is coming together. Now, I'm sure you've heard a lot today and we've got a similar view of the world as been put out by all the various analysts and in this structure. But the way we do it is we come about with three different kind of views or dimensions into this space. Starting with Smart Cloud Foundation, which is a set of enabling technologies to allow you to build on-premise solutions for infrastructure as a service and platform as a service that's highly scalable, workload aware and secure. Our next dimension is with the Smart Cloud services. So again, we want to provide infrastructure as a service and platforms as a service for you, but in this case, we want to allow you to not have to worry about the overhead of managing that. We don't want to make you have to worry about security or any of those issues there. So this is an opportunity for you to have those same benefits, but then IBM takes care of the back end of this. So it's a private cloud, public cloud, hybrid cloud, whatever you want it to be, but IBM is managing the back end of this. And last is this notion of Smart Cloud services. So one thing I don't think a lot of people realize is that IBM is actually one of the top ten software as a service companies out there. Through a series of organic growth, also through acquisition, IBM's got over 60 different software as a service solutions, some of which are around business process as a service, or business analytics as a service, collaboration as a service, variety of different entry points or solutions to help you be able to do your business better and engage your customers in a more efficient way. The other thing that's unique about our services or our solution business is that it's everywhere in the world, whether you're doing business in China, India, Brazil, South Africa or the United States, we do business there. Our solutions are there and available. This is one of those competitive advantages that IBM has just because of the size of who we are and what we do. But one of the key things that underlies this is the need to be on an open infrastructure, have an open reference model, so that we can communicate back and forth across these dimensions, and more importantly, so we can communicate with any of the other investments made yesterday or make in the future. We're trying to get away from that world where we have siloed vendor locking situations, make sure that your cloud is interoperable and communicate with the rest of the environment. As I said, I'm with Smart Cloud Foundation, so there's a couple of things that I want to highlight of what we're doing. One of our goals is to create a comprehensive set of capabilities to allow you to build your private cloud. And one of the key pieces of this is it needs to be a progressive situation. You're not going to come in the first day and say, we want to buy the Cadillac every time you come in and say, I want to buy the Cadillac. That would be a great world if that's all we built. So we want to make sure that there's a progression here for you to come in, get the capabilities you need, and then be able to add on those as you go forward. The other key to this is it has to be a situation that's heterogeneous, not just different hardware and different servers, but you're going to make different network choices. You're going to make different database choices. And those things all need to be able to work together not just today but also in the future. And ultimately have that designed in an interoperable and open model. So the overview is going to be over in just a second. Just give me a few minutes here to give you a history lesson of what the IBM products. So I said something about a progressive platform or portfolio. And this started a few years back with a product called Smart Cloud Entry that's in the market and has been in the market for a number of years. And this is a basic cloud offering. It gives you basic management capabilities. It gives you basic usage and metering capabilities. And it also gives you very powerful hardware management capabilities. The next product once you've got kind of that infrastructure and service layer done is to move to Smart Cloud Provisioning. Kind of starts to step you into this platform as a service opportunity. This product brings in the ability to do multi-tiered applications to develop in what we call patterns. You'll be able to scale out and scale down these workloads as needed as a complete entity instead of having to do single images and then string those all together. It's really about automating your business and automating your services to make it more efficient and faster. Now these products both have been in the market for a while and they're very powerful. They've got a lot of benefits. Of course I'm drinking the water from IBM but at the same time we realize there are some issues here. Well we have some great differentiators here. They weren't necessarily as interoperable as we wanted. So people like Andrew here and other distinguished engineers with an IBM and other great developers that we had said hey there's this great opportunity of starting an open stack. So we started to search that and we started to build along those lines to make those investments. Like I said we've got a coming soon label coming up with this last product which is Smart Cloud Orchestrator which takes us not only the next step in the capabilities that differentiate IBM around things like complete data center orchestration and automation and incredibly easy interfaces around your processes. Probably more importantly it allows us to take that step into complete interoperability on open stack. This is a leader within IBM not just because it's a great product and then it's a step forward in our cloud foundation portfolio but also because it's our first open stack product. Not just the first for the Smart Cloud Foundation meaning that provisioning will soon be on open stack and entry will soon be on open stack but also