 Hello My name is Nikkeler. I am with Dell IT lives in South Texas, which explains the twang and South, you know south Texas. Yes take a left swim a few hours on any home And I am part of Dell IT Which means that I am a consumer a user of all the stuff that is being talked about today and Like everybody in IT I had to basically come in with a baseball bat to try to convince people that this private closing might be a good idea and So I will talk to you about three things first one is why IT today need to innovate Okay to What are the what is the right way to innovate and finally what we are doing in our lab? If you saw the del quinoa yesterday, there was a mention of an innovation lab This is what I have built with my team and so I will tell you a bit more about that so First thing why I teen is to innovate Okay, you know a big thing that we want to make sure is make sure that Dell IT can start using new technologies so that we can hopefully get out of the 1980s Okay, and when doing that make sure that it makes sense from a financial point of view So our CIO Andy give us some very very clear directions. Okay, basically go ahead disrupt Make sure that there is a lot of value for users. Okay, incubate on steroids and It's not this is it because as you know IT is all about stability Establishing a comfort zone. Okay, if it works do not touch it period And then if you try to hire Someone who's just getting out of college. They don't know what the computer is the only device that they know is a cell phone Okay, very far from the nice green screen that you have been using to manage your customer. That is working perfectly fine So there's a lot of conflict right there Technology is evolving so fast and user experience of everything so fast Internet access people can do stuff at home watch movies Turn off their lights remotely from the internet inside their house Now control the temperature inside their house remotely. That's a lot better than what a lot of IT Can offer them so how can we make that better? Okay? That is a biggest conflict the biggest challenge when fighting against the fact that any CIO needs to provide a stable environment and The key a key metric is really to improve the user experience Because at the end of the day IT is being used by people. Okay, usually You know if you have a large at the organization in a large company Nobody likes IT Okay, because IT is the people that are going to take six weeks to give you a laptop and They're going to give you the wrong one Okay, then if you dare break it You're going to have to talk with someone who has no clue what application you're trying to run Okay, we have all and been there. Okay, so very very bad user experience We need to change that that is our prime directive and if we start to change that then we'll be able to innovate in the right direction Because it is also so easy to try to do new things that are completely wrong that makes things even worse Okay, so some of the things that our team We are both a team of 30 people Part of a 90 department that has roughly 5,000 people. Okay a few things that we have done Is that we have established a standout for device refresh What does that mean is that depending of your job? We have tried to match the right device to the right job type for example a sales guy Doesn't need a graphic workstation to go and see customers a Tablet might be a great thing for a sales guy Okay, and we need to make sure that the engineer get the right source power No, I have some engineers in our team. They get one of those laptop that has 32 gig of RAM They can run a few VMs on top of it Okay, and now that is a standard that is available for some engineers Okay, make it a lot easier We also So that you know Samsung it seems is coming out with a new phone every three weeks or something like that. Okay People want to see those new phones Okay, so not only we make sure that we have a program to refresh our corporate smartphone But also we establish an app portal So that if someone wants to file an expense report for example Whatever device they're using a laptop a cell phone a web browser They can have the same experience it is becoming more seamless for users to do their job that make it easier They don't have now, you know earlier. I mentioned a sales guy that has a tablet He doesn't have to go on to go to a web browser on a PC In order to file expense report you can do it on on his cell phone on his tablet And lastly what was mentioned the IT innovation lab as we have to create a bunch of stuff We built a business case to basically get a bunch of gear Put it in a nice lab and start to run a bunch of stuff completely isolated from the rest of the world Okay, so How are you able to do that? Okay? How can we innovate and in a very short period of time with a very small stuff? Come up with all this cool stuff the first step that you need to get executive support Okay, we got that from our CIO our CEO told us go and do stuff. That is the first step But that's not enough. There are three things that absolutely you need to do Very very important when you innovate you need to innovate on the right stuff Okay, the right stuff is never the latest on greatest technology It might be but very often it is not what is the right stuff if the stuff that matters to user community Okay, and you need to find ways to get that from your users. Okay, so You can interview people. Okay, you can ask them a what do you think is good? What do you think will be good? If you are the same stuff 20 times that might be a good idea. Okay tip Do that in a very anonymous fashion Tell them that you're going to collect their results their things. Okay, the idea is but you're not going to give their names to anybody We have a program, you know that we started like this at Dell and suddenly we're starting to have people complaining a lot About stuff that slow them down in their day-to-day work That's how we found some of the best ideas. Okay, if you interview 10 different groups And you are six