our services will be on open stack and our solution will be on open stack. So this is just the first of many investments that IBM has made that's coming into reality today around this investment that we have with open stack. So with that I want to turn over to Andrew who's going to go into deeper details and deeper dive into those investments how they're panning out for IBM and why they're important also to open stack as well. Great, thank you. So we have right now they're that's fine just they're over 250 people in IBM that are focused on open stack now. A year ago this same time I could count that many the number of people on maybe two hands that were involved in open stack. So a lot of this has been very, very, very recent. A lot of the products that Alex just showed you originally were before we embraced open stack and actually myself and a few others here in the crowd today were the early guys getting involved in open stack and we recognized certain things certain technologies that we had that we wanted to donate to the community and of course that donation is done by porting and building patches that we then contribute into the product. So I just want to step back a little bit and look at that this is an orchestrated product why are we doing an orchestrator what is that? And this chart here is a workflow that actually came from a real customer of ours when this is a workflow that they use when they deploy their systems into their data center and the point of this chart is not to show you all the little details but to show that there's a whole lot of crap that has to be done just provisioning of a VM to get it into a data center there's lots and lots of other processes and systems that you need to integrate with particularly for enterprise data centers next please So in our smart cloud orchestrator tool actually one of my buddies one of the open stack founders he used the phrase smart cloud orchestrator is like visual basic for the cloud and that's what we're trying to do is to really simplify the way that you can exploit the cloud for all the wonderful things that many of you folks have done with the cloud to make it easier and so one of the things that we do here is when you want to create an application invariably an application is going to be a collection of a bunch of virtual machines with a bunch of software on them that all need to work together so here's a simple drag and drop tool that you can just drag onto your canvas and you can drag VMs onto the canvas you can connect those VMs together and you can put software into each of those VMs next please as we move up so that's kind of we refer to three different levels of orchestration at the base level we call resource orchestration that's really about orchestrating server storage compute and network and really that's what the core of OpenStack is all about and that's why we're standardizing it there as we move up we get into what I just showed you which we refer to as patterns or workload orchestration but as you go further up we refer to it as service orchestration and this is where we can start integrating other systems like CMDBs and service desks and backup systems and monitoring systems and invariably even infrastructure like firewalls and load balancers etc and we have best in class workflow tool we acquired a company called Lombardi award winning software and internally we call it our business process manager it is a fantastic tool from the point of view of the designer so this is a nice GUI tool for creating workflows and in the workflow it can be completely unattended automation but you can also have human human steps integrated into those workflows and it's just as easy it's all drag and drop you can create little user interfaces all using this tool then on the runtime side it's a highly scalable robust platform for doing these kinds of automations next please so this is something we call our common cloud stack internally this is a collection of technologies that we've integrated together from which we provide a bunch of different offerings and Alex mentioned some of the product offerings that we deliver and the point here is that because everything's built off of this same set of technologies wherever possible we use open source and open standards to make all of our pieces interoperable not just interoperable with our own technologies but with third parties and customer products as well so at the bottom you see the typical infrastructure as a service coming from open stack we have something we call our IaaS gateway and this is something you know if there's interest to bring it into the community it's certainly something we'd be happy to do and essentially what it does is it helps us federate multiple clouds because most of our customers have lots of clouds deployed across the globe sometimes it's for geographic reasons sometimes it's for different purposes like dev test versus production usually you have to have many of these and you want to simplify the management of them and the consistency of your images across them without having to do silly things like you know sending terabytes over wan links as you move up you'll see the background in that bright colored green we have this online marketplace that we're developing and what the idea is that throughout our stack there's content automation it can be as simple as I have images and open stack compatible images that we want ourselves to deliver to our customers partners, vendors as well as open source be able to make these images available so somebody can download them and say here's an open stack compatible image that I can go and use to you know run a patchy or something as you go up there things like you might want to be able to download chef recipes, chef cookbooks another you know software stacks the next layer are these patterns and we're embracing some of the standards like TOSTA you might want to be able to download or if you're a vendor you might put a make your stacks