time people complaining about the same exact things Then that's perhaps what working on it. And then you can start talking about it to other people crowds or saying Okay, start to basically have contests. Hey, what do you think will be a great idea to improve sales productivity? I'll see what people come up with and let people vote to what the bet the best idea is Day-to-day interaction Look at ticket cues. Okay, the good old ticket cues when people come to call to complain about something that is broken If you see the same thing brain being broken 500 times do something about it This is how not only you can Generate ID But also present an ID to leadership to say hey, you know, see, you know what? This is a problem that we see a lot or an ID a lot of people asked for that Let's make a project about it. Okay the next thing as soon as you do that your leadership is going to ask to tell you Good idea, but I'm not going to get you a penny to make it do it. Okay Show me why we should do it Then you need to start the only language that your leadership will understand business case Okay What we do is that we start with a four step process When initially inception we basically say, you know what idea of the day They should serve us some money or have some benefits that was investing a bit We could play a little bit with it Then we do a proof of concept where we really bring the technology into the lab We try it out for a few months perhaps even more to really prove that it is that it works. We can size it Okay Then we start meeting with production IT To tell them a look at this brand-new thing Let's work together on making it fit. What does that mean? We need to do training for operators if it is a new technology We need to figure out, you know, what is the right gear that we want to use to deploy it in a data center And then finally We can transition into production and that's something that is very very important Okay, a mistake that a lot of people who innovate do Is that once they create something it becomes their baby and they want to keep it on a get Okay, let the baby grow get rid of it go to the next one very very important right now I spend a lot of my time working on open stack Okay, my plan is in December to get done with this phase of open stack because I will work on something else and Production IT will take over the stuff that we have done Very very important learn how to let go and build a business case for the whole for the whole step Okay, to do that that require a special type of people. Okay, because I'm talking about business case Okay, I'm talking about being able to sell leadership on your ID So you people, you know one person business Who are not afraid to speak in front of an exact who can articulate even in French Okay, we can very quickly make strategic theory decision. Is it a good ID? Is it a bad ID based off a whole high-level strategy? Okay, and also are very passionate about technology Because when you talk to people when you present them a new technology You need to show, you know if you really believe into it in it That will show that will help you gain people to use your technology So that is the third part. Okay, you have to have the right type of people You need to make sure that you've got a good business case through the whole process if you don't have a business case anything that you're going to do is going to be shut down very fast and You need to make sure that you innovate on the right stuff and innovate on the right stuff need to really you really need to Listen to what is the stuff that people want and what is the stuff that people do not like? Okay, and so one of the things that we have done is We started to do a private cloud Okay Private cloud means that I'm going to have to spend a lot of money. Okay. I'm going to buy servers I'm going to buy switches. I'm going to buy software. Okay to build a lab I'm going to spend a few hundred grants right there. That's coming very painful for the business case. Okay If you deal with open stack, there's something that is absolutely beautiful Which is object store? Okay, if you have to build a business case on open stack here's how to do it. Okay Talk with your storage team Ask them what they do today to archive data What technology they use to archive the stuff that they want to keep for five years They're going to start to talk to you about putting it on a block array Enterprise grade block array. Okay, then how they take a snapshot Then how they take the snapshot and they put it on tape Then now they use they have a beautiful contract with Aaron Mountain to Sheep the tape out and keep it in a role for five years And then now if they need to take back it only going to cost two grand per tape to bring it back in a Superfast and then they might in fact need to buy a new block array in order to Do some stuff just to make sure that everything works very quickly. You say, you know what I could replace all that by a few servers okay run some software that is affordable Okay, put a few ops guy on top of it and Be done and you're going to very easily show that using something like a swift or a Seff or something like this you can very easily show a 60 to 70 percent Money-saving. Okay, right there Leadership will listen to that and they will tell you hey, you know what? Is there some other stuff in this open-stack magic thing that's ever so much money on then you can tell them about the compute and When we talk about the compute you can start to find some example the thing that we did that at Dell, you know within the lighty Okay, that within the lighty we wanted to demonstrate that the software Define for sure could be efficient secure and give a great user experience Example I am a software dev guy okay, I Need some type of virtual machine to get something done Okay, traditional IT. I talk with the network guide to get an IP address I talk with the server guy to get an open system and operating system stood up on a VM I talk with the storage guy to get a bit of