available so people can then download them and then deploy them in our environments and lastly is the workflow at the top and that's that business process tool that I mentioned before next please I'm not going to get into too much detail here but in each of the Cinder, Quantum and Nova there's some technologies where we're we've made contributions and we're going to continue to make contributions but we also have some IBM technologies that we maintain of our own one example is over here in the storage space we have a technology called GPFS it's a cluster file system that's been around for many years and we've built the Cinder drivers but it's a technology that's been scalable robust and battle tested for many years next please another key thing that many folks have found in trying to use OpenStack is that the support for VMware is not quite on par for what people have typically deployed most people when they deploy VMware they're going to have vCenter and you're going to set up your data sources, your clusters and all those kinds of things and the current OpenStack drivers really just can't work with that as you saw we already had a number of cloud related assets so what we've done is we've taken those cloud those existing assets for talking to VMware and using the same OpenStack driver plug-in model we've plugged those in into OpenStack so that we can manage existing vCenters as in in place without making any changes there's a lot that needs to happen in this and this is something where we've been working with VMware and a collection of other partners to improve what's in the open source over time but right now this is something that we had to put out for our customers to actually that works now next please and here's one other key component actually there are a number of components that we have that I think add additional value to an OpenStack environment and we've got James here from our platform computing acquisition and these guys were for 20 some odd years doing high performance computing and I'm going to let James tell you a little bit about how we leverage that technology for OpenStack okay thanks Andrew so as Andrew said a year ago you can count the number of people that work on OpenStack in two hands so a year ago my team I'm from the platform computing team in IBM and we joined IBM as part of our acquisition a year and three months ago and a year ago our team also didn't do anything with OpenStack but now we're we're knee deep in OpenStack as part of the overall IBM effort to first contribute to OpenStack and second to also drive adoption of OpenStack and also to leverage the OpenStack so that we are also creating interoperable clouds and sharing all the good stuff around an open community so a little bit of background about this thing called EGO so this EGO technology is a resource manager and what it is is a matchmaker between supply resources and demand your applications your workload demand as I mentioned to a lot of this IBM partnerships and acquisition EGO is not a new technology it's battle proven it is in the high performance and grid computing in which platform computing have been involved with for almost 20 years and there are literally thousands, 2,000 plus customers using the grid and the high performance computing products that have been using EGO structure so it has been doing that resource management in a different market in a different context but if you look at this spirit of OpenStack it has a lot of semblance to grid computing, scale computing and even high performance computing and this EGO technology is one of the things where after we joined IBM it looks like hey this could be interesting this could be interesting in creating enterprise resource management capability for OpenStack based clouds so there are two areas of EGO that it does as it adds value to the OpenStack context one is placement VM placement, policy placement of VMs based on resources, based on load based on where it is in the rack for example and the other associated would be migration where should I migrate VMs based on SLA and based on policy so that's the runtime mobility aspects that EGO brings to an OpenStack environment which is the VM placement and replacement now the other aspect of EGO as a resource manager is being able to arbitrate the use of the resources among different tenants because at the end of the day a cloud as much as we think there's infinite resources well it's actually not infinite especially if it's inside your own data center there is a box around it so what's the best way I can prioritize the use of these resources among different departments among different projects among different groups and these map to the concept of OpenStack called tenants how can you define shares among different consumers and that's an EGO concept so that's the other aspect of EGO in terms of what it does so in terms of some of the policies that EGO already has and again a lot of it we're trying to as part of our effort to port it over to the OpenStack context includes things like packing putting as many VMs on as few hosts as possible, striping which is the office, striping that VMs as many hosts as possible both in order to meet certain SLA's for example minimizing memory fragmentation etc and then other things that we talk about being aware of low, being aware of things in the racks, all of these things are intrinsic in the EGO technology but at the same time one of the things we learn and I think in the OpenStack community for folks who are building clouds also I've learned is you can't be restricted by whatever is out of the box the customer whether it is a service provider whether it's your admin, whether it's your infrastructure team needs to be able to write your own policies because your policies map your own organization's SLA to your own organization's business needs so one of the capability of EGO is extensibility being able to expand it so that you can create your own site specific