storage One week like a one week later. I have my virtual machine, which is perfect Except for little facts. I know understand my software a bit better and my virtual machine is the wrong size Good news you can change that and one week later. You get the right one. Okay, this is a lot of IT department work Okay, so what we're going to bring is basically look We're going to bring you self-service Your software dev is going to be able to within a budget that we're going to pre-define Provision their own stuff when they want it Right there. There's also a business case. It's harder to measure But just the time that is going to be saved Will Hello your software devs to be more efficient Also since we are Dell okay, and I've heard that they'll do some good hardware As a partnership with some people We need to make sure as a light that we are a showcase for Dell. So only use the hardware They don't tell okay. That's what we need to do at the end of the day Okay, which give them, you know And so what we're going to do for our private cloud is that we're using Dell servers They're networking for Stan Dell storage, which is a mix of server with bunch of drive on Ecologic or compelent blocks systems and we're going to heavily leverage our red app partnership and Also very very important. Do not forget that there's some existing stuff Okay, in any IT shop today, you've got everything is running on top of VMware Okay, make sure that you really understand how The stuff runs on top of VMware What are the processes that are being used so that you can identify some workloads that are candidate to be moved So that once you have deployed your new infrastructure Based on open stack that is slightly different you can also have the right understanding to be able to start migrating your Cloud right now, you know and things like SDN that Kamash talk to a node Kamash talk about storage But the discussion around SDN is very very important Because if you start to have more flexible network be able to have the legacy worlds and the new world Talk together, and you will be able to little by at all split your application across or move them around Okay, so based on that we started to play in our lab Okay, and what we discover very fast Okay, that's good news. We could build something. Okay using earlier of the hardware In the middle of a bunch of virtuosism Open stack with Freddad. Okay, make sure that we keep in mind all those are stuff Which is VMware and everything Microsoft? Okay Then on top of that Start to layer some del software and work with the del software team to start to really leverage the del IT IP Intellectual property the way I like it the way I like to say to that if I take open stack We Freddad on can start to add some del thing on top of it That's really better always better for del too because we're using some of four, you know We're drinking our own champagne, which is very interesting. Okay and What we discovered is that was flexible. It was secure and we found three key use cases Okay Archive storage, which is good because it was the business or for business case a lot of demand for that Okay, from just backing up to compliance Legal team loves us. You mean that now I can store all the legal stuff that they have to keep and not have to save it On top of Amazon and I can save it in house because those lawyers they hate to have any data leave the building Okay, so that's great stuff. Okay the software devs. Okay talk with your software dev teams Show them that look you want to VM to just try an ID go for it Okay, they're going to love you then they're going to talk with your manager The manager are going to come to you and say hey, we want to play in your lab Want to try new stuff in your lab? Okay, that is great to help you move stuff forward And if you're in an innovation business, we use that a lot in our lab to just incubate on trying new stuff a lot of technology Okay, and so what we have done so far today. Okay is that we have demonstrated it works and Our next step is really we're going so we deployed open stack. Okay using OSP tree Not really worrying about the stability of it. Okay, we just wanted to show look here something that can work We got a lot of interest around it Okay, so now we're going to show that it can work for IT remember the only thing Other than spending no money that IT cares about is stability Okay, so we're going to deploy OSP for as soon as I come back To to beautiful Texas. Okay. We're going to deploy Havana. Okay. We're not, you know stable We're not going to play with 5,000. Okay, stay with Havana Make sure that we've got a chair for the controller so that the API is always available And then I'm going to pull a dirty trick on the rest of IT I'm going to tell them. Oh, no, it's okay. I can upgrade my controller whenever I want Change control window and of the month freeze. I don't care about that Your stuff is available 100% of the time. It is the end of maintenance windows. You need to show the new world Okay, we're going to also to be on the safe side use Nova networking rather than neutron I'm right now a bit uncomfortable with neutron. So we're going to wait a little bit for that Okay, and since we're there, we're going to use some del storage ecologic as a backhand behind cinder and One thing that we have discovered Is that the people who are interesting by archive another similar that there are some people that are that interesting by compute So we're going to fully separate the two aspects. We're going to build storage cloud for archive and we're going to build a compute cloud for the developers basically Okay, which by the way will make our life a lot easier and our past to success Okay, we are starting This is based on the knees Cloud reference architecture. We're going to start, you know working very tightly with red hat Defining resource abstraction and control, which is basically open stack What I'm going to put under open stack to make it work to have a cloud that works Okay, then I'm going to work