policies so how do we integrate this EGO component into OpenStack so this gives you a picture of how that's done and what a great thing about OpenStack is it's open, right you find interface around the schedule around the NOVA scheduler and so we leverage that component of NOVA to work with EGO and one of the things we did have to do was to create a Python layer and that Python layer was new but that enabled us to then talk with the NOVA scheduler so that's what you see on the left-hand side on that side so that takes care of the initial kind of the initial placement and intercepting every call that's coming in so then we can then apply our policies against whatever are the new VMs so we know what's going on so then the other aspect of it is the runtime, when do you do migration when should you grow and shrink your resources so that's where we introduce a second component called the runtime policy engine and that runtime policy is new code, it basically monitors the environment and based on some of the pre-configured policies it would then trigger migrations based on those policies so those are the two areas that we bring into this architecture in order again to add to bring enterprise level resource management to an open stack based environment and the final thing I just want to highlight is the hardware management module we follow the open stack parameters on how to gather metrics and so on but in case the customer wants additional metrics that are not being provided by open stack the ego itself also have agents and these agents you can also deploy in a complimentary way so they can collect your own site specific metrics such as application licenses as a metric to do placement so that again shows the example of embracing working with the community at the same time providing additional extensions and value add to enterprise customers so I'll just wrap up a couple of illustrations in addition to the runtime so because I have a lot of time when we first did this initial development focus was on runtime where should I do the placement and when should I do the migration a lot of focus was around there but as we getting more and more engagement with enterprise customers and enterprise service providers it becomes clear that you have to deal with the resource arbitration the priority of the different tennis in which you are serving so one of the things that comes with ego is it has a concept of reservation and this is actually based on capacity so it's not kind of like this runtime VMware type reservation where you give a time and then you run something it's not about that it's actually guaranteeing that the resources is available at that time in the future so it comes with a whole reservation based system what we call advanced reservation which is very typical that you see in the HPC and grid computing environment so we bring that reservation concept into this into the open stack world so that you can have very well defined predictable SLAs and then the other aspect just to wrap this up is the idea of again going back to the consumer aspect the multi tenancy ego itself has this concept of multiple consumers and these map to the tennis and these can represent different departments, different projects and you can then define shares among them and you can define things like minimum and maximum and what maximum means these are the amount of resources that you want to borrow from somebody else from another tenant if that tenant is not using those resources so ego comes with these concept of borrow understanding and sharing policies and so with these kinds of attributes we think that this will help enterprises take open stack to the next level which is providing enterprise level resource management so I think just before we take some questions I think one of the key things to what James was showing you is it's open right open stack provides open interfaces that allow us to plug these kinds of capabilities in this capability is open in terms of it can work across different hypervisors and it's also open in terms of there's control given to the end user so that you can define the policies that meet your needs so I think that's going to really be something we've seen a lot of our customers see is really valuable and we hope you'll find that too so we have some questions I see a question over here the open source so for specifically for the ego stuff or the whole okay so the open source components if you look in our the smart cloud orchestrator so the whole stack it's basically open stack we also enable chef and hopefully puppet in the very near future ego is proprietary code yeah the ego itself is proprietary some of the metric collections that we're working on those things open source in our deployment so our deployments in smart cloud orchestrator today is with cupid I don't have that information but we do have one of our guys from our performance team is is at the summit this week and so I can certainly hook you up with him if you want to connect with me after I'll connect you with that guy so can you be a little bit more specific in which product which one smart cloud provisioning smart cloud entry or the public cloud okay so the public cloud is alright so public cloud is a little different today the public cloud is not using open stack yet that's something that we are in the process of doing and IBM I mean IBM has a historical value which is maintain backwards compatibility so those APIs are going to continue and they're going to be maintained for a fairly long period of time I don't know precisely how long but so as the public cloud does adopt open stack those will become new APIs but the legacy ones will continue to be maintained any other questions can you speak up louder please the baseline do you want to just repeat the question the question about the reservation is it this advanced static or is it also based on predictive data that you can change over time in other words the reservation itself can change over time so out of the box the idea here is that exactly the name is actually called advanced reservation which is to guarantee a certain capacity a certain time a certain point in time just like a hotel reservation system so having said that if based on runtime information and based on different policy changes those reservations can be changed we've had these in the grid world where customers have taken some data some workload data and using that to feed and then they adjust the reservations accordingly in the future it can support both models but I think the second model is more it's more customized than the first one which is really out of the box other questions for both the schedules and the policies is there some sort of format that's explorable that would be used in other other implementations like an XML format or something that would define the policy and the scope actually right now the ego component is actually technology preview the way you configure the way you define the policy is through XML format right exactly so as a tenant I also have to give a parameter called minimum my guarantee and then because I'm playing in a cloud world at this point if you also tell me how much I could lend out to others if I'm not using it so all of that is in control of the tenant but typically in an enterprise environment it is some kind of central IT person or group being an arbitrator talking to different business units and then working out the priorities and working out the borrow and lending policies so in smart cloud orchestrator that's currently in beta to be announced whenever that's based on Folsom and it is Nova networking our plan is to start aligning our release cycle as close as possible to the open stack releases so sometime this summer we're going to have an update with Grizzly then Havana in the fall and along the way we're going to start making use of Quantum other questions how do we monitor the infrastructure so one of the things with IBM is we have no shortage of products we like to buy companies we take those companies we create new products we also like to market a lot of crap out of everything and so we've got a brand called smart cloud and so we actually have a product called smart cloud monitoring and jokes aside that's exactly what we try to make it as simple as possible so you can buy our smart cloud foundations core cloud products based on open stack you want to do monitoring great add smart cloud monitoring oh you want to do usage and accounting great add smart cloud cost management and there are a number of these half a dozen or more that kind of help round out the overall offering other questions anybody anybody ever use an orchestrator tool yeah anybody use runbook automation tools yes a few anybody got a question who's thirsty open stack is that going to be I sort of don't want to have yeah Alex let me make sure I understand the question are we going to have our own distribution of open stack yeah it would be and it would be in the context of those products that we talked about earlier so we've got smart like I said smart cloud entry provisioning an orchestrator which by the end of the year we intend to have them all on open stack and those would be on a common base open stack that's within IBM the key is that it's not going to be anything different necessarily than what open stack is we're not going in there we're not changing any of the bits of open stack but we're adding to that and extending above it capabilities that are differentiating so like I mentioned the smart cloud entry has got very detailed and powerful management hardware management capabilities provisioning ads some pattern capabilities orchestrating ads BPM but it's all on top of it a common open stack component there so if I step back for a moment so there are three products that are in the field today smart cloud entry smart cloud provisioning and smart cloud orchestrator and they kind of build on each other today smart cloud orchestrator is the first one to pick up open stack in the summer we're going to start adding for each of the other two the idea is to get them all lined up on the same technology base in terms of the open stack we are not having we have no plans right now to do an IBM distribution of open stack but we do ship it as part of any of those three products that answer your question of those three products yes but it's embedded but it's also exposed yeah I mean we're not hiding right and we're not doing anything unnatural to it we do have some plugins to it maybe maybe a little bit a little bit unnatural but that's for later any other questions before we all run out and get dry how does he well so I think he fits very nicely into what we referred to as workload orchestration so it's more about I mean the term we've been using internally are patterns and so we think that's fantastic because the most important thing is to start standardizing so that people can create those templates make them available in a marketplace or somewhere online so we can distribute them and get that kind of consistency and I think one thing to add in here is along that notion one question is how is the best way to interact with business user on how to create those templates Andrew showed a couple of screenshots of Smart Cloud Orchestrator mentioned the best of breed BPM capabilities and the pattern creation capabilities you know that's one aspect that I think is very crucial to understand here is how can we create a user interface that makes it so that this crowd doesn't always have to be right in new templates but that we can start to pass that off and use the development resources in better ways in ways that are more entertaining and exciting for the developers and I think that's one thing that is a very big differentiator for us we talk about how can we make it easier to build and consume so that people can code things that are the next gen instead of the next template Other questions? Everybody having fun?