with our IT operations to understand What are the business processes that they go through if someone is a virtual machine? Where is the money coming from? How do we check that they've got enough money who has the authorization was the right to authorize the money to be spent? How how is the showback system working? That is all basically a lot of the service orchestration What type of service catalog do we want to have to have application pre package or system pre package? We're going to build a layer of service orchestration on top of that That we are also then going to take care of the VMware integration so that we can have under the same Automation of the business processes give people access to boost the new stuff and the old stuff Okay Sorry, I don't mean to be mean by calling VMware the old stuff But the different viewers of the world. Let's put it this way. Okay, so that From the people from the developers they can really pick and choose and what we're going to see I'm ready to bet that we're going to find that people who have Application that are fairly complex that have a state full a state full backhand Fronted by a state less backhand are going to very quickly ask a Can I put my stateful stuff on the very mature technology? on top of yes, expressors and Can I put my state less tough on the top of the technology that has a very easy to manipulate API? So that I can scale it whenever I want without worries So this is also what we're going to establish and over the next few months Building that we are going to figure what we really need to monitor Okay, because it's useless to monitor everything. Okay, you just need to figure what to monitor on when to act on it For you know Paul in a previous life. I was working with some fairly large trip. Deployments. Okay You do not care if you've got a large object or deployment. You do not care if a drive fails Don't give a damn about it. Okay. We just replace it whenever This is this goes against a lot of the philosophy of IT is something break and into act on it No, we need to also teach people that we need to build for the fact that things are going to fail Okay, so this is all the green stuff, which is really the operation aspect and the final integration you know with active directory and You know identity measurement active directory ticket cues This is what we're going to take care at the end of it and by the way I have a secret objective which is to get rid of ticket cues for this Okay, I do not want people if they've got a problem using this private cloud platform I do not want them to have to open a ticket. Okay. Here's the concept of a ticket. I Go to a web interface. I Write my problem and we've seen 24 to 48 hours someone was no clue about What I'm doing is going to tell me that they're willing to help me if I give them more information That is the wrong concept if we provide self-service to the infrastructure Okay, so we're going to experiment with basically what a lot of companies are doing today Which is social media? knowledge base user groups Very polished IRC like channels Okay, make them all pretty so that you know People can hopefully get instant feedback instant supports Okay, that is what we believe is going to help us transform IT on bringing a private cloud within IT what we did one year ago when I had my first meeting with the infrastructure with the IT why John I tell IT one year ago when I had my first meeting with the infrastructure management team I was basically told that it was very nice of me to meet with them But it was probably the last meeting ever that I would have with them since My career was over in the lighty since I wanted to do all this fancy stuff. Okay then as of roughly one month ago the same team has Created a small team whose only job is to become the operators of Whatever I'm building right now. They're now go to the point where they're ready to accept Okay, and we are probably going to end them out our first Stabilization of a private cloud in July. Okay That matches all the Dell internal recommends Working on that with some of the people that you saw earlier Okay to make sure that we get the best of the Dell red-eye partnership and probably Two out of the year that will be in production within the Dell IT data center. So That's it for me. Okay any questions Yes, so we're talking about that with them right now. It will really depend of the comfort factor that they have Okay if They want to be very cautious We are basically going to stick with OSP for on On the on the Nova networking for a bit longer. However, there. There's a big. There's a big challenge in that Uprighting from Avana twice house would be a very fun exercise So it's why hopefully we're going to see if look, this is the new stuff It's not as nice. It's not as refined potentially, but it should be easier to upgrade in the future Okay, so that's what we're going to do and in June July. We're not yet Going to carry production workload. So it will still be okay to play around with it So they are I love neutron. I love neutron for the features. Okay, I Have banged my head against neutron Where you know layer 3? It can be very interesting to make sure that I do not have a single point of failure Okay, being IT. I want stability at all costs. I will sacrifice features In order to get stability Okay, right now I Believe that we still need a little bit I do not know if it will be you know to Juno or perhaps after to have a very stable Neutron where I am sure that there is no single point of failure because today if in an IT data center If you've got a server that has two nicks Okay, the network should always be up Period right now. I'm not comfortable building that with neutron. Any other question okay, well, thank you and what I hope to do is that all myself or someone from my team at the Paris summit will be able to come and say look We weren't even faster that what we thought we were doing. We're already in production. That will be the perfect world We can always dream. Thank you